diff --git a/ci-scripts/Jenkinsfile-gitlab b/ci-scripts/Jenkinsfile-gitlab index b371f6af20..6ae75d71ae 100644 --- a/ci-scripts/Jenkinsfile-gitlab +++ b/ci-scripts/Jenkinsfile-gitlab @@ -212,7 +212,7 @@ pipeline { steps { gitlabCommitStatus(name: "Build eNb-ethernet") { timeout (time: 20, unit: 'MINUTES') { - sh "./ci-scripts/oai-ci-vm-tool wait --workspace $WORKSPACE --variant enb-ethernet --job-name ${JOB_NAME} --build-id ${BUILD_ID} --keep-vm-alive" + sh "./ci-scripts/oai-ci-vm-tool wait --workspace $WORKSPACE --variant enb-ethernet --job-name ${JOB_NAME} --build-id ${BUILD_ID}" } } } @@ -221,7 +221,7 @@ pipeline { steps { gitlabCommitStatus(name: "Build UE-ethernet") { timeout (time: 20, unit: 'MINUTES') { - sh "./ci-scripts/oai-ci-vm-tool wait --workspace $WORKSPACE --variant ue-ethernet --job-name ${JOB_NAME} --build-id ${BUILD_ID} --keep-vm-alive" + sh "./ci-scripts/oai-ci-vm-tool wait --workspace $WORKSPACE --variant ue-ethernet --job-name ${JOB_NAME} --build-id ${BUILD_ID}" } } } diff --git a/ci-scripts/conf_files/enb.band7.tm1.100PRB.usrpb210.conf b/ci-scripts/conf_files/enb.band7.tm1.100PRB.usrpb210.conf index f3cb31f449..2c2ca635a6 100644 --- a/ci-scripts/conf_files/enb.band7.tm1.100PRB.usrpb210.conf +++ b/ci-scripts/conf_files/enb.band7.tm1.100PRB.usrpb210.conf @@ -197,8 +197,8 @@ MACRLCs = ( tr_s_preference = "local_L1"; tr_n_preference = "local_RRC"; phy_test_mode = 0; - puSch10xSnr = 200; - puCch10xSnr = 200; + puSch10xSnr = 160; + puCch10xSnr = 160; } ); @@ -227,7 +227,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_ENABLE"; } diff --git a/ci-scripts/conf_files/enb.band7.tm1.25PRB.usrpb210.conf b/ci-scripts/conf_files/enb.band7.tm1.25PRB.usrpb210.conf index e33daa7a8a..4dd8c1ee22 100644 --- a/ci-scripts/conf_files/enb.band7.tm1.25PRB.usrpb210.conf +++ b/ci-scripts/conf_files/enb.band7.tm1.25PRB.usrpb210.conf @@ -197,8 +197,8 @@ MACRLCs = ( tr_s_preference = "local_L1"; tr_n_preference = "local_RRC"; phy_test_mode = 0; - puSch10xSnr = 200; - puCch10xSnr = 200; + puSch10xSnr = 160; + puCch10xSnr = 160; } ); @@ -218,7 +218,7 @@ RUs = ( att_rx = 0; bands = [7]; max_pdschReferenceSignalPower = -27; - max_rxgain = 115; + max_rxgain = 125; eNB_instances = [0]; } @@ -227,7 +227,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_ENABLE"; } diff --git a/ci-scripts/conf_files/enb.band7.tm1.50PRB.usrpb210.conf b/ci-scripts/conf_files/enb.band7.tm1.50PRB.usrpb210.conf index 32e8bb9931..cae47b152c 100644 --- a/ci-scripts/conf_files/enb.band7.tm1.50PRB.usrpb210.conf +++ b/ci-scripts/conf_files/enb.band7.tm1.50PRB.usrpb210.conf @@ -197,8 +197,8 @@ MACRLCs = ( tr_s_preference = "local_L1"; tr_n_preference = "local_RRC"; phy_test_mode = 0; - puSch10xSnr = 200; - puCch10xSnr = 200; + puSch10xSnr = 160; + puCch10xSnr = 160; } ); @@ -218,7 +218,7 @@ RUs = ( att_rx = 0; bands = [7]; max_pdschReferenceSignalPower = -27; - max_rxgain = 115; + max_rxgain = 120; eNB_instances = [0]; } @@ -227,7 +227,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_ENABLE"; } diff --git a/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.100PRB.usrpb210.conf b/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.100PRB.usrpb210.conf index fc2f24b8a1..6c80a4f7cf 100644 --- a/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.100PRB.usrpb210.conf +++ b/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.100PRB.usrpb210.conf @@ -200,8 +200,8 @@ MACRLCs = ( tr_s_preference = "local_L1"; tr_n_preference = "local_RRC"; phy_test_mode = 0; - puSch10xSnr = 200; - puCch10xSnr = 200; + puSch10xSnr = 160; + puCch10xSnr = 160; } ); @@ -234,7 +234,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_ENABLE"; } diff --git a/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.25PRB.usrpb210.conf b/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.25PRB.usrpb210.conf index 010ef28061..f4be50810c 100644 --- a/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.25PRB.usrpb210.conf +++ b/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.25PRB.usrpb210.conf @@ -200,8 +200,8 @@ MACRLCs = ( tr_s_preference = "local_L1"; tr_n_preference = "local_RRC"; phy_test_mode = 0; - puSch10xSnr = 200; - puCch10xSnr = 200; + puSch10xSnr = 160; + puCch10xSnr = 160; } ); @@ -234,7 +234,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_ENABLE"; } diff --git a/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.50PRB.usrpb210.conf b/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.50PRB.usrpb210.conf index ddf01096cd..48c8370b4b 100644 --- a/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.50PRB.usrpb210.conf +++ b/ci-scripts/conf_files/rcc.band7.tm1.if4p5.lo.50PRB.usrpb210.conf @@ -200,8 +200,8 @@ MACRLCs = ( tr_s_preference = "local_L1"; tr_n_preference = "local_RRC"; phy_test_mode = 0; - puSch10xSnr = 200; - puCch10xSnr = 200; + puSch10xSnr = 160; + puCch10xSnr = 160; } ); @@ -234,7 +234,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_ENABLE"; } diff --git a/ci-scripts/conf_files/rru.fdd.band7.conf b/ci-scripts/conf_files/rru.fdd.band7.conf index ff5e13908e..dab69e1f1e 100644 --- a/ci-scripts/conf_files/rru.fdd.band7.conf +++ b/ci-scripts/conf_files/rru.fdd.band7.conf @@ -20,7 +20,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_ENABLE"; } diff --git a/ci-scripts/main.py b/ci-scripts/main.py index 24475f7eb3..5c371a99ee 100644 --- a/ci-scripts/main.py +++ b/ci-scripts/main.py @@ -1504,7 +1504,8 @@ class SSHConnection(): return ENB_PROCESS_ASSERTION if foundRealTimeIssue: logging.debug('\u001B[1;37;41m eNB faced real time issues! \u001B[0m') - return ENB_PROCESS_REALTIME_ISSUE + self.htmleNBFailureMsg += 'eNB faced real time issues!\n' + #return ENB_PROCESS_REALTIME_ISSUE return 0 def TerminateeNB(self): diff --git a/ci-scripts/oai-ci-vm-tool b/ci-scripts/oai-ci-vm-tool index 8d3dbb4ddd..940bcd2852 100755 --- a/ci-scripts/oai-ci-vm-tool +++ b/ci-scripts/oai-ci-vm-tool @@ -252,8 +252,9 @@ case $key in VM_NAME=ci-phy-sim ARCHIVES_LOC=phy_sim LOG_PATTERN=.Rel15.txt - NB_PATTERN_FILES=6 + NB_PATTERN_FILES=8 BUILD_OPTIONS="--phy_simulators" + VM_MEMORY=4096 RUN_OPTIONS="./run_exec_autotests.bash -g \"01510*\" -q -np -b" NBARGS=$[$NBARGS+256] shift @@ -330,8 +331,9 @@ case $key in VM_NAME=ci-phy-sim ARCHIVES_LOC=phy_sim LOG_PATTERN=.Rel15.txt - NB_PATTERN_FILES=6 + NB_PATTERN_FILES=8 BUILD_OPTIONS="--phy_simulators" + VM_MEMORY=4096 RUN_OPTIONS="./run_exec_autotests.bash -g \"01510*\" -q -np -b" NBARGS=$[$NBARGS+256] ;; diff --git a/ci-scripts/reportBuildLocally.sh b/ci-scripts/reportBuildLocally.sh index 2e78e6f6cd..d523c25baa 100755 --- a/ci-scripts/reportBuildLocally.sh +++ b/ci-scripts/reportBuildLocally.sh @@ -411,6 +411,10 @@ function report_build { then summary_table_row "PBCH Test - Release 15" ./archives/phy_sim/nr_pbchsim.Rel15.txt "Built target nr_pbchsim" ./phy_sim_row6.html fi + if [ -f ./archives/phy_sim/nr_dlschsim.Rel15.txt ] + then + summary_table_row "DLSCH Test - Release 15" ./archives/phy_sim/nr_dlschsim.Rel15.txt "Built target nr_dlschsim" ./phy_sim_row7.html + fi summary_table_footer diff --git a/ci-scripts/waitBuildOnVM.sh b/ci-scripts/waitBuildOnVM.sh index a6f4edace8..208bf06318 100755 --- a/ci-scripts/waitBuildOnVM.sh +++ b/ci-scripts/waitBuildOnVM.sh @@ -165,6 +165,10 @@ function check_on_vm_build { fi done - if [ $NB_PATTERN_FILES -ne $NB_FOUND_FILES ]; then STATUS=-1; fi + if [ $NB_PATTERN_FILES -ne $NB_FOUND_FILES ] + then + echo "Expecting $NB_PATTERN_FILES log files and found $NB_FOUND_FILES" + STATUS=-1 + fi } diff --git a/cmake_targets/CMakeLists.txt b/cmake_targets/CMakeLists.txt index 76553ba1fd..c4e2399351 100644 --- a/cmake_targets/CMakeLists.txt +++ b/cmake_targets/CMakeLists.txt @@ -290,6 +290,8 @@ set(protobuf_generated_dir ${OPENAIR_BIN_DIR}) add_list2_option(RRC_ASN1_VERSION "Rel15" "ASN.1 version of RRC interface" "Rel8" "Rel10" "Rel14" "CBA") +Message("RRC_ASN1_VERSION is ${RRC_ASN1_VERSION}") + if (${RRC_ASN1_VERSION} STREQUAL "Rel8") make_version(RRC_VERSION 8 6 0) set (RRC_GRAMMAR ${OPENAIR2_DIR}/RRC/LTE/MESSAGES/asn1c/ASN1_files/EUTRA-RRC-Definitions-86.asn) @@ -366,9 +368,9 @@ execute_process(COMMAND mkdir -p ${NR_RRC_FULL_DIR} OUTPUT_QUIET ERROR_QUIET) -if (NOT ${ret} STREQUAL 0) - message(FATAL_ERROR "${ret}: error") -endif (NOT ${ret} STREQUAL 0) +#if (NOT ${ret} STREQUAL 0) +# message(FATAL_ERROR "${ret}: error") +#endif (NOT ${ret} STREQUAL 0) file(GLOB nr_rrc_source ${NR_RRC_FULL_DIR}/*.c) file(GLOB nr_rrc_h ${NR_RRC_FULL_DIR}/*.h) @@ -961,7 +963,7 @@ set(UTIL_SRC ${OPENAIR_DIR}/common/utils/LOG/log.c # ${OPENAIR2_DIR}/UTIL/LOG/vcd_signal_dumper.c ${OPENAIR2_DIR}/UTIL/MATH/oml.c - ${OPENAIR2_DIR}/UTIL/MEM/mem_block.c +# ${OPENAIR2_DIR}/UTIL/MEM/mem_block.c # ${OPENAIR2_DIR}/UTIL/OCG/OCG.c # ${OPENAIR2_DIR}/UTIL/OCG/OCG_create_dir.c # ${OPENAIR2_DIR}/UTIL/OCG/OCG_detect_file.c @@ -1184,6 +1186,7 @@ set(PHY_SRC_COMMON ${OPENAIR1_DIR}/PHY/CODING/lte_segmentation.c ${OPENAIR1_DIR}/PHY/CODING/nr_segmentation.c ${OPENAIR1_DIR}/PHY/CODING/nr_compute_tbs.c + ${OPENAIR1_DIR}/PHY/CODING/nr_rate_matching.c ${OPENAIR1_DIR}/PHY/CODING/ccoding_byte.c ${OPENAIR1_DIR}/PHY/CODING/ccoding_byte_lte.c ${OPENAIR1_DIR}/PHY/CODING/3gpplte_sse.c @@ -1304,6 +1307,11 @@ set(PHY_SRC_UE ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_pbch.c ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_dci.c ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_dci_tools.c + ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_dlsch.c + ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_dlsch_tools.c + ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_dlsch_coding.c + ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_tbs_tools.c + ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_sch_dmrs.c ${OPENAIR1_DIR}/PHY/NR_REFSIG/nr_gold.c ${OPENAIR1_DIR}/PHY/TOOLS/file_output.c ${OPENAIR1_DIR}/PHY/TOOLS/cadd_vv.c @@ -1333,6 +1341,10 @@ set(PHY_SRC_UE ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/dmrs_nr.c ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/nr_initial_sync.c ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/nr_pbch.c + ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c + ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/nr_dlsch_decoding.c + ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/nr_dlsch_llr_computation.c + ${OPENAIR1_DIR}/PHY/NR_TRANSPORT/nr_tbs_tools.c ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/ ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/nr_prach.c ${OPENAIR1_DIR}/PHY/NR_UE_TRANSPORT/srs_modulation_nr.c @@ -1343,6 +1355,7 @@ set(PHY_SRC_UE ${OPENAIR1_DIR}/PHY/NR_REFSIG/nr_dmrs_rx.c ${OPENAIR1_DIR}/PHY/NR_REFSIG/nr_gold_ue.c ${OPENAIR1_DIR}/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c + ${OPENAIR1_DIR}/PHY/NR_UE_ESTIMATION/nr_adjust_synch_ue.c ${OPENAIR1_DIR}/PHY/LTE_ESTIMATION/lte_ue_measurements.c ${OPENAIR1_DIR}/PHY/TOOLS/file_output.c ${OPENAIR1_DIR}/PHY/TOOLS/cadd_vv.c @@ -1523,14 +1536,6 @@ set(NR_L2_SRC_UE ${NR_UE_RRC_DIR}/L2_interface_ue.c ${NR_UE_RRC_DIR}/main_ue.c ${NR_UE_RRC_DIR}/rrc_UE.c -# ${NR_UE_RRC_DIR}/mac_vars.c -#${RRC_DIR}/rrc_UE.c - ) - -set(L2_NR_SRC_UE - ${NR_UE_RRC_DIR}/L2_interface_ue.c - ${NR_UE_RRC_DIR}/main_ue.c - ${NR_UE_RRC_DIR}/rrc_UE.c ) set (MAC_SRC @@ -1636,26 +1641,27 @@ set ( NR_LTE_UE_REUSE_SRC ${OPENAIR1_DIR}/SCHED/phy_procedures_lte_common.c #${OPENAIR1_DIR}/SCHED_UE/phy_procedures_lte_ue.c ${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/lte_mcs.c - ${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/pmch_common.c + #${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/pmch_common.c - ${OPENAIR1_DIR}/PHY/LTE_REFSIG/lte_dl_cell_spec.c - ${OPENAIR1_DIR}/PHY/CODING/lte_segmentation.c - ${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/prach_common.c + #${OPENAIR1_DIR}/PHY/LTE_REFSIG/lte_dl_cell_spec.c + #${OPENAIR1_DIR}/PHY/CODING/lte_segmentation.c + #${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/prach_common.c # ${OPENAIR1_DIR}/PHY/INIT/lte_init_ue.c ${OPENAIR1_DIR}/PHY/CODING/crc_byte.c ${OPENAIR1_DIR}/PHY/CODING/viterbi_lte.c ${OPENAIR1_DIR}/PHY/CODING/viterbi.c - ${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/phich_common.c + #${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/phich_common.c ${OPENAIR1_DIR}/PHY/LTE_UE_TRANSPORT/dlsch_llr_computation.c - ${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dci_tools_common.c - ${OPENAIR1_DIR}/PHY/CODING/lte_rate_matching.c + +# ${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dci_tools_common.c +# ${OPENAIR1_DIR}/PHY/CODING/lte_rate_matching.c ${OPENAIR1_DIR}/PHY/CODING/ccoding_byte_lte.c ${OPENAIR1_DIR}/PHY/CODING/ccoding_byte.c ${OPENAIR1_DIR}/PHY/LTE_REFSIG/lte_gold.c ${OPENAIR1_DIR}/PHY/LTE_ESTIMATION/lte_est_freq_offset.c ${OPENAIR1_DIR}/PHY/LTE_ESTIMATION/lte_dl_channel_estimation.c ${OPENAIR1_DIR}/PHY/LTE_ESTIMATION/lte_ue_measurements.c - ${OPENAIR1_DIR}/PHY/LTE_UE_TRANSPORT/dlsch_demodulation.c + #${OPENAIR1_DIR}/PHY/LTE_UE_TRANSPORT/dlsch_demodulation.c #${OPENAIR1_DIR}/PHY/LTE_UE_TRANSPORT/dlsch_llr_computation_avx2.c ) @@ -2555,11 +2561,14 @@ target_link_libraries(polartest SIMU PHY PHY_NR PHY_COMMON m ${ATLAS_LIBRARIES}) add_executable(ldpctest ${OPENAIR1_DIR}/PHY/CODING/TESTBENCH/ldpctest.c) target_link_libraries(ldpctest SIMU PHY PHY_NR m ${ATLAS_LIBRARIES}) +add_executable(nr_dlschsim ${OPENAIR1_DIR}/SIMULATION/NR_PHY/dlschsim.c ${T_SOURCE}) +target_link_libraries(nr_dlschsim -Wl,--start-group UTIL SIMU PHY_COMMON PHY_NR PHY_NR_UE SCHED_NR_LIB CONFIG_LIB -Wl,--end-group m pthread ${ATLAS_LIBRARIES} ${T_LIB} dl) + add_executable(nr_pbchsim ${OPENAIR1_DIR}/SIMULATION/NR_PHY/pbchsim.c ${T_SOURCE}) target_link_libraries(nr_pbchsim -Wl,--start-group UTIL SIMU PHY_COMMON PHY_NR PHY_NR_UE SCHED_NR_LIB CONFIG_LIB -Wl,--end-group m pthread ${ATLAS_LIBRARIES} ${T_LIB} dl) -add_executable(nr_dlsim ${OPENAIR1_DIR}/SIMULATION/NR_PHY/dlsim.c ${T_SOURCE}) -target_link_libraries(nr_dlsim -Wl,--start-group UTIL SIMU PHY_COMMON PHY_NR PHY_NR_UE SCHED_NR_LIB SCHED_NR_UE_LIB MAC_NR MAC_UE_NR CONFIG_LIB -Wl,--end-group m pthread ${ATLAS_LIBRARIES} ${T_LIB} dl) +add_executable(nr_dlsim ${OPENAIR1_DIR}/SIMULATION/NR_PHY/dlsim.c ${T_SOURCE}) +target_link_libraries(nr_dlsim -Wl,--start-group UTIL SIMU PHY_COMMON PHY_NR PHY_NR_UE SCHED_NR_LIB SCHED_NR_UE_LIB MAC_NR MAC_UE_NR RRC_LIB NR_RRC_LIB CONFIG_LIB L2_NR -Wl,--end-group m pthread ${ATLAS_LIBRARIES} ${T_LIB} dl) foreach(myExe dlsim dlsim_tm7 ulsim pbchsim scansim mbmssim pdcchsim pucchsim prachsim syncsim) diff --git a/cmake_targets/autotests/test_case_list.xml b/cmake_targets/autotests/test_case_list.xml index 4c4ed12ca5..2f66c6af34 100644 --- a/cmake_targets/autotests/test_case_list.xml +++ b/cmake_targets/autotests/test_case_list.xml @@ -1035,14 +1035,16 @@ execution - polartest Test cases. (Test1: PBCH polar test) + polartest Test cases. (Test1: PBCH polar test), + (Test2: DCI polar test) $OPENAIR_DIR/cmake_targets/build_oai --phy_simulators -c $OPENAIR_DIR/cmake_targets/autotests/tools/free_mem.bash $OPENAIR_DIR/targets/bin/polartest.Rel15 - -q -s-10 -f0 + -q -s-10 -f0 + -q -s-10 -f0 -m1 polartest.test1 BLER= 0.000000 segmentation fault|assertion|exiting|fatal @@ -1051,16 +1053,24 @@ execution - nr_pbchsim Test cases. (Test1: PBCH-only), - (Test2: PBCH and synchronization) + nr_pbchsim Test cases. (Test1: PBCH-only, 106 PRB), + (Test2: PBCH and synchronization, 106PBR), + (Test3: PBCH-only, 217 PRB), + (Test4: PBCH and synchronization, 217 RPB), + (Test5: PBCH-only, 217 PRB), + (Test6: PBCH and synchronization, 217 PRB) $OPENAIR_DIR/cmake_targets/build_oai --phy_simulators -c $OPENAIR_DIR/cmake_targets/autotests/tools/free_mem.bash $OPENAIR_DIR/targets/bin/nr_pbchsim.Rel15 - -s-11 -S-10 -n1000 - -s-11 -S-10 -n10 -I + -s0 -S1 -n1000 -R106 + -s0 -S1 -n10 -I -R106 + -s0 -S1 -n1000 -R217 + -s0 -S1 -n10 -I -R217 + -s0 -S1 -n1000 -R273 + -s0 -S1 -n10 -I -R273 nr_pbchsim.test1 nr_pbchsim.test2 PBCH test OK segmentation fault|assertion|exiting|fatal diff --git a/cmake_targets/build_oai b/cmake_targets/build_oai index 82cdba228f..01ab01f0a7 100755 --- a/cmake_targets/build_oai +++ b/cmake_targets/build_oai @@ -339,6 +339,10 @@ function main() { FORCE_DEADLINE_SCHEDULER_FLAG_USER="True" echo_info "Enabling the usage of deadline scheduler" shift 1;; + --enable-cpu-affinity) + CPU_AFFINITY_FLAG_USER="True" + echo_info "Enabling CPU Affinity (only valid when not using deadline scheduler)" + shift 1;; --disable-cpu-affinity) CPU_AFFINITY_FLAG_USER="False" echo_info "Disabling CPU Affinity (only valid when not using deadline scheduler)" @@ -692,7 +696,7 @@ function main() { echo_info "Compiling unitary tests simulators" # TODO: fix: dlsim_tm4 pucchsim prachsim pdcchsim pbchsim mbmssim #simlist="dlsim_tm4 dlsim ulsim pucchsim prachsim pdcchsim pbchsim mbmssim" - simlist="dlsim ulsim polartest ldpctest nr_pbchsim" + simlist="dlsim ulsim polartest ldpctest nr_pbchsim nr_dlschsim nr_dlsim" for f in $simlist ; do compilations \ phy_simulators $f \ diff --git a/cmake_targets/phy_simulators/CMakeLists.txt b/cmake_targets/phy_simulators/CMakeLists.txt index 3ca13e722e..4347189bc2 100644 --- a/cmake_targets/phy_simulators/CMakeLists.txt +++ b/cmake_targets/phy_simulators/CMakeLists.txt @@ -3,11 +3,11 @@ set(PACKAGE_NAME "unitary_tests_simulators") set(PHYSIM True) set(RF_BOARD None) set(XFORMS True) -set(ENABLE_ITTI False) +set(ENABLE_ITTI True) set(DEBUG_PHY False) set(MU_RECIEVER False) set(NAS_UE False) set(MESSAGE_CHART_GENERATOR False) -set(RRC_ASN1_VERSION "Rel14") +set(RRC_ASN1_VERSION "Rel15") set(T_TRACER True) include(${CMAKE_CURRENT_SOURCE_DIR}/../CMakeLists.txt) diff --git a/common/utils/LOG/vcd_signal_dumper.c b/common/utils/LOG/vcd_signal_dumper.c index cb28be8550..12971c3761 100644 --- a/common/utils/LOG/vcd_signal_dumper.c +++ b/common/utils/LOG/vcd_signal_dumper.c @@ -74,10 +74,10 @@ const char* eurecomVariablesNames[] = { "frame_number_TX1_RU", "frame_number_RX0_RU", "frame_number_RX1_RU", - "subframe_number_TX0_RU", - "subframe_number_TX1_RU", - "subframe_number_RX0_RU", - "subframe_number_RX1_RU", + "tti_number_TX0_RU", + "tti_number_TX1_RU", + "tti_number_RX0_RU", + "tti_number_RX1_RU", "runtime_TX_eNB", "runtime_RX_eNB", "frame_number_TX0_UE", @@ -196,10 +196,10 @@ const char* eurecomVariablesNames[] = { "frame_number_TX1_gNB", "frame_number_RX0_gNB", "frame_number_RX1_gNB", - "subframe_number_TX0_gNB", - "subframe_number_TX1_gNB", - "subframe_number_RX0_gNB", - "subframe_number_RX1_gNB" + "slot_number_TX0_gNB", + "slot_number_TX1_gNB", + "slot_number_RX0_gNB", + "slot_number_RX1_gNB" }; const char* eurecomFunctionsNames[] = { diff --git a/common/utils/LOG/vcd_signal_dumper.h b/common/utils/LOG/vcd_signal_dumper.h index 5c1ba9e7fe..7b439d4dd5 100644 --- a/common/utils/LOG/vcd_signal_dumper.h +++ b/common/utils/LOG/vcd_signal_dumper.h @@ -51,10 +51,10 @@ typedef enum { VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX1_RU, VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX0_RU, VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX1_RU, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX1_RU, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_RU, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX1_RU, + VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, + VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX1_RU, + VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX0_RU, + VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX1_RU, VCD_SIGNAL_DUMPER_VARIABLES_RUNTIME_TX_ENB, VCD_SIGNAL_DUMPER_VARIABLES_RUNTIME_RX_ENB, VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_UE, @@ -173,10 +173,10 @@ typedef enum { VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX1_GNB, VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX0_GNB, VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX1_GNB, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_GNB, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX1_GNB, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_GNB, - VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX1_GNB, + VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_TX0_GNB, + VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_TX1_GNB, + VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_RX0_GNB, + VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_RX1_GNB, VCD_SIGNAL_DUMPER_VARIABLES_END } vcd_signal_dump_variables; diff --git a/common/utils/T/T_messages.txt b/common/utils/T/T_messages.txt index 6057c09f1c..10dc5fb07c 100644 --- a/common/utils/T/T_messages.txt +++ b/common/utils/T/T_messages.txt @@ -1099,26 +1099,26 @@ ID = VCD_VARIABLE_FRAME_NUMBER_RX1_RU GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value VCD_NAME = frame_number_RX1_RU -ID = VCD_VARIABLE_SUBFRAME_NUMBER_TX0_RU - DESC = VCD variable SUBFRAME_NUMBER_TX0_RU +ID = VCD_VARIABLE_TTI_NUMBER_TX0_RU + DESC = VCD variable TTI_NUMBER_TX0_RU GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_TX0_RU -ID = VCD_VARIABLE_SUBFRAME_NUMBER_TX1_RU - DESC = VCD variable SUBFRAME_NUMBER_TX1_RU + VCD_NAME = tti_number_TX0_RU +ID = VCD_VARIABLE_TTI_NUMBER_TX1_RU + DESC = VCD variable TTI_NUMBER_TX1_RU GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_TX1_RU -ID = VCD_VARIABLE_SUBFRAME_NUMBER_RX0_RU - DESC = VCD variable SUBFRAME_NUMBER_RX0_RU + VCD_NAME = tti_number_TX1_RU +ID = VCD_VARIABLE_TTI_NUMBER_RX0_RU + DESC = VCD variable TTI_NUMBER_RX0_RU GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_RX0_RU -ID = VCD_VARIABLE_SUBFRAME_NUMBER_RX1_RU - DESC = VCD variable SUBFRAME_NUMBER_RX1_RU + VCD_NAME = tti_number_RX0_RU +ID = VCD_VARIABLE_TTI_NUMBER_RX1_RU + DESC = VCD variable TTI_NUMBER_RX1_RU GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_RX1_RU + VCD_NAME = tti_number_RX1_RU ID = VCD_VARIABLE_RUNTIME_TX_ENB DESC = VCD variable RUNTIME_TX_ENB GROUP = ALL:VCD:ENB:VCD_VARIABLE @@ -1701,26 +1701,26 @@ ID = VCD_VARIABLE_FRAME_NUMBER_RX1_GNB GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value VCD_NAME = frame_number_RX1_gNB -ID = VCD_VARIABLE_SUBFRAME_NUMBER_TX0_GNB - DESC = VCD variable SUBFRAME_NUMBER_TX0_GNB +ID = VCD_VARIABLE_SLOT_NUMBER_TX0_GNB + DESC = VCD variable SLOT_NUMBER_TX0_GNB GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_TX0_gNB -ID = VCD_VARIABLE_SUBFRAME_NUMBER_TX1_GNB - DESC = VCD variable SUBFRAME_NUMBER_TX1_GNB + VCD_NAME = slot_number_TX0_gNB +ID = VCD_VARIABLE_SLOT_NUMBER_TX1_GNB + DESC = VCD variable SLOT_NUMBER_TX1_GNB GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_TX1_gNB -ID = VCD_VARIABLE_SUBFRAME_NUMBER_RX0_GNB - DESC = VCD variable SUBFRAME_NUMBER_RX0_GNB + VCD_NAME = slot_number_TX1_gNB +ID = VCD_VARIABLE_SLOT_NUMBER_RX0_GNB + DESC = VCD variable SLOT_NUMBER_RX0_GNB GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_RX0_gNB -ID = VCD_VARIABLE_SUBFRAME_NUMBER_RX1_GNB - DESC = VCD variable SUBFRAME_NUMBER_RX1_GNB + VCD_NAME = slot_number_RX0_gNB +ID = VCD_VARIABLE_SLOT_NUMBER_RX1_GNB + DESC = VCD variable SLOT_NUMBER_RX1_GNB GROUP = ALL:VCD:ENB:VCD_VARIABLE FORMAT = ulong,value - VCD_NAME = subframe_number_RX1_gNB + VCD_NAME = slot_number_RX1_gNB #functions diff --git a/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h b/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h index 1396faced8..838e52dada 100644 --- a/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h +++ b/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h @@ -24,119 +24,119 @@ #include "PHY/impl_defs_nr.h" /* -typedef unsigned int uint32_t; -typedef unsigned short uint16_t; -typedef unsigned char uint8_t; -typedef signed int int32_t; -typedef signed short int16_t; -typedef signed char int8_t; + typedef unsigned int uint32_t; + typedef unsigned short uint16_t; + typedef unsigned char uint8_t; + typedef signed int int32_t; + typedef signed short int16_t; + typedef signed char int8_t; */ typedef struct { - uint8_t identifier_dci_formats ; // 0 IDENTIFIER_DCI_FORMATS: - uint8_t carrier_ind ; // 1 CARRIER_IND: 0 or 3 bits, as defined in Subclause x.x of [5, TS38.213] - uint8_t sul_ind_0_1 ; // 2 SUL_IND_0_1: - uint8_t slot_format_ind ; // 3 SLOT_FORMAT_IND: size of DCI format 2_0 is configurable by higher layers up to 128 bits, according to Subclause 11.1.1 of [5, TS 38.213] - uint8_t pre_emption_ind ; // 4 PRE_EMPTION_IND: size of DCI format 2_1 is configurable by higher layers up to 126 bits, according to Subclause 11.2 of [5, TS 38.213]. Each pre-emption indication is 14 bits - uint8_t block_number ; // 5 BLOCK_NUMBER: starting position of a block is determined by the parameter startingBitOfFormat2_3 - uint8_t close_loop_ind ; // 6 CLOSE_LOOP_IND: - uint8_t bandwidth_part_ind ; // 7 BANDWIDTH_PART_IND: - uint8_t short_message_ind ; // 8 SHORT_MESSAGE_IND: - uint8_t short_messages ; // 9 SHORT_MESSAGES: - uint16_t freq_dom_resource_assignment_UL; // 10 FREQ_DOM_RESOURCE_ASSIGNMENT_UL: PUSCH hopping with resource allocation type 1 not considered - // (NOTE 1) If DCI format 0_0 is monitored in common search space - // and if the number of information bits in the DCI format 0_0 prior to padding - // is larger than the payload size of the DCI format 1_0 monitored in common search space - // the bitwidth of the frequency domain resource allocation field in the DCI format 0_0 - // is reduced such that the size of DCI format 0_0 equals to the size of the DCI format 1_0 - uint16_t freq_dom_resource_assignment_DL; // 11 FREQ_DOM_RESOURCE_ASSIGNMENT_DL: - uint8_t time_dom_resource_assignment ; // 12 TIME_DOM_RESOURCE_ASSIGNMENT: 0, 1, 2, 3, or 4 bits as defined in Subclause 6.1.2.1 of [6, TS 38.214]. The bitwidth for this field is determined as log2(I) bits, - // where I the number of entries in the higher layer parameter pusch-AllocationList - uint8_t vrb_to_prb_mapping ; // 13 VRB_TO_PRB_MAPPING: 0 bit if only resource allocation type 0 - uint8_t prb_bundling_size_ind ; // 14 PRB_BUNDLING_SIZE_IND:0 bit if the higher layer parameter PRB_bundling is not configured or is set to 'static', or 1 bit if the higher layer parameter PRB_bundling is set to 'dynamic' according to Subclause 5.1.2.3 of [6, TS 38.214] - uint8_t rate_matching_ind ; // 15 RATE_MATCHING_IND: 0, 1, or 2 bits according to higher layer parameter rate-match-PDSCH-resource-set - uint8_t zp_csi_rs_trigger ; // 16 ZP_CSI_RS_TRIGGER: - uint8_t freq_hopping_flag ; // 17 FREQ_HOPPING_FLAG: 0 bit if only resource allocation type 0 - uint8_t tb1_mcs ; // 18 TB1_MCS: - uint8_t tb1_ndi ; // 19 TB1_NDI: - uint8_t tb1_rv ; // 20 TB1_RV: - uint8_t tb2_mcs ; // 21 TB2_MCS: - uint8_t tb2_ndi ; // 22 TB2_NDI: - uint8_t tb2_rv ; // 23 TB2_RV: - uint8_t mcs ; // 24 MCS: - uint8_t ndi ; // 25 NDI: - uint8_t rv ; // 26 RV: - uint8_t harq_process_number ; // 27 HARQ_PROCESS_NUMBER: - uint8_t dai ; // 28 DAI: For format1_1: 4 if more than one serving cell are configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 MSB bits are the counter DAI and the 2 LSB bits are the total DAI - // 2 if one serving cell is configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 bits are the counter DAI - // 0 otherwise - uint8_t first_dai ; // 29 FIRST_DAI: (1 or 2 bits) 1 bit for semi-static HARQ-ACK - uint8_t second_dai ; // 30 SECOND_DAI: (0 or 2 bits) 2 bits for dynamic HARQ-ACK codebook with two HARQ-ACK sub-codebooks - uint8_t tb_scaling ; // 31 TB_SCALING: - uint8_t tpc_pusch ; // 32 TPC_PUSCH: - uint8_t tpc_pucch ; // 33 TPC_PUCCH: - uint8_t pucch_resource_ind ; // 34 PUCCH_RESOURCE_IND: - uint8_t pdsch_to_harq_feedback_time_ind ; // 35 PDSCH_TO_HARQ_FEEDBACK_TIME_IND: - uint8_t srs_resource_ind ; // 36 SRS_RESOURCE_IND: - uint8_t precod_nbr_layers ; // 37 PRECOD_NBR_LAYERS: - uint8_t antenna_ports ; // 38 ANTENNA_PORTS: - uint8_t tci ; // 39 TCI: 0 bit if higher layer parameter tci-PresentInDCI is not enabled; otherwise 3 bits - uint8_t srs_request ; // 40 SRS_REQUEST: - uint8_t tpc_cmd ; // 41 TPC_CMD: - uint8_t csi_request ; // 42 CSI_REQUEST: - uint8_t cbgti ; // 43 CBGTI: 0, 2, 4, 6, or 8 bits determined by higher layer parameter maxCodeBlockGroupsPerTransportBlock for the PDSCH - uint8_t cbgfi ; // 44 CBGFI: 0 or 1 bit determined by higher layer parameter codeBlockGroupFlushIndicator - uint8_t ptrs_dmrs ; // 45 PTRS_DMRS: - uint8_t beta_offset_ind ; // 46 BETA_OFFSET_IND: - uint8_t dmrs_seq_ini ; // 47 DMRS_SEQ_INI: 1 bit if the cell has two ULs and the number of bits for DCI format 1_0 before padding - // is larger than the number of bits for DCI format 0_0 before padding; 0 bit otherwise - uint8_t ul_sch_ind ; // 48 UL_SCH_IND: value of "1" indicates UL-SCH shall be transmitted on the PUSCH and a value of "0" indicates UL-SCH shall not be transmitted on the PUSCH - uint16_t padding_nr_dci ; // 49 PADDING_NR_DCI: (Note 2) If DCI format 0_0 is monitored in common search space - // and if the number of information bits in the DCI format 0_0 prior to padding - // is less than the payload size of the DCI format 1_0 monitored in common search space - // zeros shall be appended to the DCI format 0_0 - // until the payload size equals that of the DCI format 1_0 - uint8_t sul_ind_0_0 ; // 50 SUL_IND_0_0: - uint8_t ra_preamble_index ; // 51 RA_PREAMBLE_INDEX: - uint8_t sul_ind_1_0 ; // 52 SUL_IND_1_0: - uint8_t ss_pbch_index ; // 53 SS_PBCH_INDEX - uint8_t prach_mask_index ; // 54 PRACH_MASK_INDEX - uint8_t reserved_nr_dci ; // 55 RESERVED_NR_DCI + uint8_t identifier_dci_formats ; // 0 IDENTIFIER_DCI_FORMATS: + uint8_t carrier_ind ; // 1 CARRIER_IND: 0 or 3 bits, as defined in Subclause x.x of [5, TS38.213] + uint8_t sul_ind_0_1 ; // 2 SUL_IND_0_1: + uint8_t slot_format_ind ; // 3 SLOT_FORMAT_IND: size of DCI format 2_0 is configurable by higher layers up to 128 bits, according to Subclause 11.1.1 of [5, TS 38.213] + uint8_t pre_emption_ind ; // 4 PRE_EMPTION_IND: size of DCI format 2_1 is configurable by higher layers up to 126 bits, according to Subclause 11.2 of [5, TS 38.213]. Each pre-emption indication is 14 bits + uint8_t block_number ; // 5 BLOCK_NUMBER: starting position of a block is determined by the parameter startingBitOfFormat2_3 + uint8_t close_loop_ind ; // 6 CLOSE_LOOP_IND: + uint8_t bandwidth_part_ind ; // 7 BANDWIDTH_PART_IND: + uint8_t short_message_ind ; // 8 SHORT_MESSAGE_IND: + uint8_t short_messages ; // 9 SHORT_MESSAGES: + uint16_t freq_dom_resource_assignment_UL; // 10 FREQ_DOM_RESOURCE_ASSIGNMENT_UL: PUSCH hopping with resource allocation type 1 not considered + // (NOTE 1) If DCI format 0_0 is monitored in common search space + // and if the number of information bits in the DCI format 0_0 prior to padding + // is larger than the payload size of the DCI format 1_0 monitored in common search space + // the bitwidth of the frequency domain resource allocation field in the DCI format 0_0 + // is reduced such that the size of DCI format 0_0 equals to the size of the DCI format 1_0 + uint16_t freq_dom_resource_assignment_DL; // 11 FREQ_DOM_RESOURCE_ASSIGNMENT_DL: + uint8_t time_dom_resource_assignment ; // 12 TIME_DOM_RESOURCE_ASSIGNMENT: 0, 1, 2, 3, or 4 bits as defined in Subclause 6.1.2.1 of [6, TS 38.214]. The bitwidth for this field is determined as log2(I) bits, + // where I the number of entries in the higher layer parameter pusch-AllocationList + uint8_t vrb_to_prb_mapping ; // 13 VRB_TO_PRB_MAPPING: 0 bit if only resource allocation type 0 + uint8_t prb_bundling_size_ind ; // 14 PRB_BUNDLING_SIZE_IND:0 bit if the higher layer parameter PRB_bundling is not configured or is set to 'static', or 1 bit if the higher layer parameter PRB_bundling is set to 'dynamic' according to Subclause 5.1.2.3 of [6, TS 38.214] + uint8_t rate_matching_ind ; // 15 RATE_MATCHING_IND: 0, 1, or 2 bits according to higher layer parameter rate-match-PDSCH-resource-set + uint8_t zp_csi_rs_trigger ; // 16 ZP_CSI_RS_TRIGGER: + uint8_t freq_hopping_flag ; // 17 FREQ_HOPPING_FLAG: 0 bit if only resource allocation type 0 + uint8_t tb1_mcs ; // 18 TB1_MCS: + uint8_t tb1_ndi ; // 19 TB1_NDI: + uint8_t tb1_rv ; // 20 TB1_RV: + uint8_t tb2_mcs ; // 21 TB2_MCS: + uint8_t tb2_ndi ; // 22 TB2_NDI: + uint8_t tb2_rv ; // 23 TB2_RV: + uint8_t mcs ; // 24 MCS: + uint8_t ndi ; // 25 NDI: + uint8_t rv ; // 26 RV: + uint8_t harq_process_number ; // 27 HARQ_PROCESS_NUMBER: + uint8_t dai ; // 28 DAI: For format1_1: 4 if more than one serving cell are configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 MSB bits are the counter DAI and the 2 LSB bits are the total DAI + // 2 if one serving cell is configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 bits are the counter DAI + // 0 otherwise + uint8_t first_dai ; // 29 FIRST_DAI: (1 or 2 bits) 1 bit for semi-static HARQ-ACK + uint8_t second_dai ; // 30 SECOND_DAI: (0 or 2 bits) 2 bits for dynamic HARQ-ACK codebook with two HARQ-ACK sub-codebooks + uint8_t tb_scaling ; // 31 TB_SCALING: + uint8_t tpc_pusch ; // 32 TPC_PUSCH: + uint8_t tpc_pucch ; // 33 TPC_PUCCH: + uint8_t pucch_resource_ind ; // 34 PUCCH_RESOURCE_IND: + uint8_t pdsch_to_harq_feedback_time_ind ; // 35 PDSCH_TO_HARQ_FEEDBACK_TIME_IND: + uint8_t srs_resource_ind ; // 36 SRS_RESOURCE_IND: + uint8_t precod_nbr_layers ; // 37 PRECOD_NBR_LAYERS: + uint8_t antenna_ports ; // 38 ANTENNA_PORTS: + uint8_t tci ; // 39 TCI: 0 bit if higher layer parameter tci-PresentInDCI is not enabled; otherwise 3 bits + uint8_t srs_request ; // 40 SRS_REQUEST: + uint8_t tpc_cmd ; // 41 TPC_CMD: + uint8_t csi_request ; // 42 CSI_REQUEST: + uint8_t cbgti ; // 43 CBGTI: 0, 2, 4, 6, or 8 bits determined by higher layer parameter maxCodeBlockGroupsPerTransportBlock for the PDSCH + uint8_t cbgfi ; // 44 CBGFI: 0 or 1 bit determined by higher layer parameter codeBlockGroupFlushIndicator + uint8_t ptrs_dmrs ; // 45 PTRS_DMRS: + uint8_t beta_offset_ind ; // 46 BETA_OFFSET_IND: + uint8_t dmrs_seq_ini ; // 47 DMRS_SEQ_INI: 1 bit if the cell has two ULs and the number of bits for DCI format 1_0 before padding + // is larger than the number of bits for DCI format 0_0 before padding; 0 bit otherwise + uint8_t ul_sch_ind ; // 48 UL_SCH_IND: value of "1" indicates UL-SCH shall be transmitted on the PUSCH and a value of "0" indicates UL-SCH shall not be transmitted on the PUSCH + uint16_t padding_nr_dci ; // 49 PADDING_NR_DCI: (Note 2) If DCI format 0_0 is monitored in common search space + // and if the number of information bits in the DCI format 0_0 prior to padding + // is less than the payload size of the DCI format 1_0 monitored in common search space + // zeros shall be appended to the DCI format 0_0 + // until the payload size equals that of the DCI format 1_0 + uint8_t sul_ind_0_0 ; // 50 SUL_IND_0_0: + uint8_t ra_preamble_index ; // 51 RA_PREAMBLE_INDEX: + uint8_t sul_ind_1_0 ; // 52 SUL_IND_1_0: + uint8_t ss_pbch_index ; // 53 SS_PBCH_INDEX + uint8_t prach_mask_index ; // 54 PRACH_MASK_INDEX + uint8_t reserved_nr_dci ; // 55 RESERVED_NR_DCI } fapi_nr_dci_pdu_rel15_t; typedef struct { - uint8_t uci_format; - uint8_t uci_channel; - uint8_t harq_ack_bits; - uint32_t harq_ack; - uint8_t csi_bits; - uint32_t csi; - uint8_t sr_bits; - uint32_t sr; + uint8_t uci_format; + uint8_t uci_channel; + uint8_t harq_ack_bits; + uint32_t harq_ack; + uint8_t csi_bits; + uint32_t csi; + uint8_t sr_bits; + uint32_t sr; } fapi_nr_uci_pdu_rel15_t; - typedef struct { - /// frequency_domain_resource; - //uint32_t rb_start; - //uint32_t rb_end; - uint64_t frequency_domain_resource; - uint16_t rb_offset; +typedef struct { + /// frequency_domain_resource; + //uint32_t rb_start; + //uint32_t rb_end; + uint64_t frequency_domain_resource; + uint16_t rb_offset; - uint8_t duration; - uint8_t cce_reg_mapping_type; // interleaved or noninterleaved - uint8_t cce_reg_interleaved_reg_bundle_size; // valid if CCE to REG mapping type is interleaved type - uint8_t cce_reg_interleaved_interleaver_size; // valid if CCE to REG mapping type is interleaved type - uint8_t cce_reg_interleaved_shift_index; // valid if CCE to REG mapping type is interleaved type - uint8_t precoder_granularity; - uint16_t pdcch_dmrs_scrambling_id; + uint8_t duration; + uint8_t cce_reg_mapping_type; // interleaved or noninterleaved + uint8_t cce_reg_interleaved_reg_bundle_size; // valid if CCE to REG mapping type is interleaved type + uint8_t cce_reg_interleaved_interleaver_size; // valid if CCE to REG mapping type is interleaved type + uint8_t cce_reg_interleaved_shift_index; // valid if CCE to REG mapping type is interleaved type + uint8_t precoder_granularity; + uint16_t pdcch_dmrs_scrambling_id; - uint8_t tci_state_pdcch; - uint8_t tci_present_in_dci; - } fapi_nr_coreset_t; + uint8_t tci_state_pdcch; + uint8_t tci_present_in_dci; +} fapi_nr_coreset_t; // // Top level FAPI messages @@ -148,79 +148,80 @@ typedef struct { // P7 // - typedef struct { - uint16_t rnti; - uint8_t dci_format; - // n_CCE index of first CCE for PDCCH reception - int n_CCE; - // N_CCE is L, or number of CCEs for DCI - int N_CCE; - fapi_nr_dci_pdu_rel15_t dci; - } fapi_nr_dci_indication_pdu_t; +typedef struct { + uint16_t rnti; + uint8_t dci_format; + // n_CCE index of first CCE for PDCCH reception + int n_CCE; + // N_CCE is L, or number of CCEs for DCI + int N_CCE; + fapi_nr_dci_pdu_rel15_t dci; +} fapi_nr_dci_indication_pdu_t; /// typedef struct { - uint32_t sfn_slot; - uint16_t number_of_dcis; - fapi_nr_dci_indication_pdu_t dci_list[10]; + uint32_t sfn_slot; + uint16_t number_of_dcis; + fapi_nr_dci_indication_pdu_t dci_list[10]; } fapi_nr_dci_indication_t; - typedef struct { - uint32_t pdu_length; - uint8_t* pdu; - } fapi_nr_pdsch_pdu_t; +typedef struct { + uint32_t pdu_length; + uint8_t* pdu; +} fapi_nr_pdsch_pdu_t; - typedef struct { - uint8_t* pdu; // 3bytes - uint8_t additional_bits; - uint8_t ssb_index; - uint8_t ssb_length; - uint16_t cell_id; +typedef struct { + uint8_t* pdu; // 3bytes + uint8_t additional_bits; + uint8_t ssb_index; + uint8_t ssb_length; + uint16_t cell_id; - } fapi_nr_mib_pdu_t; +} fapi_nr_mib_pdu_t; - typedef struct { - uint32_t pdu_length; - uint8_t* pdu; - uint32_t sibs_mask; - } fapi_nr_sib_pdu_t; +typedef struct { + uint32_t pdu_length; + uint8_t* pdu; + uint32_t sibs_mask; +} fapi_nr_sib_pdu_t; - typedef struct { - uint8_t pdu_type; - union { - fapi_nr_pdsch_pdu_t pdsch_pdu; - fapi_nr_mib_pdu_t mib_pdu; - fapi_nr_sib_pdu_t sib_pdu; - }; - } fapi_nr_rx_indication_body_t; +typedef struct { + uint8_t pdu_type; + union { + fapi_nr_pdsch_pdu_t pdsch_pdu; + fapi_nr_mib_pdu_t mib_pdu; + fapi_nr_sib_pdu_t sib_pdu; + }; +} fapi_nr_rx_indication_body_t; /// +#define NFAPI_RX_IND_MAX_PDU 100 typedef struct { - uint32_t sfn_slot; - uint16_t number_pdus; - fapi_nr_rx_indication_body_t *rx_indication_body; + uint32_t sfn_slot; + uint16_t number_pdus; + fapi_nr_rx_indication_body_t rx_indication_body[NFAPI_RX_IND_MAX_PDU]; } fapi_nr_rx_indication_t; - typedef struct { - uint8_t ul_cqi; - uint16_t timing_advance; - uint16_t rnti; - } fapi_nr_tx_config_t; +typedef struct { + uint8_t ul_cqi; + uint16_t timing_advance; + uint16_t rnti; +} fapi_nr_tx_config_t; - typedef struct { - uint16_t pdu_length; - uint16_t pdu_index; - uint8_t* pdu; - } fapi_nr_tx_request_body_t; +typedef struct { + uint16_t pdu_length; + uint16_t pdu_index; + uint8_t* pdu; +} fapi_nr_tx_request_body_t; /// typedef struct { - uint32_t sfn_slot; - fapi_nr_tx_config_t tx_config; - uint16_t number_of_pdus; - fapi_nr_tx_request_body_t *tx_request_body; + uint32_t sfn_slot; + fapi_nr_tx_config_t tx_config; + uint16_t number_of_pdus; + fapi_nr_tx_request_body_t *tx_request_body; } fapi_nr_tx_request_t; typedef struct { @@ -240,7 +241,7 @@ typedef struct { uint16_t prach_freq_offset; } fapi_nr_ul_config_prach_pdu; - typedef struct { +typedef struct { pucch_format_nr_t format; /* format 0 1 2 3 4 */ uint8_t initialCyclicShift; /* x x */ @@ -353,104 +354,105 @@ typedef struct { uint8_t beta_offset_ind; } fapi_nr_ul_config_pusch_pdu_rel15_t; - typedef struct { - uint16_t rnti; - fapi_nr_ul_config_pusch_pdu_rel15_t ulsch_pdu_rel15; - } fapi_nr_ul_config_pusch_pdu; - - typedef struct { - - } fapi_nr_ul_config_srs_pdu; - - typedef struct { - uint8_t pdu_type; - union { - fapi_nr_ul_config_prach_pdu prach_config_pdu; - fapi_nr_ul_config_pucch_pdu pucch_config_pdu; - fapi_nr_ul_config_pusch_pdu ulsch_config_pdu; - fapi_nr_ul_config_srs_pdu srs_config_pdu; - }; - } fapi_nr_ul_config_request_pdu_t; +typedef struct { + uint16_t rnti; + fapi_nr_ul_config_pusch_pdu_rel15_t ulsch_pdu_rel15; +} fapi_nr_ul_config_pusch_pdu; typedef struct { - uint32_t sfn_slot; - uint8_t number_pdus; - fapi_nr_ul_config_request_pdu_t ul_config_list[FAPI_NR_UL_CONFIG_LIST_NUM]; + +} fapi_nr_ul_config_srs_pdu; + +typedef struct { + uint8_t pdu_type; + union { + fapi_nr_ul_config_prach_pdu prach_config_pdu; + fapi_nr_ul_config_pucch_pdu pucch_config_pdu; + fapi_nr_ul_config_pusch_pdu ulsch_config_pdu; + fapi_nr_ul_config_srs_pdu srs_config_pdu; + }; +} fapi_nr_ul_config_request_pdu_t; + +typedef struct { + uint32_t sfn_slot; + uint8_t number_pdus; + fapi_nr_ul_config_request_pdu_t ul_config_list[FAPI_NR_UL_CONFIG_LIST_NUM]; } fapi_nr_ul_config_request_t; - typedef struct { - uint16_t rnti; - - fapi_nr_coreset_t coreset; - uint32_t duration; - uint8_t number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 - uint16_t monitoring_symbols_within_slot; - // DCI foramt-specific - uint8_t format_2_0_number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 - uint8_t format_2_3_monitorying_periodicity; - uint8_t format_2_3_number_of_candidates; - } fapi_nr_dl_config_dci_dl_pdu_rel15_t; - - typedef struct { - fapi_nr_dl_config_dci_dl_pdu_rel15_t dci_config_rel15; - } fapi_nr_dl_config_dci_pdu; - typedef struct{ - uint8_t aperiodicSRS_ResourceTrigger; - } fapi_nr_dl_srs_config_t; - - typedef enum{vrb_to_prb_mapping_non_interleaved = 0, vrb_to_prb_mapping_interleaved = 1} vrb_to_prb_mapping_t; - //typedef fapi_nr_dci_pdu_rel15_t fapi_nr_dl_config_dlsch_pdu_rel15_t; - typedef struct { - uint8_t bandwidth_part_ind; - uint16_t number_rbs; - uint16_t start_rb; - uint8_t frame_offset; - uint16_t number_symbols; - uint16_t start_symbol; - uint8_t prb_bundling_size_ind; - uint8_t rate_matching_ind; - uint8_t zp_csi_rs_trigger; - uint8_t mcs; - uint8_t ndi; - uint8_t rv; - uint8_t tb2_mcs; - uint8_t tb2_ndi; - uint8_t tb2_rv; - uint8_t harq_process_nbr; - vrb_to_prb_mapping_t vrb_to_prb_mapping; - uint8_t dai; - double scaling_factor_S; - int8_t accumulated_delta_PUCCH; - uint8_t pucch_resource_id; - uint8_t pdsch_to_harq_feedback_time_ind; - uint8_t n_dmrs_cdm_groups; - uint8_t dmrs_ports[10]; - uint8_t n_front_load_symb; - uint8_t tci_state; - fapi_nr_dl_srs_config_t srs_config; - uint8_t cbgti; - uint8_t codeBlockGroupFlushIndicator; - // to be check the fields needed to L1 with NR_DL_UE_HARQ_t and NR_UE_DLSCH_t - } fapi_nr_dl_config_dlsch_pdu_rel15_t; - - typedef struct { - uint16_t rnti; - fapi_nr_dl_config_dlsch_pdu_rel15_t dlsch_config_rel15; - } fapi_nr_dl_config_dlsch_pdu; - - typedef struct { - uint8_t pdu_type; - union { - fapi_nr_dl_config_dci_pdu dci_config_pdu; - fapi_nr_dl_config_dlsch_pdu dlsch_config_pdu; - }; - } fapi_nr_dl_config_request_pdu_t; +typedef struct { + uint16_t rnti; + + fapi_nr_coreset_t coreset; + uint16_t N_RB_BWP; + uint32_t duration; + uint8_t number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 + uint16_t monitoring_symbols_within_slot; + // DCI foramt-specific + uint8_t format_2_0_number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 + uint8_t format_2_3_monitorying_periodicity; + uint8_t format_2_3_number_of_candidates; +} fapi_nr_dl_config_dci_dl_pdu_rel15_t; typedef struct { - uint32_t sfn_slot; - uint8_t number_pdus; - fapi_nr_dl_config_request_pdu_t dl_config_list[FAPI_NR_DL_CONFIG_LIST_NUM]; + fapi_nr_dl_config_dci_dl_pdu_rel15_t dci_config_rel15; +} fapi_nr_dl_config_dci_pdu; +typedef struct{ + uint8_t aperiodicSRS_ResourceTrigger; +} fapi_nr_dl_srs_config_t; + +typedef enum{vrb_to_prb_mapping_non_interleaved = 0, vrb_to_prb_mapping_interleaved = 1} vrb_to_prb_mapping_t; +//typedef fapi_nr_dci_pdu_rel15_t fapi_nr_dl_config_dlsch_pdu_rel15_t; +typedef struct { + uint8_t bandwidth_part_ind; + uint16_t number_rbs; + uint16_t start_rb; + uint8_t frame_offset; + uint16_t number_symbols; + uint16_t start_symbol; + uint8_t prb_bundling_size_ind; + uint8_t rate_matching_ind; + uint8_t zp_csi_rs_trigger; + uint8_t mcs; + uint8_t ndi; + uint8_t rv; + uint8_t tb2_mcs; + uint8_t tb2_ndi; + uint8_t tb2_rv; + uint8_t harq_process_nbr; + vrb_to_prb_mapping_t vrb_to_prb_mapping; + uint8_t dai; + double scaling_factor_S; + int8_t accumulated_delta_PUCCH; + uint8_t pucch_resource_id; + uint8_t pdsch_to_harq_feedback_time_ind; + uint8_t n_dmrs_cdm_groups; + uint8_t dmrs_ports[10]; + uint8_t n_front_load_symb; + uint8_t tci_state; + fapi_nr_dl_srs_config_t srs_config; + uint8_t cbgti; + uint8_t codeBlockGroupFlushIndicator; + // to be check the fields needed to L1 with NR_DL_UE_HARQ_t and NR_UE_DLSCH_t +} fapi_nr_dl_config_dlsch_pdu_rel15_t; + +typedef struct { + uint16_t rnti; + fapi_nr_dl_config_dlsch_pdu_rel15_t dlsch_config_rel15; +} fapi_nr_dl_config_dlsch_pdu; + +typedef struct { + uint8_t pdu_type; + union { + fapi_nr_dl_config_dci_pdu dci_config_pdu; + fapi_nr_dl_config_dlsch_pdu dlsch_config_pdu; + }; +} fapi_nr_dl_config_request_pdu_t; + +typedef struct { + uint32_t sfn_slot; + uint8_t number_pdus; + fapi_nr_dl_config_request_pdu_t dl_config_list[FAPI_NR_DL_CONFIG_LIST_NUM]; } fapi_nr_dl_config_request_t; @@ -460,444 +462,444 @@ typedef struct { - typedef struct { - fapi_nr_coreset_t coreset; +typedef struct { + fapi_nr_coreset_t coreset; - uint8_t monitoring_slot_peridicity; - uint8_t monitoring_slot_offset; - uint16_t duration; - uint16_t monitoring_symbols_within_slot; - uint8_t number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 + uint8_t monitoring_slot_peridicity; + uint8_t monitoring_slot_offset; + uint16_t duration; + uint16_t monitoring_symbols_within_slot; + uint8_t number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 - uint8_t dci_2_0_number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 - uint8_t dci_2_3_monitorying_periodicity; - uint8_t dci_2_3_number_of_candidates; + uint8_t dci_2_0_number_of_candidates[5]; // aggregation level 1, 2, 4, 8, 16 + uint8_t dci_2_3_monitorying_periodicity; + uint8_t dci_2_3_number_of_candidates; - } fapi_nr_search_space_t; +} fapi_nr_search_space_t; - typedef struct { - fapi_nr_search_space_t search_space_sib1; - fapi_nr_search_space_t search_space_others_sib; - fapi_nr_search_space_t search_space_paging; - //fapi_nr_coreset_t coreset_ra; // common coreset - fapi_nr_search_space_t search_space_ra; - } fapi_nr_pdcch_config_common_t; +typedef struct { + fapi_nr_search_space_t search_space_sib1; + fapi_nr_search_space_t search_space_others_sib; + fapi_nr_search_space_t search_space_paging; + //fapi_nr_coreset_t coreset_ra; // common coreset + fapi_nr_search_space_t search_space_ra; +} fapi_nr_pdcch_config_common_t; - typedef struct { - uint8_t k0; - uint8_t mapping_type; - uint8_t symbol_starting; - uint8_t symbol_length; - } fapi_nr_pdsch_time_domain_resource_allocation_t; +typedef struct { + uint8_t k0; + uint8_t mapping_type; + uint8_t symbol_starting; + uint8_t symbol_length; +} fapi_nr_pdsch_time_domain_resource_allocation_t; - typedef struct { - fapi_nr_pdsch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_DL_ALLOCATIONS]; - } fapi_nr_pdsch_config_common_t; +typedef struct { + fapi_nr_pdsch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_DL_ALLOCATIONS]; +} fapi_nr_pdsch_config_common_t; - typedef struct { - uint8_t prach_configuration_index; - uint8_t msg1_fdm; - uint8_t msg1_frequency_start; - uint8_t zero_correlation_zone_config; - uint8_t preamble_received_target_power; - uint8_t preamble_transmission_max; - uint8_t power_ramping_step; - uint8_t ra_window_size; +typedef struct { + uint8_t prach_configuration_index; + uint8_t msg1_fdm; + uint8_t msg1_frequency_start; + uint8_t zero_correlation_zone_config; + uint8_t preamble_received_target_power; + uint8_t preamble_transmission_max; + uint8_t power_ramping_step; + uint8_t ra_window_size; - uint8_t total_number_of_preamble; - uint8_t ssb_occasion_per_rach; - uint8_t cb_preamble_per_ssb; + uint8_t total_number_of_preamble; + uint8_t ssb_occasion_per_rach; + uint8_t cb_preamble_per_ssb; - uint8_t group_a_msg3_size; - uint8_t group_a_number_of_preamble; - uint8_t group_b_power_offset; - uint8_t contention_resolution_timer; - uint8_t rsrp_threshold_ssb; - uint8_t rsrp_threshold_ssb_sul; - uint8_t prach_length; // l839, l139 - uint8_t prach_root_sequence_index; // 0 - 837 for l839, 0 - 137 for l139 - uint8_t msg1_subcarrier_spacing; - uint8_t restrictedset_config; - uint8_t msg3_transform_precoding; - } fapi_nr_rach_config_common_t; + uint8_t group_a_msg3_size; + uint8_t group_a_number_of_preamble; + uint8_t group_b_power_offset; + uint8_t contention_resolution_timer; + uint8_t rsrp_threshold_ssb; + uint8_t rsrp_threshold_ssb_sul; + uint8_t prach_length; // l839, l139 + uint8_t prach_root_sequence_index; // 0 - 837 for l839, 0 - 137 for l139 + uint8_t msg1_subcarrier_spacing; + uint8_t restrictedset_config; + uint8_t msg3_transform_precoding; +} fapi_nr_rach_config_common_t; - typedef struct { - uint8_t k2; - uint8_t mapping_type; - uint8_t symbol_starting; - uint8_t symbol_length; - } fapi_nr_pusch_time_domain_resource_allocation_t; +typedef struct { + uint8_t k2; + uint8_t mapping_type; + uint8_t symbol_starting; + uint8_t symbol_length; +} fapi_nr_pusch_time_domain_resource_allocation_t; - typedef struct { - uint8_t group_hopping_enabled_transform_precoding; - fapi_nr_pusch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_UL_ALLOCATIONS]; - uint8_t msg3_delta_preamble; - uint8_t p0_nominal_with_grant; - } fapi_nr_pusch_config_common_t; +typedef struct { + uint8_t group_hopping_enabled_transform_precoding; + fapi_nr_pusch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_UL_ALLOCATIONS]; + uint8_t msg3_delta_preamble; + uint8_t p0_nominal_with_grant; +} fapi_nr_pusch_config_common_t; - typedef struct { - uint8_t pucch_resource_common; - uint8_t pucch_group_hopping; - uint8_t hopping_id; - uint8_t p0_nominal; - } fapi_nr_pucch_config_common_t; +typedef struct { + uint8_t pucch_resource_common; + uint8_t pucch_group_hopping; + uint8_t hopping_id; + uint8_t p0_nominal; +} fapi_nr_pucch_config_common_t; - typedef struct { +typedef struct { - uint8_t subcarrier_spacing_common; - uint8_t ssb_subcarrier_offset; - uint8_t dmrs_type_a_position; - uint8_t pdcch_config_sib1; - uint8_t cell_barred; - uint8_t intra_frequency_reselection; + uint8_t subcarrier_spacing_common; + uint8_t ssb_subcarrier_offset; + uint8_t dmrs_type_a_position; + uint8_t pdcch_config_sib1; + uint8_t cell_barred; + uint8_t intra_frequency_reselection; - uint16_t system_frame_number; - uint8_t ssb_index; - uint8_t half_frame_bit; - } fapi_nr_pbch_config_t; + uint16_t system_frame_number; + uint8_t ssb_index; + uint8_t half_frame_bit; +} fapi_nr_pbch_config_t; - typedef struct { +typedef struct { - fapi_nr_pdcch_config_common_t pdcch_config_common; - fapi_nr_pdsch_config_common_t pdsch_config_common; + fapi_nr_pdcch_config_common_t pdcch_config_common; + fapi_nr_pdsch_config_common_t pdsch_config_common; - } fapi_nr_dl_bwp_common_config_t; +} fapi_nr_dl_bwp_common_config_t; - typedef struct { - uint16_t int_rnti; - uint8_t time_frequency_set; - uint8_t dci_payload_size; - uint8_t serving_cell_id[FAPI_NR_MAX_NUM_SERVING_CELLS]; // interrupt configuration per serving cell - uint8_t position_in_dci[FAPI_NR_MAX_NUM_SERVING_CELLS]; // interrupt configuration per serving cell - } fapi_nr_downlink_preemption_t; +typedef struct { + uint16_t int_rnti; + uint8_t time_frequency_set; + uint8_t dci_payload_size; + uint8_t serving_cell_id[FAPI_NR_MAX_NUM_SERVING_CELLS]; // interrupt configuration per serving cell + uint8_t position_in_dci[FAPI_NR_MAX_NUM_SERVING_CELLS]; // interrupt configuration per serving cell +} fapi_nr_downlink_preemption_t; - typedef struct { - uint8_t tpc_index; - uint8_t tpc_index_sul; - uint8_t target_cell; - } fapi_nr_pusch_tpc_command_config_t; +typedef struct { + uint8_t tpc_index; + uint8_t tpc_index_sul; + uint8_t target_cell; +} fapi_nr_pusch_tpc_command_config_t; - typedef struct { - uint8_t tpc_index_pcell; - uint8_t tpc_index_pucch_scell; - } fapi_nr_pucch_tpc_command_config_t; +typedef struct { + uint8_t tpc_index_pcell; + uint8_t tpc_index_pucch_scell; +} fapi_nr_pucch_tpc_command_config_t; - typedef struct { - uint8_t starting_bit_of_format_2_3; - uint8_t feild_type_format_2_3; - } fapi_nr_srs_tpc_command_config_t; +typedef struct { + uint8_t starting_bit_of_format_2_3; + uint8_t feild_type_format_2_3; +} fapi_nr_srs_tpc_command_config_t; - typedef struct { - fapi_nr_downlink_preemption_t downlink_preemption; - fapi_nr_pusch_tpc_command_config_t tpc_pusch; - fapi_nr_pucch_tpc_command_config_t tpc_pucch; - fapi_nr_srs_tpc_command_config_t tpc_srs; - } fapi_nr_pdcch_config_dedicated_t; +typedef struct { + fapi_nr_downlink_preemption_t downlink_preemption; + fapi_nr_pusch_tpc_command_config_t tpc_pusch; + fapi_nr_pucch_tpc_command_config_t tpc_pucch; + fapi_nr_srs_tpc_command_config_t tpc_srs; +} fapi_nr_pdcch_config_dedicated_t; - typedef struct { - uint8_t dmrs_type; - uint8_t dmrs_addition_position; - uint8_t max_length; - uint16_t scrambling_id0; - uint16_t scrambling_id1; - uint8_t ptrs_frequency_density[2]; // phase tracking rs - uint8_t ptrs_time_density[3]; // phase tracking rs - uint8_t ptrs_epre_ratio; // phase tracking rs - uint8_t ptrs_resource_element_offset; // phase tracking rs - } fapi_nr_dmrs_downlink_config_t; +typedef struct { + uint8_t dmrs_type; + uint8_t dmrs_addition_position; + uint8_t max_length; + uint16_t scrambling_id0; + uint16_t scrambling_id1; + uint8_t ptrs_frequency_density[2]; // phase tracking rs + uint8_t ptrs_time_density[3]; // phase tracking rs + uint8_t ptrs_epre_ratio; // phase tracking rs + uint8_t ptrs_resource_element_offset; // phase tracking rs +} fapi_nr_dmrs_downlink_config_t; - typedef struct { - uint8_t bwp_or_cell_level; - uint8_t pattern_type; - uint32_t resource_blocks[9]; // bitmaps type 275 bits - uint8_t slot_type; // bitmaps type one/two slot(s) - uint32_t symbols_in_resouece_block; // bitmaps type 14/28 bits - uint8_t periodic; // bitmaps type - uint32_t pattern[2]; // bitmaps type 2/4/5/8/10/20/40 bits +typedef struct { + uint8_t bwp_or_cell_level; + uint8_t pattern_type; + uint32_t resource_blocks[9]; // bitmaps type 275 bits + uint8_t slot_type; // bitmaps type one/two slot(s) + uint32_t symbols_in_resouece_block; // bitmaps type 14/28 bits + uint8_t periodic; // bitmaps type + uint32_t pattern[2]; // bitmaps type 2/4/5/8/10/20/40 bits - fapi_nr_coreset_t coreset; // coreset + fapi_nr_coreset_t coreset; // coreset - uint8_t subcarrier_spacing; - uint8_t mode; - } fapi_nr_rate_matching_pattern_group_t; + uint8_t subcarrier_spacing; + uint8_t mode; +} fapi_nr_rate_matching_pattern_group_t; - typedef struct { - // resource mapping - uint8_t row; // row1/row2/row4/other - uint16_t frequency_domain_allocation; // 4/12/3/6 bits - uint8_t number_of_ports; - uint8_t first_ofdm_symbol_in_time_domain; - uint8_t first_ofdm_symbol_in_time_domain2; - uint8_t cdm_type; - uint8_t density; // .5/1/3 - uint8_t density_dot5_type; // even/odd PRBs +typedef struct { + // resource mapping + uint8_t row; // row1/row2/row4/other + uint16_t frequency_domain_allocation; // 4/12/3/6 bits + uint8_t number_of_ports; + uint8_t first_ofdm_symbol_in_time_domain; + uint8_t first_ofdm_symbol_in_time_domain2; + uint8_t cdm_type; + uint8_t density; // .5/1/3 + uint8_t density_dot5_type; // even/odd PRBs - uint8_t frequency_band_starting_rb; // freqBand - uint8_t frequency_band_number_of_rb; // freqBand + uint8_t frequency_band_starting_rb; // freqBand + uint8_t frequency_band_number_of_rb; // freqBand - // periodicityAndOffset - uint8_t periodicity; // slot4/5/8/10/16/20/32/40/64/80/160/320/640 - uint32_t offset; // 0..639 bits - } fapi_nr_zp_csi_rs_resource_t; + // periodicityAndOffset + uint8_t periodicity; // slot4/5/8/10/16/20/32/40/64/80/160/320/640 + uint32_t offset; // 0..639 bits +} fapi_nr_zp_csi_rs_resource_t; - typedef struct { - uint16_t data_scrambling_id_pdsch; - fapi_nr_dmrs_downlink_config_t dmrs_dl_for_pdsch_mapping_type_a; - fapi_nr_dmrs_downlink_config_t dmrs_dl_for_pdsch_mapping_type_b; - uint8_t vrb_to_prb_interleaver; - uint8_t resource_allocation; - fapi_nr_pdsch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_DL_ALLOCATIONS]; - uint8_t pdsch_aggregation_factor; - fapi_nr_rate_matching_pattern_group_t rate_matching_pattern_group1; - fapi_nr_rate_matching_pattern_group_t rate_matching_pattern_group2; - uint8_t rbg_size; - uint8_t mcs_table; - uint8_t max_num_of_code_word_scheduled_by_dci; - uint8_t bundle_size; // prb_bundling static - uint8_t bundle_size_set1; // prb_bundling dynamic - uint8_t bundle_size_set2; // prb_bundling dynamic - fapi_nr_zp_csi_rs_resource_t periodically_zp_csi_rs_resource_set[FAPI_NR_MAX_NUM_ZP_CSI_RS_RESOURCE_PER_SET]; - } fapi_nr_pdsch_config_dedicated_t; +typedef struct { + uint16_t data_scrambling_id_pdsch; + fapi_nr_dmrs_downlink_config_t dmrs_dl_for_pdsch_mapping_type_a; + fapi_nr_dmrs_downlink_config_t dmrs_dl_for_pdsch_mapping_type_b; + uint8_t vrb_to_prb_interleaver; + uint8_t resource_allocation; + fapi_nr_pdsch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_DL_ALLOCATIONS]; + uint8_t pdsch_aggregation_factor; + fapi_nr_rate_matching_pattern_group_t rate_matching_pattern_group1; + fapi_nr_rate_matching_pattern_group_t rate_matching_pattern_group2; + uint8_t rbg_size; + uint8_t mcs_table; + uint8_t max_num_of_code_word_scheduled_by_dci; + uint8_t bundle_size; // prb_bundling static + uint8_t bundle_size_set1; // prb_bundling dynamic + uint8_t bundle_size_set2; // prb_bundling dynamic + fapi_nr_zp_csi_rs_resource_t periodically_zp_csi_rs_resource_set[FAPI_NR_MAX_NUM_ZP_CSI_RS_RESOURCE_PER_SET]; +} fapi_nr_pdsch_config_dedicated_t; - typedef struct { - uint16_t starting_prb; - uint8_t intra_slot_frequency_hopping; - uint16_t second_hop_prb; - uint8_t format; // pucch format 0..4 - uint8_t initial_cyclic_shift; - uint8_t number_of_symbols; - uint8_t starting_symbol_index; - uint8_t time_domain_occ; - uint8_t number_of_prbs; - uint8_t occ_length; - uint8_t occ_index; - } fapi_nr_pucch_resource_t; +typedef struct { + uint16_t starting_prb; + uint8_t intra_slot_frequency_hopping; + uint16_t second_hop_prb; + uint8_t format; // pucch format 0..4 + uint8_t initial_cyclic_shift; + uint8_t number_of_symbols; + uint8_t starting_symbol_index; + uint8_t time_domain_occ; + uint8_t number_of_prbs; + uint8_t occ_length; + uint8_t occ_index; +} fapi_nr_pucch_resource_t; - typedef struct { - uint8_t periodicity; - uint8_t number_of_harq_process; - fapi_nr_pucch_resource_t n1_pucch_an; - } fapi_nr_sps_config_t; +typedef struct { + uint8_t periodicity; + uint8_t number_of_harq_process; + fapi_nr_pucch_resource_t n1_pucch_an; +} fapi_nr_sps_config_t; - typedef struct { - uint8_t beam_failure_instance_max_count; - uint8_t beam_failure_detection_timer; - } fapi_nr_radio_link_monitoring_config_t; +typedef struct { + uint8_t beam_failure_instance_max_count; + uint8_t beam_failure_detection_timer; +} fapi_nr_radio_link_monitoring_config_t; - typedef struct { - fapi_nr_pdcch_config_dedicated_t pdcch_config_dedicated; - fapi_nr_pdsch_config_dedicated_t pdsch_config_dedicated; - fapi_nr_sps_config_t sps_config; - fapi_nr_radio_link_monitoring_config_t radio_link_monitoring_config; +typedef struct { + fapi_nr_pdcch_config_dedicated_t pdcch_config_dedicated; + fapi_nr_pdsch_config_dedicated_t pdsch_config_dedicated; + fapi_nr_sps_config_t sps_config; + fapi_nr_radio_link_monitoring_config_t radio_link_monitoring_config; - } fapi_nr_dl_bwp_dedicated_config_t; +} fapi_nr_dl_bwp_dedicated_config_t; - typedef struct { - fapi_nr_rach_config_common_t rach_config_common; - fapi_nr_pusch_config_common_t pusch_config_common; - fapi_nr_pucch_config_common_t pucch_config_common; +typedef struct { + fapi_nr_rach_config_common_t rach_config_common; + fapi_nr_pusch_config_common_t pusch_config_common; + fapi_nr_pucch_config_common_t pucch_config_common; - } fapi_nr_ul_bwp_common_config_t; +} fapi_nr_ul_bwp_common_config_t; - typedef struct { - uint8_t inter_slot_frequency_hopping; - uint8_t additional_dmrs; - uint8_t max_code_rate; - uint8_t number_of_slots; - uint8_t pi2bpsk; - uint8_t simultaneous_harq_ack_csi; - } fapi_nr_pucch_format_config_t; +typedef struct { + uint8_t inter_slot_frequency_hopping; + uint8_t additional_dmrs; + uint8_t max_code_rate; + uint8_t number_of_slots; + uint8_t pi2bpsk; + uint8_t simultaneous_harq_ack_csi; +} fapi_nr_pucch_format_config_t; - typedef struct { - fapi_nr_pucch_format_config_t format1; - fapi_nr_pucch_format_config_t format2; - fapi_nr_pucch_format_config_t format3; - fapi_nr_pucch_format_config_t format4; - fapi_nr_pucch_resource_t multi_csi_pucch_resources[2]; - uint8_t dl_data_to_ul_ack[8]; - // pucch power control - uint8_t deltaF_pucch_f0; - uint8_t deltaF_pucch_f1; - uint8_t deltaF_pucch_f2; - uint8_t deltaF_pucch_f3; - uint8_t deltaF_pucch_f4; - uint8_t two_pucch_pc_adjusment_states; - } fapi_nr_pucch_config_dedicated_t; +typedef struct { + fapi_nr_pucch_format_config_t format1; + fapi_nr_pucch_format_config_t format2; + fapi_nr_pucch_format_config_t format3; + fapi_nr_pucch_format_config_t format4; + fapi_nr_pucch_resource_t multi_csi_pucch_resources[2]; + uint8_t dl_data_to_ul_ack[8]; + // pucch power control + uint8_t deltaF_pucch_f0; + uint8_t deltaF_pucch_f1; + uint8_t deltaF_pucch_f2; + uint8_t deltaF_pucch_f3; + uint8_t deltaF_pucch_f4; + uint8_t two_pucch_pc_adjusment_states; +} fapi_nr_pucch_config_dedicated_t; - typedef struct { - uint8_t dmrs_type; - uint8_t dmrs_addition_position; - uint8_t ptrs_uplink_config; // to indicate if PTRS Uplink is configured of not - uint8_t ptrs_type; //cp-OFDM, dft-S-OFDM - uint16_t ptrs_frequency_density[2]; - uint8_t ptrs_time_density[3]; - uint8_t ptrs_max_number_of_ports; - uint8_t ptrs_resource_element_offset; - uint8_t ptrs_power; - uint16_t ptrs_sample_density[5]; - uint8_t ptrs_time_density_transform_precoding; +typedef struct { + uint8_t dmrs_type; + uint8_t dmrs_addition_position; + uint8_t ptrs_uplink_config; // to indicate if PTRS Uplink is configured of not + uint8_t ptrs_type; //cp-OFDM, dft-S-OFDM + uint16_t ptrs_frequency_density[2]; + uint8_t ptrs_time_density[3]; + uint8_t ptrs_max_number_of_ports; + uint8_t ptrs_resource_element_offset; + uint8_t ptrs_power; + uint16_t ptrs_sample_density[5]; + uint8_t ptrs_time_density_transform_precoding; - uint8_t max_length; - uint16_t scrambling_id0; - uint16_t scrambling_id1; - uint8_t npusch_identity; - uint8_t disable_sequence_group_hopping; - uint8_t sequence_hopping_enable; - } fapi_nr_dmrs_uplink_config_t; + uint8_t max_length; + uint16_t scrambling_id0; + uint16_t scrambling_id1; + uint8_t npusch_identity; + uint8_t disable_sequence_group_hopping; + uint8_t sequence_hopping_enable; +} fapi_nr_dmrs_uplink_config_t; - typedef struct { - uint8_t tpc_accmulation; - uint8_t msg3_alpha; - uint8_t p0_nominal_with_grant; - uint8_t two_pusch_pc_adjustments_states; - uint8_t delta_mcs; - } fapi_nr_pusch_power_control_t; +typedef struct { + uint8_t tpc_accmulation; + uint8_t msg3_alpha; + uint8_t p0_nominal_with_grant; + uint8_t two_pusch_pc_adjustments_states; + uint8_t delta_mcs; +} fapi_nr_pusch_power_control_t; - typedef enum {tx_config_codebook = 1, tx_config_nonCodebook = 2} tx_config_t; - typedef enum {transform_precoder_disabled = 0, transform_precoder_enabled = 1} transform_precoder_t; - typedef enum { - codebook_subset_fullyAndPartialAndNonCoherent = 1, - codebook_subset_partialAndNonCoherent = 2, - codebook_subset_nonCoherent = 3} codebook_subset_t; - typedef struct { - uint16_t data_scrambling_identity; - tx_config_t tx_config; - fapi_nr_dmrs_uplink_config_t dmrs_ul_for_pusch_mapping_type_a; - fapi_nr_dmrs_uplink_config_t dmrs_ul_for_pusch_mapping_type_b; - fapi_nr_pusch_power_control_t pusch_power_control; - uint8_t frequency_hopping; - uint16_t frequency_hopping_offset_lists[4]; - uint8_t resource_allocation; - fapi_nr_pusch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_UL_ALLOCATIONS]; - uint8_t pusch_aggregation_factor; - uint8_t mcs_table; - uint8_t mcs_table_transform_precoder; - transform_precoder_t transform_precoder; - codebook_subset_t codebook_subset; - uint8_t max_rank; - uint8_t rbg_size; +typedef enum {tx_config_codebook = 1, tx_config_nonCodebook = 2} tx_config_t; +typedef enum {transform_precoder_disabled = 0, transform_precoder_enabled = 1} transform_precoder_t; +typedef enum { + codebook_subset_fullyAndPartialAndNonCoherent = 1, + codebook_subset_partialAndNonCoherent = 2, + codebook_subset_nonCoherent = 3} codebook_subset_t; +typedef struct { + uint16_t data_scrambling_identity; + tx_config_t tx_config; + fapi_nr_dmrs_uplink_config_t dmrs_ul_for_pusch_mapping_type_a; + fapi_nr_dmrs_uplink_config_t dmrs_ul_for_pusch_mapping_type_b; + fapi_nr_pusch_power_control_t pusch_power_control; + uint8_t frequency_hopping; + uint16_t frequency_hopping_offset_lists[4]; + uint8_t resource_allocation; + fapi_nr_pusch_time_domain_resource_allocation_t allocation_list[FAPI_NR_MAX_NUM_UL_ALLOCATIONS]; + uint8_t pusch_aggregation_factor; + uint8_t mcs_table; + uint8_t mcs_table_transform_precoder; + transform_precoder_t transform_precoder; + codebook_subset_t codebook_subset; + uint8_t max_rank; + uint8_t rbg_size; - //uci-OnPUSCH - uint8_t uci_on_pusch_type; //dynamic, semi-static - uint8_t beta_offset_ack_index1[4]; - uint8_t beta_offset_ack_index2[4]; - uint8_t beta_offset_ack_index3[4]; - uint8_t beta_offset_csi_part1_index1[4]; - uint8_t beta_offset_csi_part1_index2[4]; - uint8_t beta_offset_csi_part2_index1[4]; - uint8_t beta_offset_csi_part2_index2[4]; + //uci-OnPUSCH + uint8_t uci_on_pusch_type; //dynamic, semi-static + uint8_t beta_offset_ack_index1[4]; + uint8_t beta_offset_ack_index2[4]; + uint8_t beta_offset_ack_index3[4]; + uint8_t beta_offset_csi_part1_index1[4]; + uint8_t beta_offset_csi_part1_index2[4]; + uint8_t beta_offset_csi_part2_index1[4]; + uint8_t beta_offset_csi_part2_index2[4]; - uint8_t tp_pi2BPSK; - } fapi_nr_pusch_config_dedicated_t; + uint8_t tp_pi2BPSK; +} fapi_nr_pusch_config_dedicated_t; - typedef struct { - uint8_t frequency_hopping; - fapi_nr_dmrs_uplink_config_t cg_dmrs_configuration; - uint8_t mcs_table; - uint8_t mcs_table_transform_precoder; +typedef struct { + uint8_t frequency_hopping; + fapi_nr_dmrs_uplink_config_t cg_dmrs_configuration; + uint8_t mcs_table; + uint8_t mcs_table_transform_precoder; - //uci-OnPUSCH - uint8_t uci_on_pusch_type; //dynamic, semi-static - uint8_t beta_offset_ack_index1[4]; - uint8_t beta_offset_ack_index2[4]; - uint8_t beta_offset_ack_index3[4]; - uint8_t beta_offset_csi_part1_index1[4]; - uint8_t beta_offset_csi_part1_index2[4]; - uint8_t beta_offset_csi_part2_index1[4]; - uint8_t beta_offset_csi_part2_index2[4]; + //uci-OnPUSCH + uint8_t uci_on_pusch_type; //dynamic, semi-static + uint8_t beta_offset_ack_index1[4]; + uint8_t beta_offset_ack_index2[4]; + uint8_t beta_offset_ack_index3[4]; + uint8_t beta_offset_csi_part1_index1[4]; + uint8_t beta_offset_csi_part1_index2[4]; + uint8_t beta_offset_csi_part2_index1[4]; + uint8_t beta_offset_csi_part2_index2[4]; - uint8_t resource_allocation; - // rgb-Size structure missing in spec. - uint8_t power_control_loop_to_use; - // p0-PUSCH-Alpha - uint8_t p0; - uint8_t alpha; + uint8_t resource_allocation; + // rgb-Size structure missing in spec. + uint8_t power_control_loop_to_use; + // p0-PUSCH-Alpha + uint8_t p0; + uint8_t alpha; - uint8_t transform_precoder; - uint8_t number_of_harq_process; - uint8_t rep_k; - uint8_t rep_k_rv; - uint8_t periodicity; - uint8_t configured_grant_timer; - // rrc-ConfiguredUplinkGrant - uint16_t time_domain_offset; - uint8_t time_domain_allocation; - uint32_t frequency_domain_allocation; - uint8_t antenna_ports; - uint8_t dmrs_seq_initialization; - uint8_t precoding_and_number_of_layers; - uint8_t srs_resource_indicator; - uint8_t mcs_and_tbs; - uint8_t frequency_hopping_offset; - uint8_t path_loss_reference_index; + uint8_t transform_precoder; + uint8_t number_of_harq_process; + uint8_t rep_k; + uint8_t rep_k_rv; + uint8_t periodicity; + uint8_t configured_grant_timer; + // rrc-ConfiguredUplinkGrant + uint16_t time_domain_offset; + uint8_t time_domain_allocation; + uint32_t frequency_domain_allocation; + uint8_t antenna_ports; + uint8_t dmrs_seq_initialization; + uint8_t precoding_and_number_of_layers; + uint8_t srs_resource_indicator; + uint8_t mcs_and_tbs; + uint8_t frequency_hopping_offset; + uint8_t path_loss_reference_index; - } fapi_nr_configured_grant_config_t; +} fapi_nr_configured_grant_config_t; - typedef struct { - uint8_t qcl_type1_serving_cell_index; - uint8_t qcl_type1_bwp_id; - uint8_t qcl_type1_rs_type; // csi-rs or ssb - uint8_t qcl_type1_nzp_csi_rs_resource_id; - uint8_t qcl_type1_ssb_index; - uint8_t qcl_type1_type; +typedef struct { + uint8_t qcl_type1_serving_cell_index; + uint8_t qcl_type1_bwp_id; + uint8_t qcl_type1_rs_type; // csi-rs or ssb + uint8_t qcl_type1_nzp_csi_rs_resource_id; + uint8_t qcl_type1_ssb_index; + uint8_t qcl_type1_type; - uint8_t qcl_type2_serving_cell_index; - uint8_t qcl_type2_bwp_id; - uint8_t qcl_type2_rs_type; // csi-rs or ssb - uint8_t qcl_type2_nzp_csi_rs_resource_id; - uint8_t qcl_type2_ssb_index; - uint8_t qcl_type2_type; + uint8_t qcl_type2_serving_cell_index; + uint8_t qcl_type2_bwp_id; + uint8_t qcl_type2_rs_type; // csi-rs or ssb + uint8_t qcl_type2_nzp_csi_rs_resource_id; + uint8_t qcl_type2_ssb_index; + uint8_t qcl_type2_type; - } fapi_nr_tci_state_t; +} fapi_nr_tci_state_t; - typedef struct { - uint8_t root_sequence_index; - // rach genertic - uint8_t prach_configuration_index; - uint8_t msg1_fdm; - uint8_t msg1_frequency_start; - uint8_t zero_correlation_zone_config; - uint8_t preamble_received_target_power; - uint8_t preamble_transmission_max; - uint8_t power_ramping_step; - uint8_t ra_window_size; +typedef struct { + uint8_t root_sequence_index; + // rach genertic + uint8_t prach_configuration_index; + uint8_t msg1_fdm; + uint8_t msg1_frequency_start; + uint8_t zero_correlation_zone_config; + uint8_t preamble_received_target_power; + uint8_t preamble_transmission_max; + uint8_t power_ramping_step; + uint8_t ra_window_size; - uint8_t rsrp_threshold_ssb; - // PRACH-ResourceDedicatedBFR - uint8_t bfr_ssb_index[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint8_t bfr_ssb_ra_preamble_index[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - // NZP-CSI-RS-Resource - uint8_t bfr_csi_rs_nzp_resource_mapping[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint8_t bfr_csi_rs_power_control_offset[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint8_t bfr_csi_rs_power_control_offset_ss[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint16_t bfr_csi_rs_scrambling_id[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint8_t bfr_csi_rs_resource_periodicity[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint16_t bfr_csi_rs_resource_offset[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - fapi_nr_tci_state_t qcl_infomation_periodic_csi_rs[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint8_t rsrp_threshold_ssb; + // PRACH-ResourceDedicatedBFR + uint8_t bfr_ssb_index[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint8_t bfr_ssb_ra_preamble_index[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + // NZP-CSI-RS-Resource + uint8_t bfr_csi_rs_nzp_resource_mapping[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint8_t bfr_csi_rs_power_control_offset[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint8_t bfr_csi_rs_power_control_offset_ss[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint16_t bfr_csi_rs_scrambling_id[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint8_t bfr_csi_rs_resource_periodicity[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint16_t bfr_csi_rs_resource_offset[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + fapi_nr_tci_state_t qcl_infomation_periodic_csi_rs[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint8_t bfr_csirs_ra_occasions[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; - uint8_t bfr_csirs_ra_preamble_index[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS][FAPI_NR_MAX_RA_OCCASION_PER_CSIRS]; + uint8_t bfr_csirs_ra_occasions[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS]; + uint8_t bfr_csirs_ra_preamble_index[FAPI_NR_MAX_NUM_CANDIDATE_BEAMS][FAPI_NR_MAX_RA_OCCASION_PER_CSIRS]; - uint8_t ssb_per_rach_occasion; - uint8_t ra_ssb_occasion_mask_index; - fapi_nr_search_space_t recovery_search_space; - // RA-Prioritization - uint8_t power_ramping_step_high_priority; - uint8_t scaling_factor_bi; - uint8_t beam_failure_recovery_timer; - } fapi_nr_beam_failure_recovery_config_t; + uint8_t ssb_per_rach_occasion; + uint8_t ra_ssb_occasion_mask_index; + fapi_nr_search_space_t recovery_search_space; + // RA-Prioritization + uint8_t power_ramping_step_high_priority; + uint8_t scaling_factor_bi; + uint8_t beam_failure_recovery_timer; +} fapi_nr_beam_failure_recovery_config_t; - typedef struct { - fapi_nr_pucch_config_dedicated_t pucch_config_dedicated; - fapi_nr_pusch_config_dedicated_t pusch_config_dedicated; - fapi_nr_configured_grant_config_t configured_grant_config; - // SRS-Config - uint8_t srs_tpc_accumulation; - fapi_nr_beam_failure_recovery_config_t beam_failure_recovery_config; +typedef struct { + fapi_nr_pucch_config_dedicated_t pucch_config_dedicated; + fapi_nr_pusch_config_dedicated_t pusch_config_dedicated; + fapi_nr_configured_grant_config_t configured_grant_config; + // SRS-Config + uint8_t srs_tpc_accumulation; + fapi_nr_beam_failure_recovery_config_t beam_failure_recovery_config; - } fapi_nr_ul_bwp_dedicated_config_t; +} fapi_nr_ul_bwp_dedicated_config_t; #define FAPI_NR_CONFIG_REQUEST_MASK_PBCH 0x01 #define FAPI_NR_CONFIG_REQUEST_MASK_DL_BWP_COMMON 0x02 @@ -906,15 +908,15 @@ typedef struct { #define FAPI_NR_CONFIG_REQUEST_MASK_UL_BWP_DEDICATED 0x10 typedef struct { - uint32_t config_mask; + uint32_t config_mask; - fapi_nr_pbch_config_t pbch_config; // MIB + fapi_nr_pbch_config_t pbch_config; // MIB - fapi_nr_dl_bwp_common_config_t dl_bwp_common; - fapi_nr_dl_bwp_dedicated_config_t dl_bwp_dedicated; + fapi_nr_dl_bwp_common_config_t dl_bwp_common; + fapi_nr_dl_bwp_dedicated_config_t dl_bwp_dedicated; - fapi_nr_ul_bwp_common_config_t ul_bwp_common; - fapi_nr_ul_bwp_dedicated_config_t ul_bwp_dedicated; + fapi_nr_ul_bwp_common_config_t ul_bwp_common; + fapi_nr_ul_bwp_dedicated_config_t ul_bwp_dedicated; } fapi_nr_config_request_t; diff --git a/nfapi/open-nFAPI/nfapi/public_inc/nfapi_nr_interface.h b/nfapi/open-nFAPI/nfapi/public_inc/nfapi_nr_interface.h index 23bd09dcc9..4273186dcf 100644 --- a/nfapi/open-nFAPI/nfapi/public_inc/nfapi_nr_interface.h +++ b/nfapi/open-nFAPI/nfapi/public_inc/nfapi_nr_interface.h @@ -72,7 +72,7 @@ typedef struct // These TLVs are used by the VNF to configure the RF in the PNF // nfapi_uint16_tlv_t max_transmit_power; - nfapi_uint16_tlv_t earfcn; + nfapi_uint16_tlv_t nrarfcn; // nfapi_nmm_frequency_bands_t nmm_gsm_frequency_bands; // nfapi_nmm_frequency_bands_t nmm_umts_frequency_bands; @@ -95,7 +95,7 @@ typedef struct #define NFAPI_NR_NFAPI_TIMING_INFO_MODE_TAG 0x511F #define NFAPI_NR_NFAPI_TIMING_INFO_PERIOD_TAG 0x5120 #define NFAPI_NR_NFAPI_MAXIMUM_TRANSMIT_POWER_TAG 0x5128 -#define NFAPI_NR_NFAPI_EARFCN_TAG 0x5129 +#define NFAPI_NR_NFAPI_NRARFCN_TAG 0x5129 #define NFAPI_NR_NFAPI_NMM_GSM_FREQUENCY_BANDS_TAG 0x5130 #define NFAPI_NR_NFAPI_NMM_UMTS_FREQUENCY_BANDS_TAG 0x5131 #define NFAPI_NR_NFAPI_NMM_LTE_FREQUENCY_BANDS_TAG 0x5132 @@ -149,8 +149,6 @@ typedef struct { #define NFAPI_NR_SUBFRAME_CONFIG_NUMEROLOGY_INDEX_MU_TAG 0x5006 typedef struct { - nfapi_uint16_tlv_t tx_antenna_ports; - nfapi_uint16_tlv_t rx_antenna_ports; nfapi_uint16_tlv_t dl_carrier_bandwidth; nfapi_uint16_tlv_t ul_carrier_bandwidth; nfapi_uint16_tlv_t dl_bwp_subcarrierspacing; @@ -161,18 +159,17 @@ typedef struct { nfapi_uint16_tlv_t ul_absolutefrequencypointA; nfapi_uint16_tlv_t dl_offsettocarrier; nfapi_uint16_tlv_t ul_offsettocarrier; - nfapi_uint16_tlv_t dl_scs_subcarrierspacing; - nfapi_uint16_tlv_t ul_scs_subcarrierspacing; - nfapi_uint16_tlv_t dl_scs_specificcarrier_k0; - nfapi_uint16_tlv_t ul_scs_specificcarrier_k0; + nfapi_uint16_tlv_t dl_subcarrierspacing; + nfapi_uint16_tlv_t ul_subcarrierspacing; + nfapi_uint16_tlv_t dl_specificcarrier_k0; + nfapi_uint16_tlv_t ul_specificcarrier_k0; nfapi_uint16_tlv_t NIA_subcarrierspacing; } nfapi_nr_rf_config_t; -#define NFAPI_NR_RF_CONFIG_DL_CHANNEL_BANDWIDTH_TAG 0x500A -#define NFAPI_NR_RF_CONFIG_UL_CHANNEL_BANDWIDTH_TAG 0x500B -#define NFAPI_NR_RF_CONFIG_REFERENCE_SIGNAL_POWER_TAG 0x500C -#define NFAPI_NR_RF_CONFIG_TX_ANTENNA_PORTS_TAG 0x500D -#define NFAPI_NR_RF_CONFIG_RX_ANTENNA_PORTS_TAG 0x500E +#define NFAPI_NR_RF_CONFIG_DL_CARRIER_BANDWIDTH_TAG 0x500A +#define NFAPI_NR_RF_CONFIG_UL_CARRIER_BANDWIDTH_TAG 0x500B +#define NFAPI_NR_RF_CONFIG_DL_SUBCARRIERSPACING_TAG 0x500C +#define NFAPI_NR_RF_CONFIG_UL_SUBCARRIERSPACING_TAG 0x500D typedef struct { nfapi_uint16_tlv_t physical_cell_id; @@ -194,6 +191,27 @@ typedef struct { #define NFAPI_NR_SCH_CONFIG_SS_PBCH_BLOCK_POWER 0x5023 #define NFAPI_NR_SCH_CONFIG_N_SSB_CRB 0x5024 +typedef struct { + nfapi_uint16_tlv_t data_scrambling_id; + nfapi_uint16_tlv_t dmrs_typeA_position; + nfapi_uint16_tlv_t dmrs_additional_position; + nfapi_uint16_tlv_t dmrs_type; + nfapi_uint16_tlv_t dmrs_max_length; + nfapi_uint16_tlv_t mapping_type; + nfapi_uint16_tlv_t resource_allocation; + nfapi_uint16_tlv_t time_allocation_list; + nfapi_uint16_tlv_t mcs_table; + nfapi_uint16_tlv_t aggregation_factor; + nfapi_uint16_tlv_t prb_bundling_type; + nfapi_uint16_tlv_t rbg_size; + nfapi_uint16_tlv_t vrb_to_prb_interleaver; + nfapi_uint16_tlv_t code_block_groug_transmission; + nfapi_uint16_tlv_t x_overhead; + nfapi_uint16_tlv_t max_dci_codewords; + +} nfapi_nr_pdsch_config_t; +#define NFAPI_NR_PDSCH_CONFIG_TAG + typedef struct { nfapi_uint16_tlv_t prach_RootSequenceIndex; ///// L1 parameter 'PRACHRootSequenceIndex' @@ -208,12 +226,6 @@ typedef struct { nfapi_uint16_tlv_t preambleReceivedTargetPower; } nfapi_nr_rach_config_t; -typedef struct { - nfapi_uint16_tlv_t dmrs_TypeA_Position; ///// Position of (first) DL DM-RS - nfapi_uint16_tlv_t TimeDomainResourceAllocation_k0; ///// L1 parameter 'K0' - nfapi_uint16_tlv_t TimeDomainResourceAllocation_mappingType; ///// L1 parameter 'Mapping-type' -} nfapi_nr_pdsch_config_t; - typedef struct { nfapi_uint16_tlv_t groupHoppingEnabledTransformPrecoding; ///// L1 parameter 'Group-hopping-enabled-Transform-precoding' nfapi_uint16_tlv_t msg3_DeltaPreamble; ///// L1 parameter 'Delta-preamble-msg3' @@ -272,8 +284,8 @@ typedef struct { nfapi_nr_subframe_config_t subframe_config; nfapi_nr_rf_config_t rf_config; nfapi_nr_sch_config_t sch_config; - nfapi_nr_rach_config_t rach_config; nfapi_nr_pdsch_config_t pdsch_config; + nfapi_nr_rach_config_t rach_config; nfapi_nr_pusch_config_t pusch_config; nfapi_nr_pucch_config_t pucch_config; nfapi_nr_pdcch_config_t pdcch_config; @@ -324,6 +336,13 @@ typedef enum { NFAPI_NR_SEARCH_SPACE_TYPE_UE_SPECIFIC } nfapi_nr_search_space_type_e; +typedef enum { + NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_0=0, + NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_0A, + NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_1, + NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_2 +} nfapi_nr_common_search_space_type_e; + typedef enum { NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1=0, NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE2, @@ -345,6 +364,44 @@ typedef enum { NFAPI_NR_CSET_ALL_CONTIGUOUS_RBS } nfapi_nr_coreset_precoder_granularity_type_e; +typedef enum { + NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_A=0, + NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_B, + NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_C, + NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_ALLOC_LIST +} nfapi_nr_pdsch_time_domain_alloc_type_e; + +typedef enum { + NFAPI_NR_PDSCH_MAPPING_TYPE_A=0, + NFAPI_NR_PDSCH_MAPPING_TYPE_B +} nfapi_nr_pdsch_mapping_type_e; + +typedef enum { + NFAPI_NR_PDSCH_RBG_CONFIG_TYPE1=0, + NFAPI_NR_PDSCH_RBG_CONFIG_TYPE2 +} nfapi_nr_pdsch_rbg_config_type_e; + +typedef enum { + NFAPI_NR_PRG_GRANULARITY_2=2, + NFAPI_NR_PRG_GRANULARITY_4=4, + NFAPI_NR_PRG_GRANULARITY_WIDEBAND +} nfapi_nr_prg_granularity_e; + +typedef enum { + NFAPI_NR_PRB_BUNDLING_TYPE_STATIC=0, + NFAPI_NR_PRB_BUNDLING_TYPE_DYNAMIC +} nfapi_nr_prb_bundling_type_e; + +typedef enum { + NFAPI_NR_MCS_TABLE_QAM64_LOW_SE=0, + NFAPI_NR_MCS_TABLE_QAM256 +} nfapi_nr_pdsch_mcs_table_e; + +typedef enum { + NFAPI_NR_DMRS_TYPE1=0, + NFAPI_NR_DMRS_TYPE2 +} nfapi_nr_dmrs_type_e; + // P7 Sub Structures typedef struct { @@ -448,11 +505,14 @@ typedef struct{ typedef struct { nfapi_tl_t tl; - uint8_t rnti; + uint16_t rnti; uint8_t rnti_type; uint8_t dci_format; + /// Number of CRB in BWP that this DCI configures + uint16_t n_RB_BWP; uint8_t config_type; - uint8_t search_space_type; + uint8_t search_space_type; + uint8_t common_search_space_type; uint8_t aggregation_level; uint8_t n_rb; uint8_t n_symb; @@ -485,14 +545,50 @@ typedef struct { } nfapi_nr_dl_config_bch_pdu; typedef struct { - nfapi_tl_t tl; - nfapi_nr_search_space_t sib1searchSpace; - nfapi_nr_coreset_t sib1ControlResourceSets; - nfapi_nr_search_space_t sibssearchSpace; - nfapi_nr_coreset_t sibsControlResourceSets; - nfapi_nr_search_space_t ra_SearchSpace; - nfapi_nr_coreset_t ra_ControlResourceSets; + nfapi_tl_t tl; + uint16_t length; + uint8_t pdu_index; + uint16_t rnti; + uint16_t nb_mod_symbols; + uint8_t time_allocation_type; + uint8_t freq_allocation_type; + uint8_t start_prb; + uint8_t n_prb; + uint8_t start_symbol; + uint8_t nb_symbols; + uint8_t mcs_idx; + uint8_t ndi; + uint8_t nb_codewords; + uint8_t nb_layers; + uint16_t coding_rate; + uint8_t modulation; + uint8_t modulation_order; + uint16_t transport_block_size; + uint8_t nb_re_dmrs; + uint8_t time_alloc_list_flag; + uint8_t time_alloc_list; + uint8_t rbg_list; + uint8_t virtual_resource_block_assignment_flag; + uint32_t resource_block_coding; + uint8_t redundancy_version; + uint8_t transport_blocks; + uint8_t transmission_scheme; + uint8_t number_of_subbands; + uint8_t codebook_index[NFAPI_MAX_NUM_SUBBANDS]; + uint8_t ue_category_capacity; + uint8_t pa; + uint8_t delta_power_offset_index; + uint8_t ngap; + uint8_t transmission_mode; + uint8_t num_bf_prb_per_subband; + uint8_t num_bf_vector; + nfapi_bf_vector_t bf_vector[NFAPI_MAX_BF_VECTORS]; }nfapi_nr_dl_config_dlsch_pdu_rel15_t; +#define NFAPI_NR_DL_CONFIG_REQUEST_DLSCH_PDU_REL15_TAG + +typedef struct { + nfapi_nr_dl_config_dlsch_pdu_rel15_t dlsch_pdu_rel15; +} nfapi_nr_dl_config_dlsch_pdu; typedef struct { nfapi_tl_t tl; @@ -502,15 +598,15 @@ typedef struct { typedef struct { -} nfapi_nr_dl_config_nbch_pdu; +} nfapi_nr_dl_config_nbch_pdu_rel15_t; typedef struct { -} nfapi_nr_dl_config_npdcch_pdu; +} nfapi_nr_dl_config_npdcch_pdu_rel15_t; typedef struct { -} nfapi_nr_dl_config_ndlsch_pdu; +} nfapi_nr_dl_config_ndlsch_pdu_rel15_t; typedef struct { nfapi_nr_dl_config_dci_dl_pdu_rel15_t dci_dl_pdu_rel15; @@ -525,11 +621,11 @@ typedef struct { union { nfapi_nr_dl_config_dci_dl_pdu dci_dl_pdu; nfapi_nr_dl_config_bch_pdu_rel15_t bch_pdu_rel15; - nfapi_nr_dl_config_dlsch_pdu_rel15_t dlsch_pdu_rel15; + nfapi_nr_dl_config_dlsch_pdu dlsch_pdu; nfapi_nr_dl_config_pch_pdu_rel15_t pch_pdu_rel15; - //nfapi_nr_dl_config_nbch_pdu_rel15_t nbch_pdu_rel15; - //nfapi_nr_dl_config_npdcch_pdu_rel15_t npdcch_pdu_rel15; - //nfapi_nr_dl_config_ndlsch_pdu_rel15_t ndlsch_pdu_rel15; + nfapi_nr_dl_config_nbch_pdu_rel15_t nbch_pdu_rel15; + nfapi_nr_dl_config_npdcch_pdu_rel15_t npdcch_pdu_rel15; + nfapi_nr_dl_config_ndlsch_pdu_rel15_t ndlsch_pdu_rel15; }; } nfapi_nr_dl_config_request_pdu_t; diff --git a/oaienv b/oaienv index 4cee684b3d..61b607c605 100644 --- a/oaienv +++ b/oaienv @@ -18,6 +18,6 @@ alias oailte='cd $OPENAIR_TARGETS/RT/USER' alias oais='cd $OPENAIR_TARGETS/SIMU/USER' alias oaiex='cd $OPENAIR_TARGETS/SIMU/EXAMPLES' -#export IIOD_REMOTE=192.168.121.32 -export IIOD_REMOTE=192.168.1.11 +export IIOD_REMOTE=192.168.121.32 +#export IIOD_REMOTE=192.168.1.11 diff --git a/openair1/PHY/CODING/TESTBENCH/ldpctest.c b/openair1/PHY/CODING/TESTBENCH/ldpctest.c index 158be941c7..6db9548dbe 100644 --- a/openair1/PHY/CODING/TESTBENCH/ldpctest.c +++ b/openair1/PHY/CODING/TESTBENCH/ldpctest.c @@ -24,6 +24,7 @@ #include #include +#include "assertions.h" #include "SIMULATION/TOOLS/sim.h" #include "PHY/CODING/nrLDPC_encoder/defs.h" #include "PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.h" @@ -38,6 +39,10 @@ # endif #endif +#define NR_LDPC_PROFILER_DETAIL +#define NR_LDPC_ENABLE_PARITY_CHECK + + // 4-bit quantizer char quantize4bit(double D,double x) { @@ -53,13 +58,33 @@ char quantize4bit(double D,double x) return((char)qxd); } +char quantize8bit(double D,double x) +{ + double qxd; + //char maxlev; + qxd = floor(x/D); + + //maxlev = 1<<(B-1); + + //printf("x=%f,qxd=%f,maxlev=%d\n",x,qxd, maxlev); + + if (qxd <= -128) + qxd = -128; + else if (qxd >= 128) + qxd = 127; + + return((char)qxd); +} + char quantize(double D,double x,unsigned char B) { double qxd; - char maxlev; + short maxlev; qxd = floor(x/D); - // printf("x=%f,qxd=%f\n",x,qxd); - maxlev = 1<<(B-1); + + maxlev = 1<<(B-1);//(char)(pow(2,B-1)); + + //printf("x=%f,qxd=%f,maxlev=%d\n",x,qxd, maxlev); if (qxd <= -maxlev) qxd = -maxlev; @@ -69,45 +94,80 @@ char quantize(double D,double x,unsigned char B) return((char)qxd); } +typedef struct { +double n_iter_mean; +double n_iter_std; +int n_iter_max; +} n_iter_stats_t; + #define MAX_BLOCK_LENGTH 8448 +short lift_size[51]= {2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,18,20,22,24,26,28,30,32,36,40,44,48,52,56,60,64,72,80,88,96,104,112,120,128,144,160,176,192,208,224,240,256,288,320,352,384}; + int test_ldpc(short No_iteration, - //int nom_rate, - //int denom_rate, + int nom_rate, + int denom_rate, double SNR, unsigned char qbits, - unsigned short block_length, + short block_length, unsigned int ntrials, + int n_segments, unsigned int *errors, - unsigned int *crc_misses) + unsigned int *errors_bit, + double *errors_bit_uncoded, + unsigned int *crc_misses, + time_stats_t *time_optim, + time_stats_t *time_decoder, + n_iter_stats_t * dec_iter) { //clock initiate - time_stats_t time,time_optim,tinput,tprep,tparity,toutput; + //time_stats_t time,time_optim,tinput,tprep,tparity,toutput, time_decoder; + time_stats_t time, tinput,tprep,tparity,toutput; + double n_iter_mean = 0; + double n_iter_std = 0; + int n_iter_max = 0; + unsigned int segment_bler = 0; + + double sigma; + sigma = 1.0/sqrt(2*SNR); + opp_enabled=1; cpu_freq_GHz = get_cpu_freq_GHz(); //short test_input[block_length]; unsigned char *test_input[MAX_NUM_DLSCH_SEGMENTS]={NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL};; //short *c; //padded codeword - unsigned char *estimated_output; + unsigned char *estimated_output[MAX_NUM_DLSCH_SEGMENTS]; + unsigned char *estimated_output_bit[MAX_NUM_DLSCH_SEGMENTS]; + unsigned char *test_input_bit; unsigned char *channel_input[MAX_NUM_DLSCH_SEGMENTS]; + unsigned char *channel_output_uncoded[MAX_NUM_DLSCH_SEGMENTS]; unsigned char *channel_input_optim[MAX_NUM_DLSCH_SEGMENTS]; double *channel_output; - double *modulated_input; - char *channel_output_fixed; + double *modulated_input[MAX_NUM_DLSCH_SEGMENTS]; + char *channel_output_fixed[MAX_NUM_DLSCH_SEGMENTS]; unsigned int i,j,trial=0; - short BG,Zc,Kb; - //short nrows,ncols; - //int no_punctured_columns,removed_bit; + short BG,Zc,Kb,nrows;//,ncols; + int no_punctured_columns,removed_bit; int i1; + int R_ind = 0; //Table of possible lifting sizes - short lift_size[51]= {2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,18,20,22,24,26,28,30,32,36,40,44,48,52,56,60,64,72,80,88,96,104,112,120,128,144,160,176,192,208,224,240,256,288,320,352,384}; - int n_segments=8; - int N; //length encoded block +//short lift_size[51]= {2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,18,20,22,24,26,28,30,32,36,40,44,48,52,56,60,64,72,80,88,96,104,112,120,128,144,160,176,192,208,224,240,256,288,320,352,384}; + //int n_segments=1; + int code_rate_vec[8] = {15, 13, 25, 12, 23, 34, 56, 89}; + //double code_rate_actual_vec[8] = {0.2, 0.33333, 0.4, 0.5, 0.66667, 0.73333, 0.81481, 0.88}; t_nrLDPC_dec_params decParams; - int n_iter; + t_nrLDPC_procBuf nrLDPC_procBuf; + t_nrLDPC_procBuf* p_nrLDPC_procBuf = &nrLDPC_procBuf; + + t_nrLDPC_time_stats decoder_profiler; + t_nrLDPC_time_stats* p_decoder_profiler =&decoder_profiler ; + + int32_t n_iter = 0; *errors=0; + *errors_bit=0; + *errors_bit_uncoded=0; *crc_misses=0; // generate input block @@ -115,37 +175,459 @@ int test_ldpc(short No_iteration, test_input[j]=(unsigned char *)malloc16(sizeof(unsigned char) * block_length/8); channel_input[j] = (unsigned char *)malloc16(sizeof(unsigned char) * 68*384); channel_input_optim[j] = (unsigned char *)malloc16(sizeof(unsigned char) * 68*384); + channel_output_uncoded[j] = (unsigned char *)malloc16(sizeof(unsigned char) * 68*384); + estimated_output[j] = (unsigned char*) malloc16(sizeof(unsigned char) * block_length); + estimated_output_bit[j] = (unsigned char*) malloc16(sizeof(unsigned char) * block_length); + modulated_input[j] = (double *)malloc16(sizeof(double) * 68*384); + channel_output_fixed[j] = (char *)malloc16(sizeof( char) * 68*384); } - modulated_input = (double *)malloc(sizeof(double) * 68*384); - channel_output = (double *)malloc(sizeof(double) * 68*384); - channel_output_fixed = (char *)malloc16(sizeof(char) * 68*384); - estimated_output = (unsigned char*) malloc16(sizeof(unsigned char) * block_length); - - memset(channel_output_fixed,0,sizeof(char) * 68*384); + //modulated_input = (double *)malloc(sizeof(double) * 68*384); + //channel_output = (double *)malloc(sizeof(double) * 68*384); + //channel_output_fixed = (char *)malloc16(sizeof(char) * 68*384); + //modulated_input = (double *)calloc(68*384, sizeof(double)); + channel_output = (double *)calloc(68*384, sizeof(double)); + //channel_output_fixed = (double *)calloc(68*384, sizeof(double)); + //channel_output_fixed = (unsigned char*)calloc(68*384, sizeof(unsigned char*)); + //estimated_output = (unsigned char*) malloc16(sizeof(unsigned char) * block_length);///8); + //estimated_output_bit = (unsigned char*) malloc16(sizeof(unsigned char) * block_length); + test_input_bit = (unsigned char*) malloc16(sizeof(unsigned char) * block_length); reset_meas(&time); - reset_meas(&time_optim); + reset_meas(time_optim); + reset_meas(time_decoder); reset_meas(&tinput); reset_meas(&tprep); reset_meas(&tparity); reset_meas(&toutput); + //Reset Decoder profiles + reset_meas(&decoder_profiler.llr2llrProcBuf); + reset_meas(&decoder_profiler.llr2CnProcBuf); + reset_meas(&decoder_profiler.cnProc); + reset_meas(&decoder_profiler.cnProcPc); + reset_meas(&decoder_profiler.bnProc); + reset_meas(&decoder_profiler.bnProcPc); + reset_meas(&decoder_profiler.cn2bnProcBuf); + reset_meas(&decoder_profiler.bn2cnProcBuf); + reset_meas(&decoder_profiler.llrRes2llrOut); + reset_meas(&decoder_profiler.llr2bit); + //reset_meas(&decoder_profiler.total); + + // Allocate LDPC decoder buffers + p_nrLDPC_procBuf = nrLDPC_init_mem(); for (j=0;j3840) && (block_length <= MAX_BLOCK_LENGTH)) + if (block_length>3840) + { + BG=1; + Kb = 22; + nrows=46; //parity check bits + //ncols=22; //info bits + } + else if (block_length<=3840) + { + BG=2; + nrows=42; //parity check bits + //ncols=10; // info bits + + if (block_length>640) + Kb = 10; + else if (block_length>560) + Kb = 9; + else if (block_length>192) + Kb = 8; + else + Kb = 6; + } + + if (nom_rate == 1) + if (denom_rate == 5) + if (BG == 2) + R_ind = 0; + else + printf("Not supported"); + else if (denom_rate == 3) + R_ind = 1; + else if (denom_rate == 2) + //R_ind = 3; + printf("Not supported"); + else + printf("Not supported"); + + else if (nom_rate == 2) + if (denom_rate == 5) + //R_ind = 2; + printf("Not supported"); + else if (denom_rate == 3) + R_ind = 4; + else + printf("Not supported"); + + else if ((nom_rate == 22) && (denom_rate == 30)) + //R_ind = 5; + printf("Not supported"); + else if ((nom_rate == 22) && (denom_rate == 27)) + //R_ind = 6; + printf("Not supported"); + else if ((nom_rate == 22) && (denom_rate == 25)) + if (BG == 1) + R_ind = 7; + else + printf("Not supported"); + else + printf("Not supported"); + + //find minimum value in all sets of lifting size + Zc=0; + + for (i1=0; i1 < 51; i1++) + { + if (lift_size[i1] >= (double) block_length/Kb) + { + Zc = lift_size[i1]; + //printf("%d\n",Zc); + break; + } + } + + printf("ldpc_test: codeword_length %d, n_segments %d, block_length %d, BG %d, Zc %d, Kb %d\n",n_segments *block_length, n_segments, block_length,BG, Zc, Kb); + no_punctured_columns=(int)((nrows-2)*Zc+block_length-block_length*(1/((float)nom_rate/(float)denom_rate)))/Zc; + // printf("puncture:%d\n",no_punctured_columns); + removed_bit=(nrows-no_punctured_columns-2) * Zc+block_length-(int)(block_length/((float)nom_rate/(float)denom_rate)); + if (ntrials==0) + ldpc_encoder_orig(test_input[0],channel_input[0], block_length, BG, 1); + + for (trial=0; trial < ntrials; trial++) + { + segment_bler = 0; + //// encoder + start_meas(&time); + for(j=0;j>3; i++) + { + //printf("block_length>>3: %d \n",block_length>>3); + /// printf("i: %d \n",i); + ///printf("estimated_output[%d]: %d \n",i,estimated_output[i]); + ///printf("test_input[0][%d]: %d \n",i,test_input[0][i]); + if (estimated_output[j][i] != test_input[j][i]) + { + //////printf("error pos %d (%d, %d)\n\n",i,estimated_output[i],test_input[0][i]); + segment_bler = segment_bler + 1; + break; + } + } + + for (i=0; i>(i&7); + test_input_bit[i] = (test_input[j][i/8]&(1<<(i&7)))>>(i&7); // Further correct for multiple segments + if (estimated_output_bit[j][i] != test_input_bit[i]) + { + *errors_bit = (*errors_bit) + 1; + } + } + + //if (*errors == 1000) + //break; + + n_iter_mean = n_iter_mean + n_iter; + n_iter_std = n_iter_std + pow(n_iter-1,2); + + if ( n_iter > n_iter_max ) + n_iter_max = n_iter; + + } // end segments + + if (segment_bler != 0) + *errors = (*errors) + 1; + + } + /*else if (trial==0) + printf("decoder is not supported\n");*/ + } + + + dec_iter->n_iter_mean = n_iter_mean/(double)ntrials/(double)n_segments - 1; + dec_iter->n_iter_std = sqrt(n_iter_std/(double)ntrials/(double)n_segments - pow(n_iter_mean/(double)ntrials/(double)n_segments - 1,2)); + dec_iter->n_iter_max = n_iter_max -1; + + *errors_bit_uncoded = *errors_bit_uncoded / (double)((Kb+nrows-no_punctured_columns-2) * Zc-removed_bit); + + for(j=0;j 1280, 2401 -> 2560 + + short No_iteration=5; + int n_segments=1; + //double rate=0.333; + int nom_rate=1; + int denom_rate=3; + double SNR0=-2.0,SNR,SNR_lin; + unsigned char qbits=8; + unsigned int decoded_errors[10000]; // initiate the size of matrix equivalent to size of SNR + int c,i=0, i1 = 0; + + int n_trials = 1; + double SNR_step = 0.1; + + randominit(0); + int test_uncoded= 0; + + time_stats_t time_optim[10], time_decoder[10]; + n_iter_stats_t dec_iter[3]; + + short BG,Zc,Kb; + + while ((c = getopt (argc, argv, "q:r:s:S:l:n:d:i:t:u:h")) != -1) + switch (c) + { + case 'q': + qbits = atoi(optarg); + break; + + case 'r': + nom_rate = atoi(optarg); + break; + + case 'd': + denom_rate = atoi(optarg); + break; + + case 'l': + block_length = atoi(optarg); + break; + + case 'n': + n_trials = atoi(optarg); + break; + + case 's': + SNR0 = atof(optarg); + break; + + case 'S': + n_segments = atof(optarg); + break; + + case 't': + SNR_step = atof(optarg); + break; + + case 'i': + No_iteration = atoi(optarg); + break; + + case 'u': + test_uncoded = atoi(optarg); + break; + + case 'h': + default: + printf("CURRENTLY SUPPORTED CODE RATES: \n"); + printf("BG1 (blocklength > 3840): 1/3, 2/3, 22/25 (8/9) \n"); + printf("BG2 (blocklength <= 3840): 1/5, 1/3, 2/3 \n\n"); + printf("-h This message\n"); + printf("-q Quantization bits, Default: 8\n"); + printf("-r Nominator rate, (1, 2, 22), Default: 1\n"); + printf("-d Denominator rate, (3, 5, 25), Default: 1\n"); + printf("-l Block length (l > 3840 -> BG1, rest BG2 ), Default: 8448\n"); + printf("-n Number of simulation trials, Default: 1\n"); + //printf("-M MCS2 for TB 2\n"); + printf("-s SNR per information bit (EbNo) in dB, Default: -2\n"); + printf("-S Number of segments (Maximum: 8), Default: 1\n"); + printf("-t SNR simulation step, Default: 0.1\n"); + printf("-i Max decoder iterations, Default: 5\n"); + printf("-u Set SNR per coded bit, Default: 0\n"); + exit(1); + break; + } + cpu_freq_GHz = get_cpu_freq_GHz(); + //printf("the decoder supports BG2, Kb=10, Z=128 & 256\n"); + //printf(" range of blocklength: 1201 -> 1280, 2401 -> 2560\n"); + printf("block length %d: \n", block_length); + printf("n_trials %d: \n", n_trials); + printf("SNR0 %f: \n", SNR0); + + + + + //for (block_length=8;block_length<=MAX_BLOCK_LENGTH;block_length+=8) + + + //determine number of bits in codeword + if (block_length>3840) { BG=1; Kb = 22; //nrows=46; //parity check bits //ncols=22; //info bits } - else if (block_length<=3840) + else if (block_length<=3840) { BG=2; //nrows=42; //parity check bits @@ -159,14 +641,11 @@ int test_ldpc(short No_iteration, Kb = 8; else Kb = 6; - } - else { - printf("block_length %d not supported\n",block_length); - exit(-1); - } - + } + //find minimum value in all sets of lifting size Zc=0; + for (i1=0; i1 < 51; i1++) { if (lift_size[i1] >= (double) block_length/Kb) @@ -177,192 +656,87 @@ int test_ldpc(short No_iteration, } } - if (BG==1) - N=66*Zc; - else if (BG==2) - N=50*Zc; + char fname[200]; + sprintf(fname,"ldpctest_BG_%d_Zc_%d_rate_%d-%d_block_length_%d_maxit_%d.txt",BG,Zc,nom_rate,denom_rate,block_length, No_iteration); + FILE *fd=fopen(fname,"w"); + AssertFatal(fd!=NULL,"cannot open %s\n",fname); - printf("ldpc_test: block_length %d, BG %d, Zc %d, Kb %d, N %d\n",block_length,BG, Zc, Kb, N); - //no_punctured_columns=(int)((nrows-2)*Zc+block_length-block_length*(1/((float)nom_rate/(float)denom_rate)))/Zc; - // printf("puncture:%d\n",no_punctured_columns); - //removed_bit=(nrows-no_punctured_columns-2) * Zc+block_length-(int)(block_length/((float)nom_rate/(float)denom_rate)); - if (ntrials==0) - ldpc_encoder_orig(test_input[0],channel_input[0], block_length, BG, 1); - - for (trial=0; trial < ntrials; trial++) + fprintf(fd,"SNR BLER BER UNCODED_BER ENCODER_MEAN ENCODER_STD ENCODER_MAX DECODER_TIME_MEAN DECODER_TIME_STD DECODER_TIME_MAX DECODER_ITER_MEAN DECODER_ITER_STD DECODER_ITER_MAX\n"); + + for (SNR=SNR0;SNR1.0) - channel_output_fixed[i]=127; - else if (channel_output[i]<-1.0) - channel_output_fixed[i]=-128; - else - channel_output_fixed[i] = (char) ((channel_output[i]*128)<0?(channel_output[i]*128-0.5):(channel_output[i]*128+0.5)); //fixed point 9-7 - //printf("llr[%d]=%d (%f)\n",i,channel_output_fixed[i],channel_output[i]); - } - - /* - the LDPC decoder supports several rates and the input has to be adapted accordingly (by padding with zeros to the next lower supported rate). Moreover, the first 2*Z LLRs are zero (since they are punctured and not transmitted) then you have the LLRs corresponding to the systematic bits. After that come the filler bits, the encoder uses 0 for the filler bits hence for the decoder you need to put 127 (max LLR for bit 0) at their place. Then come the LLRs corresponding to the punctured bits followed by zeros to pad for the supported decoder rate. - */ - for (i=0;i<2*Zc;i++) - channel_output_fixed[i] = 0; - - decParams.BG=BG; - decParams.Z=Zc; - decParams.R=13; - decParams.numMaxIter=6; - decParams.outMode = nrLDPC_outMode_BIT; - - // decode the sequence - // decoder supports BG2, Z=128 & 256 - //estimated_output=ldpc_decoder(channel_output_fixed, block_length, No_iteration, (double)((float)nom_rate/(float)denom_rate)); - - n_iter = nrLDPC_decoder(&decParams, (int8_t*) channel_output_fixed, (int8_t*) estimated_output, NULL); - - //for (i=(Kb+nrows) * Zc-5;i<(Kb+nrows) * Zc;i++) - // printf("esimated_output[%d]=%d\n",i,esimated_output[i]); - - //count errors - for (i=0; i>3; i++) - { - if (estimated_output[i] != test_input[0][i]) - { - printf("error pos %d (%d, %d)\n",i,estimated_output[i],test_input[0][i]); - *errors = (*errors) + 1; - break; - } - } - - } - - for(j=0;j 1280, 2401 -> 2560 - short No_iteration=25; - //double rate=0.333; - //int nom_rate=1; - //int denom_rate=3; - double SNR0=-2.0,SNR,SNR_lin; - unsigned char qbits=4; - unsigned int decoded_errors[100]; // initiate the size of matrix equivalent to size of SNR - int c,i=0; - - int n_trials = 1; - - randominit(0); - - while ((c = getopt (argc, argv, "q:s:l:n:h")) != -1) - switch (c) - { - case 'q': - qbits = atoi(optarg); - break; - - case 'l': - block_length = atoi(optarg); - break; - - case 'n': - n_trials = atoi(optarg); - break; - - case 's': - SNR0 = atoi(optarg); - break; - - default: - case 'h': - printf("usage ldpctest -h -l blocklength -n n_trials -s SNR\n"); - exit(0); - } - - //printf("the decoder supports BG2, Kb=10, Z=128 & 256\n"); - //printf(" range of blocklength: 1201 -> 1280, 2401 -> 2560\n"); - //printf("rate: %d/%d\n",nom_rate,denom_rate); - printf("block length %d: \n", block_length); - printf("n_trials %d: \n", n_trials); - printf("SNR0 %f: \n", SNR0); - - //for (block_length=8;block_length<=MAX_BLOCK_LENGTH;block_length+=8) - for (SNR=SNR0;SNRn_iter_mean); + printf("SNR %f, Std iterations: %f\n",SNR, dec_iter->n_iter_std); + printf("SNR %f, Max iterations: %d\n",SNR, dec_iter->n_iter_max); + printf("\n"); + printf("Encoding time mean: %15.3f us\n",(double)time_optim->diff/time_optim->trials/1000.0/cpu_freq_GHz); + printf("Encoding time std: %15.3f us\n",sqrt((double)time_optim->diff_square/time_optim->trials/pow(1000,2)/pow(cpu_freq_GHz,2)-pow((double)time_optim->diff/time_optim->trials/1000.0/cpu_freq_GHz,2))); + printf("Encoding time max: %15.3f us\n",(double)time_optim->max/1000.0/cpu_freq_GHz); + printf("\n"); + printf("Decoding time mean: %15.3f us\n",(double)time_decoder->diff/time_decoder->trials/1000.0/cpu_freq_GHz); + printf("Decoding time std: %15.3f us\n",sqrt((double)time_decoder->diff_square/time_decoder->trials/pow(1000,2)/pow(cpu_freq_GHz,2)-pow((double)time_decoder->diff/time_decoder->trials/1000.0/cpu_freq_GHz,2))); + printf("Decoding time max: %15.3f us\n",(double)time_decoder->max/1000.0/cpu_freq_GHz); + + fprintf(fd,"%f %f %f %f %f %f %f %f %f %f %f %f %d \n", + SNR, + (double)decoded_errors[i]/(double)n_trials , + (double)errors_bit/(double)n_trials/(double)block_length/(double)n_segments , + errors_bit_uncoded/(double)n_trials/(double)n_segments , + (double)time_optim->diff/time_optim->trials/1000.0/cpu_freq_GHz, + sqrt((double)time_optim->diff_square/time_optim->trials/pow(1000,2)/pow(cpu_freq_GHz,2)-pow((double)time_optim->diff/time_optim->trials/1000.0/cpu_freq_GHz,2)), + (double)time_optim->max/1000.0/cpu_freq_GHz, + (double)time_decoder->diff/time_decoder->trials/1000.0/cpu_freq_GHz, + sqrt((double)time_decoder->diff_square/time_decoder->trials/pow(1000,2)/pow(cpu_freq_GHz,2)-pow((double)time_decoder->diff/time_decoder->trials/1000.0/cpu_freq_GHz,2)), + (double)time_decoder->max/1000.0/cpu_freq_GHz, + dec_iter->n_iter_mean, + dec_iter->n_iter_std, + dec_iter->n_iter_max + ); + + if (decoded_errors[i] == 0) + break; i=i+1; + + } + fclose(fd); return(0); } + + + + + + + diff --git a/openair1/PHY/CODING/TESTBENCH/polartest.c b/openair1/PHY/CODING/TESTBENCH/polartest.c index ddc720f0e1..6d50070b86 100644 --- a/openair1/PHY/CODING/TESTBENCH/polartest.c +++ b/openair1/PHY/CODING/TESTBENCH/polartest.c @@ -7,6 +7,8 @@ #include #include "PHY/CODING/nrPolar_tools/nr_polar_defs.h" +#include "PHY/CODING/nrPolar_tools/nr_polar_pbch_defs.h" +#include "PHY/CODING/nrPolar_tools/nr_polar_uci_defs.h" #include "PHY/CODING/coding_defs.h" #include "SIMULATION/TOOLS/sim.h" @@ -18,11 +20,8 @@ int main(int argc, char *argv[]) { //Initiate timing. (Results depend on CPU Frequency. Therefore, might change due to performance variances during simulation.) time_stats_t timeEncoder,timeDecoder; - time_stats_t polar_decoder_init,polar_rate_matching,decoding,bit_extraction,deinterleaving; - time_stats_t path_metric,sorting,update_LLR; opp_enabled=1; int decoder_int16=0; - int generate_optim_code=0; cpu_freq_GHz = get_cpu_freq_GHz(); reset_meas(&timeEncoder); reset_meas(&timeDecoder); @@ -35,7 +34,7 @@ int main(int argc, char *argv[]) { double SNR, SNR_lin; int16_t nBitError = 0; // -1 = Decoding failed (All list entries have failed the CRC checks). - int8_t decoderState=0, blockErrorState=0; //0 = Success, -1 = Decoding failed, 1 = Block Error. + uint32_t decoderState=0, blockErrorState=0; //0 = Success, -1 = Decoding failed, 1 = Block Error. uint16_t testLength = 0, coderLength = 0, blockErrorCumulative=0, bitErrorCumulative=0; double timeEncoderCumulative = 0, timeDecoderCumulative = 0; uint8_t aggregation_level = 8, decoderListSize = 8, pathMetricAppr = 0; @@ -71,21 +70,20 @@ int main(int argc, char *argv[]) { pathMetricAppr = (uint8_t) atoi(optarg); break; - case 'q': - decoder_int16=1; - break; + case 'q': + decoder_int16 = 1; + break; - case 'g': - generate_optim_code=1; - iterations=1; - SNRstart=-6.0; - SNRstop =-6.0; - decoder_int16=1; - break; + case 'g': + iterations = 1; + SNRstart = -6.0; + SNRstop = -6.0; + decoder_int16 = 1; + break; - case 'h': - printf("./polartest -s SNRstart -d SNRinc -f SNRstop -m [0=PBCH|1=DCI|2=UCI] -i iterations -l decoderListSize -a pathMetricAppr -q (use fixed point decoder)\n"); - exit(-1); + case 'h': + printf("./polartest -s SNRstart -d SNRinc -f SNRstop -m [0=PBCH|1=DCI|2=UCI] -i iterations -l decoderListSize -a pathMetricAppr\n"); + exit(-1); default: perror("[polartest.c] Problem at argument parsing with getopt"); @@ -93,13 +91,13 @@ int main(int argc, char *argv[]) { } if (polarMessageType == 0) { //PBCH - testLength = NR_POLAR_PBCH_PAYLOAD_BITS; + testLength = 64;//NR_POLAR_PBCH_PAYLOAD_BITS; coderLength = NR_POLAR_PBCH_E; aggregation_level = NR_POLAR_PBCH_AGGREGATION_LEVEL; } else if (polarMessageType == 1) { //DCI //testLength = nr_get_dci_size(params_rel15->dci_format, params_rel15->rnti_type, &fp->initial_bwp_dl, cfg); - testLength = 20; - coderLength = 108; //to be changed by aggregate level function. + testLength = 41; //20; + coderLength = 108*8; //to be changed by aggregate level function. } else if (polarMessageType == -1) { //UCI //testLength = ; //coderLength = ; @@ -147,18 +145,18 @@ int main(int argc, char *argv[]) { uint8_t testArrayLength = ceil(testLength / 32.0); uint8_t coderArrayLength = ceil(coderLength / 32.0); - uint32_t *testInput = malloc(sizeof(uint32_t) * testArrayLength); //generate randomly - uint32_t *encoderOutput = malloc(sizeof(uint32_t) * coderArrayLength); - uint32_t *estimatedOutput = malloc(sizeof(uint32_t) * testArrayLength); //decoder output + uint32_t testInput[testArrayLength]; //generate randomly + uint32_t encoderOutput[coderArrayLength]; + uint32_t estimatedOutput[testArrayLength]; //decoder output memset(testInput,0,sizeof(uint32_t) * testArrayLength); memset(encoderOutput,0,sizeof(uint32_t) * coderArrayLength); memset(estimatedOutput,0,sizeof(uint32_t) * testArrayLength); - uint8_t *encoderOutputByte = malloc(sizeof(uint8_t) * coderLength); - double *modulatedInput = malloc (sizeof(double) * coderLength); //channel input - double *channelOutput = malloc (sizeof(double) * coderLength); //add noise - int16_t *channelOutput_int16; - if (decoder_int16 == 1) channelOutput_int16 = (int16_t*)malloc (sizeof(int16_t) * coderLength); + uint8_t encoderOutputByte[coderLength]; + double modulatedInput[coderLength]; //channel input + double channelOutput[coderLength]; //add noise + + int16_t channelOutput_int16[coderLength]; t_nrPolar_paramsPtr nrPolar_params = NULL, currentPtr = NULL; nr_polar_init(&nrPolar_params, polarMessageType, testLength, aggregation_level); @@ -209,6 +207,8 @@ int main(int argc, char *argv[]) { rnti); printf("dci_estimation: [0]->0x%08x \t [1]->0x%08x \t [2]->0x%08x \t [3]->0x%08x\n", dci_estimation[0], dci_estimation[1], dci_estimation[2], dci_estimation[3]); + free(encoder_outputByte); + free(channel_output); return 0; #endif @@ -268,7 +268,7 @@ int main(int argc, char *argv[]) { uint8_t nr_polar_A[32] = {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,1,1,1,1,1,1,1,1,0,0,0,1}; uint8_t nr_polar_crc[24]; uint8_t **crc_generator_matrix = crc24c_generator_matrix(32); - nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(nr_polar_A, + nr_matrix_multiplication_uint8_1D_uint8_2D(nr_polar_A, crc_generator_matrix, nr_polar_crc, 32, @@ -324,10 +324,18 @@ int main(int argc, char *argv[]) { for (int i=0; i<32; i++) printf("%d\n",(testInput[0]>>i)&1);*/ - start_meas(&timeEncoder); - polar_encoder(testInput, encoderOutput, currentPtr); - stop_meas(&timeEncoder); + + int len_mod64=currentPtr->payloadBits&63; + ((uint64_t*)testInput)[currentPtr->payloadBits/64]&=((((uint64_t)1)<0x%08x\n", encoderOutput[0]); printf("encoderOutput: [1]->0x%08x\n", encoderOutput[1]);*/ @@ -369,7 +377,7 @@ int main(int argc, char *argv[]) { aPrioriArray); else decoderState = polar_decoder_int16(channelOutput_int16, - estimatedOutput, + (uint64_t*)estimatedOutput, currentPtr); @@ -379,17 +387,20 @@ int main(int argc, char *argv[]) { //calculate errors - if (decoderState==-1) { + if (decoderState!=0) { blockErrorState=-1; nBitError=-1; } else { - for (int i = 0; i < testArrayLength; i++) { - for (int j = 0; j < (sizeof(testInput[0])*8); j++) { - if (((estimatedOutput[i]>>j) & 1) != ((testInput[i]>>j) & 1)) nBitError++; - } - } + for (int j = 0; j < currentPtr->payloadBits; j++) { + if (((estimatedOutput[0]>>j) & 1) != ((testInput[0]>>j) & 1)) nBitError++; + // printf("bit %d: %d => %d\n",j,(testInput[0]>>j)&1,(estimatedOutput[0]>>j)&1); + } + - if (nBitError>0) blockErrorState=1; + if (nBitError>0) { + blockErrorState=1; + // printf("Error: Input %x, Output %x\n",testInput[0],estimatedOutput[0]); + } } //Iteration times are in microseconds. @@ -415,7 +426,8 @@ int main(int argc, char *argv[]) { printf("[ListSize=%d, Appr=%d] SNR=%+8.3f, BLER=%9.6f, BER=%12.9f, t_Encoder=%9.3fus, t_Decoder=%9.3fus\n", decoderListSize, pathMetricAppr, SNR, ((double)blockErrorCumulative/iterations), ((double)bitErrorCumulative / (iterations*testLength)), - (timeEncoderCumulative/iterations),timeDecoderCumulative/iterations); + (double)timeEncoder.diff/timeEncoder.trials/(cpu_freq_GHz*1000.0),(double)timeDecoder.diff/timeDecoder.trials/(cpu_freq_GHz*1000.0)); + //(timeEncoderCumulative/iterations),timeDecoderCumulative/iterations); if (blockErrorCumulative==0 && bitErrorCumulative==0) break; @@ -428,14 +440,5 @@ int main(int argc, char *argv[]) { print_meas(&timeDecoder,"polar_decoder",NULL,NULL); fclose(logFile); - //Bit - free(testInput); - free(encoderOutput); - free(estimatedOutput); - //Byte - free(encoderOutputByte); - free(modulatedInput); - free(channelOutput); - return (0); } diff --git a/openair1/PHY/CODING/coding_defs.h b/openair1/PHY/CODING/coding_defs.h index 1842762cd1..0f316e7869 100644 --- a/openair1/PHY/CODING/coding_defs.h +++ b/openair1/PHY/CODING/coding_defs.h @@ -38,6 +38,11 @@ #define MAX_TURBO_ITERATIONS_MBSFN 8 #define MAX_TURBO_ITERATIONS max_turbo_iterations +#define MAX_LDPC_ITERATIONS 4 +#define MAX_LDPC_ITERATIONS_MBSFN 4 + +#define LTE_NULL 2 +#define NR_NULL 2 typedef struct { unsigned short nb_bits; unsigned short f1; @@ -337,7 +342,7 @@ void ccodedab_init_inv(void); /*!\fn void crcTableInit(void) \brief This function initializes the different crc tables.*/ -//void crcTableInit (void); +void crcTableInit (void); @@ -382,6 +387,8 @@ unsigned int crc12 (unsigned char * inptr, int bitlen); @param inPtr Pointer to input byte stream @param bitlen length of inputs in bits*/ unsigned int crc8 (unsigned char * inptr, int bitlen); + +int check_crc(uint8_t* decoded_bytes, uint32_t n, uint32_t F, uint8_t crc_type); /*!\fn void phy_viterbi_dot11_sse2(int8_t *y, uint8_t *decoded_bytes, uint16_t n,int offset,int traceback) \brief This routine performs a SIMD optmized Viterbi decoder for the 802.11 64-state convolutional code. It can be @@ -436,23 +443,16 @@ int32_t rate_matching_lte(uint32_t N_coded, uint8_t *inPtr, uint32_t off); -void crcTableInit (void); - unsigned int crcbit (unsigned char * inputptr, int octetlen, unsigned int poly); int16_t reverseBits(int32_t ,int32_t); void phy_viterbi_dot11(int8_t *,uint8_t *,uint16_t); -decoder_if_t phy_threegpplte_turbo_decoder; -decoder_if_t phy_threegpplte_turbo_decoder8; -decoder_if_t phy_threegpplte_turbo_decoder16; - int32_t nr_segmentation(unsigned char *input_buffer, unsigned char **output_buffers, unsigned int B, unsigned int *C, - unsigned int *Kplus, - unsigned int *Kminus, + unsigned int *K, unsigned int *Zout, unsigned int *F); @@ -463,4 +463,33 @@ uint32_t nr_compute_tbs(uint8_t mcs, uint16_t length_dmrs, uint8_t Nl); +void nr_interleaving_ldpc(uint32_t E, uint8_t Qm, uint8_t *e,uint8_t *f); + +void nr_deinterleaving_ldpc(uint32_t E, uint8_t Qm, int16_t *e,int16_t *f); + +int nr_rate_matching_ldpc(uint8_t Ilbrm, + uint32_t Tbslbrm, + uint8_t BG, + uint16_t Z, + uint8_t *w, + uint8_t *e, + uint8_t C, + uint8_t rvidx, + uint32_t E); + +int nr_rate_matching_ldpc_rx(uint8_t Ilbrm, + uint32_t Tbslbrm, + uint8_t BG, + uint16_t Z, + int16_t *w, + int16_t *soft_input, + uint8_t C, + uint8_t rvidx, + uint8_t clear, + uint32_t E); + +decoder_if_t phy_threegpplte_turbo_decoder; +decoder_if_t phy_threegpplte_turbo_decoder8; +decoder_if_t phy_threegpplte_turbo_decoder16; + #endif diff --git a/openair1/PHY/CODING/crc_byte.c b/openair1/PHY/CODING/crc_byte.c index 6176e1f92f..d728d6b9ad 100644 --- a/openair1/PHY/CODING/crc_byte.c +++ b/openair1/PHY/CODING/crc_byte.c @@ -35,12 +35,14 @@ /*ref 36-212 v8.6.0 , pp 8-9 */ /* the highest degree is set by default */ + unsigned int poly24a = 0x864cfb00; // 1000 0110 0100 1100 1111 1011 // D^24 + D^23 + D^18 + D^17 + D^14 + D^11 + D^10 + D^7 + D^6 + D^5 + D^4 + D^3 + D + 1 unsigned int poly24b = 0x80006300; // 1000 0000 0000 0000 0110 0011 // D^24 + D^23 + D^6 + D^5 + D + 1 unsigned int poly24c = 0xb2b11700; // 1011 0010 1011 0001 0001 0111 // D^24+D^23+D^21+D^20+D^17+D^15+D^13+D^12+D^8+D^4+D^2+D+1 + unsigned int poly16 = 0x10210000; // 0001 0000 0010 0001 D^16 + D^12 + D^5 + 1 unsigned int poly12 = 0x80F00000; // 1000 0000 1111 D^12 + D^11 + D^3 + D^2 + D + 1 unsigned int poly8 = 0x9B000000; // 1001 1011 D^8 + D^7 + D^4 + D^3 + D + 1 @@ -56,8 +58,8 @@ The first bit is in the MSB of each byte *********************************************************/ unsigned int crcbit (unsigned char * inputptr, - int octetlen, - unsigned int poly) + int octetlen, + unsigned int poly) { unsigned int i, crc = 0, c; @@ -168,14 +170,12 @@ unsigned int crc24c (unsigned char * inptr, resbit = (bitlen % 8); while (octetlen-- > 0) { -/*#ifdef DEBUG_CRC24C - printf("crc24c: in %x => crc %x (%x)\n",crc,*inptr,crc24cTable[(*inptr) ^ (crc >> 24)]); -#endif*/ crc = (crc << 8) ^ crc24cTable[(*inptr++) ^ (crc >> 24)]; } - if (resbit > 0) + if (resbit > 0) { crc = (crc << resbit) ^ crc24cTable[((*inptr) >> (8 - resbit)) ^ (crc >> (32 - resbit))]; + } return crc; } @@ -255,6 +255,68 @@ unsigned int crcPayload(unsigned char * inptr, int bitlen, uint32_t* crc256Table return crc; } +int check_crc(uint8_t* decoded_bytes, uint32_t n, uint32_t F, uint8_t crc_type) +{ + uint32_t crc=0,oldcrc=0; + uint8_t crc_len,temp; + + switch (crc_type) { + case CRC24_A: + case CRC24_B: + crc_len=3; + break; + + case CRC16: + crc_len=2; + break; + + case CRC8: + crc_len=1; + break; + + default: + crc_len=3; + } + + for (int i=0; i>3)-crc_len+i])<<((crc_len-1-i)<<3); + + switch (crc_type) { + + case CRC24_A: + oldcrc&=0x00ffffff; + crc = crc24a(decoded_bytes, + n-24)>>8; + + break; + + case CRC24_B: + oldcrc&=0x00ffffff; + crc = crc24b(decoded_bytes, + n-24)>>8; + + break; + + case CRC16: + oldcrc&=0x0000ffff; + crc = crc16(decoded_bytes, + n-16)>>16; + + break; + + case CRC8: + oldcrc&=0x000000ff; + crc = crc8(decoded_bytes, + n-8)>>24; + break; + } + + if (crc == oldcrc) + return(1); + else + return(0); + +} #ifdef DEBUG_CRC diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_bnProc.h b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_bnProc.h index 96d5c157ec..00e8fd98d5 100644 --- a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_bnProc.h +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_bnProc.h @@ -37,12 +37,17 @@ \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_bnProcPc(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_bnProcPc(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint8_t* lut_numBnInBnGroups = p_lut->numBnInBnGroups; const uint32_t* lut_startAddrBnGroups = p_lut->startAddrBnGroups; const uint16_t* lut_startAddrBnGroupsLlr = p_lut->startAddrBnGroupsLlr; + int8_t* bnProcBuf = p_procBuf->bnProcBuf; + int8_t* bnProcBufRes = p_procBuf->bnProcBufRes; + int8_t* llrRes = p_procBuf->llrRes; + int8_t* llrProcBuf = p_procBuf->llrProcBuf; + __m128i* p_bnProcBuf; __m256i* p_bnProcBufRes; __m128i* p_llrProcBuf; @@ -1676,7 +1681,7 @@ static inline void nrLDPC_bnProcPc(t_nrLDPC_lut* p_lut, uint16_t Z) \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_bnProc(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_bnProc(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { // BN Processing calculating the values to send back to the CNs for next iteration // bnProcBufRes contains the sum of all edges to each BN at the start of each group @@ -1685,6 +1690,10 @@ static inline void nrLDPC_bnProc(t_nrLDPC_lut* p_lut, uint16_t Z) const uint32_t* lut_startAddrBnGroups = p_lut->startAddrBnGroups; const uint16_t* lut_startAddrBnGroupsLlr = p_lut->startAddrBnGroupsLlr; + int8_t* bnProcBuf = p_procBuf->bnProcBuf; + int8_t* bnProcBufRes = p_procBuf->bnProcBufRes; + int8_t* llrRes = p_procBuf->llrRes; + __m256i* p_bnProcBuf; __m256i* p_bnProcBufRes; __m256i* p_llrRes; diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_cnProc.h b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_cnProc.h index 300624ae76..5b96d6226e 100644 --- a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_cnProc.h +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_cnProc.h @@ -36,11 +36,14 @@ \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_cnProc_BG2(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_cnProc_BG2(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBuf = p_procBuf->cnProcBuf; + int8_t* cnProcBufRes = p_procBuf->cnProcBufRes; + __m256i* p_cnProcBuf; __m256i* p_cnProcBufRes; @@ -362,11 +365,14 @@ static inline void nrLDPC_cnProc_BG2(t_nrLDPC_lut* p_lut, uint16_t Z) \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_cnProc_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_cnProc_BG1(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBuf = p_procBuf->cnProcBuf; + int8_t* cnProcBufRes = p_procBuf->cnProcBufRes; + __m256i* p_cnProcBuf; __m256i* p_cnProcBufRes; @@ -858,11 +864,14 @@ static inline void nrLDPC_cnProc_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) \param Z Lifting size \return 32-bit parity check indicator */ -static inline uint32_t nrLDPC_cnProcPc_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline uint32_t nrLDPC_cnProcPc_BG1(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBuf = p_procBuf->cnProcBuf; + int8_t* cnProcBufRes = p_procBuf->cnProcBufRes; + __m256i* p_cnProcBuf; __m256i* p_cnProcBufRes; @@ -1490,11 +1499,14 @@ static inline uint32_t nrLDPC_cnProcPc_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) \param Z Lifting size \return 32-bit parity check indicator */ -static inline uint32_t nrLDPC_cnProcPc_BG2(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline uint32_t nrLDPC_cnProcPc_BG2(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBuf = p_procBuf->cnProcBuf; + int8_t* cnProcBufRes = p_procBuf->cnProcBufRes; + __m256i* p_cnProcBuf; __m256i* p_cnProcBufRes; diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.c b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.c index 53966befff..bbcd34c368 100644 --- a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.c +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.c @@ -38,24 +38,29 @@ #include "nrLDPC_cnProc.h" #include "nrLDPC_bnProc.h" +#define NR_LDPC_ENABLE_PARITY_CHECK +//#define NR_LDPC_PROFILER_DETAIL + #ifdef NR_LDPC_DEBUG_MODE #include "nrLDPC_tools/nrLDPC_debug.h" #endif -static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_t numLLR, t_nrLDPC_lut* p_lut, t_nrLDPC_dec_params* p_decParams, t_nrLDPC_time_stats* p_profiler); +static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, t_nrLDPC_procBuf* p_procBuf, uint32_t numLLR, t_nrLDPC_lut* p_lut, t_nrLDPC_dec_params* p_decParams, t_nrLDPC_time_stats* p_profiler); -int32_t nrLDPC_decoder(t_nrLDPC_dec_params* p_decParams, int8_t* p_llr, int8_t* p_out, t_nrLDPC_time_stats* p_profiler) +int32_t nrLDPC_decoder(t_nrLDPC_dec_params* p_decParams, int8_t* p_llr, int8_t* p_out, t_nrLDPC_procBuf* p_procBuf, t_nrLDPC_time_stats* p_profiler) { uint32_t numLLR; uint32_t numIter = 0; t_nrLDPC_lut lut; t_nrLDPC_lut* p_lut = &lut; + printf("p_procBuf->cnProcBuf = %p\n", p_procBuf->cnProcBuf); + // Initialize decoder core(s) with correct LUTs numLLR = nrLDPC_init(p_decParams, p_lut); // Launch LDPC decoder core for one segment - numIter = nrLDPC_decoder_core(p_llr, p_out, numLLR, p_lut, p_decParams, p_profiler); + numIter = nrLDPC_decoder_core(p_llr, p_out, p_procBuf, numLLR, p_lut, p_decParams, p_profiler); return numIter; } @@ -69,7 +74,7 @@ int32_t nrLDPC_decoder(t_nrLDPC_dec_params* p_decParams, int8_t* p_llr, int8_t* \param p_decParams LDPC decoder parameters \param p_profiler LDPC profiler statistics */ -static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_t numLLR, t_nrLDPC_lut* p_lut, t_nrLDPC_dec_params* p_decParams, t_nrLDPC_time_stats* p_profiler) +static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, t_nrLDPC_procBuf* p_procBuf, uint32_t numLLR, t_nrLDPC_lut* p_lut, t_nrLDPC_dec_params* p_decParams, t_nrLDPC_time_stats* p_profiler) { uint16_t Z = p_decParams->Z; uint8_t BG = p_decParams->BG; @@ -90,7 +95,7 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ else { // Use LLR processing buffer as temporary output buffer - p_llrOut = (int8_t*) llrProcBuf; + p_llrOut = p_procBuf->llrProcBuf; } @@ -98,27 +103,27 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->llr2llrProcBuf); #endif - nrLDPC_llr2llrProcBuf(p_lut, p_llr, Z, BG); + nrLDPC_llr2llrProcBuf(p_lut, p_llr, p_procBuf, Z, BG); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->llr2llrProcBuf); #endif #ifdef NR_LDPC_DEBUG_MODE nrLDPC_debug_initBuffer2File(nrLDPC_buffers_LLR_PROC); - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_PROC, p_procBuf); #endif #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->llr2CnProcBuf); #endif - nrLDPC_llr2CnProcBuf(p_lut, p_llr, numLLR, Z, BG); + nrLDPC_llr2CnProcBuf(p_lut, p_llr, p_procBuf, numLLR, Z, BG); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->llr2CnProcBuf); #endif #ifdef NR_LDPC_DEBUG_MODE nrLDPC_debug_initBuffer2File(nrLDPC_buffers_CN_PROC); - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC, p_procBuf); #endif // First iteration @@ -129,11 +134,11 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_cnProc_BG1(p_lut, Z); + nrLDPC_cnProc_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_cnProc_BG2(p_lut, Z); + nrLDPC_cnProc_BG2(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cnProc); @@ -141,7 +146,7 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #ifdef NR_LDPC_DEBUG_MODE nrLDPC_debug_initBuffer2File(nrLDPC_buffers_CN_PROC_RES); - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC_RES, p_procBuf); #endif #ifdef NR_LDPC_PROFILER_DETAIL @@ -149,11 +154,11 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_cn2bnProcBuf_BG1(p_lut, Z); + nrLDPC_cn2bnProcBuf_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_cn2bnProcBuf(p_lut, Z); + nrLDPC_cn2bnProcBuf(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cn2bnProcBuf); @@ -161,34 +166,34 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #ifdef NR_LDPC_DEBUG_MODE nrLDPC_debug_initBuffer2File(nrLDPC_buffers_BN_PROC); - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC, p_procBuf); #endif // BN processing #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->bnProcPc); #endif - nrLDPC_bnProcPc(p_lut, Z); + nrLDPC_bnProcPc(p_lut, p_procBuf, Z); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bnProcPc); #endif #ifdef NR_LDPC_DEBUG_MODE nrLDPC_debug_initBuffer2File(nrLDPC_buffers_LLR_RES); - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_RES, p_procBuf); #endif #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->bnProc); #endif - nrLDPC_bnProc(p_lut, Z); + nrLDPC_bnProc(p_lut, p_procBuf, Z); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bnProc); #endif #ifdef NR_LDPC_DEBUG_MODE nrLDPC_debug_initBuffer2File(nrLDPC_buffers_BN_PROC_RES); - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC_RES, p_procBuf); #endif // BN results to CN processing buffer @@ -197,18 +202,18 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_bn2cnProcBuf_BG1(p_lut, Z); + nrLDPC_bn2cnProcBuf_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_bn2cnProcBuf(p_lut, Z); + nrLDPC_bn2cnProcBuf(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bn2cnProcBuf); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC, p_procBuf); #endif // Parity Check not necessary here since it will fail @@ -228,18 +233,18 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_cnProc_BG1(p_lut, Z); + nrLDPC_cnProc_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_cnProc_BG2(p_lut, Z); + nrLDPC_cnProc_BG2(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cnProc); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC_RES, p_procBuf); #endif // Send CN results back to BNs @@ -248,43 +253,43 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_cn2bnProcBuf_BG1(p_lut, Z); + nrLDPC_cn2bnProcBuf_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_cn2bnProcBuf(p_lut, Z); + nrLDPC_cn2bnProcBuf(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cn2bnProcBuf); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC, p_procBuf); #endif // BN Processing #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->bnProcPc); #endif - nrLDPC_bnProcPc(p_lut, Z); + nrLDPC_bnProcPc(p_lut, p_procBuf, Z); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bnProcPc); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_RES, p_procBuf); #endif #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->bnProc); #endif - nrLDPC_bnProc(p_lut, Z); + nrLDPC_bnProc(p_lut, p_procBuf, Z); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bnProc); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC_RES, p_procBuf); #endif // BN results to CN processing buffer @@ -293,18 +298,18 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_bn2cnProcBuf_BG1(p_lut, Z); + nrLDPC_bn2cnProcBuf_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_bn2cnProcBuf(p_lut, Z); + nrLDPC_bn2cnProcBuf(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bn2cnProcBuf); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC, p_procBuf); #endif // Parity Check @@ -314,11 +319,11 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - pcRes = nrLDPC_cnProcPc_BG1(p_lut, Z); + pcRes = nrLDPC_cnProcPc_BG1(p_lut, p_procBuf, Z); } else { - pcRes = nrLDPC_cnProcPc_BG2(p_lut, Z); + pcRes = nrLDPC_cnProcPc_BG2(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cnProcPc); @@ -339,18 +344,18 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_cnProc_BG1(p_lut, Z); + nrLDPC_cnProc_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_cnProc_BG2(p_lut, Z); + nrLDPC_cnProc_BG2(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cnProc); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC_RES, p_procBuf); #endif // Send CN results back to BNs @@ -359,31 +364,31 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_cn2bnProcBuf_BG1(p_lut, Z); + nrLDPC_cn2bnProcBuf_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_cn2bnProcBuf(p_lut, Z); + nrLDPC_cn2bnProcBuf(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cn2bnProcBuf); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC, p_procBuf); #endif // BN Processing #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->bnProcPc); #endif - nrLDPC_bnProcPc(p_lut, Z); + nrLDPC_bnProcPc(p_lut, p_procBuf, Z); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bnProcPc); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_LLR_RES, p_procBuf); #endif // If parity check not enabled, no need to send the BN proc results @@ -392,13 +397,13 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->bnProc); #endif - nrLDPC_bnProc(p_lut, Z); + nrLDPC_bnProc(p_lut, p_procBuf, Z); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bnProc); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC_RES); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_BN_PROC_RES, p_procBuf); #endif // BN results to CN processing buffer @@ -407,18 +412,18 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - nrLDPC_bn2cnProcBuf_BG1(p_lut, Z); + nrLDPC_bn2cnProcBuf_BG1(p_lut, p_procBuf, Z); } else { - nrLDPC_bn2cnProcBuf(p_lut, Z); + nrLDPC_bn2cnProcBuf(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->bn2cnProcBuf); #endif #ifdef NR_LDPC_DEBUG_MODE - nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC); + nrLDPC_debug_writeBuffer2File(nrLDPC_buffers_CN_PROC, p_procBuf); #endif // Parity Check @@ -427,11 +432,11 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #endif if (BG == 1) { - pcRes = nrLDPC_cnProcPc_BG1(p_lut, Z); + pcRes = nrLDPC_cnProcPc_BG1(p_lut, p_procBuf, Z); } else { - pcRes = nrLDPC_cnProcPc_BG2(p_lut, Z); + pcRes = nrLDPC_cnProcPc_BG2(p_lut, p_procBuf, Z); } #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->cnProcPc); @@ -451,7 +456,7 @@ static inline uint32_t nrLDPC_decoder_core(int8_t* p_llr, int8_t* p_out, uint32_ #ifdef NR_LDPC_PROFILER_DETAIL start_meas(&p_profiler->llrRes2llrOut); #endif - nrLDPC_llrRes2llrOut(p_lut, p_llrOut, numLLR); + nrLDPC_llrRes2llrOut(p_lut, p_llrOut, p_procBuf, numLLR); #ifdef NR_LDPC_PROFILER_DETAIL stop_meas(&p_profiler->llrRes2llrOut); #endif diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.h b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.h index e1dff38723..4d2401db83 100644 --- a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.h +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_decoder.h @@ -32,6 +32,7 @@ #define __NR_LDPC_DECODER__H__ #include "nrLDPC_types.h" +#include "nrLDPC_init_mem.h" /** \brief LDPC decoder @@ -40,6 +41,6 @@ \param p_llrOut Output vector \param p_profiler LDPC profiler statistics */ -int32_t nrLDPC_decoder(t_nrLDPC_dec_params* p_decParams, int8_t* p_llr, int8_t* p_llrOut, t_nrLDPC_time_stats* p_profiler); +int32_t nrLDPC_decoder(t_nrLDPC_dec_params* p_decParams, int8_t* p_llr, int8_t* p_llrOut, t_nrLDPC_procBuf* p_procBuf, t_nrLDPC_time_stats* p_profiler); #endif diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_defs.h b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_defs.h index 6dbd38e681..940d842550 100644 --- a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_defs.h +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_defs.h @@ -100,23 +100,7 @@ #define NR_LDPC_MAX_NUM_LLR 26112 // ============================================================================== -// GLOBAL VARIABLES - -// Aligned on 32 bytes = 256 bits for AVX2 -/** CN processing buffer */ -static int8_t cnProcBuf [NR_LDPC_SIZE_CN_PROC_BUF] __attribute__ ((aligned(32))); -/** Buffer for CN processing results */ -static int8_t cnProcBufRes[NR_LDPC_SIZE_CN_PROC_BUF] __attribute__ ((aligned(32))); - -/** BN processing buffer */ -static int8_t bnProcBuf [NR_LDPC_SIZE_BN_PROC_BUF] __attribute__ ((aligned(32))); -/** Buffer for BN processing results */ -static int8_t bnProcBufRes[NR_LDPC_SIZE_BN_PROC_BUF] __attribute__ ((aligned(32))); - -/** Buffer for LLR results */ -static int8_t llrRes [NR_LDPC_MAX_NUM_LLR] __attribute__ ((aligned(32))); -/** LLR processing buffer */ -static int8_t llrProcBuf[NR_LDPC_MAX_NUM_LLR] __attribute__ ((aligned(32))); +// GLOBAL CONSTANT VARIABLES /** Start addresses for the cnProcBuf for each CN group in BG1*/ static const uint32_t lut_startAddrCnGroups_BG1[NR_LDPC_NUM_CN_GROUPS_BG1] = {0, 1152, 8832, 43392, 61824, 75264, 81408, 88320, 92160}; diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_init_mem.h b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_init_mem.h new file mode 100644 index 0000000000..bdaa118e1f --- /dev/null +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_init_mem.h @@ -0,0 +1,86 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*!\file nrLDPC_init_mem.h + * \brief Defines the function to initialize the LDPC decoder and sets correct LUTs. + * \author Sebastian Wagner (TCL Communications) Email: + * \date 07-12-2018 + * \version 1.0 + * \note + * \warning + */ + +#ifndef __NR_LDPC_INIT_MEM__H__ +#define __NR_LDPC_INIT_MEM__H__ + +#include +#include "nrLDPC_defs.h" +#include "nrLDPC_types.h" + +#ifndef malloc32_clear +/** + \brief Allocates 32 byte aligned memory and initializes to zero + \param size Input size in bytes + \return Pointer to memory +*/ +static inline void* malloc32_clear(size_t size) +{ + void* ptr = (void*) memalign(32, size+32); + memset(ptr, 0, size); + return ptr; +} +#endif + +/** + \brief Allocates and initializes the internal decoder processing buffers + \param p_decParams Pointer to decoder parameters + \param p_lut Pointer to decoder LUTs + \return Number of LLR values +*/ +static inline t_nrLDPC_procBuf* nrLDPC_init_mem(void) +{ + t_nrLDPC_procBuf* p_procBuf = (t_nrLDPC_procBuf*) malloc32_clear(sizeof(t_nrLDPC_procBuf)); + + if (p_procBuf) + { + p_procBuf->cnProcBuf = (int8_t*) malloc32_clear(NR_LDPC_SIZE_CN_PROC_BUF*sizeof(int8_t)); + p_procBuf->cnProcBufRes = (int8_t*) malloc32_clear(NR_LDPC_SIZE_CN_PROC_BUF*sizeof(int8_t)); + p_procBuf->bnProcBuf = (int8_t*) malloc32_clear(NR_LDPC_SIZE_BN_PROC_BUF*sizeof(int8_t)); + p_procBuf->bnProcBufRes = (int8_t*) malloc32_clear(NR_LDPC_SIZE_BN_PROC_BUF*sizeof(int8_t)); + p_procBuf->llrRes = (int8_t*) malloc32_clear(NR_LDPC_MAX_NUM_LLR *sizeof(int8_t)); + p_procBuf->llrProcBuf = (int8_t*) malloc32_clear(NR_LDPC_MAX_NUM_LLR *sizeof(int8_t)); + } + + return(p_procBuf); +} + +static inline void nrLDPC_free_mem(t_nrLDPC_procBuf* p_procBuf) +{ + free(p_procBuf->cnProcBuf); + free(p_procBuf->cnProcBufRes); + free(p_procBuf->bnProcBuf); + free(p_procBuf->bnProcBufRes); + free(p_procBuf->llrRes); + free(p_procBuf->llrProcBuf); + + free(p_procBuf); +} +#endif diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_mPass.h b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_mPass.h index 3e6c4c0c7c..f3431eba24 100644 --- a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_mPass.h +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_mPass.h @@ -41,13 +41,15 @@ \param Z Lifting size \param BG Base graph */ -static inline void nrLDPC_llr2llrProcBuf(t_nrLDPC_lut* p_lut, int8_t* llr, uint16_t Z, uint8_t BG) +static inline void nrLDPC_llr2llrProcBuf(t_nrLDPC_lut* p_lut, int8_t* llr, t_nrLDPC_procBuf* p_procBuf, uint16_t Z, uint8_t BG) { const uint16_t* lut_llr2llrProcBuf = p_lut->llr2llrProcBuf; uint32_t i; const uint8_t numBn2CnG1 = p_lut->numBnInBnGroups[0]; uint32_t colG1 = NR_LDPC_START_COL_PARITY_BG1*Z; + int8_t* llrProcBuf = p_procBuf->llrProcBuf; + if (BG == 2) { colG1 = NR_LDPC_START_COL_PARITY_BG2*Z; @@ -74,11 +76,13 @@ static inline void nrLDPC_llr2llrProcBuf(t_nrLDPC_lut* p_lut, int8_t* llr, uint1 \param Z Lifting size \param BG Base graph */ -static inline void nrLDPC_llr2CnProcBuf(t_nrLDPC_lut* p_lut, int8_t* llr, uint16_t numLLR, uint16_t Z, uint8_t BG) +static inline void nrLDPC_llr2CnProcBuf(t_nrLDPC_lut* p_lut, int8_t* llr, t_nrLDPC_procBuf* p_procBuf, uint16_t numLLR, uint16_t Z, uint8_t BG) { const uint32_t* lut_llr2CnProcBuf = p_lut->llr2CnProcBuf; const uint8_t* lut_numEdgesPerBn = p_lut->numEdgesPerBn; + int8_t* cnProcBuf = p_procBuf->cnProcBuf; + int8_t curLLR; uint8_t numEdges; uint32_t i; @@ -123,12 +127,15 @@ static inline void nrLDPC_llr2CnProcBuf(t_nrLDPC_lut* p_lut, int8_t* llr, uint16 \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_cn2bnProcBuf(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_cn2bnProcBuf(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint32_t* lut_cn2bnProcBuf = p_lut->cn2bnProcBuf; const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBufRes = p_procBuf->cnProcBufRes; + int8_t* bnProcBuf = p_procBuf->bnProcBuf; + const uint32_t* p_lut_cn2bn; int8_t* p_cnProcBufRes; uint32_t bitOffsetInGroup; @@ -245,12 +252,15 @@ static inline void nrLDPC_cn2bnProcBuf(t_nrLDPC_lut* p_lut, uint16_t Z) \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_cn2bnProcBuf_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_cn2bnProcBuf_BG1(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint32_t* lut_cn2bnProcBuf = p_lut->cn2bnProcBuf; const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBufRes = p_procBuf->cnProcBufRes; + int8_t* bnProcBuf = p_procBuf->bnProcBuf; + const uint32_t* p_lut_cn2bn; int8_t* p_cnProcBufRes; uint32_t bitOffsetInGroup; @@ -418,12 +428,15 @@ static inline void nrLDPC_cn2bnProcBuf_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_bn2cnProcBuf(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_bn2cnProcBuf(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint32_t* lut_cn2bnProcBuf = p_lut->cn2bnProcBuf; const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBuf = p_procBuf->cnProcBuf; + int8_t* bnProcBufRes = p_procBuf->bnProcBufRes; + int8_t* p_cnProcBuf; const uint32_t* p_lut_cn2bn; uint32_t bitOffsetInGroup; @@ -543,12 +556,15 @@ static inline void nrLDPC_bn2cnProcBuf(t_nrLDPC_lut* p_lut, uint16_t Z) \param p_lut Pointer to decoder LUTs \param Z Lifting size */ -static inline void nrLDPC_bn2cnProcBuf_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) +static inline void nrLDPC_bn2cnProcBuf_BG1(t_nrLDPC_lut* p_lut, t_nrLDPC_procBuf* p_procBuf, uint16_t Z) { const uint32_t* lut_cn2bnProcBuf = p_lut->cn2bnProcBuf; const uint8_t* lut_numCnInCnGroups = p_lut->numCnInCnGroups; const uint32_t* lut_startAddrCnGroups = p_lut->startAddrCnGroups; + int8_t* cnProcBuf = p_procBuf->cnProcBuf; + int8_t* bnProcBufRes = p_procBuf->bnProcBufRes; + int8_t* p_cnProcBuf; const uint32_t* p_lut_cn2bn; uint32_t bitOffsetInGroup; @@ -720,11 +736,13 @@ static inline void nrLDPC_bn2cnProcBuf_BG1(t_nrLDPC_lut* p_lut, uint16_t Z) \param llrOut Pointer to output LLRs \param numLLR Number of LLR values */ -static inline void nrLDPC_llrRes2llrOut(t_nrLDPC_lut* p_lut, int8_t* llrOut, uint16_t numLLR) +static inline void nrLDPC_llrRes2llrOut(t_nrLDPC_lut* p_lut, int8_t* llrOut, t_nrLDPC_procBuf* p_procBuf, uint16_t numLLR) { const uint16_t* lut_llr2llrProcBuf = p_lut->llr2llrProcBuf; uint32_t i; + int8_t* llrRes = p_procBuf->llrRes; + for (i=0; illrProcBuf, NR_LDPC_MAX_NUM_LLR); break; } case nrLDPC_buffers_CN_PROC: { - nrLDPC_writeFile("cnProcBuf.txt", cnProcBuf, NR_LDPC_SIZE_CN_PROC_BUF); + nrLDPC_writeFile("cnProcBuf.txt", p_procBuf->cnProcBuf, NR_LDPC_SIZE_CN_PROC_BUF); break; } case nrLDPC_buffers_CN_PROC_RES: { - nrLDPC_writeFile("cnProcBufRes.txt", cnProcBufRes, NR_LDPC_SIZE_CN_PROC_BUF); + nrLDPC_writeFile("cnProcBufRes.txt", p_procBuf->cnProcBufRes, NR_LDPC_SIZE_CN_PROC_BUF); break; } case nrLDPC_buffers_BN_PROC: { - nrLDPC_writeFile("bnProcBuf.txt", bnProcBuf, NR_LDPC_SIZE_BN_PROC_BUF); + nrLDPC_writeFile("bnProcBuf.txt", p_procBuf->bnProcBuf, NR_LDPC_SIZE_BN_PROC_BUF); break; } case nrLDPC_buffers_BN_PROC_RES: { - nrLDPC_writeFile("bnProcBufRes.txt", bnProcBufRes, NR_LDPC_SIZE_BN_PROC_BUF); + nrLDPC_writeFile("bnProcBufRes.txt", p_procBuf->bnProcBufRes, NR_LDPC_SIZE_BN_PROC_BUF); break; } case nrLDPC_buffers_LLR_RES: { - nrLDPC_writeFile("llrRes.txt", llrRes, NR_LDPC_MAX_NUM_LLR); + nrLDPC_writeFile("llrRes.txt", p_procBuf->llrRes, NR_LDPC_MAX_NUM_LLR); break; } } diff --git a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_types.h b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_types.h index 3c5128d1a1..ae32f6ffb6 100644 --- a/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_types.h +++ b/openair1/PHY/CODING/nrLDPC_decoder/nrLDPC_types.h @@ -88,4 +88,16 @@ typedef struct nrLDPC_time_stats { time_stats_t total; /**< Statistics for total processing time */ } t_nrLDPC_time_stats; +/** + Structure containing the processing buffers + */ +typedef struct nrLDPC_procBuf { + int8_t* cnProcBuf; /**< CN processing buffer */ + int8_t* cnProcBufRes; /**< Buffer for CN processing results */ + int8_t* bnProcBuf; /**< BN processing buffer */ + int8_t* bnProcBufRes; /**< Buffer for BN processing results */ + int8_t* llrRes; /**< Buffer for LLR results */ + int8_t* llrProcBuf; /**< LLR processing buffer */ +} t_nrLDPC_procBuf; + #endif diff --git a/openair1/PHY/CODING/nrLDPC_encoder/defs.h b/openair1/PHY/CODING/nrLDPC_encoder/defs.h index 629d50c79c..4230056f31 100644 --- a/openair1/PHY/CODING/nrLDPC_encoder/defs.h +++ b/openair1/PHY/CODING/nrLDPC_encoder/defs.h @@ -38,6 +38,7 @@ int encode_parity_check_part_orig(unsigned char *c,unsigned char *d, short BG,sh void encode_parity_check_part_optim(uint8_t *c,uint8_t *d, short BG,short Zc,short Kb); int ldpc_encoder_optim(unsigned char *test_input,unsigned char *channel_input,short block_length,short BG,time_stats_t *tinput,time_stats_t *tprep,time_stats_t *tparity,time_stats_t *toutput); int ldpc_encoder_optim_8seg(unsigned char **test_input,unsigned char **channel_input,short block_length,short BG,int n_segments,time_stats_t *tinput,time_stats_t *tprep,time_stats_t *tparity,time_stats_t *toutput); +int ldpc_encoder_optim_8seg_multi(unsigned char **test_input,unsigned char **channel_input,short block_length, short BG, int n_segments,unsigned int macro_num, time_stats_t *tinput,time_stats_t *tprep,time_stats_t *tparity,time_stats_t *toutput); /*ldpc_generate_coefficient.c*/ int ldpc_encoder_orig(unsigned char *test_input,unsigned char *channel_input,short block_length,short BG,unsigned char gen_code); diff --git a/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder.c b/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder.c index b97b780333..a225730845 100644 --- a/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder.c +++ b/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder.c @@ -68,7 +68,7 @@ int encode_parity_check_part_orig(unsigned char *c,unsigned char *d, short BG,sh printf("problem with BG\n"); return(-1); } - + no_punctured_columns=(int)((nrows-2)*Zc+block_length-block_length*3)/Zc; diff --git a/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder2.c b/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder2.c index 037c19f88e..4466fc8eea 100644 --- a/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder2.c +++ b/openair1/PHY/CODING/nrLDPC_encoder/ldpc_encoder2.c @@ -201,7 +201,7 @@ void encode_parity_check_part_optim(uint8_t *c,uint8_t *d, short BG,short Zc,sho int ldpc_encoder_optim(unsigned char *test_input,unsigned char *channel_input,short block_length,short BG,time_stats_t *tinput,time_stats_t *tprep,time_stats_t *tparity,time_stats_t *toutput) { - short Kb,Zc,nrows,ncols; + short Zc,Kb=0,nrows=0,ncols=0; int i,i1; int no_punctured_columns,removed_bit; @@ -211,26 +211,30 @@ int ldpc_encoder_optim(unsigned char *test_input,unsigned char *channel_input,sh int simd_size; //determine number of bits in codeword - if (BG==1) - { - Kb = 22; - nrows=46; //parity check bits - ncols=22; //info bits - } - else if (BG==2) - { - nrows=42; //parity check bits - ncols=10; // info bits - - if (block_length>640) - Kb = 10; - else if (block_length>560) - Kb = 9; - else if (block_length>192) - Kb = 8; - else - Kb = 6; - } + //if (block_length>3840) + if (BG==1) + { + //BG=1; + Kb = 22; + nrows=46; //parity check bits + ncols=22; //info bits + } + //else if (block_length<=3840) + else if (BG==2) + { + //BG=2; + nrows=42; //parity check bits + ncols=10; // info bits + + if (block_length>640) + Kb = 10; + else if (block_length>560) + Kb = 9; + else if (block_length>192) + Kb = 8; + else + Kb = 6; + } //find minimum value in all sets of lifting size Zc=0; @@ -251,7 +255,7 @@ int ldpc_encoder_optim(unsigned char *test_input,unsigned char *channel_input,sh #endif if ((Zc&31) > 0) simd_size = 16; - else simd_size = 32; + else simd_size = 32; unsigned char c[22*Zc] __attribute__((aligned(32))); //padded input, unpacked, max size unsigned char d[46*Zc] __attribute__((aligned(32))); //coded parity part output, unpacked, max size @@ -316,7 +320,7 @@ int ldpc_encoder_optim(unsigned char *test_input,unsigned char *channel_input,sh int ldpc_encoder_optim_8seg(unsigned char **test_input,unsigned char **channel_input,short block_length,short BG,int n_segments,time_stats_t *tinput,time_stats_t *tprep,time_stats_t *tparity,time_stats_t *toutput) { - short Kb,Zc,nrows,ncols; + short Zc,Kb=0,nrows=0,ncols=0; int i,i1,j; int no_punctured_columns,removed_bit; //Table of possible lifting sizes @@ -343,14 +347,18 @@ int ldpc_encoder_optim_8seg(unsigned char **test_input,unsigned char **channel_i AssertFatal(n_segments>0&&n_segments<=8,"0 < n_segments %d <= 8\n",n_segments); //determine number of bits in codeword + //if (block_length>3840) if (BG==1) { + //BG=1; Kb = 22; nrows=46; //parity check bits ncols=22; //info bits } - else if (BG==2) + //else if (block_length<=3840) + else if (BG==2) { + //BG=2; nrows=42; //parity check bits ncols=10; // info bits @@ -362,7 +370,7 @@ int ldpc_encoder_optim_8seg(unsigned char **test_input,unsigned char **channel_i Kb = 8; else Kb = 6; - } + } //find minimum value in all sets of lifting size Zc=0; @@ -481,12 +489,12 @@ int ldpc_encoder_optim_8seg(unsigned char **test_input,unsigned char **channel_i // if (((block_length-(2*Zc))&31)>0) l1++; for (i=0;i0) l2++; for (i1=0;i1>j)&1; for (i=0;i<((nrows-no_punctured_columns) * Zc-removed_bit);i++) - for (j=0; j>j)&1; } @@ -508,3 +516,217 @@ int ldpc_encoder_optim_8seg(unsigned char **test_input,unsigned char **channel_i stop_meas(toutput); return 0; } + +int ldpc_encoder_optim_8seg_multi(unsigned char **test_input,unsigned char **channel_input,short block_length, short BG, int n_segments,unsigned int macro_num, time_stats_t *tinput,time_stats_t *tprep,time_stats_t *tparity,time_stats_t *toutput) +{ + + short Zc,Kb=0,nrows=0,ncols=0; + int i,i1,j; + int no_punctured_columns,removed_bit; + //Table of possible lifting sizes + short lift_size[51]= {2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,18,20,22,24,26,28,30,32,36,40,44,48,52,56,60,64,72,80,88,96,104,112,120,128,144,160,176,192,208,224,240,256,288,320,352,384}; + char temp; + int simd_size; + int macro_segment, macro_segment_end; + + + macro_segment = 8*macro_num; + macro_segment_end = (n_segments > 8*(macro_num+1)) ? 8*(macro_num+1) : n_segments; + ///printf("macro_segment: %d\n", macro_segment); + ///printf("macro_segment_end: %d\n", macro_segment_end ); + +#ifdef __AVX2__ + __m256i shufmask = _mm256_set_epi64x(0x0303030303030303, 0x0202020202020202,0x0101010101010101, 0x0000000000000000); + __m256i andmask = _mm256_set1_epi64x(0x8040201008040201); // every 8 bits -> 8 bytes, pattern repeats. + __m256i zero256 = _mm256_setzero_si256(); + __m256i masks[8]; + register __m256i c256; + masks[0] = _mm256_set1_epi8(0x1); + masks[1] = _mm256_set1_epi8(0x2); + masks[2] = _mm256_set1_epi8(0x4); + masks[3] = _mm256_set1_epi8(0x8); + masks[4] = _mm256_set1_epi8(0x10); + masks[5] = _mm256_set1_epi8(0x20); + masks[6] = _mm256_set1_epi8(0x40); + masks[7] = _mm256_set1_epi8(0x80); +#endif + + ///AssertFatal(n_segments>0&&n_segments<=8,"0 < n_segments %d <= 8\n",n_segments); + + //determine number of bits in codeword + if (BG==1) + { + BG=1; + Kb = 22; + nrows=46; //parity check bits + ncols=22; //info bits + } + else if (BG==2) + { + BG=2; + nrows=42; //parity check bits + ncols=10; // info bits + + if (block_length>640) + Kb = 10; + else if (block_length>560) + Kb = 9; + else if (block_length>192) + Kb = 8; + else + Kb = 6; + } + + //find minimum value in all sets of lifting size + Zc=0; + for (i1=0; i1 < 51; i1++) + { + if (lift_size[i1] >= (double) block_length/Kb) + { + Zc = lift_size[i1]; + //printf("%d\n",Zc); + break; + } + } + +#ifdef DEBUG_LDPC + LOG_D(PHY,"ldpc_encoder_optim_8seg: BG %d, Zc %d, Kb %d, block_length %d, segments %d\n",BG,Zc,Kb,block_length,n_segments); + LOG_D(PHY,"ldpc_encoder_optim_8seg: PDU (seg 0) %x %x %x %x\n",test_input[0][0],test_input[0][1],test_input[0][2],test_input[0][3]); +#endif + + AssertFatal(Zc>0,"no valid Zc found for block length %d\n",block_length); + + if ((Zc&31) > 0) simd_size = 16; + else simd_size = 32; + + unsigned char c[22*Zc] __attribute__((aligned(32))); //padded input, unpacked, max size + unsigned char d[46*Zc] __attribute__((aligned(32))); //coded parity part output, unpacked, max size + + unsigned char c_extension[2*22*Zc*simd_size] __attribute__((aligned(32))); //double size matrix of c + + // calculate number of punctured bits + no_punctured_columns=(int)((nrows-2)*Zc+block_length-block_length*3)/Zc; + removed_bit=(nrows-no_punctured_columns-2) * Zc+block_length-(int)(block_length*3); + // printf("%d\n",no_punctured_columns); + // printf("%d\n",removed_bit); + // unpack input + memset(c,0,sizeof(unsigned char) * ncols * Zc); + memset(d,0,sizeof(unsigned char) * nrows * Zc); + + start_meas(tinput); +#if 0 + for (i=0; i>(i&7); + //printf("c(%d,%d)=%d\n",j,i,temp); + c[i] |= (temp << (j-macro_segment)); + } + } +#else +#ifdef __AVX2__ + for (i=0; i>5; i++) { + c256 = _mm256_and_si256(_mm256_cmpeq_epi8(_mm256_andnot_si256(_mm256_shuffle_epi8(_mm256_set1_epi32(((uint32_t*)test_input[macro_segment])[i]), shufmask),andmask),zero256),masks[0]); + //for (j=1; j>5)<<5;i>(i&7); + //printf("c(%d,%d)=%d\n",j,i,temp); + c[i] |= (temp << (j-macro_segment)); + } + } +#else + AssertFatal(1==0,"Need AVX2 for this\n"); +#endif +#endif + + stop_meas(tinput); + + if ((BG==1 && Zc>176) || (BG==2 && Zc>64)) { + // extend matrix + start_meas(tprep); + for (i1=0; i1 < ncols; i1++) + { + memcpy(&c_extension[2*i1*Zc], &c[i1*Zc], Zc*sizeof(unsigned char)); + memcpy(&c_extension[(2*i1+1)*Zc], &c[i1*Zc], Zc*sizeof(unsigned char)); + } + for (i1=1;i1>5; + uint32_t l2 = ((nrows-no_punctured_columns) * Zc-removed_bit)>>5; + __m256i *c256p = (__m256i *)&c[2*Zc]; + __m256i *d256p = (__m256i *)&d[0]; + // if (((block_length-(2*Zc))&31)>0) l1++; + + for (i=0;i0) l2++; + + for (i1=0;i1>(j-macro_segment))&1; + for (i=0;i<((nrows-no_punctured_columns) * Zc-removed_bit);i++) + //for (j=0; j>(j-macro_segment))&1; + } + +#else + AssertFatal(1==0,"Need AVX2 for now\n"); +#endif + + stop_meas(toutput); + return 0; +} + diff --git a/openair1/PHY/CODING/nrLDPC_encoder/ldpc_generate_coefficient.c b/openair1/PHY/CODING/nrLDPC_encoder/ldpc_generate_coefficient.c index 93edc734de..13318c9d5d 100644 --- a/openair1/PHY/CODING/nrLDPC_encoder/ldpc_generate_coefficient.c +++ b/openair1/PHY/CODING/nrLDPC_encoder/ldpc_generate_coefficient.c @@ -361,13 +361,19 @@ short *choose_generator_matrix(short BG,short Zc) return Gen_shift_values; } -int ldpc_encoder_orig(unsigned char *test_input,unsigned char *channel_input,short block_length,short BG,unsigned char gen_code) +int ldpc_encoder_orig(unsigned char *test_input,unsigned char *channel_input,short block_length, short BG,unsigned char gen_code) { unsigned char c[22*384]; //padded input, unpacked, max size unsigned char d[68*384]; //coded output, unpacked, max size unsigned char channel_temp,temp; short *Gen_shift_values, *no_shift_values, *pointer_shift_values; - short Zc,Kb,nrows,ncols; + short Zc; + //initialize for BG == 1 + short Kb = 22; + short nrows = 46;//parity check bits + short ncols = 22;//info bits + + int i,i1,i2,i3,i4,i5,temp_prime,var; int no_punctured_columns,removed_bit; //Table of possible lifting sizes @@ -378,26 +384,30 @@ int ldpc_encoder_orig(unsigned char *test_input,unsigned char *channel_input,sho int indlist2[1000]; //determine number of bits in codeword - if (BG==1) - { - Kb = 22; - nrows=46; //parity check bits - ncols=22; //info bits - } - else if (BG==2) - { - nrows=42; //parity check bits - ncols=10; // info bits + //if (block_length>3840) + if (BG==1) + { + //BG=1; + Kb = 22; + nrows=46; //parity check bits + ncols=22; //info bits + } + //else if (block_length<=3840) + else if (BG==2) + { + //BG=2; + nrows=42; //parity check bits + ncols=10; // info bits - if (block_length>640) - Kb = 10; - else if (block_length>560) - Kb = 9; - else if (block_length>192) - Kb = 8; - else - Kb = 6; - } + if (block_length>640) + Kb = 10; + else if (block_length>560) + Kb = 9; + else if (block_length>192) + Kb = 8; + else + Kb = 6; + } //find minimum value in all sets of lifting size Zc=0; @@ -415,8 +425,6 @@ int ldpc_encoder_orig(unsigned char *test_input,unsigned char *channel_input,sho return(-1); } - int K = ncols*Zc; - Gen_shift_values=choose_generator_matrix(BG,Zc); if (Gen_shift_values==NULL) { printf("ldpc_encoder_orig: could not find generator matrix\n"); @@ -436,17 +444,14 @@ int ldpc_encoder_orig(unsigned char *test_input,unsigned char *channel_input,sho no_shift_values=(short *) no_shift_values_BG2; pointer_shift_values=(short *) pointer_shift_values_BG2; } + else { + AssertFatal(0,"BG %d is not supported yet\n",BG); + } no_punctured_columns=(int)((nrows-2)*Zc+block_length-block_length*3)/Zc; - //nrows - no_punctured_columns = 2 +2*block_length/Zc removed_bit=(nrows-no_punctured_columns-2) * Zc+block_length-(block_length*3); - // ((nrows-no_punctured_columns) * Zc-removed_bit) = - // 2Zc + 2*block_length //printf("%d\n",no_punctured_columns); //printf("%d\n",removed_bit); - //printf("%d\n",nrows-no_punctured_columns); - //printf("%d\n",((nrows-no_punctured_columns) * Zc-removed_bit)); - // unpack input memset(c,0,sizeof(unsigned char) * ncols * Zc); memset(d,0,sizeof(unsigned char) * nrows * Zc); @@ -614,8 +619,8 @@ int ldpc_encoder_orig(unsigned char *test_input,unsigned char *channel_input,sho } // information part and puncture columns - memcpy(&channel_input[0], &c[2*Zc], (block_length-2*Zc)*sizeof(unsigned char)); //systematic bits - memcpy(&channel_input[block_length-2*Zc], &d[0], ((nrows-no_punctured_columns) * Zc-removed_bit)*sizeof(unsigned char)); //systematic bits 2Zc + 2*block_length + memcpy(&channel_input[0], &c[2*Zc], (block_length-2*Zc)*sizeof(unsigned char)); + memcpy(&channel_input[block_length-2*Zc], &d[0], ((nrows-no_punctured_columns) * Zc-removed_bit)*sizeof(unsigned char)); //memcpy(channel_input,c,Kb*Zc*sizeof(unsigned char)); return 0; } diff --git a/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoder.c b/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoder.c index 2c9f8ad985..305a9c6266 100644 --- a/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoder.c +++ b/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoder.c @@ -37,7 +37,7 @@ */ #include "PHY/CODING/nrPolar_tools/nr_polar_defs.h" - +#include "assertions.h" int8_t polar_decoder( double *input, @@ -48,11 +48,11 @@ int8_t polar_decoder( { //Assumes no a priori knowledge. - uint8_t ***bit = nr_alloc_uint8_t_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **bitUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True - uint8_t **llrUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t ***bit = nr_alloc_uint8_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); + uint8_t **bitUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t **llrUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True double ***llr = nr_alloc_double_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **crcChecksum = nr_alloc_uint8_t_2D_array(polarParams->crcParityBits, 2*listSize); + uint8_t **crcChecksum = nr_alloc_uint8_2D_array(polarParams->crcParityBits, 2*listSize); double *pathMetric = malloc(sizeof(double)*(2*listSize)); uint8_t *crcState = malloc(sizeof(uint8_t)*(2*listSize)); //0=False, 1=True @@ -231,9 +231,9 @@ int8_t polar_decoder( free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); return(-1); } @@ -256,6 +256,7 @@ int8_t polar_decoder( //Deinterleaving (ĉ to b) nr_polar_deinterleaver(polarParams->nr_polar_CPrime, polarParams->nr_polar_B, polarParams->interleaving_pattern, polarParams->K); + //Remove the CRC (â) for (int j = 0; j < polarParams->payloadBits; j++) polarParams->nr_polar_A[j]=polarParams->nr_polar_B[j]; @@ -266,11 +267,11 @@ int8_t polar_decoder( free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); - nr_free_uint8_t_2D_array(extended_crc_generator_matrix, polarParams->K); - nr_free_uint8_t_2D_array(tempECGM, polarParams->K); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(extended_crc_generator_matrix, polarParams->K); + nr_free_uint8_2D_array(tempECGM, polarParams->K); /* * Return bits. @@ -280,17 +281,17 @@ int8_t polar_decoder( } int8_t polar_decoder_aPriori(double *input, - uint32_t *out, - t_nrPolar_paramsPtr polarParams, - uint8_t listSize, - uint8_t pathMetricAppr, - double *aPrioriPayload) + uint32_t *out, + t_nrPolar_paramsPtr polarParams, + uint8_t listSize, + uint8_t pathMetricAppr, + double *aPrioriPayload) { - uint8_t ***bit = nr_alloc_uint8_t_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **bitUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True - uint8_t **llrUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t ***bit = nr_alloc_uint8_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); + uint8_t **bitUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t **llrUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True double ***llr = nr_alloc_double_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **crcChecksum = nr_alloc_uint8_t_2D_array(polarParams->crcParityBits, 2*listSize); + uint8_t **crcChecksum = nr_alloc_uint8_2D_array(polarParams->crcParityBits, 2*listSize); double *pathMetric = malloc(sizeof(double)*(2*listSize)); uint8_t *crcState = malloc(sizeof(uint8_t)*(2*listSize)); //0=False, 1=True @@ -483,9 +484,9 @@ int8_t polar_decoder_aPriori(double *input, free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); return(-1); } @@ -509,8 +510,8 @@ int8_t polar_decoder_aPriori(double *input, nr_polar_deinterleaver(polarParams->nr_polar_CPrime, polarParams->nr_polar_B, polarParams->interleaving_pattern, polarParams->K); //Remove the CRC (â) - for (int j = 0; j < polarParams->payloadBits; j++) polarParams->nr_polar_A[j]=polarParams->nr_polar_B[j]; - + for (int j = 0; j < polarParams->payloadBits; j++) + polarParams->nr_polar_A[j]=polarParams->nr_polar_B[j]; break; } } @@ -518,16 +519,17 @@ int8_t polar_decoder_aPriori(double *input, free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); - nr_free_uint8_t_2D_array(extended_crc_generator_matrix, polarParams->K); - nr_free_uint8_t_2D_array(tempECGM, polarParams->K); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(extended_crc_generator_matrix, polarParams->K); + nr_free_uint8_2D_array(tempECGM, polarParams->K); /* * Return bits. */ nr_byte2bit_uint8_32_t(polarParams->nr_polar_A, polarParams->payloadBits, out); + return(0); } @@ -544,11 +546,11 @@ int8_t polar_decoder_aPriori_timing(double *input, FILE* logFile) { - uint8_t ***bit = nr_alloc_uint8_t_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **bitUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True - uint8_t **llrUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t ***bit = nr_alloc_uint8_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); + uint8_t **bitUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t **llrUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True double ***llr = nr_alloc_double_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **crcChecksum = nr_alloc_uint8_t_2D_array(polarParams->crcParityBits, 2*listSize); + uint8_t **crcChecksum = nr_alloc_uint8_2D_array(polarParams->crcParityBits, 2*listSize); double *pathMetric = malloc(sizeof(double)*(2*listSize)); uint8_t *crcState = malloc(sizeof(uint8_t)*(2*listSize)); //0=False, 1=True @@ -740,9 +742,9 @@ int8_t polar_decoder_aPriori_timing(double *input, free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); return(-1); } @@ -775,11 +777,11 @@ int8_t polar_decoder_aPriori_timing(double *input, free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); - nr_free_uint8_t_2D_array(extended_crc_generator_matrix, polarParams->K); - nr_free_uint8_t_2D_array(tempECGM, polarParams->K); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(extended_crc_generator_matrix, polarParams->K); + nr_free_uint8_2D_array(tempECGM, polarParams->K); /* * Return bits. @@ -797,11 +799,11 @@ int8_t polar_decoder_dci(double *input, uint16_t n_RNTI) { - uint8_t ***bit = nr_alloc_uint8_t_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **bitUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True - uint8_t **llrUpdated = nr_alloc_uint8_t_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t ***bit = nr_alloc_uint8_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); + uint8_t **bitUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True + uint8_t **llrUpdated = nr_alloc_uint8_2D_array(polarParams->N, (polarParams->n+1)); //0=False, 1=True double ***llr = nr_alloc_double_3D_array(polarParams->N, (polarParams->n+1), 2*listSize); - uint8_t **crcChecksum = nr_alloc_uint8_t_2D_array(polarParams->crcParityBits, 2*listSize); + uint8_t **crcChecksum = nr_alloc_uint8_2D_array(polarParams->crcParityBits, 2*listSize); double *pathMetric = malloc(sizeof(double)*(2*listSize)); uint8_t *crcState = malloc(sizeof(uint8_t)*(2*listSize)); //0=False, 1=True uint8_t extended_crc_scrambling_pattern[polarParams->crcParityBits]; @@ -989,9 +991,9 @@ int8_t polar_decoder_dci(double *input, free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); return(-1); } @@ -1024,11 +1026,11 @@ int8_t polar_decoder_dci(double *input, free(d_tilde); free(pathMetric); free(crcState); - nr_free_uint8_t_3D_array(bit, polarParams->N, (polarParams->n+1)); + nr_free_uint8_3D_array(bit, polarParams->N, (polarParams->n+1)); nr_free_double_3D_array(llr, polarParams->N, (polarParams->n+1)); - nr_free_uint8_t_2D_array(crcChecksum, polarParams->crcParityBits); - nr_free_uint8_t_2D_array(extended_crc_generator_matrix, polarParams->K); - nr_free_uint8_t_2D_array(tempECGM, polarParams->K); + nr_free_uint8_2D_array(crcChecksum, polarParams->crcParityBits); + nr_free_uint8_2D_array(extended_crc_generator_matrix, polarParams->K); + nr_free_uint8_2D_array(tempECGM, polarParams->K); /* * Return bits. @@ -1037,49 +1039,130 @@ int8_t polar_decoder_dci(double *input, return(0); } +void init_polar_deinterleaver_table(t_nrPolar_params *polarParams) { -int8_t polar_decoder_int16(int16_t *input, - uint8_t *out, - t_nrPolar_params *polarParams) + AssertFatal(polarParams->K > 32, "K = %d < 33, is not supported yet\n",polarParams->K); + AssertFatal(polarParams->K < 129, "K = %d > 128, is not supported yet\n",polarParams->K); + + int bit_i,ip,ipmod64; + + int numbytes = polarParams->K>>3; + int residue = polarParams->K&7; + int numbits; + if (residue>0) numbytes++; + for (int byte=0;byteK>>3)) numbits=8; + else numbits=residue; + for (int i=0;iK - 1 - polarParams->interleaving_pattern[(8*byte)+i]; +#if 0 + printf("byte %d, i %d => ip %d\n",byte,i,ip); +#endif + ipmod64 = ip&63; + AssertFatal(ip<128,"ip = %d\n",ip); + for (int val=0;val<256;val++) { + bit_i=(val>>i)&1; + if (ip<64) polarParams->B_tab0[byte][val] |= (((uint64_t)bit_i)<B_tab1[byte][val] |= (((uint64_t)bit_i)<N];// = malloc(sizeof(double) * polarParams->N); + nr_polar_rate_matching_int16(input, d_tilde, polarParams->rate_matching_pattern, polarParams->K, polarParams->N, polarParams->encoderLength); for (int i=0;iN;i++) { if (d_tilde[i]<-128) d_tilde[i]=-128; else if (d_tilde[i]>127) d_tilde[i]=128; } + memcpy((void*)&polarParams->tree.root->alpha[0],(void*)&d_tilde[0],sizeof(int16_t)*polarParams->N); - /* - * SCL polar decoder. - */ - - generic_polar_decoder(polarParams,polarParams->tree.root); //Extract the information bits (û to ĉ) - nr_polar_info_bit_extraction(polarParams->nr_polar_U, polarParams->nr_polar_CPrime, polarParams->information_bit_pattern, polarParams->N); + uint64_t Cprime[4]={0,0,0,0}; + uint64_t B[4]={0,0,0,0}; + for (int i=0;iK;i++) Cprime[i>>6] = Cprime[i>>6] | ((uint64_t)polarParams->nr_polar_U[polarParams->Q_I_N[i]])<<(i&63); //Deinterleaving (ĉ to b) - - nr_polar_deinterleaver(polarParams->nr_polar_CPrime, polarParams->nr_polar_B, polarParams->interleaving_pattern, polarParams->K); - - //Remove the CRC (â) - //for (int j = 0; j < polarParams->payloadBits; j++) polarParams->nr_polar_A[j]=polarParams->nr_polar_B[j]; - - // Check the CRC - for (int j=0;jcrcParityBits;j++) { - int crcbit=0; - for (int i=0;ipayloadBits;i++) - crcbit = crcbit ^ (polarParams->crc_generator_matrix[i][j] & polarParams->nr_polar_B[i]); - if (crcbit != polarParams->nr_polar_B[polarParams->payloadBits+j]) return(-1); + uint8_t *Cprimebyte = (uint8_t*)Cprime; + if (polarParams->K<65) { + B[0] = polarParams->B_tab0[0][Cprimebyte[0]] | + polarParams->B_tab0[1][Cprimebyte[1]] | + polarParams->B_tab0[2][Cprimebyte[2]] | + polarParams->B_tab0[3][Cprimebyte[3]] | + polarParams->B_tab0[4][Cprimebyte[4]] | + polarParams->B_tab0[5][Cprimebyte[5]] | + polarParams->B_tab0[6][Cprimebyte[6]] | + polarParams->B_tab0[7][Cprimebyte[7]]; } - // pack into ceil(payloadBits/32) 32 bit words, lowest index in MSB - // nr_byte2bit_uint8_32_t(polarParams->nr_polar_A, polarParams->payloadBits, out); - nr_byte2bit_uint8_32_t(polarParams->nr_polar_B, polarParams->payloadBits, (unsigned int *)out); - return(0); + else if (polarParams->K<129) { + int len = polarParams->K/8; + if ((polarParams->K&7) > 0) len++; + for (int k=0;kB_tab0[k][Cprimebyte[k]]; + B[1] |= polarParams->B_tab1[k][Cprimebyte[k]]; + } + } + + + int len=polarParams->payloadBits; + //int len_mod64=len&63; + int crclen = polarParams->crcParityBits; + uint64_t rxcrc=B[0]&((1<>crclen); + uint8_t A32_flip[4]; + uint32_t Aprime= (uint32_t)(Ar<<(32-len)); + A32_flip[0]=((uint8_t*)&Aprime)[3]; + A32_flip[1]=((uint8_t*)&Aprime)[2]; + A32_flip[2]=((uint8_t*)&Aprime)[1]; + A32_flip[3]=((uint8_t*)&Aprime)[0]; + crc = (uint64_t)(crc24c(A32_flip,len)>>8); + } + else if (len<=64) { + Ar = (B[0]>>crclen) | (B[1]<<(64-crclen));; + uint8_t A64_flip[8]; + uint64_t Aprime= (uint32_t)(Ar<<(64-len)); + A64_flip[0]=((uint8_t*)&Aprime)[7]; + A64_flip[1]=((uint8_t*)&Aprime)[6]; + A64_flip[2]=((uint8_t*)&Aprime)[5]; + A64_flip[3]=((uint8_t*)&Aprime)[4]; + A64_flip[4]=((uint8_t*)&Aprime)[3]; + A64_flip[5]=((uint8_t*)&Aprime)[2]; + A64_flip[6]=((uint8_t*)&Aprime)[1]; + A64_flip[7]=((uint8_t*)&Aprime)[0]; + crc = (uint64_t)(crc24c(A64_flip,len)>>8); + } + +#if 0 + printf("A %llx B %llx|%llx Cprime %llx|%llx (crc %x,rxcrc %llx %d)\n", + Ar, + B[1],B[0],Cprime[1],Cprime[0],crc, + rxcrc,polarParams->payloadBits); +#endif + + out[0]=Ar; + + + return(crc^rxcrc); + + + } diff --git a/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoding_tools.c b/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoding_tools.c index 1967d3934c..bec322b946 100644 --- a/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoding_tools.c +++ b/openair1/PHY/CODING/nrPolar_tools/nr_polar_decoding_tools.c @@ -34,18 +34,18 @@ #include "PHY/sse_intrin.h" #include "PHY/impl_defs_top.h" -//#define DEBUG_NEW_IMPL +//#define DEBUG_NEW_IMPL 1 void updateLLR(double ***llr, - uint8_t **llrU, - uint8_t ***bit, - uint8_t **bitU, - uint8_t listSize, - uint16_t row, - uint16_t col, - uint16_t xlen, - uint8_t ylen, - uint8_t approximation) + uint8_t **llrU, + uint8_t ***bit, + uint8_t **bitU, + uint8_t listSize, + uint16_t row, + uint16_t col, + uint16_t xlen, + uint8_t ylen, + uint8_t approximation) { uint16_t offset = (xlen/(pow(2,(ylen-col-1)))); for (uint8_t i=0; itree.num_nodes,Nv,level); - pp->tree.num_nodes++; #endif + pp->tree.num_nodes++; if (level==0) { #ifdef DEBUG_NEW_IMPL printf("leaf %d (%s)\n",first_leaf_index,pp->information_bit_pattern[first_leaf_index]==1 ? "information or crc" : "frozen"); @@ -233,16 +233,19 @@ decoder_node_t *add_nodes(int level,int first_leaf_index,t_nrPolar_params *pp) { for (int i=0;iinformation_bit_pattern[i+first_leaf_index]>0) all_frozen_below=0; } -if (all_frozen_below==0) new_node->left=add_nodes(level-1,first_leaf_index,pp); - else { + if (all_frozen_below==0) new_node->left=add_nodes(level-1,first_leaf_index,pp); + else { #ifdef DEBUG_NEW_IMPL - printf("aggregating frozen bits %d ... %d at level %d (%s)\n",first_leaf_index,first_leaf_index+Nv-1,level,((first_leaf_index/Nv)&1)==0?"left":"right"); + printf("aggregating frozen bits %d ... %d at level %d (%s)\n",first_leaf_index,first_leaf_index+Nv-1,level,((first_leaf_index/Nv)&1)==0?"left":"right"); #endif new_node->leaf=1; new_node->all_frozen=1; } if (all_frozen_below==0) new_node->right=add_nodes(level-1,first_leaf_index+(Nv/2),pp); +#ifdef DEBUG_NEW_IMPL + printf("new_node (%d): first_leaf_index %d, left %p, right %p\n",Nv,first_leaf_index,new_node->left,new_node->right); +#endif return(new_node); } @@ -251,7 +254,9 @@ void build_decoder_tree(t_nrPolar_params *pp) { pp->tree.num_nodes=0; pp->tree.root = add_nodes(pp->n,0,pp); - +#ifdef DEBUG_NEW_IMPL + printf("root : left %p, right %p\n",pp->tree.root->left,pp->tree.root->right); +#endif } #if defined(__arm__) || defined(__aarch64__) diff --git a/openair1/PHY/CODING/nrPolar_tools/nr_polar_defs.h b/openair1/PHY/CODING/nrPolar_tools/nr_polar_defs.h index 30584c7ddc..68e9ef3571 100644 --- a/openair1/PHY/CODING/nrPolar_tools/nr_polar_defs.h +++ b/openair1/PHY/CODING/nrPolar_tools/nr_polar_defs.h @@ -78,54 +78,62 @@ typedef struct decoder_tree_t_s { } decoder_tree_t; struct nrPolar_params { - //messageType: 0=PBCH, 1=DCI, -1=UCI - int idx; //idx = (messageType * messageLength * aggregation_prime); - struct nrPolar_params *nextPtr; - - uint8_t n_max; - uint8_t i_il; - uint8_t i_seg; - uint8_t n_pc; - uint8_t n_pc_wm; - uint8_t i_bil; - uint16_t payloadBits; - uint16_t encoderLength; - uint8_t crcParityBits; - uint8_t crcCorrectionBits; - uint16_t K; - uint16_t N; - uint8_t n; - uint32_t crcBit; - - uint16_t *interleaving_pattern; - uint16_t *rate_matching_pattern; - const uint16_t *Q_0_Nminus1; - int16_t *Q_I_N; - int16_t *Q_F_N; - int16_t *Q_PC_N; - uint8_t *information_bit_pattern; - uint16_t *channel_interleaver_pattern; - uint32_t crc_polynomial; - - uint8_t **crc_generator_matrix; //G_P - uint8_t **G_N; - uint32_t* crc256Table; + //messageType: 0=PBCH, 1=DCI, -1=UCI + int idx; //idx = (messageType * messageLength * aggregation_prime); + struct nrPolar_params *nextPtr; + + uint8_t n_max; + uint8_t i_il; + uint8_t i_seg; + uint8_t n_pc; + uint8_t n_pc_wm; + uint8_t i_bil; + uint16_t payloadBits; + uint16_t encoderLength; + uint8_t crcParityBits; + uint8_t crcCorrectionBits; + uint16_t K; + uint16_t N; + uint8_t n; + uint32_t crcBit; + + uint16_t *interleaving_pattern; + uint16_t *deinterleaving_pattern; + uint16_t *rate_matching_pattern; + const uint16_t *Q_0_Nminus1; + int16_t *Q_I_N; + int16_t *Q_F_N; + int16_t *Q_PC_N; + uint8_t *information_bit_pattern; + uint16_t *channel_interleaver_pattern; + uint32_t crc_polynomial; + + uint8_t **crc_generator_matrix; //G_P + uint8_t **G_N; + uint64_t **G_N_tab; + int groupsize; + int *rm_tab; + uint64_t cprime_tab0[32][256]; + uint64_t cprime_tab1[32][256]; + uint64_t B_tab0[32][256]; + uint64_t B_tab1[32][256]; + uint32_t* crc256Table; uint8_t **extended_crc_generator_matrix; - //lowercase: bits, Uppercase: Bits stored in bytes - //polar_encoder vectors - uint8_t *nr_polar_crc; - uint8_t *nr_polar_aPrime; - uint8_t *nr_polar_APrime; - uint8_t *nr_polar_D; - uint8_t *nr_polar_E; - - //Polar Coding vectors - uint8_t *nr_polar_A; - uint8_t *nr_polar_CPrime; - uint8_t *nr_polar_B; - uint8_t *nr_polar_U; - -decoder_tree_t tree; + //lowercase: bits, Uppercase: Bits stored in bytes + //polar_encoder vectors + uint8_t *nr_polar_crc; + uint8_t *nr_polar_aPrime; + uint8_t *nr_polar_APrime; + uint8_t *nr_polar_D; + uint8_t *nr_polar_E; + + //Polar Coding vectors + uint8_t *nr_polar_A; + uint8_t *nr_polar_CPrime; + uint8_t *nr_polar_B; + uint8_t *nr_polar_U; + + decoder_tree_t tree; } __attribute__ ((__packed__)); typedef struct nrPolar_params t_nrPolar_params; typedef t_nrPolar_params *t_nrPolar_paramsPtr; @@ -139,11 +147,10 @@ void polar_encoder_dci(uint32_t *in, t_nrPolar_paramsPtr polarParams, uint16_t n_RNTI); -void polar_encoder_timing(uint32_t *in, - uint32_t *out, - t_nrPolar_paramsPtr polarParams, - double cpuFreqGHz, - FILE* logFile); +void polar_encoder_fast(uint64_t *A, + uint32_t *out, + int32_t crcmask, + t_nrPolar_paramsPtr polarParams); int8_t polar_decoder(double *input, uint8_t *output, @@ -151,6 +158,10 @@ int8_t polar_decoder(double *input, uint8_t listSize, uint8_t pathMetricAppr); +uint32_t polar_decoder_int16(int16_t *input, + uint64_t *out, + t_nrPolar_params *polarParams); + int8_t polar_decoder_aPriori(double *input, uint32_t *output, t_nrPolar_paramsPtr polarParams, @@ -174,7 +185,12 @@ int8_t polar_decoder_dci(double *input, uint8_t pathMetricAppr, uint16_t n_RNTI); -void generic_polar_decoder(t_nrPolar_params *,decoder_node_t *); +void generic_polar_decoder(t_nrPolar_params *, + decoder_node_t *); + +void build_decoder_tree(t_nrPolar_params *pp); +void build_polar_tables(t_nrPolar_paramsPtr polarParams); +void init_polar_deinterleaver_table(t_nrPolar_params *polarParams); void nr_polar_init(t_nrPolar_paramsPtr *polarParams, int8_t messageType, @@ -200,153 +216,168 @@ const uint16_t* nr_polar_sequence_pattern(uint8_t n); * @param E * @param n_max */ uint32_t nr_polar_output_length(uint16_t K, - uint16_t E, - uint8_t n_max); + uint16_t E, + uint8_t n_max); void nr_polar_channel_interleaver_pattern(uint16_t *cip, - uint8_t I_BIL, - uint16_t E); + uint8_t I_BIL, + uint16_t E); void nr_polar_rate_matching_pattern(uint16_t *rmp, - uint16_t *J, - const uint8_t *P_i_, - uint16_t K, - uint16_t N, - uint16_t E); + uint16_t *J, + const uint8_t *P_i_, + uint16_t K, + uint16_t N, + uint16_t E); void nr_polar_rate_matching(double *input, - double *output, - uint16_t *rmp, - uint16_t K, - uint16_t N, - uint16_t E); + double *output, + uint16_t *rmp, + uint16_t K, + uint16_t N, + uint16_t E); -void nr_polar_rate_matching_int16(int16_t *input, int16_t *output, uint16_t *rmp, uint16_t K, uint16_t N, uint16_t E); +void nr_polar_rate_matching_int16(int16_t *input, + int16_t *output, + uint16_t *rmp, + uint16_t K, + uint16_t N, + uint16_t E); void nr_polar_interleaving_pattern(uint16_t K, - uint8_t I_IL, - uint16_t *PI_k_); + uint8_t I_IL, + uint16_t *PI_k_); void nr_polar_info_bit_pattern(uint8_t *ibp, - int16_t *Q_I_N, - int16_t *Q_F_N, - uint16_t *J, - const uint16_t *Q_0_Nminus1, - uint16_t K, - uint16_t N, - uint16_t E, - uint8_t n_PC); + int16_t *Q_I_N, + int16_t *Q_F_N, + uint16_t *J, + const uint16_t *Q_0_Nminus1, + uint16_t K, + uint16_t N, + uint16_t E, + uint8_t n_PC); void nr_polar_info_bit_extraction(uint8_t *input, - uint8_t *output, - uint8_t *pattern, - uint16_t size); + uint8_t *output, + uint8_t *pattern, + uint16_t size); void nr_bit2byte_uint32_8_t(uint32_t *in, - uint16_t arraySize, - uint8_t *out); + uint16_t arraySize, + uint8_t *out); void nr_byte2bit_uint8_32_t(uint8_t *in, - uint16_t arraySize, - uint32_t *out); + uint16_t arraySize, + uint32_t *out); void nr_crc_bit2bit_uint32_8_t(uint32_t *in, - uint16_t arraySize, - uint8_t *out); + uint16_t arraySize, + uint8_t *out); void nr_polar_bit_insertion(uint8_t *input, - uint8_t *output, - uint16_t N, - uint16_t K, - int16_t *Q_I_N, - int16_t *Q_PC_N, - uint8_t n_PC); + uint8_t *output, + uint16_t N, + uint16_t K, + int16_t *Q_I_N, + int16_t *Q_PC_N, + uint8_t n_PC); -void nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(uint8_t *matrix1, - uint8_t **matrix2, - uint8_t *output, - uint16_t row, - uint16_t col); +void nr_matrix_multiplication_uint8_1D_uint8_2D(uint8_t *matrix1, + uint8_t **matrix2, + uint8_t *output, + uint16_t row, + uint16_t col); -uint8_t ***nr_alloc_uint8_t_3D_array(uint16_t xlen, - uint16_t ylen, - uint16_t zlen); +uint8_t ***nr_alloc_uint8_3D_array(uint16_t xlen, + uint16_t ylen, + uint16_t zlen); -uint8_t **nr_alloc_uint8_t_2D_array(uint16_t xlen, - uint16_t ylen); +uint8_t **nr_alloc_uint8_2D_array(uint16_t xlen, + uint16_t ylen); double ***nr_alloc_double_3D_array(uint16_t xlen, - uint16_t ylen, - uint16_t zlen); + uint16_t ylen, + uint16_t zlen); -void nr_free_uint8_t_3D_array(uint8_t ***input, - uint16_t xlen, - uint16_t ylen); - -void nr_free_uint8_t_2D_array(uint8_t **input, - uint16_t xlen); +double **nr_alloc_double_2D_array(uint16_t xlen, + uint16_t ylen); void nr_free_double_3D_array(double ***input, - uint16_t xlen, - uint16_t ylen); + uint16_t xlen, + uint16_t ylen); + +void nr_free_double_2D_array(double **input, + uint16_t xlen); + +void nr_free_uint8_3D_array(uint8_t ***input, + uint16_t xlen, + uint16_t ylen); + +void nr_free_uint8_2D_array(uint8_t **input, + uint16_t xlen); + +void nr_sort_asc_double_1D_array_ind(double *matrix, + uint8_t *ind, + uint8_t len); + +void nr_sort_asc_int16_1D_array_ind(int32_t *matrix, + int *ind, + int len); void nr_free_double_2D_array(double **input, uint16_t xlen); void updateLLR(double ***llr, - uint8_t **llrU, - uint8_t ***bit, - uint8_t **bitU, - uint8_t listSize, - uint16_t row, - uint16_t col, - uint16_t xlen, - uint8_t ylen, - uint8_t approximation); + uint8_t **llrU, + uint8_t ***bit, + uint8_t **bitU, + uint8_t listSize, + uint16_t row, + uint16_t col, + uint16_t xlen, + uint8_t ylen, + uint8_t approximation); void updateBit(uint8_t ***bit, - uint8_t **bitU, - uint8_t listSize, - uint16_t row, - uint16_t col, - uint16_t xlen, - uint8_t ylen); + uint8_t **bitU, + uint8_t listSize, + uint16_t row, + uint16_t col, + uint16_t xlen, + uint8_t ylen); void updatePathMetric(double *pathMetric, - double ***llr, - uint8_t listSize, - uint8_t bitValue, - uint16_t row, - uint8_t approximation); + double ***llr, + uint8_t listSize, + uint8_t bitValue, + uint16_t row, + uint8_t approximation); void updatePathMetric2(double *pathMetric, - double ***llr, - uint8_t listSize, - uint16_t row, - uint8_t approximation); + double ***llr, + uint8_t listSize, + uint16_t row, + uint8_t approximation); void computeLLR(double ***llr, - uint16_t row, - uint16_t col, - uint8_t i, - uint16_t offset, - uint8_t approximation); + uint16_t row, + uint16_t col, + uint8_t i, + uint16_t offset, + uint8_t approximation); void updateCrcChecksum(uint8_t **crcChecksum, - uint8_t **crcGen, - uint8_t listSize, - uint32_t i2, - uint8_t len); + uint8_t **crcGen, + uint8_t listSize, + uint32_t i2, + uint8_t len); void updateCrcChecksum2(uint8_t **crcChecksum, - uint8_t **crcGen, - uint8_t listSize, - uint32_t i2, - uint8_t len); - -void nr_sort_asc_double_1D_array_ind(double *matrix, - uint8_t *ind, - uint8_t len); + uint8_t **crcGen, + uint8_t listSize, + uint32_t i2, + uint8_t len); uint8_t **crc24c_generator_matrix(uint16_t payloadSizeBits); @@ -356,11 +387,11 @@ uint8_t **crc6_generator_matrix(uint16_t payloadSizeBits); //Also nr_polar_rate_matcher static inline void nr_polar_interleaver(uint8_t *input, - uint8_t *output, - uint16_t *pattern, - uint16_t size) + uint8_t *output, + uint16_t *pattern, + uint16_t size) { - for (int i=0; iidx == 0){//PBCH - nr_bit2byte_uint32_8_t(in, polarParams->payloadBits, polarParams->nr_polar_A); - /* - * Bytewise operations - */ - //Calculate CRC. - nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(polarParams->nr_polar_A, - polarParams->crc_generator_matrix, - polarParams->nr_polar_crc, - polarParams->payloadBits, - polarParams->crcParityBits); - for (uint8_t i = 0; i < polarParams->crcParityBits; i++) - polarParams->nr_polar_crc[i] = (polarParams->nr_polar_crc[i] % 2); - - //Attach CRC to the Transport Block. (a to b) - for (uint16_t i = 0; i < polarParams->payloadBits; i++) - polarParams->nr_polar_B[i] = polarParams->nr_polar_A[i]; - for (uint16_t i = polarParams->payloadBits; i < polarParams->K; i++) - polarParams->nr_polar_B[i]= polarParams->nr_polar_crc[i-(polarParams->payloadBits)]; - } else { //UCI - - } - - //Interleaving (c to c') - nr_polar_interleaver(polarParams->nr_polar_B, - polarParams->nr_polar_CPrime, - polarParams->interleaving_pattern, - polarParams->K); - - //Bit insertion (c' to u) - nr_polar_bit_insertion(polarParams->nr_polar_CPrime, - polarParams->nr_polar_U, - polarParams->N, - polarParams->K, - polarParams->Q_I_N, - polarParams->Q_PC_N, - polarParams->n_pc); - - //Encoding (u to d) - nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(polarParams->nr_polar_U, - polarParams->G_N, - polarParams->nr_polar_D, - polarParams->N, - polarParams->N); - for (uint16_t i = 0; i < polarParams->N; i++) - polarParams->nr_polar_D[i] = (polarParams->nr_polar_D[i] % 2); - - //Rate matching - //Sub-block interleaving (d to y) and Bit selection (y to e) - nr_polar_interleaver(polarParams->nr_polar_D, - polarParams->nr_polar_E, - polarParams->rate_matching_pattern, - polarParams->encoderLength); - - /* - * Return bits. - */ + if (polarParams->idx == 0){//PBCH + /* + uint64_t B = (((uint64_t)*in)&((((uint64_t)1)<<32)-1)) | (((uint64_t)crc24c((uint8_t*)in,polarParams->payloadBits)>>8)<payloadBits); #ifdef DEBUG_POLAR_ENCODER - for (int i=0; i< polarParams->encoderLength;i++) printf("f[%d]=%d\n", i, polarParams->nr_polar_E[i]); + printf("polar_B %llx (crc %x)\n",B,crc24c((uint8_t*)in,polarParams->payloadBits)>>8); +#endif +nr_bit2byte_uint32_8_t((uint32_t*)&B, polarParams->K, polarParams->nr_polar_B);*/ + + nr_bit2byte_uint32_8_t((uint32_t*)in, polarParams->payloadBits, polarParams->nr_polar_A); + /* + * Bytewise operations + */ + //Calculate CRC. + + nr_matrix_multiplication_uint8_1D_uint8_2D(polarParams->nr_polar_A, + polarParams->crc_generator_matrix, + polarParams->nr_polar_crc, + polarParams->payloadBits, + polarParams->crcParityBits); + for (uint8_t i = 0; i < polarParams->crcParityBits; i++) + polarParams->nr_polar_crc[i] = (polarParams->nr_polar_crc[i] % 2); + + + //Attach CRC to the Transport Block. (a to b) + for (uint16_t i = 0; i < polarParams->payloadBits; i++) + polarParams->nr_polar_B[i] = polarParams->nr_polar_A[i]; + for (uint16_t i = polarParams->payloadBits; i < polarParams->K; i++) + polarParams->nr_polar_B[i]= polarParams->nr_polar_crc[i-(polarParams->payloadBits)]; + +#ifdef DEBUG_POLAR_ENCODER + uint64_t B2=0; + for (int i = 0;iK;i++) B2 = B2 | ((uint64_t)polarParams->nr_polar_B[i] << i); + printf("polar_B %llx\n",B2); +#endif + /* for (int j=0;jcrcParityBits;j++) { + for (int i=0;ipayloadBits;i++) + printf("%1d.%1d+",polarParams->crc_generator_matrix[i][j],polarParams->nr_polar_A[i]); + printf(" => %d\n",polarParams->nr_polar_crc[j]); + }*/ + + } else { //UCI + + } + + //Interleaving (c to c') + nr_polar_interleaver(polarParams->nr_polar_B, + polarParams->nr_polar_CPrime, + polarParams->interleaving_pattern, + polarParams->K); + + +#ifdef DEBUG_POLAR_ENCODER + uint64_t Cprime=0; + for (int i = 0;iK;i++) { + Cprime = Cprime | ((uint64_t)polarParams->nr_polar_CPrime[i] << i); + if (polarParams->nr_polar_CPrime[i] == 1) printf("pos %d : %llx\n",i,Cprime); + } + printf("polar_Cprime %llx\n",Cprime); +#endif + //Bit insertion (c' to u) + nr_polar_bit_insertion(polarParams->nr_polar_CPrime, + polarParams->nr_polar_U, + polarParams->N, + polarParams->K, + polarParams->Q_I_N, + polarParams->Q_PC_N, + polarParams->n_pc); + + //Encoding (u to d) + /* memset(polarParams->nr_polar_U,0,polarParams->N); + polarParams->nr_polar_U[247]=1; + polarParams->nr_polar_U[253]=1;*/ + nr_matrix_multiplication_uint8_1D_uint8_2D(polarParams->nr_polar_U, + polarParams->G_N, + polarParams->nr_polar_D, + polarParams->N, + polarParams->N); + for (uint16_t i = 0; i < polarParams->N; i++) + polarParams->nr_polar_D[i] = (polarParams->nr_polar_D[i] % 2); + + uint64_t D[8]; + memset((void*)D,0,8*sizeof(int64_t)); +#ifdef DEBUG_POLAR_ENCODER + for (int i=0;iN;i++) D[i/64] |= ((uint64_t)polarParams->nr_polar_D[i])<<(i&63); + printf("D %llx,%llx,%llx,%llx,%llx,%llx,%llx,%llx\n", + D[0],D[1],D[2],D[3],D[4],D[5],D[6],D[7]); #endif - nr_byte2bit_uint8_32_t(polarParams->nr_polar_E, polarParams->encoderLength, out); + //Rate matching + //Sub-block interleaving (d to y) and Bit selection (y to e) + nr_polar_interleaver(polarParams->nr_polar_D, + polarParams->nr_polar_E, + polarParams->rate_matching_pattern, + polarParams->encoderLength); + + /* + * Return bits. + */ +#ifdef DEBUG_POLAR_ENCODER + for (int i=0; i< polarParams->encoderLength;i++) printf("f[%d]=%d\n", i, polarParams->nr_polar_E[i]); +#endif + + nr_byte2bit_uint8_32_t(polarParams->nr_polar_E, polarParams->encoderLength, out); } void polar_encoder_dci(uint32_t *in, - uint32_t *out, - t_nrPolar_paramsPtr polarParams, - uint16_t n_RNTI) + uint32_t *out, + t_nrPolar_paramsPtr polarParams, + uint16_t n_RNTI) { #ifdef DEBUG_POLAR_ENCODER_DCI - printf("[polar_encoder_dci] in: [0]->0x%08x \t [1]->0x%08x \t [2]->0x%08x \t [3]->0x%08x\n", in[0], in[1], in[2], in[3]); + printf("[polar_encoder_dci] in: [0]->0x%08x \t [1]->0x%08x \t [2]->0x%08x \t [3]->0x%08x\n", in[0], in[1], in[2], in[3]); #endif - /* - * Bytewise operations - */ - //(a to a') - nr_bit2byte_uint32_8_t(in, polarParams->payloadBits, polarParams->nr_polar_A); - for (int i=0; icrcParityBits; i++) polarParams->nr_polar_APrime[i]=1; - for (int i=0; ipayloadBits; i++) polarParams->nr_polar_APrime[i+(polarParams->crcParityBits)]=polarParams->nr_polar_A[i]; + /* + * Bytewise operations + */ + //(a to a') + nr_bit2byte_uint32_8_t(in, polarParams->payloadBits, polarParams->nr_polar_A); + for (int i=0; icrcParityBits; i++) polarParams->nr_polar_APrime[i]=1; + for (int i=0; ipayloadBits; i++) polarParams->nr_polar_APrime[i+(polarParams->crcParityBits)]=polarParams->nr_polar_A[i]; #ifdef DEBUG_POLAR_ENCODER_DCI - printf("[polar_encoder_dci] A: "); - for (int i=0; ipayloadBits; i++) printf("%d-", polarParams->nr_polar_A[i]); - printf("\n"); - printf("[polar_encoder_dci] APrime: "); - for (int i=0; iK; i++) printf("%d-", polarParams->nr_polar_APrime[i]); - printf("\n"); - printf("[polar_encoder_dci] GP: "); - for (int i=0; icrcParityBits; i++) printf("%d-", polarParams->crc_generator_matrix[0][i]); - printf("\n"); + printf("[polar_encoder_dci] A: "); + for (int i=0; ipayloadBits; i++) printf("%d-", polarParams->nr_polar_A[i]); + printf("\n"); + printf("[polar_encoder_dci] APrime: "); + for (int i=0; iK; i++) printf("%d-", polarParams->nr_polar_APrime[i]); + printf("\n"); + printf("[polar_encoder_dci] GP: "); + for (int i=0; icrcParityBits; i++) printf("%d-", polarParams->crc_generator_matrix[0][i]); + printf("\n"); #endif - //Calculate CRC. - nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(polarParams->nr_polar_APrime, - polarParams->crc_generator_matrix, - polarParams->nr_polar_crc, - polarParams->K, - polarParams->crcParityBits); - for (uint8_t i = 0; i < polarParams->crcParityBits; i++) polarParams->nr_polar_crc[i] = (polarParams->nr_polar_crc[i] % 2); + //Calculate CRC. + nr_matrix_multiplication_uint8_1D_uint8_2D(polarParams->nr_polar_APrime, + polarParams->crc_generator_matrix, + polarParams->nr_polar_crc, + polarParams->K, + polarParams->crcParityBits); + for (uint8_t i = 0; i < polarParams->crcParityBits; i++) polarParams->nr_polar_crc[i] = (polarParams->nr_polar_crc[i] % 2); #ifdef DEBUG_POLAR_ENCODER_DCI - printf("[polar_encoder_dci] CRC: "); - for (int i=0; icrcParityBits; i++) printf("%d-", polarParams->nr_polar_crc[i]); - printf("\n"); + printf("[polar_encoder_dci] CRC: "); + for (int i=0; icrcParityBits; i++) printf("%d-", polarParams->nr_polar_crc[i]); + printf("\n"); #endif - //Attach CRC to the Transport Block. (a to b) - for (uint16_t i = 0; i < polarParams->payloadBits; i++) - polarParams->nr_polar_B[i] = polarParams->nr_polar_A[i]; - for (uint16_t i = polarParams->payloadBits; i < polarParams->K; i++) - polarParams->nr_polar_B[i]= polarParams->nr_polar_crc[i-(polarParams->payloadBits)]; - //Scrambling (b to c) - for (int i=0; i<16; i++) { - polarParams->nr_polar_B[polarParams->payloadBits+8+i] = - ( polarParams->nr_polar_B[polarParams->payloadBits+8+i] + ((n_RNTI>>(15-i))&1) ) % 2; - } + //Attach CRC to the Transport Block. (a to b) + for (uint16_t i = 0; i < polarParams->payloadBits; i++) + polarParams->nr_polar_B[i] = polarParams->nr_polar_A[i]; + for (uint16_t i = polarParams->payloadBits; i < polarParams->K; i++) + polarParams->nr_polar_B[i]= polarParams->nr_polar_crc[i-(polarParams->payloadBits)]; + //Scrambling (b to c) + for (int i=0; i<16; i++) { + polarParams->nr_polar_B[polarParams->payloadBits+8+i] = + ( polarParams->nr_polar_B[polarParams->payloadBits+8+i] + ((n_RNTI>>(15-i))&1) ) % 2; + } -/* //(a to a') + /* //(a to a') nr_crc_bit2bit_uint32_8_t(in, polarParams->payloadBits, polarParams->nr_polar_aPrime); //Parity bits computation (p) polarParams->crcBit = crc24c(polarParams->nr_polar_aPrime, (polarParams->payloadBits+polarParams->crcParityBits)); -#ifdef DEBUG_POLAR_ENCODER_DCI + #ifdef DEBUG_POLAR_ENCODER_DCI printf("[polar_encoder_dci] crc: 0x%08x\n", polarParams->crcBit); for (int i=0; i<32; i++) { - printf("%d\n",((polarParams->crcBit)>>i)&1); + printf("%d\n",((polarParams->crcBit)>>i)&1); } -#endif + #endif //(a to b) // // Bytewise operations // uint8_t arrayInd = ceil(polarParams->payloadBits / 8.0); for (int i=0; inr_polar_B[j+(i*8)] = ((polarParams->nr_polar_aPrime[3+i]>>(7-j)) & 1); - } + for (int j=0; j<8; j++) { + polarParams->nr_polar_B[j+(i*8)] = ((polarParams->nr_polar_aPrime[3+i]>>(7-j)) & 1); + } } for (int i=0; i<((polarParams->payloadBits)%8); i++) { - polarParams->nr_polar_B[i+(arrayInd-1)*8] = ((polarParams->nr_polar_aPrime[3+(arrayInd-1)]>>(7-i)) & 1); + polarParams->nr_polar_B[i+(arrayInd-1)*8] = ((polarParams->nr_polar_aPrime[3+(arrayInd-1)]>>(7-i)) & 1); } for (int i=0; i<8; i++) { - polarParams->nr_polar_B[polarParams->payloadBits+i] = ((polarParams->crcBit)>>(31-i))&1; + polarParams->nr_polar_B[polarParams->payloadBits+i] = ((polarParams->crcBit)>>(31-i))&1; } //Scrambling (b to c) for (int i=0; i<16; i++) { - polarParams->nr_polar_B[polarParams->payloadBits+8+i] = - ( (((polarParams->crcBit)>>(23-i))&1) + ((n_RNTI>>(15-i))&1) ) % 2; + polarParams->nr_polar_B[polarParams->payloadBits+8+i] = + ( (((polarParams->crcBit)>>(23-i))&1) + ((n_RNTI>>(15-i))&1) ) % 2; }*/ #ifdef DEBUG_POLAR_ENCODER_DCI - printf("[polar_encoder_dci] B: "); - for (int i = 0; i < polarParams->K; i++) printf("%d-", polarParams->nr_polar_B[i]); - printf("\n"); + printf("[polar_encoder_dci] B: "); + for (int i = 0; i < polarParams->K; i++) printf("%d-", polarParams->nr_polar_B[i]); + printf("\n"); #endif - //Interleaving (c to c') - nr_polar_interleaver(polarParams->nr_polar_B, - polarParams->nr_polar_CPrime, - polarParams->interleaving_pattern, - polarParams->K); + //Interleaving (c to c') + nr_polar_interleaver(polarParams->nr_polar_B, + polarParams->nr_polar_CPrime, + polarParams->interleaving_pattern, + polarParams->K); - //Bit insertion (c' to u) - nr_polar_bit_insertion(polarParams->nr_polar_CPrime, - polarParams->nr_polar_U, - polarParams->N, - polarParams->K, - polarParams->Q_I_N, - polarParams->Q_PC_N, - polarParams->n_pc); + //Bit insertion (c' to u) + nr_polar_bit_insertion(polarParams->nr_polar_CPrime, + polarParams->nr_polar_U, + polarParams->N, + polarParams->K, + polarParams->Q_I_N, + polarParams->Q_PC_N, + polarParams->n_pc); - //Encoding (u to d) - nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(polarParams->nr_polar_U, - polarParams->G_N, - polarParams->nr_polar_D, - polarParams->N, - polarParams->N); - for (uint16_t i = 0; i < polarParams->N; i++) - polarParams->nr_polar_D[i] = (polarParams->nr_polar_D[i] % 2); + //Encoding (u to d) + nr_matrix_multiplication_uint8_1D_uint8_2D(polarParams->nr_polar_U, + polarParams->G_N, + polarParams->nr_polar_D, + polarParams->N, + polarParams->N); + for (uint16_t i = 0; i < polarParams->N; i++) + polarParams->nr_polar_D[i] = (polarParams->nr_polar_D[i] % 2); - //Rate matching - //Sub-block interleaving (d to y) and Bit selection (y to e) - nr_polar_interleaver(polarParams->nr_polar_D, - polarParams->nr_polar_E, - polarParams->rate_matching_pattern, - polarParams->encoderLength); + //Rate matching + //Sub-block interleaving (d to y) and Bit selection (y to e) + nr_polar_interleaver(polarParams->nr_polar_D, + polarParams->nr_polar_E, + polarParams->rate_matching_pattern, + polarParams->encoderLength); - /* - * Return bits. - */ - nr_byte2bit_uint8_32_t(polarParams->nr_polar_E, polarParams->encoderLength, out); + /* + * Return bits. + */ + nr_byte2bit_uint8_32_t(polarParams->nr_polar_E, polarParams->encoderLength, out); #ifdef DEBUG_POLAR_ENCODER_DCI - printf("[polar_encoder_dci] E: "); - for (int i = 0; i < polarParams->encoderLength; i++) printf("%d-", polarParams->nr_polar_E[i]); - uint8_t outputInd = ceil(polarParams->encoderLength / 32.0); - printf("\n[polar_encoder_dci] out: "); - for (int i = 0; i < outputInd; i++) { - printf("[%d]->0x%08x\t", i, out[i]); - } + printf("[polar_encoder_dci] E: "); + for (int i = 0; i < polarParams->encoderLength; i++) printf("%d-", polarParams->nr_polar_E[i]); + uint8_t outputInd = ceil(polarParams->encoderLength / 32.0); + printf("\n[polar_encoder_dci] out: "); + for (int i = 0; i < outputInd; i++) { + printf("[%d]->0x%08x\t", i, out[i]); + } #endif } -void polar_encoder_timing(uint32_t *in, - uint32_t *out, - t_nrPolar_paramsPtr polarParams, - double cpuFreqGHz, - FILE* logFile) -{ - //Initiate timing. - time_stats_t timeEncoderCRCByte, timeEncoderCRCBit, timeEncoderInterleaver, timeEncoderBitInsertion, timeEncoder1, timeEncoder2, timeEncoderRateMatching, timeEncoderByte2Bit; - reset_meas(&timeEncoderCRCByte); reset_meas(&timeEncoderCRCBit); reset_meas(&timeEncoderInterleaver); reset_meas(&timeEncoderBitInsertion); reset_meas(&timeEncoder1); reset_meas(&timeEncoder2); reset_meas(&timeEncoderRateMatching); reset_meas(&timeEncoderByte2Bit); - uint16_t n_RNTI=0x0000; +static inline void polar_rate_matching(t_nrPolar_paramsPtr polarParams,void *in,void *out) __attribute__((always_inline)); - start_meas(&timeEncoderCRCByte); - nr_crc_bit2bit_uint32_8_t(in, polarParams->payloadBits, polarParams->nr_polar_aPrime); //(a to a') - polarParams->crcBit = crc24c(polarParams->nr_polar_aPrime, (polarParams->payloadBits+polarParams->crcParityBits)); //Parity bits computation (p) - uint8_t arrayInd = ceil(polarParams->payloadBits / 8.0); //(a to b) - for (int i=0; inr_polar_B[j+(i*8)] = ((polarParams->nr_polar_aPrime[3+i]>>(7-j)) & 1); - for (int i=0; i<((polarParams->payloadBits)%8); i++) polarParams->nr_polar_B[i+(arrayInd-1)*8] = ((polarParams->nr_polar_aPrime[3+(arrayInd-1)]>>(7-i)) & 1); - for (int i=0; i<8; i++) polarParams->nr_polar_B[polarParams->payloadBits+i] = ((polarParams->crcBit)>>(31-i))&1; - for (int i=0; i<16; i++) polarParams->nr_polar_B[polarParams->payloadBits+8+i] = ( (((polarParams->crcBit)>>(23-i))&1) + ((n_RNTI>>(15-i))&1) ) % 2; //Scrambling (b to c) - stop_meas(&timeEncoderCRCByte); +static inline void polar_rate_matching(t_nrPolar_paramsPtr polarParams,void *in,void *out) { + + if (polarParams->groupsize == 8) + for (int i=0;iencoderLength>>3;i++) ((uint8_t*)out)[i] = ((uint8_t *)in)[polarParams->rm_tab[i]]; + else + for (int i=0;iencoderLength>>4;i++) { + ((uint16_t*)out)[i] = ((uint16_t *)in)[polarParams->rm_tab[i]]; + } +} + +void build_polar_tables(t_nrPolar_paramsPtr polarParams) { + + // build table b -> c' + + AssertFatal(polarParams->K > 32, "K = %d < 33, is not supported yet\n",polarParams->K); + AssertFatal(polarParams->K < 129, "K = %d > 64, is not supported yet\n",polarParams->K); + + int bit_i,ip; + + int numbytes = polarParams->K>>3; + int residue = polarParams->K&7; + int numbits; + if (residue>0) numbytes++; + for (int byte=0;byteK>>3)) numbits=8; + else numbits=residue; + for (int val=0;val<256;val++) { + polarParams->cprime_tab0[byte][val] = 0; + polarParams->cprime_tab1[byte][val] = 0; + for (int i=0;ideinterleaving_pattern[polarParams->K-1-((8*byte)+i)]; + AssertFatal(ip<128,"ip = %d\n",ip); + bit_i=(val>>i)&1; + if (ip<64) polarParams->cprime_tab0[byte][val] |= (((uint64_t)bit_i)<cprime_tab1[byte][val] |= (((uint64_t)bit_i)<<(ip&63)); + } + } + } + + AssertFatal(polarParams->N==512,"N = %d, not done yet\n",polarParams->N); + + // build G bit vectors for information bit positions and convert the bit as bytes tables in nr_polar_kronecker_power_matrices.c to 64 bit packed vectors. + // keep only rows of G which correspond to information/crc bits + polarParams->G_N_tab = (uint64_t**)malloc(polarParams->K * sizeof(int64_t*)); + int k=0; + for (int i=0;iN;i++) { + if (polarParams->information_bit_pattern[i] > 0) { + polarParams->G_N_tab[k] = (uint64_t*)memalign(32,(polarParams->N/64)*sizeof(uint64_t)); + memset((void*)polarParams->G_N_tab[k],0,(polarParams->N/64)*sizeof(uint64_t)); + for (int j=0;jN;j++) + polarParams->G_N_tab[k][j/64] |= ((uint64_t)polarParams->G_N[i][j])<<(j&63); +#ifdef DEBUG_POLAR_ENCODER + printf("Bit %d Selecting row %d of G : ",k,i); + for (int j=0;jN;j+=4) printf("%1x",polarParams->G_N[i][j]+(polarParams->G_N[i][j+1]*2)+(polarParams->G_N[i][j+2]*4)+(polarParams->G_N[i][j+3]*8)); + printf("\n"); +#endif + k++; + } + } + + // rate matching table + int iplast=polarParams->rate_matching_pattern[0]; + int ccnt=0; + int groupcnt=0; + int firstingroup_out=0; + int firstingroup_in=iplast; + int mingroupsize = 1024; + // compute minimum group size of rate-matching pattern + for (int outpos=1; outposencoderLength; outpos++) { + + ip=polarParams->rate_matching_pattern[outpos]; + if ((ip - iplast) == 1) ccnt++; + else { + groupcnt++; +#ifdef DEBUG_POLAR_ENCODER + printf("group %d (size %d): (%d:%d) => (%d:%d)\n",groupcnt,ccnt+1, + firstingroup_in,firstingroup_in+ccnt, + firstingroup_out,firstingroup_out+ccnt); +#endif + if ((ccnt+1)groupsize=mingroupsize; + int shift=3; + if (mingroupsize == 16) shift=4; + polarParams->rm_tab=(int*)malloc(sizeof(int)*polarParams->encoderLength/mingroupsize); + // rerun again to create groups + int tcnt=0; + for (int outpos=0;outposencoderLength; outpos+=mingroupsize,tcnt++) + polarParams->rm_tab[tcnt] = polarParams->rate_matching_pattern[outpos]>>shift; + +} + +void polar_encoder_fast(uint64_t *A, + uint32_t *out, + int32_t crcmask, + t_nrPolar_paramsPtr polarParams) { + + AssertFatal(polarParams->K > 32, "K = %d < 33, is not supported yet\n",polarParams->K); + AssertFatal(polarParams->K < 129, "K = %d > 128, is not supported yet\n",polarParams->K); + AssertFatal(polarParams->payloadBits < 65, "payload bits = %d > 64, is not supported yet\n",polarParams->payloadBits); + + uint64_t B[4]={0,0,0,0},Cprime[4]={0,0,0,0}; + int bitlen = polarParams->payloadBits; + + // append crc + + AssertFatal(bitlen<129,"support for payloads <= 128 bits\n"); + AssertFatal(polarParams->crcParityBits == 24,"support for 24-bit crc only for now\n"); + + //int bitlen0=bitlen; + + uint64_t tcrc=0; + + // A bitstring should be stored as a_{N-1} a_{N-2} ... a_{N-A} 0 .... 0, where N=64,128,192,..., N is smallest multiple of 64 greater than or equal to A + + // First flip A bitstring byte endian for CRC routines (optimized for DLSCH/ULSCH, not PBCH/PDCCH) + // CRC reads in each byte in bit positions 7 downto 0, for PBCH/PDCCH we need to read in a_{A-1} downto a_{0}, A = length of bit string (e.g. 32 for PBCH) + if (bitlen<=32) { + uint8_t A32_flip[4]; + uint32_t Aprime= (uint32_t)(((uint32_t)*A)<<(32-bitlen)); + A32_flip[0]=((uint8_t*)&Aprime)[3]; + A32_flip[1]=((uint8_t*)&Aprime)[2]; + A32_flip[2]=((uint8_t*)&Aprime)[1]; + A32_flip[3]=((uint8_t*)&Aprime)[0]; + tcrc = (uint64_t)((crcmask^(crc24c(A32_flip,bitlen)>>8))); + } + else if (bitlen<=64) { + uint8_t A64_flip[8]; + uint64_t Aprime= (uint32_t)(((uint64_t)*A)<<(64-bitlen)); + A64_flip[0]=((uint8_t*)&Aprime)[7]; + A64_flip[1]=((uint8_t*)&Aprime)[6]; + A64_flip[2]=((uint8_t*)&Aprime)[5]; + A64_flip[3]=((uint8_t*)&Aprime)[4]; + A64_flip[4]=((uint8_t*)&Aprime)[3]; + A64_flip[5]=((uint8_t*)&Aprime)[2]; + A64_flip[6]=((uint8_t*)&Aprime)[1]; + A64_flip[7]=((uint8_t*)&Aprime)[0]; + tcrc = (uint64_t)((crcmask^(crc24c(A64_flip,bitlen)>>8))); + } + + int n; + // this is number of quadwords in the bit string + int quadwlen = (polarParams->K>>6); + if ((polarParams->K&63) > 0) quadwlen++; + + // Create the B bitstring as + // b_{N'-1} b_{N'-2} ... b_{N'-A} b_{N'-A-1} ... b_{N'-A-Nparity} = a_{N-1} a_{N-2} ... a_{N-A} p_{N_parity-1} ... p_0 + + for (n=0;ncrcParityBits) | tcrc; + else B[n] = (A[n] << polarParams->crcParityBits) | (A[n-1]>>(64-polarParams->crcParityBits)); + + + uint8_t *Bbyte = (uint8_t*)B; + // for each byte of B, lookup in corresponding table for 64-bit word corresponding to that byte and its position + if (polarParams->K<65) + Cprime[0] = polarParams->cprime_tab0[0][Bbyte[0]] | + polarParams->cprime_tab0[1][Bbyte[1]] | + polarParams->cprime_tab0[2][Bbyte[2]] | + polarParams->cprime_tab0[3][Bbyte[3]] | + polarParams->cprime_tab0[4][Bbyte[4]] | + polarParams->cprime_tab0[5][Bbyte[5]] | + polarParams->cprime_tab0[6][Bbyte[6]] | + polarParams->cprime_tab0[7][Bbyte[7]]; + + else if (polarParams->K < 129) { + for (int i=0;i<1+(polarParams->K/8);i++) { + Cprime[0] |= polarParams->cprime_tab0[i][Bbyte[i]]; + Cprime[1] |= polarParams->cprime_tab1[i][Bbyte[i]]; + } + } + +#ifdef DEBUG_POLAR_ENCODER + if (polarParams->K<65) + printf("A %llx B %llx Cprime %llx (payload bits %d,crc %x)\n", + (unsigned long long)(A[0]&(((uint64_t)1<payloadBits, + tcrc); + else if (polarParams->K<129) { + if (bitlen<64) + printf("A %llx B %llx|%llx Cprime %llx|%llx (payload bits %d,crc %x)\n", + (unsigned long long)(A[0]&(((uint64_t)1<payloadBits, + tcrc); + else + printf("A %llx|%llx B %llx|%llx Cprime %llx|%llx (payload bits %d,crc %x)\n", + (unsigned long long)(A[1]&(((uint64_t)1<<(bitlen-64))-1)),(unsigned long long)(A[0]), + (unsigned long long)(B[1]),(unsigned long long)(B[0]), + (unsigned long long)(Cprime[1]),(unsigned long long)(Cprime[0]), + polarParams->payloadBits, + crc24c((uint8_t*)A,bitlen)>>8); + } +#endif + /* printf("Bbytes : %x.%x.%x.%x.%x.%x.%x.%x\n",Bbyte[0],Bbyte[1],Bbyte[2],Bbyte[3],Bbyte[4],Bbyte[5],Bbyte[6],Bbyte[7]); + printf("%llx,%llx,%llx,%llx,%llx,%llx,%llx,%llx\n",polarParams->cprime_tab[0][Bbyte[0]] , + polarParams->cprime_tab[1][Bbyte[1]] , + polarParams->cprime_tab[2][Bbyte[2]] , + polarParams->cprime_tab[3][Bbyte[3]] , + polarParams->cprime_tab[4][Bbyte[4]] , + polarParams->cprime_tab[5][Bbyte[5]] , + polarParams->cprime_tab[6][Bbyte[6]] , + polarParams->cprime_tab[7][Bbyte[7]]);*/ + + // now do Gu product (here using 64-bit XORs, we can also do with SIMD after) + // here we're reading out the bits LSB -> MSB, is this correct w.r.t. 3GPP ? + + uint64_t Cprime_i; + /* printf("%llx Cprime_0 (%llx) G %llx,%llx,%llx,%llx,%llx,%llx,%llx,%llx\n",Cprime_i,Cprime &1, + polarParams->G_N_tab[0][0], + polarParams->G_N_tab[0][1], + polarParams->G_N_tab[0][2], + polarParams->G_N_tab[0][3], + polarParams->G_N_tab[0][4], + polarParams->G_N_tab[0][5], + polarParams->G_N_tab[0][6], + polarParams->G_N_tab[0][7]);*/ + uint64_t D[8]={0,0,0,0,0,0,0,0}; + int off=0; + int len=polarParams->K; + for (int j=0;j<(1+(polarParams->K>>6));j++,off+=64,len-=64) { + for (int i=0;i<((len>63) ? 64 : len);i++) { + + Cprime_i = -((Cprime[j]>>i)&1); // this converts bit 0 as, 0 => 0000x00, 1 => 1111x11 + /* +#ifdef DEBUG_POLAR_ENCODER + printf("%llx Cprime_%d (%llx) G %llx,%llx,%llx,%llx,%llx,%llx,%llx,%llx\n", + Cprime_i,off+i,(Cprime[j]>>i) &1, + polarParams->G_N_tab[off+i][0], + polarParams->G_N_tab[off+i][1], + polarParams->G_N_tab[off+i][2], + polarParams->G_N_tab[off+i][3], + polarParams->G_N_tab[off+i][4], + polarParams->G_N_tab[off+i][5], + polarParams->G_N_tab[off+i][6], + polarParams->G_N_tab[off+i][7]); +#endif + */ + uint64_t *Gi=polarParams->G_N_tab[off+i]; + D[0] ^= (Cprime_i & Gi[0]); + D[1] ^= (Cprime_i & Gi[1]); + D[2] ^= (Cprime_i & Gi[2]); + D[3] ^= (Cprime_i & Gi[3]); + D[4] ^= (Cprime_i & Gi[4]); + D[5] ^= (Cprime_i & Gi[5]); + D[6] ^= (Cprime_i & Gi[6]); + D[7] ^= (Cprime_i & Gi[7]); + } + } +#ifdef DEBUG_POLAR_ENCODER + printf("D %llx,%llx,%llx,%llx,%llx,%llx,%llx,%llx\n", + D[0], + D[1], + D[2], + D[3], + D[4], + D[5], + D[6], + D[7]); +#endif + + polar_rate_matching(polarParams,(void*)D,(void*)out); - - start_meas(&timeEncoderCRCBit); - nr_bit2byte_uint32_8_t(in, polarParams->payloadBits, polarParams->nr_polar_A); - nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(polarParams->nr_polar_A, polarParams->crc_generator_matrix, polarParams->nr_polar_crc, polarParams->payloadBits, polarParams->crcParityBits); //Calculate CRC. - for (uint8_t i = 0; i < polarParams->crcParityBits; i++) polarParams->nr_polar_crc[i] = (polarParams->nr_polar_crc[i] % 2); - for (uint16_t i = 0; i < polarParams->payloadBits; i++) polarParams->nr_polar_B[i] = polarParams->nr_polar_A[i]; //Attach CRC to the Transport Block. (a to b) - for (uint16_t i = polarParams->payloadBits; i < polarParams->K; i++) polarParams->nr_polar_B[i]= polarParams->nr_polar_crc[i-(polarParams->payloadBits)]; - stop_meas(&timeEncoderCRCBit); - - - start_meas(&timeEncoderInterleaver); //Interleaving (c to c') - nr_polar_interleaver(polarParams->nr_polar_B, polarParams->nr_polar_CPrime, polarParams->interleaving_pattern, polarParams->K); - stop_meas(&timeEncoderInterleaver); - - - start_meas(&timeEncoderBitInsertion); //Bit insertion (c' to u) - nr_polar_bit_insertion(polarParams->nr_polar_CPrime, polarParams->nr_polar_U, polarParams->N, polarParams->K, polarParams->Q_I_N, polarParams->Q_PC_N, polarParams->n_pc); - stop_meas(&timeEncoderBitInsertion); - - - start_meas(&timeEncoder1); //Encoding (u to d) - nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(polarParams->nr_polar_U, polarParams->G_N, polarParams->nr_polar_D, polarParams->N, polarParams->N); - stop_meas(&timeEncoder1); - - - start_meas(&timeEncoder2); - for (uint16_t i = 0; i < polarParams->N; i++) polarParams->nr_polar_D[i] = (polarParams->nr_polar_D[i] % 2); - stop_meas(&timeEncoder2); - - - start_meas(&timeEncoderRateMatching);//Rate matching //Sub-block interleaving (d to y) and Bit selection (y to e) - nr_polar_interleaver(polarParams->nr_polar_D, polarParams->nr_polar_E, polarParams->rate_matching_pattern, polarParams->encoderLength); - stop_meas(&timeEncoderRateMatching); - - - start_meas(&timeEncoderByte2Bit); //Return bits. - nr_byte2bit_uint8_32_t(polarParams->nr_polar_E, polarParams->encoderLength, out); - stop_meas(&timeEncoderByte2Bit); - - fprintf(logFile,",%f,%f,%f,%f,%f,%f,%f,%f\n", - (timeEncoderCRCByte.diff/(cpuFreqGHz*1000.0)), - (timeEncoderCRCBit.diff/(cpuFreqGHz*1000.0)), - (timeEncoderInterleaver.diff/(cpuFreqGHz*1000.0)), - (timeEncoderBitInsertion.diff/(cpuFreqGHz*1000.0)), - (timeEncoder1.diff/(cpuFreqGHz*1000.0)), - (timeEncoder2.diff/(cpuFreqGHz*1000.0)), - (timeEncoderRateMatching.diff/(cpuFreqGHz*1000.0)), - (timeEncoderByte2Bit.diff/(cpuFreqGHz*1000.0))); } diff --git a/openair1/PHY/CODING/nrPolar_tools/nr_polar_matrix_and_array.c b/openair1/PHY/CODING/nrPolar_tools/nr_polar_matrix_and_array.c index 81c4125faf..c4c68e4e0e 100644 --- a/openair1/PHY/CODING/nrPolar_tools/nr_polar_matrix_and_array.c +++ b/openair1/PHY/CODING/nrPolar_tools/nr_polar_matrix_and_array.c @@ -32,7 +32,7 @@ #include "PHY/CODING/nrPolar_tools/nr_polar_defs.h" -void nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(uint8_t *matrix1, uint8_t **matrix2, +void nr_matrix_multiplication_uint8_1D_uint8_2D(uint8_t *matrix1, uint8_t **matrix2, uint8_t *output, uint16_t row, uint16_t col) { for (uint16_t i = 0; i < col; i++) { @@ -43,12 +43,12 @@ void nr_matrix_multiplication_uint8_t_1D_uint8_t_2D(uint8_t *matrix1, uint8_t ** } } -uint8_t ***nr_alloc_uint8_t_3D_array(uint16_t xlen, uint16_t ylen, uint16_t zlen) { +uint8_t ***nr_alloc_uint8_3D_array(uint16_t xlen, uint16_t ylen, uint16_t zlen) { uint8_t ***output; int i, j; if ((output = malloc(xlen * sizeof(*output))) == NULL) { - perror("[nr_alloc_uint8_t_3D_array] Problem at 1D allocation"); + perror("[nr_alloc_uint8_3D_array] Problem at 1D allocation"); return NULL; } for (i = 0; i < xlen; i++) @@ -57,8 +57,8 @@ uint8_t ***nr_alloc_uint8_t_3D_array(uint16_t xlen, uint16_t ylen, uint16_t zlen for (i = 0; i < xlen; i++) if ((output[i] = malloc(ylen * sizeof *output[i])) == NULL) { - perror("[nr_alloc_uint8_t_3D_array] Problem at 2D allocation"); - nr_free_uint8_t_3D_array(output, xlen, ylen); + perror("[nr_alloc_uint8_3D_array] Problem at 2D allocation"); + nr_free_uint8_3D_array(output, xlen, ylen); return NULL; } for (i = 0; i < xlen; i++) @@ -69,14 +69,39 @@ uint8_t ***nr_alloc_uint8_t_3D_array(uint16_t xlen, uint16_t ylen, uint16_t zlen for (i = 0; i < xlen; i++) for (j = 0; j < ylen; j++) if ((output[i][j] = malloc(zlen * sizeof *output[i][j])) == NULL) { - perror("[nr_alloc_uint8_t_3D_array] Problem at 3D allocation"); - nr_free_uint8_t_3D_array(output, xlen, ylen); + perror("[nr_alloc_uint8_3D_array] Problem at 3D allocation"); + nr_free_uint8_3D_array(output, xlen, ylen); return NULL; } return output; } +uint8_t **nr_alloc_uint8_2D_array(uint16_t xlen, uint16_t ylen) { + uint8_t **output; + int i, j; + + if ((output = malloc(xlen * sizeof(*output))) == NULL) { + perror("[nr_alloc_uint8_2D_array] Problem at 1D allocation"); + return NULL; + } + for (i = 0; i < xlen; i++) + output[i] = NULL; + + + for (i = 0; i < xlen; i++) + if ((output[i] = malloc(ylen * sizeof *output[i])) == NULL) { + perror("[nr_alloc_uint8_2D_array] Problem at 2D allocation"); + nr_free_uint8_2D_array(output, xlen); + return NULL; + } + for (i = 0; i < xlen; i++) + for (j = 0; j < ylen; j++) + output[i][j] = 0; + + return output; +} + double ***nr_alloc_double_3D_array(uint16_t xlen, uint16_t ylen, uint16_t zlen) { double ***output; int i, j; @@ -137,31 +162,6 @@ double **nr_alloc_double_2D_array(uint16_t xlen, uint16_t ylen) { return output; } -uint8_t **nr_alloc_uint8_t_2D_array(uint16_t xlen, uint16_t ylen) { - uint8_t **output; - int i, j; - - if ((output = malloc(xlen * sizeof(*output))) == NULL) { - perror("[nr_alloc_uint8_t_2D_array] Problem at 1D allocation"); - return NULL; - } - for (i = 0; i < xlen; i++) - output[i] = NULL; - - - for (i = 0; i < xlen; i++) - if ((output[i] = malloc(ylen * sizeof *output[i])) == NULL) { - perror("[nr_alloc_uint8_t_2D_array] Problem at 2D allocation"); - nr_free_uint8_t_2D_array(output, xlen); - return NULL; - } - for (i = 0; i < xlen; i++) - for (j = 0; j < ylen; j++) - output[i][j] = 0; - - return output; -} - void nr_free_double_3D_array(double ***input, uint16_t xlen, uint16_t ylen) { int i, j; @@ -174,7 +174,16 @@ void nr_free_double_3D_array(double ***input, uint16_t xlen, uint16_t ylen) { free(input); } -void nr_free_uint8_t_3D_array(uint8_t ***input, uint16_t xlen, uint16_t ylen) { +void nr_free_double_2D_array(double **input, uint16_t xlen) { + int i; + + for (i = 0; i < xlen; i++) { + free(input[i]); + } + free(input); +} + +void nr_free_uint8_3D_array(uint8_t ***input, uint16_t xlen, uint16_t ylen) { int i, j; for (i = 0; i < xlen; i++) { @@ -186,20 +195,11 @@ void nr_free_uint8_t_3D_array(uint8_t ***input, uint16_t xlen, uint16_t ylen) { free(input); } -void nr_free_uint8_t_2D_array(uint8_t **input, uint16_t xlen) { +void nr_free_uint8_2D_array(uint8_t **input, uint16_t xlen) { for (int i = 0; i < xlen; i++) free(input[i]); free(input); } -void nr_free_double_2D_array(double **input, uint16_t xlen) { - int i; - - for (i = 0; i < xlen; i++) { - free(input[i]); - } - free(input); -} - // Modified Bubble Sort. void nr_sort_asc_double_1D_array_ind(double *matrix, uint8_t *ind, uint8_t len) { int swaps; diff --git a/openair1/PHY/CODING/nrPolar_tools/nr_polar_procedures.c b/openair1/PHY/CODING/nrPolar_tools/nr_polar_procedures.c index 95fc79ecb1..985266447b 100644 --- a/openair1/PHY/CODING/nrPolar_tools/nr_polar_procedures.c +++ b/openair1/PHY/CODING/nrPolar_tools/nr_polar_procedures.c @@ -33,15 +33,15 @@ #include "PHY/CODING/nrPolar_tools/nr_polar_defs.h" void nr_polar_bit_insertion(uint8_t *input, - uint8_t *output, - uint16_t N, - uint16_t K, - int16_t *Q_I_N, - int16_t *Q_PC_N, - uint8_t n_PC) + uint8_t *output, + uint16_t N, + uint16_t K, + int16_t *Q_I_N, + int16_t *Q_PC_N, + uint8_t n_PC) { - uint16_t k=0; - uint8_t flag; + uint16_t k=0; + uint8_t flag; if (n_PC>0) { /* @@ -325,7 +325,7 @@ void nr_polar_rate_matching_int16(int16_t *input, int16_t *output, uint16_t *rmp if ( (K/(double)E) <= (7.0/16) ) { //puncturing for (int i=0; i<=N-1; i++) output[i]=0; } else { //shortening - for (int i=0; i<=N-1; i++) output[i]=INFINITY; + for (int i=0; i<=N-1; i++) output[i]=32767;//instead of INFINITY, to prevent [-Woverflow] } for (int i=0; i<=E-1; i++){ diff --git a/openair1/PHY/CODING/nrPolar_tools/nr_polar_uci_defs.h b/openair1/PHY/CODING/nrPolar_tools/nr_polar_uci_defs.h index 724da73fa0..42051eb1b5 100644 --- a/openair1/PHY/CODING/nrPolar_tools/nr_polar_uci_defs.h +++ b/openair1/PHY/CODING/nrPolar_tools/nr_polar_uci_defs.h @@ -35,5 +35,27 @@ #define NR_POLAR_UCI_MESSAGE_TYPE -1 //int8_t #define NR_POLAR_UCI_CRC_ERROR_CORRECTION_BITS 3 +#define NR_POLAR_PUCCH_PAYLOAD_BITS 32 +#define NR_POLAR_PUCCH_E 32 + +//Ref. 38-212, Section 6.3.1.2.1 +#define NR_POLAR_PUCCH_CRC_PARITY_BITS_SHORT 6 +#define NR_POLAR_PUCCH_CRC_PARITY_BITS_LONG 11 + +#define NR_POLAR_PUCCH_I_SEG_LONG 1 +#define NR_POLAR_PUCCH_I_SEG_SHORT 0 + +//Ref. 38-212, Section 6.3.1.3.1 +#define NR_POLAR_PUCCH_N_MAX 10 +#define NR_POLAR_PUCCH_I_IL 0 +#define NR_POLAR_PUCCH_N_PC_SHORT 3 +#define NR_POLAR_PUCCH_N_PC_LONG 0 +#define NR_POLAR_PUCCH_N_PC_WM_LONG 0 +#define NR_POLAR_PUCCH_N_PC_WM_SHORT 1 + +//Ref. 38-212, Section 6.3.1.4.1 +#define NR_POLAR_PUCCH_I_BIL 1 + + #endif diff --git a/openair1/PHY/CODING/nr_compute_tbs.c b/openair1/PHY/CODING/nr_compute_tbs.c index 4db253f56a..e1c15d4a86 100644 --- a/openair1/PHY/CODING/nr_compute_tbs.c +++ b/openair1/PHY/CODING/nr_compute_tbs.c @@ -28,35 +28,8 @@ #define INDEX_MAX_TBS_TABLE (93) //Table 5.1.3.1-1 -uint16_t Mcsindextable1[29][3] = {{2,120,0.2344}, - {2,157,0.3066}, - {2,193,0.3770}, - {2,251,0.4902}, - {2,308,0.6016}, - {2,379,0.7402}, - {2,449,0.8770}, - {2,526,1.0273}, - {2,602,1.1758}, - {2,679,1.3262}, - {4,340,1.3281}, - {4,378,1.4766}, - {4,434,1.6953}, - {4,490,1.9141}, - {4,553,2.1602}, - {4,616,2.4063}, - {4,658,2.5703}, - {6,438,2.5664}, - {6,466,2.7305}, - {6,517,3.0293}, - {6,567,3.3223}, - {6,616,3.6094}, - {6,666,3.9023}, - {6,719,4.2129}, - {6,772,4.5234}, - {6,822,4.8164}, - {6,873,5.1152}, - {6,910,5.3320}, - {6,948,5.5547}}; +uint16_t Mcsindextable1[29][2] = {{2,120},{2,157},{2,193},{2,251},{2,308},{2,379},{2,449},{2,526},{2,602},{2,679},{4,340},{4,378},{4,434},{4,490},{4,553},{4,616}, + {4,658},{6,438},{6,466},{6,517},{6,567},{6,616},{6,666},{6,719},{6,772},{6,822},{6,873}, {6,910}, {6,948}}; //Table 5.1.2.2-2 uint16_t Tbstable_nr[INDEX_MAX_TBS_TABLE] = {24,32,40,48,56,64,72,80,88,96,104,112,120,128,136,144,152,160,168,176,184,192,208,224,240,256,272,288,304,320,336,352,368,384,408,432,456,480,504,528,552,576,608,640,672,704,736,768,808,848,888,928,984,1032,1064,1128,1160,1192,1224,1256,1288,1320,1352,1416,1480,1544,1608,1672,1736,1800,1864,1928,2024,2088,2152,2216,2280,2408,2472,2536,2600,2664,2728,2792,2856,2976,3104,3240,3368,3496,3624,3752,3824}; @@ -67,8 +40,9 @@ uint32_t nr_compute_tbs(uint8_t mcs, uint16_t length_dmrs, uint8_t Nl) { - uint16_t nbp_re, nb_re, nb_dmrs_prb, nb_rb_oh, Ninfo,Np_info,n,Qm,R,C; + uint16_t nbp_re, nb_re, nb_dmrs_prb, nb_rb_oh,Qm,R; uint32_t nr_tbs=0; + double Ninfo,Np_info,n,C; nb_rb_oh = 0; //set to 0 if not configured by higher layer Qm = Mcsindextable1[mcs][0]; @@ -80,17 +54,17 @@ uint32_t nr_compute_tbs(uint8_t mcs, nb_re = min(156, nbp_re) * nb_rb; // Intermediate number of information bits - Ninfo = (nb_re * R * Qm * Nl)/1024; + Ninfo = (double)((nb_re * R * Qm * Nl)/1024); - //printf("Ninfo %d nbp_re %d nb_re %d Qm %d, R %d\n", Ninfo, nbp_re, nb_re, Qm, R); + //printf("Ninfo %lf nbp_re %d nb_re %d mcs %d Qm %d, R %d\n", Ninfo, nbp_re, nb_re,mcs, Qm, R); if (Ninfo <=3824) { n = max(3, floor(log2(Ninfo)) - 6); - Np_info = max(24, pow(n,2) * floor(Ninfo/pow(n,2))); + Np_info = max(24, pow(2,n) * floor(Ninfo/pow(2,n))); for (int i=0; i= Np_info){ + if ((double)Tbstable_nr[i] >= Np_info){ nr_tbs = Tbstable_nr[i]; break; } @@ -98,19 +72,21 @@ uint32_t nr_compute_tbs(uint8_t mcs, } else { n = floor(log2(Ninfo-24)) - 5; - Np_info = max(3840, pow(n,2) * round((Ninfo - 24)/pow(n,2))); + Np_info = max(3840, pow(2,n) * round((Ninfo - 24)/pow(2,n))); - if (R <= 1024/4) { + if (R <= 256) { //1/4 C = ceil( (Np_info + 24)/3816 ); - nr_tbs = 8 * C * ceil( (Np_info + 24)/(8*C) ) - 24; + nr_tbs = (uint32_t)(8 * C * ceil( (Np_info + 24)/(8*C) ) - 24); } else { if (Np_info > 8424){ C = ceil( (Np_info + 24)/8424 ); - nr_tbs = 8 * C * ceil( (Np_info + 24)/(8*C) ) - 24; + nr_tbs = (uint32_t)(8 * C * ceil( (Np_info + 24)/(8*C) ) - 24); + } + else { + nr_tbs = (uint32_t)(8 * ceil( (Np_info + 24)/8 ) - 24); + //printf("n %lf Np_info %f pow %f ceil %f \n",n, Np_info,pow(2,6),ceil( (Np_info + 24)/8 )); } - else - nr_tbs = 8 * ceil( (Np_info + 24)/8 ) - 24; } diff --git a/openair1/PHY/CODING/nr_polar_init.c b/openair1/PHY/CODING/nr_polar_init.c index 895bace829..93bcdecd6e 100644 --- a/openair1/PHY/CODING/nr_polar_init.c +++ b/openair1/PHY/CODING/nr_polar_init.c @@ -36,20 +36,27 @@ #include "PHY/CODING/nrPolar_tools/nr_polar_pbch_defs.h" #include "PHY/NR_TRANSPORT/nr_dci.h" +static int intcmp(const void *p1,const void *p2) { + + return(*(int16_t*)p1 > *(int16_t*)p2); +} + void nr_polar_init(t_nrPolar_paramsPtr *polarParams, - int8_t messageType, - uint16_t messageLength, - uint8_t aggregation_level) + int8_t messageType, + uint16_t messageLength, + uint8_t aggregation_level) { t_nrPolar_paramsPtr currentPtr = *polarParams; uint16_t aggregation_prime = nr_polar_aggregation_prime(aggregation_level); //Parse the list. If the node is already created, return without initialization. while (currentPtr != NULL) { - if (currentPtr->idx == (messageType * messageLength * aggregation_prime)) return; - else currentPtr = currentPtr->nextPtr; + //printf("currentPtr->idx %d, (%d,%d)\n",currentPtr->idx,currentPtr->payloadBits,currentPtr->encoderLength); + if (currentPtr->idx == (messageType * messageLength * aggregation_prime)) return; + else currentPtr = currentPtr->nextPtr; } + // printf("currentPtr %p (polarParams %p)\n",currentPtr,polarParams); //Else, initialize and add node to the end of the linked list. t_nrPolar_paramsPtr newPolarInitNode = malloc(sizeof(t_nrPolar_params)); @@ -57,6 +64,7 @@ void nr_polar_init(t_nrPolar_paramsPtr *polarParams, newPolarInitNode->idx = (messageType * messageLength * aggregation_prime); newPolarInitNode->nextPtr = NULL; + //printf("newPolarInitNode->idx %d, (%d,%d,%d:%d)\n",newPolarInitNode->idx,messageType,messageLength,aggregation_prime,aggregation_level); if (messageType == 0) { //PBCH newPolarInitNode->n_max = NR_POLAR_PBCH_N_MAX; @@ -70,6 +78,7 @@ void nr_polar_init(t_nrPolar_paramsPtr *polarParams, newPolarInitNode->encoderLength = NR_POLAR_PBCH_E; newPolarInitNode->crcCorrectionBits = NR_POLAR_PBCH_CRC_ERROR_CORRECTION_BITS; newPolarInitNode->crc_generator_matrix = crc24c_generator_matrix(newPolarInitNode->payloadBits);//G_P + //printf("Initializing polar parameters for PBCH (K %d, E %d)\n",newPolarInitNode->payloadBits,newPolarInitNode->encoderLength); } else if (messageType == 1) { //DCI newPolarInitNode->n_max = NR_POLAR_DCI_N_MAX; newPolarInitNode->i_il = NR_POLAR_DCI_I_IL; @@ -82,6 +91,7 @@ void nr_polar_init(t_nrPolar_paramsPtr *polarParams, newPolarInitNode->encoderLength = aggregation_level*108; newPolarInitNode->crcCorrectionBits = NR_POLAR_DCI_CRC_ERROR_CORRECTION_BITS; newPolarInitNode->crc_generator_matrix=crc24c_generator_matrix(newPolarInitNode->payloadBits+newPolarInitNode->crcParityBits);//G_P + //printf("Initializing polar parameters for DCI (K %d, E %d, L %d)\n",newPolarInitNode->payloadBits,newPolarInitNode->encoderLength,aggregation_level); } else if (messageType == -1) { //UCI } else { @@ -115,6 +125,10 @@ void nr_polar_init(t_nrPolar_paramsPtr *polarParams, newPolarInitNode->i_il, newPolarInitNode->interleaving_pattern); + newPolarInitNode->deinterleaving_pattern = malloc(sizeof(uint16_t) * newPolarInitNode->K); + for (int i=0;iK;i++) + newPolarInitNode->deinterleaving_pattern[newPolarInitNode->interleaving_pattern[i]] = i; + newPolarInitNode->rate_matching_pattern = malloc(sizeof(uint16_t) * newPolarInitNode->encoderLength); uint16_t *J = malloc(sizeof(uint16_t) * newPolarInitNode->N); nr_polar_rate_matching_pattern(newPolarInitNode->rate_matching_pattern, @@ -139,7 +153,9 @@ void nr_polar_init(t_nrPolar_paramsPtr *polarParams, newPolarInitNode->N, newPolarInitNode->encoderLength, newPolarInitNode->n_pc); - + // sort the Q_I_N array in ascending order (first K positions) + qsort((void*)newPolarInitNode->Q_I_N,newPolarInitNode->K,sizeof(int16_t),intcmp); + newPolarInitNode->channel_interleaver_pattern = malloc(sizeof(uint16_t) * newPolarInitNode->encoderLength); nr_polar_channel_interleaver_pattern(newPolarInitNode->channel_interleaver_pattern, newPolarInitNode->i_bil, @@ -148,6 +164,9 @@ void nr_polar_init(t_nrPolar_paramsPtr *polarParams, free(J); build_decoder_tree(newPolarInitNode); + build_polar_tables(newPolarInitNode); + init_polar_deinterleaver_table(newPolarInitNode); + //printf("decoder tree nodes %d\n",newPolarInitNode->tree.num_nodes); } else { @@ -159,13 +178,18 @@ void nr_polar_init(t_nrPolar_paramsPtr *polarParams, if (currentPtr == NULL) { *polarParams = newPolarInitNode; + //printf("Creating first polarParams entry index %d, %p\n",newPolarInitNode->idx,*polarParams); return; } //Else, add node to the end of the linked list. while (currentPtr->nextPtr != NULL) { - currentPtr = currentPtr->nextPtr; + currentPtr = currentPtr->nextPtr; } currentPtr->nextPtr= newPolarInitNode; + printf("Adding new polarParams entry to list index %d,%p\n", + newPolarInitNode->idx, + currentPtr->nextPtr); + return; } diff --git a/openair1/PHY/CODING/nr_rate_matching.c b/openair1/PHY/CODING/nr_rate_matching.c new file mode 100644 index 0000000000..0bce31630f --- /dev/null +++ b/openair1/PHY/CODING/nr_rate_matching.c @@ -0,0 +1,205 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/* file: nr_rate_matching.c + purpose: Procedures for rate matching/interleaving for NR LDPC + author: hongzhi.wang@tcl.com +*/ + +#include "PHY/defs_gNB.h" +#include "PHY/defs_nr_UE.h" + +//#define RM_DEBUG 1 + +uint8_t index_k0[2][4] = {{0,17,33,56},{0,13,25,43}}; + + +void nr_interleaving_ldpc(uint32_t E, uint8_t Qm, uint8_t *e,uint8_t *f) +{ + uint32_t EQm; + + EQm = E/Qm; + memset(f,0,E*sizeof(uint8_t)); + + for (int j = 0; j< EQm; j++){ + for (int i = 0; i< Qm; i++){ + f[(i+j*Qm)] = e[(i*EQm + j)]; + } + } + +} + + +void nr_deinterleaving_ldpc(uint32_t E, uint8_t Qm, int16_t *e,int16_t *f) +{ + + uint32_t EQm; + + EQm = E/Qm; + + for (int j = 0; j< EQm; j++){ + for (int i = 0; i< Qm; i++){ + e[(i*EQm + j)] = f[(i+j*Qm)]; + } + } +} + + +int nr_rate_matching_ldpc(uint8_t Ilbrm, + uint32_t Tbslbrm, + uint8_t BG, + uint16_t Z, + uint8_t *w, + uint8_t *e, + uint8_t C, + uint8_t rvidx, + uint32_t E) +{ + uint32_t Ncb,ind,k,Nref,N; + + if (C==0) { + printf("nr_rate_matching: invalid parameters (C %d\n",C); + return -1; + } + + //Bit selection + N = (BG==1)?(66*Z):(50*Z); + + if (Ilbrm == 0) + Ncb = N; + else { + Nref = 3*Tbslbrm/(2*C); //R_LBRM = 2/3 + Ncb = min(N, Nref); + } + + ind = (index_k0[BG-1][rvidx]*Ncb/N)*Z; + +#ifdef RM_DEBUG + printf("nr_rate_matching_ldpc: E %d, k0 %d, Ncb %d, rvidx %d\n", E, ind, Ncb, rvidx); +#endif + + k=0; + + for (; (indMAX_NUM_DLSCH_SEGMENTS) { - LOG_E(PHY,"lte_segmentation.c: too many segments %d, B %d, L %d, Bprime %d\n",*C,B,L,Bprime); + LOG_E(PHY,"nr_segmentation.c: too many segments %d, B %d, L %d, Bprime %d\n",*C,B,L,Bprime); return(-1); } // Find K+ Bprime_by_C = Bprime/(*C); - /*if (Bprime <=192) { + if (Bprime <=192) { Kb = 6; } else if (Bprime <=560) { Kb = 8; @@ -73,9 +72,9 @@ int32_t nr_segmentation(unsigned char *input_buffer, Kb = 9; } else if (Bprime <=3840) { Kb = 10;; - } else {*/ + } else { Kb = 22; - //} + } if ((Bprime_by_C%Kb) > 0) @@ -88,97 +87,85 @@ else #endif if (Z <= 2) { - *Kplus = 2; - *Kminus = 0; + *K = 2; } else if (Z<=16) { // increase by 1 byte til here - *Kplus = Z; - *Kminus = Z-1; + *K = Z; } else if (Z <=32) { // increase by 2 bytes til here - *Kplus = (Z>>1)<<1; + *K = (Z>>1)<<1; - if (*Kplus < Z) - *Kplus = *Kplus + 2; + if (*K < Z) + *K = *K + 2; - *Kminus = (*Kplus - 2); } else if (Z <= 64) { // increase by 4 bytes til here - *Kplus = (Z>>2)<<2; + *K = (Z>>2)<<2; - if (*Kplus < Z) - *Kplus = *Kplus + 4; + if (*K < Z) + *K = *K + 4; - *Kminus = (*Kplus - 4); } else if (Z <=128 ) { // increase by 8 bytes til here - *Kplus = (Z>>3)<<3; + *K = (Z>>3)<<3; - if (*Kplus < Z) - *Kplus = *Kplus + 8; + if (*K < Z) + *K = *K + 8; #ifdef DEBUG_SEGMENTATION - printf("Z_by_C %d , Kplus2 %d\n",Z,*Kplus); + printf("Z_by_C %d , K2 %d\n",Z,*K); #endif - *Kminus = (*Kplus - 8); } else if (Z <= 256) { // increase by 4 bytes til here - *Kplus = (Z>>4)<<4; + *K = (Z>>4)<<4; - if (*Kplus < Z) - *Kplus = *Kplus + 16; + if (*K < Z) + *K = *K + 16; - *Kminus = (*Kplus - 16); } else if (Z <= 384) { // increase by 4 bytes til here - *Kplus = (Z>>5)<<5; + *K = (Z>>5)<<5; - if (*Kplus < Z) - *Kplus = *Kplus + 32; + if (*K < Z) + *K = *K + 32; - *Kminus = (*Kplus - 32); } else { //msg("nr_segmentation.c: Illegal codeword size !!!\n"); return(-1); } - *Zout = *Kplus; - *Kplus = *Kplus*Kb; - *Kminus = *Kminus*Kb; + *Zout = *K; + *K = *K*Kb; - - - *F = ((*C)*(*Kplus) - (Bprime)); + *F = ((*K) - Bprime_by_C); #ifdef DEBUG_SEGMENTATION - printf("final nr seg output Z %d Kplus %d F %d \n", *Zout, *Kplus, *F); - printf("C %d, Kplus %d, Kminus %d, Bprime_bytes %d, Bprime %d, F %d\n",*C,*Kplus,*Kminus,Bprime>>3,Bprime,*F); + printf("final nr seg output Z %d K %d F %d \n", *Zout, *K, *F); + printf("C %d, K %d, Bprime_bytes %d, Bprime %d, F %d\n",*C,*K,Bprime>>3,Bprime,*F); #endif if ((input_buffer) && (output_buffers)) { - for (k=0; k<*F>>3; k++) { - output_buffers[0][k] = 0; - } - s=0; for (r=0; r<*C; r++) { - //if (r<(B%(*C))) - Kprime = *Kplus; - //else - // Kprime = *Kminus; + Kprime = Bprime_by_C; while (k<((Kprime - L)>>3)) { output_buffers[r][k] = input_buffer[s]; - // printf("encoding segment %d : byte %d (%d) => %d\n",r,k,Kr>>3,input_buffer[s]); + //printf("encoding segment %d : byte %d (%d) => %d\n",r,k,(Kprime-L)>>3,input_buffer[s]); k++; s++; } - /* - if (*C > 1) { // add CRC - crc = crc24b(output_buffers[r],Kprime-24)>>8; - output_buffers[r][(Kprime-24)>>3] = ((uint8_t*)&crc)[2]; - output_buffers[r][1+((Kprime-24)>>3)] = ((uint8_t*)&crc)[1]; - output_buffers[r][2+((Kprime-24)>>3)] = ((uint8_t*)&crc)[0]; + if (*F>0) + for (k=Kprime>>3; k<(*K)>>3; k++) { + output_buffers[r][k] = NR_NULL; + //printf("r %d filler bits [%d] = %d Kprime %d \n", r,k, output_buffers[r][k], Kprime); } - */ + + if (*C > 1) { // add CRC + crc = crc24b(output_buffers[r],Kprime-L)>>8; + output_buffers[r][(Kprime-L)>>3] = ((uint8_t*)&crc)[2]; + output_buffers[r][1+((Kprime-L)>>3)] = ((uint8_t*)&crc)[1]; + output_buffers[r][2+((Kprime-L)>>3)] = ((uint8_t*)&crc)[0]; + } + k=0; } } @@ -192,12 +179,12 @@ else main() { - unsigned int Kplus,Kminus,C,F,Bbytes; + unsigned int K,C,F,Bbytes, Zout; for (Bbytes=5; Bbytes<8; Bbytes++) { - nr_segmentation(0,0,Bbytes<<3,&C,&Kplus,&Kminus, &F); - printf("Bbytes %d : C %d, Kplus %d, Kminus %d, F %d\n", - Bbytes, C, Kplus, Kminus, F); + nr_segmentation(0,0,Bbytes<<3,&C,&K,&Zout, &F); + printf("Bbytes %d : C %d, K %d, F %d\n", + Bbytes, C, K, F); } } #endif diff --git a/openair1/PHY/INIT/nr_init.c b/openair1/PHY/INIT/nr_init.c index ba9f497e6d..1866eac96d 100644 --- a/openair1/PHY/INIT/nr_init.c +++ b/openair1/PHY/INIT/nr_init.c @@ -24,6 +24,7 @@ #include "PHY/NR_REFSIG/nr_refsig.h" #include "PHY/INIT/phy_init.h" #include "PHY/CODING/nrPolar_tools/nr_polar_pbch_defs.h" +#include "PHY/NR_TRANSPORT/nr_transport.h" #include "RadioResourceConfigCommonSIB.h" #include "RadioResourceConfigDedicated.h" #include "TDD-Config.h" @@ -36,8 +37,8 @@ #include "PHY/LTE_REFSIG/lte_refsig.h" #include "SCHED_NR/fapi_nr_l1.h" -extern uint32_t from_earfcn(int eutra_bandP,uint32_t dl_earfcn); -extern int32_t get_uldl_offset(int eutra_bandP); +extern uint32_t from_nrarfcn(int nr_bandP,uint32_t dl_nrarfcn); +extern int32_t get_uldl_offset(int nr_bandP); int l1_north_init_gNB() { @@ -124,10 +125,11 @@ int phy_init_nr_gNB(PHY_VARS_gNB *gNB, // PBCH DMRS gold sequences generation nr_init_pbch_dmrs(gNB); // Polar encoder init for PBCH + nr_polar_init(&gNB->nrPolar_params, - NR_POLAR_PBCH_MESSAGE_TYPE, - NR_POLAR_PBCH_PAYLOAD_BITS, - NR_POLAR_PBCH_AGGREGATION_LEVEL); + NR_POLAR_PBCH_MESSAGE_TYPE, + NR_POLAR_PBCH_PAYLOAD_BITS, + NR_POLAR_PBCH_AGGREGATION_LEVEL); //PDCCH DMRS init gNB->nr_gold_pdcch_dmrs = (uint32_t ***)malloc16(fp->slots_per_frame*sizeof(uint32_t**)); @@ -145,6 +147,27 @@ int phy_init_nr_gNB(PHY_VARS_gNB *gNB, nr_init_pdcch_dmrs(gNB, cfg->sch_config.physical_cell_id.value); nr_init_pbch_interleaver(gNB->nr_pbch_interleaver); + + //PDSCH DMRS init + gNB->nr_gold_pdsch_dmrs = (uint32_t ****)malloc16(fp->slots_per_frame*sizeof(uint32_t***)); + uint32_t ****pdsch_dmrs = gNB->nr_gold_pdsch_dmrs; + for (int slot=0; slotslots_per_frame; slot++) { + pdsch_dmrs[slot] = (uint32_t ***)malloc16(fp->symbols_per_slot*sizeof(uint32_t**)); + AssertFatal(pdsch_dmrs[slot]!=NULL, "NR init: pdsch_dmrs for slot %d - malloc failed\n", slot); + for (int symb=0; symbsymbols_per_slot; symb++){ + pdsch_dmrs[slot][symb] = (uint32_t **)malloc16(NR_MAX_NB_CODEWORDS*sizeof(uint32_t*)); + AssertFatal(pdsch_dmrs[slot][symb]!=NULL, "NR init: pdsch_dmrs for slot %d symbol %d - malloc failed\n", slot, symb); + for (int q=0; qsch_config.physical_cell_id.value); + /// Transport init necessary for NR synchro + init_nr_transport(gNB); + /* lte_gold(fp,gNB->lte_gold_table,fp->Nid_cell); generate_pcfich_reg_mapping(fp); @@ -164,20 +187,16 @@ int phy_init_nr_gNB(PHY_VARS_gNB *gNB, common_vars->rxdata = (int32_t **)NULL; - common_vars->txdataF = (int32_t **)malloc16(NB_ANTENNA_PORTS_ENB*sizeof(int32_t*)); + common_vars->txdataF = (int32_t **)malloc16(15*sizeof(int32_t*)); common_vars->rxdataF = (int32_t **)malloc16(64*sizeof(int32_t*)); - LOG_D(PHY,"[INIT] NB_ANTENNA_PORTS_ENB:%d fp->nb_antenna_ports_gNB:%d\n", NB_ANTENNA_PORTS_ENB, cfg->rf_config.tx_antenna_ports.value); - - for (i=0; irf_config.tx_antenna_ports.value || i==5) { + for (i=0;i<15;i++){ common_vars->txdataF[i] = (int32_t*)malloc16_clear(fp->samples_per_frame_wCP*sizeof(int32_t) ); LOG_D(PHY,"[INIT] common_vars->txdataF[%d] = %p (%lu bytes)\n", i,common_vars->txdataF[i], fp->samples_per_frame_wCP*sizeof(int32_t)); - } - } + } // Channel estimates for SRS @@ -295,8 +314,8 @@ void phy_config_request(PHY_Config_t *phy_config) { void phy_free_nr_gNB(PHY_VARS_gNB *gNB) { -// NR_DL_FRAME_PARMS* const fp = &gNB->frame_parms; - nfapi_nr_config_request_t *cfg = &gNB->gNB_config; + //NR_DL_FRAME_PARMS* const fp = &gNB->frame_parms; + //nfapi_nr_config_request_t *cfg = &gNB->gNB_config; NR_gNB_COMMON* const common_vars = &gNB->common_vars; LTE_eNB_PUSCH** const pusch_vars = gNB->pusch_vars; LTE_eNB_SRS* const srs_vars = gNB->srs_vars; @@ -305,11 +324,9 @@ void phy_free_nr_gNB(PHY_VARS_gNB *gNB) int i, UE_id; - for (i = 0; i < NB_ANTENNA_PORTS_ENB; i++) { - if (i < cfg->rf_config.tx_antenna_ports.value || i == 5) { - free_and_zero(common_vars->txdataF[i]); + for (i = 0; i < 15; i++) { + free_and_zero(common_vars->txdataF[i]); /* rxdataF[i] is not allocated -> don't free */ - } } free_and_zero(common_vars->txdataF); free_and_zero(common_vars->rxdataF); @@ -370,28 +387,28 @@ void install_schedule_handlers(IF_Module_t *if_inst) /// this function is a temporary addition for NR configuration -void nr_phy_config_request_sim(PHY_VARS_gNB *gNB,int N_RB_DL,int N_RB_UL,int mu) +void nr_phy_config_request_sim(PHY_VARS_gNB *gNB,int N_RB_DL,int N_RB_UL,int mu,int Nid_cell) { NR_DL_FRAME_PARMS *fp = &gNB->frame_parms; nfapi_nr_config_request_t *gNB_config = &gNB->gNB_config; //overwrite for new NR parameters - gNB_config->nfapi_config.rf_bands.rf_band[0] = 22; - gNB_config->nfapi_config.earfcn.value = 6600; + gNB_config->nfapi_config.rf_bands.rf_band[0] = 78; + gNB_config->nfapi_config.nrarfcn.value = 620000; gNB_config->subframe_config.numerology_index_mu.value = mu; gNB_config->subframe_config.duplex_mode.value = TDD; - gNB_config->rf_config.tx_antenna_ports.value = 1; gNB_config->rf_config.dl_carrier_bandwidth.value = N_RB_DL; gNB_config->rf_config.ul_carrier_bandwidth.value = N_RB_UL; gNB_config->sch_config.half_frame_index.value = 0; gNB_config->sch_config.ssb_subcarrier_offset.value = 0; gNB_config->sch_config.n_ssb_crb.value = (N_RB_DL-20); gNB_config->sch_config.ssb_subcarrier_offset.value = 0; + gNB_config->sch_config.physical_cell_id.value=Nid_cell; gNB->mac_enabled = 1; - fp->dl_CarrierFreq = from_earfcn(gNB_config->nfapi_config.rf_bands.rf_band[0],gNB_config->nfapi_config.earfcn.value); + fp->dl_CarrierFreq = from_nrarfcn(gNB_config->nfapi_config.rf_bands.rf_band[0],gNB_config->nfapi_config.nrarfcn.value); fp->ul_CarrierFreq = fp->dl_CarrierFreq - (get_uldl_offset(gNB_config->nfapi_config.rf_bands.rf_band[0])*100000); fp->threequarter_fs = 0; @@ -413,9 +430,8 @@ void nr_phy_config_request(NR_PHY_Config_t *phy_config) gNB_config->nfapi_config.rf_bands.rf_band[0] = phy_config->cfg->nfapi_config.rf_bands.rf_band[0]; //22 - gNB_config->nfapi_config.earfcn.value = phy_config->cfg->nfapi_config.earfcn.value; //6600 + gNB_config->nfapi_config.nrarfcn.value = phy_config->cfg->nfapi_config.nrarfcn.value; //6600 gNB_config->subframe_config.numerology_index_mu.value = phy_config->cfg->subframe_config.numerology_index_mu.value;//1 - gNB_config->rf_config.tx_antenna_ports.value = phy_config->cfg->rf_config.tx_antenna_ports.value; //1 gNB_config->rf_config.dl_carrier_bandwidth.value = phy_config->cfg->rf_config.dl_carrier_bandwidth.value;//106; gNB_config->rf_config.ul_carrier_bandwidth.value = phy_config->cfg->rf_config.ul_carrier_bandwidth.value;//106; gNB_config->sch_config.half_frame_index.value = 0; @@ -432,20 +448,19 @@ void nr_phy_config_request(NR_PHY_Config_t *phy_config) RC.gNB[Mod_id][CC_id]->mac_enabled = 1; - fp->dl_CarrierFreq = from_earfcn(gNB_config->nfapi_config.rf_bands.rf_band[0],gNB_config->nfapi_config.earfcn.value); + fp->dl_CarrierFreq = from_nrarfcn(gNB_config->nfapi_config.rf_bands.rf_band[0],gNB_config->nfapi_config.nrarfcn.value); fp->ul_CarrierFreq = fp->dl_CarrierFreq - (get_uldl_offset(gNB_config->nfapi_config.rf_bands.rf_band[0])*100000); fp->threequarter_fs = 0; - LOG_I(PHY,"Configuring MIB for instance %d, CCid %d : (band %d,N_RB_DL %d, N_RB_UL %d, Nid_cell %d,gNB_tx_antenna_ports %d,DL freq %u)\n", - Mod_id, - CC_id, - gNB_config->nfapi_config.rf_bands.rf_band[0], - gNB_config->rf_config.dl_carrier_bandwidth.value, - gNB_config->rf_config.ul_carrier_bandwidth.value, - gNB_config->sch_config.physical_cell_id.value, - gNB_config->rf_config.tx_antenna_ports.value, - fp->dl_CarrierFreq ); - + LOG_I(PHY,"Configuring MIB for instance %d, CCid %d : (band %d,N_RB_DL %d, N_RB_UL %d, Nid_cell %d,DL freq %u)\n", + Mod_id, + CC_id, + gNB_config->nfapi_config.rf_bands.rf_band[0], + gNB_config->rf_config.dl_carrier_bandwidth.value, + gNB_config->rf_config.ul_carrier_bandwidth.value, + gNB_config->sch_config.physical_cell_id.value, + fp->dl_CarrierFreq ); + nr_init_frame_parms(gNB_config, fp); if (RC.gNB[Mod_id][CC_id]->configured == 1){ @@ -459,3 +474,69 @@ void nr_phy_config_request(NR_PHY_Config_t *phy_config) } + +void init_nr_transport(PHY_VARS_gNB *gNB) { + + int i; + int j; + NR_DL_FRAME_PARMS *fp = &gNB->frame_parms; + nfapi_nr_config_request_t *cfg = &gNB->gNB_config; + + LOG_I(PHY, "Initialise nr transport\n"); + + for (i=0; idlsch[i][j] = new_gNB_dlsch(1,16,NSOFT,0,fp,cfg); + if (!gNB->dlsch[i][j]) { + LOG_E(PHY,"Can't get gNB dlsch structures for UE %d \n", i); + exit(-1); + }/* else { + gNB->dlsch[i][j]->rnti=0; + LOG_D(PHY,"dlsch[%d][%d] => %p rnti:%d\n",i,j,gNB->dlsch[i][j], gNB->dlsch[i][j]->rnti); + }*/ + } + + //LOG_I(PHY,"Allocating Transport Channel Buffer for ULSCH, UE %d\n",i); + //gNB->ulsch[1+i] = new_gNB_ulsch(MAX_TURBO_ITERATIONS,fp->N_RB_UL, 0); + + /*if (!gNB->ulsch[1+i]) { + LOG_E(PHY,"Can't get gNB ulsch structures\n"); + exit(-1); + }*/ + + // this is the transmission mode for the signalling channels + // this will be overwritten with the real transmission mode by the RRC once the UE is connected + //gNB->transmission_mode[i] = fp->nb_antenna_ports_gNB==1 ? 1 : 2; + } + // ULSCH for RA + //gNB->ulsch[0] = new_gNB_ulsch(MAX_TURBO_ITERATIONS, fp->N_RB_UL, 0); + + /*if (!gNB->ulsch[0]) { + LOG_E(PHY,"Can't get gNB ulsch structures\n"); + exit(-1); + }*/ + gNB->dlsch_SI = new_gNB_dlsch(1,8,NSOFT, 0, fp, cfg); + LOG_D(PHY,"gNB %d.%d : SI %p\n",gNB->Mod_id,gNB->CC_id,gNB->dlsch_SI); + gNB->dlsch_ra = new_gNB_dlsch(1,8,NSOFT, 0, fp, cfg); + LOG_D(PHY,"gNB %d.%d : RA %p\n",gNB->Mod_id,gNB->CC_id,gNB->dlsch_ra); + gNB->dlsch_MCH = new_gNB_dlsch(1,8,NSOFT, 0, fp, cfg); + LOG_D(PHY,"gNB %d.%d : MCH %p\n",gNB->Mod_id,gNB->CC_id,gNB->dlsch_MCH); + + + gNB->rx_total_gain_dB=130; + + for(i=0; imu_mimo_mode[i].dl_pow_off = 2; + + gNB->check_for_total_transmissions = 0; + + gNB->check_for_MUMIMO_transmissions = 0; + + gNB->FULL_MUMIMO_transmissions = 0; + + gNB->check_for_SUMIMO_transmissions = 0; + + //fp->pucch_config_common.deltaPUCCH_Shift = 1; + +} diff --git a/openair1/PHY/INIT/nr_init_ru.c b/openair1/PHY/INIT/nr_init_ru.c index 7cd1f1b58f..3ff24e343a 100644 --- a/openair1/PHY/INIT/nr_init_ru.c +++ b/openair1/PHY/INIT/nr_init_ru.c @@ -107,9 +107,7 @@ int nr_phy_init_RU(RU_t *ru) { for (i=0; igNB_list[i]->gNB_config.rf_config.tx_antenna_ports.value); - if (pgNB_list[i]->gNB_config.rf_config.tx_antenna_ports.value || p==5) { - LOG_D(PHY,"[INIT] %s() DO BEAM WEIGHTS nb_antenna_ports_eNB:%d nb_tx:%d\n", __FUNCTION__, ru->gNB_list[i]->gNB_config.rf_config.tx_antenna_ports.value, ru->nb_tx); + if (p == 0|| p==5) { ru->beam_weights[i][p] = (int32_t **)malloc16_clear(ru->nb_tx*sizeof(int32_t*)); for (j=0; jnb_tx; j++) { ru->beam_weights[i][p][j] = (int32_t *)malloc16_clear(fp->ofdm_symbol_size*sizeof(int32_t)); @@ -180,7 +178,7 @@ void nr_phy_free_RU(RU_t *ru) for (i = 0; i < RC.nb_nr_L1_inst; i++) { for (p = 0; p < 15; p++) { - if (p < ru->gNB_list[i]->gNB_config.rf_config.tx_antenna_ports.value || p == 5) { + if (p == 0 || p == 5) { for (j=0; jnb_tx; j++) free_and_zero(ru->beam_weights[i][p][j]); free_and_zero(ru->beam_weights[i][p]); } diff --git a/openair1/PHY/INIT/nr_init_ue.c b/openair1/PHY/INIT/nr_init_ue.c index a2385b998c..fc207cee84 100644 --- a/openair1/PHY/INIT/nr_init_ue.c +++ b/openair1/PHY/INIT/nr_init_ue.c @@ -32,7 +32,7 @@ #include #include "PHY/NR_UE_TRANSPORT/nr_transport_ue.h" #include "PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h" -#include "PHY/LTE_REFSIG/lte_refsig.h" +//#include "PHY/LTE_REFSIG/lte_refsig.h" #include "PHY/CODING/nrPolar_tools/nr_polar_pbch_defs.h" #include "PHY/INIT/phy_init.h" #include "PHY/NR_REFSIG/pss_nr.h" @@ -656,6 +656,7 @@ int init_nr_ue_signal(PHY_VARS_NR_UE *ue, int eNB_id; int th_id; int n_ssb_crb=(fp->N_RB_DL-20); + int k_ssb=0; abstraction_flag = 0; fp->nb_antennas_tx = 1; fp->nb_antennas_rx=1; @@ -664,7 +665,7 @@ int init_nr_ue_signal(PHY_VARS_NR_UE *ue, printf("Initializing UE vars (abstraction %"PRIu8") for eNB TXant %"PRIu8", UE RXant %"PRIu8"\n",abstraction_flag,fp->nb_antennas_tx,fp->nb_antennas_rx); //LOG_D(PHY,"[MSC_NEW][FRAME 00000][PHY_UE][MOD %02u][]\n", ue->Mod_id+NB_eNB_INST); - nr_init_frame_parms_ue(fp,NR_MU_1,NORMAL,fp->N_RB_DL,n_ssb_crb,0); + nr_init_frame_parms_ue(fp,NR_MU_1,NORMAL,fp->N_RB_DL,n_ssb_crb,k_ssb); phy_init_nr_top(ue); // many memory allocation sizes are hard coded @@ -816,10 +817,10 @@ int init_nr_ue_signal(PHY_VARS_NR_UE *ue, // 100 PRBs * 12 REs/PRB * 4 PDCCH SYMBOLS * 2 LLRs/RE for (th_id=0; th_idllr = (uint16_t*)malloc16_clear( 2*4*100*12*sizeof(uint16_t) ); - (*pdcch_vars_th)[th_id][eNB_id]->llr16 = (uint16_t*)malloc16_clear( 2*4*100*12*sizeof(uint16_t) ); - (*pdcch_vars_th)[th_id][eNB_id]->wbar = (uint16_t*)malloc16_clear( 2*4*100*12*sizeof(uint16_t) ); - (*pdcch_vars_th)[th_id][eNB_id]->e_rx = (int8_t*)malloc16_clear( 4*2*100*12 ); + (*pdcch_vars_th)[th_id][eNB_id]->llr = (int16_t*)malloc16_clear( 2*4*100*12*sizeof(uint16_t) ); + (*pdcch_vars_th)[th_id][eNB_id]->llr16 = (int16_t*)malloc16_clear( 2*4*100*12*sizeof(uint16_t) ); + (*pdcch_vars_th)[th_id][eNB_id]->wbar = (int16_t*)malloc16_clear( 2*4*100*12*sizeof(uint16_t) ); + (*pdcch_vars_th)[th_id][eNB_id]->e_rx = (int16_t*)malloc16_clear( 4*2*100*12 ); (*pdcch_vars_th)[th_id][eNB_id]->rxdataF_comp = (int32_t**)malloc16_clear( 8*sizeof(int32_t*) ); (*pdcch_vars_th)[th_id][eNB_id]->dl_ch_rho_ext = (int32_t**)malloc16_clear( 8*sizeof(int32_t*) ); @@ -912,7 +913,6 @@ int init_nr_ue_signal(PHY_VARS_NR_UE *ue, } -/* void init_nr_ue_transport(PHY_VARS_NR_UE *ue,int abstraction_flag) { int i,j,k; @@ -920,32 +920,30 @@ void init_nr_ue_transport(PHY_VARS_NR_UE *ue,int abstraction_flag) { for (i=0; idlsch[k][i][j] = new_ue_dlsch(1,NUMBER_OF_HARQ_PID_MAX,NSOFT,MAX_TURBO_ITERATIONS,ue->frame_parms.N_RB_DL, abstraction_flag))!=NULL,"Can't get ue dlsch structures\n"); + AssertFatal((ue->dlsch[k][i][j] = new_nr_ue_dlsch(1,NUMBER_OF_HARQ_PID_MAX,NSOFT,MAX_LDPC_ITERATIONS,ue->frame_parms.N_RB_DL, abstraction_flag))!=NULL,"Can't get ue dlsch structures\n"); LOG_D(PHY,"dlsch[%d][%d][%d] => %p\n",k,i,j,ue->dlsch[i][j]); } } - AssertFatal((ue->ulsch[i] = new_ue_ulsch(ue->frame_parms.N_RB_UL, abstraction_flag))!=NULL,"Can't get ue ulsch structures\n"); + //AssertFatal((ue->ulsch[i] = new_ue_ulsch(ue->frame_parms.N_RB_UL, abstraction_flag))!=NULL,"Can't get ue ulsch structures\n"); - ue->dlsch_SI[i] = new_ue_dlsch(1,1,NSOFT,MAX_TURBO_ITERATIONS,ue->frame_parms.N_RB_DL, abstraction_flag); - ue->dlsch_ra[i] = new_ue_dlsch(1,1,NSOFT,MAX_TURBO_ITERATIONS,ue->frame_parms.N_RB_DL, abstraction_flag); + ue->dlsch_SI[i] = new_nr_ue_dlsch(1,1,NSOFT,MAX_LDPC_ITERATIONS,ue->frame_parms.N_RB_DL, abstraction_flag); + ue->dlsch_ra[i] = new_nr_ue_dlsch(1,1,NSOFT,MAX_LDPC_ITERATIONS,ue->frame_parms.N_RB_DL, abstraction_flag); ue->transmission_mode[i] = ue->frame_parms.nb_antenna_ports_eNB==1 ? 1 : 2; } //ue->frame_parms.pucch_config_common.deltaPUCCH_Shift = 1; - ue->dlsch_MCH[0] = new_ue_dlsch(1,NUMBER_OF_HARQ_PID_MAX,NSOFT,MAX_TURBO_ITERATIONS_MBSFN,ue->frame_parms.N_RB_DL,0); + ue->dlsch_MCH[0] = new_nr_ue_dlsch(1,NUMBER_OF_HARQ_PID_MAX,NSOFT,MAX_LDPC_ITERATIONS_MBSFN,ue->frame_parms.N_RB_DL,0); -}*/ +} void phy_init_nr_top(PHY_VARS_NR_UE *ue) { NR_DL_FRAME_PARMS *frame_parms = &ue->frame_parms; - NR_UE_DLSCH_t *dlsch0 = ue->dlsch[0][0][0]; - dlsch0 =(NR_UE_DLSCH_t *)malloc16(sizeof(NR_UE_DLSCH_t)); - + crcTableInit(); init_dfts(); @@ -955,12 +953,7 @@ void phy_init_nr_top(PHY_VARS_NR_UE *ue) generate_ul_reference_signal_sequences(SHRT_MAX); // Polar encoder init for PBCH - - ue->nrPolar_params = NULL; - nr_polar_init(&ue->nrPolar_params, - NR_POLAR_PBCH_MESSAGE_TYPE, - NR_POLAR_PBCH_PAYLOAD_BITS, - NR_POLAR_PBCH_AGGREGATION_LEVEL); + //lte_sync_time_init(frame_parms); //generate_ul_ref_sigs(); @@ -990,8 +983,6 @@ void set_default_frame_parms_single(nfapi_nr_config_request_t *config, NR_DL_FRA config->subframe_config.dl_cyclic_prefix_type.value = 0; //NORMAL config->rf_config.dl_carrier_bandwidth.value = 106; config->rf_config.ul_carrier_bandwidth.value = 106; - config->rf_config.tx_antenna_ports.value = 1; - config->rf_config.rx_antenna_ports.value = 1; config->sch_config.physical_cell_id.value = 0; frame_parms->frame_type = FDD; diff --git a/openair1/PHY/INIT/nr_parms.c b/openair1/PHY/INIT/nr_parms.c index 0cc33861ac..484ac50dda 100644 --- a/openair1/PHY/INIT/nr_parms.c +++ b/openair1/PHY/INIT/nr_parms.c @@ -159,6 +159,8 @@ int nr_init_frame_parms0(NR_DL_FRAME_PARMS *fp, fp->symbols_per_slot = ((Ncp == NORMAL)? 14 : 12); // to redefine for different slot formats fp->samples_per_subframe_wCP = fp->ofdm_symbol_size * fp->symbols_per_slot * fp->slots_per_subframe; fp->samples_per_frame_wCP = 10 * fp->samples_per_subframe_wCP; + fp->samples_per_slot_wCP = fp->symbols_per_slot*fp->ofdm_symbol_size; + fp->samples_per_slot = fp->nb_prefix_samples0 + ((fp->symbols_per_slot-1)*fp->nb_prefix_samples) + (fp->symbols_per_slot*fp->ofdm_symbol_size); fp->samples_per_subframe = (fp->samples_per_subframe_wCP + (fp->nb_prefix_samples0 * fp->slots_per_subframe) + (fp->nb_prefix_samples * fp->slots_per_subframe * (fp->symbols_per_slot - 1))); fp->samples_per_frame = 10 * fp->samples_per_subframe; @@ -176,7 +178,9 @@ int nr_init_frame_parms0(NR_DL_FRAME_PARMS *fp, } int nr_init_frame_parms(nfapi_nr_config_request_t* config, - NR_DL_FRAME_PARMS *fp) { + NR_DL_FRAME_PARMS *fp) +{ + return nr_init_frame_parms0(fp, config->subframe_config.numerology_index_mu.value, @@ -191,10 +195,9 @@ int nr_init_frame_parms_ue(NR_DL_FRAME_PARMS *fp, int n_ssb_crb, int ssb_subcarrier_offset) { - + /*n_ssb_crb and ssb_subcarrier_offset are given in 15kHz SCS*/ nr_init_frame_parms0(fp,mu,Ncp,N_RB_DL); - int start_rb = n_ssb_crb / (1<ssb_start_subcarrier = 12 * start_rb + ssb_subcarrier_offset; + fp->ssb_start_subcarrier = (12 * n_ssb_crb + ssb_subcarrier_offset)/(1<Ncp)*(Ns&1)); ///symbol within sub-frame unsigned int nb_prefix_samples = (no_prefix ? 0 : frame_parms->nb_prefix_samples); unsigned int nb_prefix_samples0 = (no_prefix ? 0 : frame_parms->nb_prefix_samples0); - unsigned int subframe_offset;//,subframe_offset_F; + //unsigned int subframe_offset;//,subframe_offset_F; unsigned int slot_offset; //int i; unsigned int frame_length_samples = frame_parms->samples_per_subframe * 10; unsigned int rx_offset; - //NR_UE_PDCCH *pdcch_vars = ue->pdcch_vars[ue->current_thread_id[Ns>>1]][0]; - uint16_t coreset_start_subcarrier = frame_parms->first_carrier_offset+516; + NR_UE_PDCCH *pdcch_vars = ue->pdcch_vars[ue->current_thread_id[Ns>>1]][0]; + uint16_t coreset_start_subcarrier = frame_parms->first_carrier_offset+((int)floor(frame_parms->ssb_start_subcarrier/NR_NB_SC_PER_RB)+pdcch_vars->coreset[0].rb_offset)*NR_NB_SC_PER_RB; uint16_t nb_rb_coreset = 24; - uint16_t bwp_start_subcarrier = frame_parms->first_carrier_offset; - uint16_t nb_rb_pdsch = 100; - - /*LTE_UE_DLSCH_t **dlsch_ue = phy_vars_ue->dlsch_ue[eNB_id]; - unsigned char harq_pid = dlsch_ue[0]->current_harq_pid; - LTE_DL_UE_HARQ_t *dlsch0_harq = dlsch_ue[0]->harq_processes[harq_pid]; - int uespec_pilot[9][1200];*/ + uint16_t bwp_start_subcarrier = frame_parms->first_carrier_offset+516; + uint16_t nb_rb_pdsch = 50; + uint8_t p=0; + uint8_t l0 = 2; void (*dft)(int16_t *,int16_t *, int); int tmp_dft_in[8192] __attribute__ ((aligned (32))); // This is for misalignment issues for 6 and 15 PRBs @@ -102,11 +104,9 @@ int nr_slot_fep(PHY_VARS_NR_UE *ue, } if (no_prefix) { - subframe_offset = frame_parms->ofdm_symbol_size * frame_parms->symbols_per_tti * (Ns>>1); - slot_offset = frame_parms->ofdm_symbol_size * (frame_parms->symbols_per_tti>>1) * (Ns%2); + slot_offset = frame_parms->ofdm_symbol_size * (frame_parms->symbols_per_slot*frame_parms->slots_per_subframe) * (Ns>>1); } else { - subframe_offset = frame_parms->samples_per_tti * (Ns>>1); - slot_offset = (frame_parms->samples_per_tti>>1) * (Ns%2); + slot_offset = (frame_parms->samples_per_subframe) * (Ns>>1); } /*if (l<0 || l>=7-frame_parms->Ncp) { @@ -124,10 +124,16 @@ int nr_slot_fep(PHY_VARS_NR_UE *ue, for (aa=0; aanb_antennas_rx; aa++) { memset(&common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].rxdataF[aa][frame_parms->ofdm_symbol_size*symbol],0,frame_parms->ofdm_symbol_size*sizeof(int)); - rx_offset = sample_offset + slot_offset + nb_prefix_samples0 + subframe_offset - SOFFSET; + rx_offset = sample_offset + slot_offset + nb_prefix_samples0 - SOFFSET; // Align with 256 bit // rx_offset = rx_offset&0xfffffff8; +#ifdef DEBUG_FEP + // if (ue->frame <100) + /*LOG_I(PHY,*/printf("slot_fep: frame %d: slot %d, symbol %d, nb_prefix_samples %d, nb_prefix_samples0 %d, slot_offset %d, sample_offset %d,rx_offset %d, frame_length_samples %d\n", ue->proc.proc_rxtx[(Ns>>1)&1].frame_rx,Ns, symbol, + nb_prefix_samples,nb_prefix_samples0,slot_offset,sample_offset,rx_offset,frame_length_samples); +#endif + if (l==0) { if (rx_offset > (frame_length_samples - frame_parms->ofdm_symbol_size)) @@ -156,12 +162,6 @@ int nr_slot_fep(PHY_VARS_NR_UE *ue, rx_offset += (frame_parms->ofdm_symbol_size+nb_prefix_samples)*l;// + // (frame_parms->ofdm_symbol_size+nb_prefix_samples)*(l-1); -#ifdef DEBUG_FEP - // if (ue->frame <100) - LOG_I(PHY,"slot_fep: frame %d: slot %d, symbol %d, nb_prefix_samples %d, nb_prefix_samples0 %d, slot_offset %d, subframe_offset %d, sample_offset %d,rx_offset %d, frame_length_samples %d\n", ue->proc.proc_rxtx[(Ns>>1)&1].frame_rx,Ns, symbol, - nb_prefix_samples,nb_prefix_samples0,slot_offset,subframe_offset,sample_offset,rx_offset,frame_length_samples); -#endif - if (rx_offset > (frame_length_samples - frame_parms->ofdm_symbol_size)) memcpy((void *)&common_vars->rxdata[aa][frame_length_samples], (void *)&common_vars->rxdata[aa][0], @@ -190,7 +190,7 @@ int nr_slot_fep(PHY_VARS_NR_UE *ue, #ifdef DEBUG_FEP // if (ue->frame <100) - printf("slot_fep: frame %d: symbol %d rx_offset %d\n", ue->proc.proc_rxtx[(Ns>>1)&1].frame_rx, symbol,rx_offset); + printf("slot_fep: frame %d: symbol %d rx_offset %d\n", ue->proc.proc_rxtx[(Ns)&1].frame_rx, symbol,rx_offset); #endif } @@ -266,12 +266,18 @@ int nr_slot_fep(PHY_VARS_NR_UE *ue, #if UE_TIMING_TRACE start_meas(&ue->dlsch_channel_estimation_stats); #endif + + ue->frame_parms.nushift = (p>>1)&1;; + + if (symbol ==l0) nr_pdsch_channel_estimation(ue,eNB_id,0, Ns, + p, l, symbol, bwp_start_subcarrier, nb_rb_pdsch); + #if UE_TIMING_TRACE stop_meas(&ue->dlsch_channel_estimation_stats); #endif diff --git a/openair1/PHY/NR_REFSIG/nr_dmrs_rx.c b/openair1/PHY/NR_REFSIG/nr_dmrs_rx.c index 7472501316..2e5ad47d91 100644 --- a/openair1/PHY/NR_REFSIG/nr_dmrs_rx.c +++ b/openair1/PHY/NR_REFSIG/nr_dmrs_rx.c @@ -20,29 +20,21 @@ */ /*! \file PHY/NR_REFSIG/nr_dl_dmrs.c -* \brief Top-level routines for generating DMRS from 38-211 -* \author -* \date 2018 -* \version 0.1 -* \company Eurecom -* \email: -* \note -* \warning -*/ + * \brief Top-level routines for generating DMRS from 38-211 + * \author + * \date 2018 + * \version 0.1 + * \company Eurecom + * \email: + * \note + * \warning + */ -//#define DEBUG_DL_DMRS //#define NR_PBCH_DMRS_LENGTH_DWORD 5 //#define NR_PBCH_DMRS_LENGTH 144 -#ifdef USER_MODE -#include -#include -#endif - #include "refsig_defs_ue.h" #include "PHY/defs_nr_UE.h" -//#include "nr_mod_table.h" -#include "common/utils/LOG/log.h" /*Table 7.4.1.1.2-1/2 from 38.211 */ int wf1[8][2] = {{1,1},{1,-1},{1,1},{1,-1},{1,1},{1,-1},{1,1},{1,1}}; @@ -51,135 +43,56 @@ int wf2[12][2] = {{1,1},{1,-1},{1,1},{1,-1},{1,1},{1,-1},{1,1},{1,1},{1,1},{1,-1 int wt2[12][2] = {{1,1},{1,1},{1,1},{1,1},{1,1},{1,1},{1,-1},{1,-1},{1,-1},{1,-1},{1,-1},{1,-1}}; -short nr_rx_mod_table[NR_MOD_TABLE_SIZE_SHORT] = {0,0,23170,-23170,-23170,23170,23170,-23170,23170,23170,-23170,-23170,-23170,23170}; +short nr_rx_mod_table[14] = {0,0,23170,-23170,-23170,23170,23170,-23170,23170,23170,-23170,-23170,-23170,23170}; +short nr_rx_nmod_table[14] = {0,0,-23170,23170,23170,-23170,-23170,23170,-23170,-23170,23170,23170,23170,-23170}; -/*int nr_pdcch_dmrs_rx(PHY_VARS_NR_UE *ue, - uint8_t eNB_offset, - unsigned int Ns, - unsigned int nr_gold_pdcch[7][20][3][10], - int32_t *output, - unsigned short p, - int length_dmrs, - unsigned short nb_rb_coreset) -{ - int32_t qpsk[4],n; - int w,ind,l,ind_dword,ind_qpsk_symb,kp,k; - short pamp; - pamp = ONE_OVER_SQRT2_Q15; - // This includes complex conjugate for channel estimation - ((short *)&qpsk[0])[0] = pamp; - ((short *)&qpsk[0])[1] = -pamp; - ((short *)&qpsk[1])[0] = -pamp; - ((short *)&qpsk[1])[1] = -pamp; - ((short *)&qpsk[2])[0] = pamp; - ((short *)&qpsk[2])[1] = pamp; - ((short *)&qpsk[3])[0] = -pamp; - ((short *)&qpsk[3])[1] = pamp; - - if (p==2000) { - for (n=0; n>4; - ind_qpsk_symb = ind&0xf; - - output[k] = qpsk[(nr_gold_pdcch[eNB_offset][Ns][l][ind_dword]>>(2*ind_qpsk_symb))&3]; - -#ifdef DEBUG_DL_DMRS - LOG_I(PHY,"Ns %d, p %d, ind_dword %d, ind_qpsk_symbol %d\n", - Ns,p,idx_dword,idx_qpsk_symb); - LOG_I(PHY,"index = %d\n",(nr_gold_pdsch[0][Ns][lprime][ind_dword]>>(2*ind_qpsk_symb))&3); -#endif - - k++; - } - } - } - } else { - LOG_E(PHY,"Illegal PDCCH DMRS port %d\n",p); - } - - return(0); -}*/ int nr_pdsch_dmrs_rx(PHY_VARS_NR_UE *ue, - uint8_t eNB_offset, - unsigned int Ns, - unsigned int nr_gold_pdsch[2][20][2][21], - int32_t *output, - unsigned short p, - int length_dmrs, - unsigned short nb_rb_pdsch) + unsigned int Ns, + unsigned int *nr_gold_pdsch, + int32_t *output, + unsigned short p, + unsigned char lp, + unsigned short nb_pdsch_rb) { - int32_t qpsk[4],nqpsk[4],*qpsk_p, n; - //int w,mprime,ind,l,ind_dword,ind_qpsk_symb,kp,lp, config_type, k; - int w,ind,ind_dword,ind_qpsk_symb,kp,lp, config_type, k=0; - short pamp; + int8_t w,config_type; + short *mod_table; + unsigned char idx=0; typedef int array_of_w[2]; array_of_w *wf; array_of_w *wt; - config_type = 1; - printf("dmrs config type %d port %d\n", config_type, p); - - // Compute the correct pilot amplitude, sqrt_rho_b = Q3.13 - pamp = ONE_OVER_SQRT2_Q15; - - // This includes complex conjugate for channel estimation - ((short *)&qpsk[0])[0] = pamp; - ((short *)&qpsk[0])[1] = -pamp; - ((short *)&qpsk[1])[0] = -pamp; - ((short *)&qpsk[1])[1] = -pamp; - ((short *)&qpsk[2])[0] = pamp; - ((short *)&qpsk[2])[1] = pamp; - ((short *)&qpsk[3])[0] = -pamp; - ((short *)&qpsk[3])[1] = pamp; - - ((short *)&nqpsk[0])[0] = -pamp; - ((short *)&nqpsk[0])[1] = pamp; - ((short *)&nqpsk[1])[0] = pamp; - ((short *)&nqpsk[1])[1] = pamp; - ((short *)&nqpsk[2])[0] = -pamp; - ((short *)&nqpsk[2])[1] = -pamp; - ((short *)&nqpsk[3])[0] = pamp; - ((short *)&nqpsk[3])[1] = -pamp; + config_type = 0; //to be updated by higher layer wf = (config_type==0) ? wf1 : wf2; wt = (config_type==0) ? wt1 : wt2; if (config_type > 1) - LOG_E(PHY,"Bad PDSCH DMRS config type %d\n", config_type); + LOG_E(PHY,"Bad PDSCH DMRS config type %d\n", config_type); if ((p>=1000) && (p<((config_type==0) ? 1008 : 1012))) { + if (ue->frame_parms.Ncp == NORMAL) { - // r_n from 38.211 7.4.1.1 - for (n=0; n>4; - ind_qpsk_symb = ind&0xf; + w = (wf[p-1000][i&1])*(wt[p-1000][lp]); + mod_table = (w==1) ? nr_rx_mod_table : nr_rx_nmod_table; - output[k] = qpsk_p[(nr_gold_pdsch[0][Ns][lp][ind_dword]>>(2*ind_qpsk_symb))&3]; - - -#ifdef DEBUG_DL_DMRS - LOG_I(PHY,"Ns %d, p %d, ind_dword %d, ind_qpsk_symbol %d\n", - Ns,p,idx_dword,idx_qpsk_symb); - LOG_I(PHY,"index = %d\n",(nr_gold_pdsch[0][Ns][lprime][ind_dword]>>(2*ind_qpsk_symb))&3); + idx = ((((nr_gold_pdsch[(i<<1)>>5])>>((i<<1)&0x1f))&1)<<1) ^ (((nr_gold_pdsch[((i<<1)+1)>>5])>>(((i<<1)+1)&0x1f))&1); + ((int16_t*)output)[i<<1] = mod_table[(NR_MOD_TABLE_QPSK_OFFSET + idx)<<1]; + ((int16_t*)output)[(i<<1)+1] = mod_table[((NR_MOD_TABLE_QPSK_OFFSET + idx)<<1) + 1]; +#ifdef DEBUG_PDSCH + printf("nr_pdsch_dmrs_rx dmrs config type %d port %d nb_pdsch_rb %d\n", config_type, p, nb_pdsch_rb); + printf("wf[%d] = %d wt[%d]= %d\n", i&1, wf[p-1000][i&1], lp, wt[p-1000][lp]); + printf("i %d idx %d pdsch gold %u b0-b1 %d-%d mod_dmrs %d %d\n", i, idx, nr_gold_pdsch[(i<<1)>>5], (((nr_gold_pdsch[(i<<1)>>5])>>((i<<1)&0x1f))&1), + (((nr_gold_pdsch[((i<<1)+1)>>5])>>(((i<<1)+1)&0x1f))&1), ((int16_t*)output)[i<<1], ((int16_t*)output)[(i<<1)+1]); #endif - - k++; - } - } - } + } + } else { + LOG_E(PHY,"extended cp not supported for PDSCH DMRS yet\n"); + } } else { LOG_E(PHY,"Illegal p %d PDSCH DMRS port\n",p); } @@ -188,32 +101,32 @@ int nr_pdsch_dmrs_rx(PHY_VARS_NR_UE *ue, } int nr_pdcch_dmrs_rx(PHY_VARS_NR_UE *ue, - uint8_t eNB_offset, - unsigned int Ns, - unsigned int *nr_gold_pdcch, - int32_t *output, - unsigned short p, - unsigned short nb_rb_coreset) + uint8_t eNB_offset, + unsigned int Ns, + unsigned int *nr_gold_pdcch, + int32_t *output, + unsigned short p, + unsigned short nb_rb_coreset) { - uint8_t idx=0; - //uint8_t pdcch_rb_offset =0; - //nr_gold_pdcch += ((int)floor(ue->frame_parms.ssb_start_subcarrier/12)+pdcch_rb_offset)*3/32; + uint8_t idx=0; + //uint8_t pdcch_rb_offset =0; + //nr_gold_pdcch += ((int)floor(ue->frame_parms.ssb_start_subcarrier/12)+pdcch_rb_offset)*3/32; - if (p==2000) { - for (int i=0; i<((nb_rb_coreset*6)>>1); i++) { - idx = ((((nr_gold_pdcch[(i<<1)>>5])>>((i<<1)&0x1f))&1)<<1) ^ (((nr_gold_pdcch[((i<<1)+1)>>5])>>(((i<<1)+1)&0x1f))&1); - ((int16_t*)output)[i<<1] = nr_rx_mod_table[(NR_MOD_TABLE_QPSK_OFFSET + idx)<<1]; - ((int16_t*)output)[(i<<1)+1] = nr_rx_mod_table[((NR_MOD_TABLE_QPSK_OFFSET + idx)<<1) + 1]; + if (p==2000) { + for (int i=0; i<((nb_rb_coreset*6)>>1); i++) { + idx = ((((nr_gold_pdcch[(i<<1)>>5])>>((i<<1)&0x1f))&1)<<1) ^ (((nr_gold_pdcch[((i<<1)+1)>>5])>>(((i<<1)+1)&0x1f))&1); + ((int16_t*)output)[i<<1] = nr_rx_mod_table[(NR_MOD_TABLE_QPSK_OFFSET + idx)<<1]; + ((int16_t*)output)[(i<<1)+1] = nr_rx_mod_table[((NR_MOD_TABLE_QPSK_OFFSET + idx)<<1) + 1]; #ifdef DEBUG_PDCCH if (i<8) - printf("i %d idx %d pdcch gold %u b0-b1 %d-%d mod_dmrs %d %d\n", i, idx, nr_gold_pdcch[(i<<1)>>5], (((nr_gold_pdcch[(i<<1)>>5])>>((i<<1)&0x1f))&1), - (((nr_gold_pdcch[((i<<1)+1)>>5])>>(((i<<1)+1)&0x1f))&1), ((int16_t*)output)[i<<1], ((int16_t*)output)[(i<<1)+1],&output[0]); + printf("i %d idx %d pdcch gold %u b0-b1 %d-%d mod_dmrs %d %d\n", i, idx, nr_gold_pdcch[(i<<1)>>5], (((nr_gold_pdcch[(i<<1)>>5])>>((i<<1)&0x1f))&1), + (((nr_gold_pdcch[((i<<1)+1)>>5])>>(((i<<1)+1)&0x1f))&1), ((int16_t*)output)[i<<1], ((int16_t*)output)[(i<<1)+1],&output[0]); #endif - } - } + } + } - return(0); + return(0); } int nr_pbch_dmrs_rx(int symbol,unsigned int *nr_gold_pbch,int32_t *output ) @@ -251,3 +164,5 @@ int nr_pbch_dmrs_rx(int symbol,unsigned int *nr_gold_pbch,int32_t *output ) return(0); } + + diff --git a/openair1/PHY/NR_REFSIG/nr_gen_mod_table.m b/openair1/PHY/NR_REFSIG/nr_gen_mod_table.m index a081ad1baa..13a8d4ad09 100644 --- a/openair1/PHY/NR_REFSIG/nr_gen_mod_table.m +++ b/openair1/PHY/NR_REFSIG/nr_gen_mod_table.m @@ -3,12 +3,38 @@ % 0 .. "0" % 1,2 .. BPSK(0),BPSK(1) +K = 1/sqrt(2); + % Amplitude for BPSK (\f$ 2^15 \times 1/\sqrt{2}\f$) BPSK = 23170; % Amplitude for QPSK (\f$ 2^15 \times 1/\sqrt{2}\f$) QPSK = 23170; + +% First Amplitude for QAM16 (\f$ 2^15 \times 2/\sqrt{10}\f$) +QAM16_n1 = 20724; +% Second Amplitude for QAM16 (\f$ 2^15 \times 1/\sqrt{10}\f$) +QAM16_n2 = 10362; + +% First Amplitude for QAM64 (\f$ 2^15 \times 4/\sqrt{42}\f$) +QAM64_n1 = 20225; +% Second Amplitude for QAM64 (\f$ 2^15 \times 2/\sqrt{42}\f$) +QAM64_n2 = 10112; +% Third Amplitude for QAM64 (\f$ 2^15 \times 1/\sqrt{42}\f$) +QAM64_n3 = 5056; + + +% First Amplitude for QAM256 (\f$ 2^15 \times 8/\sqrt{170}\f$) +QAM256_n1 = 20105; +% Second Amplitude for QAM256 (\f$ 2^15 \times 4/\sqrt{170}\f$) +QAM256_n2 = 10052; +% Third Amplitude for QAM256 (\f$ 2^15 \times 2/\sqrt{170}\f$) +QAM256_n3 = 5026; +% Third Amplitude for QAM256 (\f$ 2^15 \times 1/\sqrt{170}\f$) +QAM256_n4 = 2513; + + % BPSK for b = 0:1 bpsk_table(b+1) = (1 - 2*b)*BPSK + 1j*(1-2*b)*BPSK; @@ -17,12 +43,91 @@ end %% QPSK for r=0:1 %0 -- 1 LS for j=0:1 %0 -- 1 MS + %% Formula is dispalayed + qpsk_table(2*r+j+1) = ((1-r*2)*QPSK + 1j*(1-2*j)*QPSK); end end -table = [ 0; bpsk_table(:); qpsk_table(:) ]; +%% QAM16 + for a=-1:2:1 + for b=-1:2:1 + index = (1+a) + (1+b)/2; + qam16_table(index+1) = -a*(QAM16_n1 + (b*QAM16_n2)); + end + end +for b0=0:1 +for b1=0:1 +for b2=0:1 +for b3=0:1 +qam16_table2(b0*8+b1*4+b2*2+b3*1+1) = qam16_table(b0*2+b2*1+1) + 1j*qam16_table(b1*2+b3*1+1); +end +end +end +end + + +%% QAM64 + for a=-1:2:1 + for b=-1:2:1 + for c=-1:2:1 + index = (1+a)*2 + (1+b) + (1+c)/2; + qam64_table(index+1) = -a*(QAM64_n1 + b*(QAM64_n2 + (c*QAM64_n3))); + end + end + end + +for b0=0:1 +for b1=0:1 +for b2=0:1 +for b3=0:1 +for b4=0:1 +for b5=0:1 +qam64_table2(b0*32+b1*16+b2*8+b3*4+b4*2+b5*1+1) = qam64_table(b0*4+b2*2+b4*1+1) + 1j*qam64_table(b1*4+b3*2+b5*1+1); +end +end +end +end +end +end + + +%%256QAM +%% QAM256 ******************************************************************************************* + for a=-1:2:1 + for b=-1:2:1 + for c=-1:2:1 + for d = -1:2:1 + + index = (1+a)*4 + (1+b)*2 + (1+c)+(1+d)/2; + qam256_table(index+1) = -a*(QAM256_n1 + b*(QAM256_n2 + (c*(QAM256_n3+d*QAM256_n4)))); + end + end + end +end +for b0=0:1 +for b1=0:1 +for b2=0:1 +for b3=0:1 +for b4=0:1 +for b5=0:1 +for b6=0:1 +for b7=0:1 + +%%qam64_table2(b0*32+b1*16+b2*8+b3*4+b4*2+b5*1+1) = qam64_table(b0*4+b2*2+b4*1+1) + 1j*qam64_table(b1*4+b3*2+b5*1+1); +qam256_table2(b0*128+b1*64+b2*32+b3*16+b4*8+b5*4+b6*2+b7*1+1) = qam256_table(b0*8+b2*4+b4*2+b6*1+1) + 1j*qam256_table(b1*8+b3*4+b5*2+b7*1+1); +end +end +end +end +end +end +end +end + +table = round(K * [ 0; bpsk_table(:); qpsk_table(:);qam16_table2(:); qam64_table2(:);qam256_table2(:) ]); +%scatter (real(qam256_table2), imag(qam256_table2), 'x'); save mod_table.mat table table2 = zeros(1,length(table)*2); @@ -33,7 +138,10 @@ fd = fopen("nr_mod_table.h","w"); fprintf(fd,"#define NR_MOD_TABLE_SIZE_SHORT %d\n", length(table)*2); fprintf(fd,"#define NR_MOD_TABLE_BPSK_OFFSET %d\n", 1); fprintf(fd,"#define NR_MOD_TABLE_QPSK_OFFSET %d\n", 3); -fprintf(fd,"short nr_mod_table[MOD_TABLE_SIZE_SHORT] = {"); +fprintf(fd,"#define NR_MOD_TABLE_QAM16_OFFSET %d\n", 7); +fprintf(fd,"#define NR_MOD_TABLE_QAM64_OFFSET %d\n", 23); +fprintf(fd,"#define NR_MOD_TABLE_QAM256_OFFSET %d\n", 87); +fprintf(fd,"short nr_mod_table[NR_MOD_TABLE_SIZE_SHORT] = {"); fprintf(fd,"%d,",table2(1:end-1)); fprintf(fd,"%d};\n",table2(end)); fclose(fd); diff --git a/openair1/PHY/NR_REFSIG/nr_gold.c b/openair1/PHY/NR_REFSIG/nr_gold.c index c4e59f907a..968f7fa7a4 100644 --- a/openair1/PHY/NR_REFSIG/nr_gold.c +++ b/openair1/PHY/NR_REFSIG/nr_gold.c @@ -65,7 +65,7 @@ void nr_init_pdcch_dmrs(PHY_VARS_gNB* gNB, uint32_t Nid) for (uint8_t symb=0; symbsymbols_per_slot; symb++) { reset = 1; - x2 = ((1<<17) * (14*slot+symb+1) * ((Nid<<1)+1) + (Nid<<1))&(((uint32_t)1<<31)-1); + x2 = ((1<<17) * (14*slot+symb+1) * ((Nid<<1)+1) + (Nid<<1)); for (uint32_t n=0; nframe_parms; + uint32_t ****pdsch_dmrs = gNB->nr_gold_pdsch_dmrs; + + uint16_t N_n_scid[NR_MAX_NB_CODEWORDS]={Nid, Nid}; // Not correct, appropriate scrambling IDs have to be updated to support DCI 1_1 + uint8_t n_scid=0; // again works only for 1_0 + for (uint8_t slot=0; slotslots_per_frame; slot++) { + for (uint8_t symb=0; symbsymbols_per_slot; symb++) { + for (uint8_t q=0; qsymbols_per_slot*slot+symb+1) * ((N_n_scid[n_scid]<<1)+1) +((N_n_scid[n_scid]<<1)+n_scid)); + for (uint32_t n=0; nframe_parms.Nid_cell; + + //printf("gold pdsch nid %d lbar %d\n",nid,lbar); for (ns=0; ns<20; ns++) { @@ -130,6 +132,8 @@ void nr_gold_pdsch(PHY_VARS_NR_UE* ue,unsigned short lbar,unsigned short *n_idDM x2tmp0 = ((14*ns+(lbar+l)+1)*((nid<<1)+1))<<17; x2 = (x2tmp0+(nid<<1)+nscid)%(1<<31); //cinit + + //printf("ns %d gold pdsch x2 %d\n",ns,x2); x1 = 1+ (1<<31); x2=x2 ^ ((x2 ^ (x2>>1) ^ (x2>>2) ^ (x2>>3))<<31); @@ -143,12 +147,13 @@ void nr_gold_pdsch(PHY_VARS_NR_UE* ue,unsigned short lbar,unsigned short *n_idDM //printf("x1 : %x, x2 : %x\n",x1,x2); } - for (n=0; n<14; n++) { + for (n=0; n<52; n++) { x1 = (x1>>1) ^ (x1>>4); x1 = x1 ^ (x1<<31) ^ (x1<<28); x2 = (x2>>1) ^ (x2>>2) ^ (x2>>3) ^ (x2>>4); x2 = x2 ^ (x2<<31) ^ (x2<<30) ^ (x2<<29) ^ (x2<<28); ue->nr_gold_pdsch[nscid][ns][l][n] = x1^x2; + // if ((ns==2)&&(l==0)) //printf("n=%d : c %x\n",n,x1^x2); } diff --git a/openair1/PHY/NR_REFSIG/nr_mod_table.h b/openair1/PHY/NR_REFSIG/nr_mod_table.h index 3c0c808d1d..a998429699 100644 --- a/openair1/PHY/NR_REFSIG/nr_mod_table.h +++ b/openair1/PHY/NR_REFSIG/nr_mod_table.h @@ -1,4 +1,7 @@ -#define NR_MOD_TABLE_SIZE_SHORT 14 +#define NR_MOD_TABLE_SIZE_SHORT 686 #define NR_MOD_TABLE_BPSK_OFFSET 1 #define NR_MOD_TABLE_QPSK_OFFSET 3 -short nr_mod_table[NR_MOD_TABLE_SIZE_SHORT] = {0,0,23170,23170,-23170,-23170,23170,23170,23170,-23170,-23170,23170,-23170,-23170}; +#define NR_MOD_TABLE_QAM16_OFFSET 7 +#define NR_MOD_TABLE_QAM64_OFFSET 23 +#define NR_MOD_TABLE_QAM256_OFFSET 87 +short nr_mod_table[NR_MOD_TABLE_SIZE_SHORT] = {0,0,16384,16384,-16384,-16384,16384,16384,16384,-16384,-16384,16384,-16384,-16384,7327,7327,7327,21981,21981,7327,21981,21981,7327,-7327,7327,-21981,21981,-7327,21981,-21981,-7327,7327,-7327,21981,-21981,7327,-21981,21981,-7327,-7327,-7327,-21981,-21981,-7327,-21981,-21981,10726,10726,10726,3576,3576,10726,3576,3576,10726,17876,10726,25027,3576,17876,3576,25027,17876,10726,17876,3576,25027,10726,25027,3576,17876,17876,17876,25027,25027,17876,25027,25027,10726,-10726,10726,-3576,3576,-10726,3576,-3576,10726,-17876,10726,-25027,3576,-17876,3576,-25027,17876,-10726,17876,-3576,25027,-10726,25027,-3576,17876,-17876,17876,-25027,25027,-17876,25027,-25027,-10726,10726,-10726,3576,-3576,10726,-3576,3576,-10726,17876,-10726,25027,-3576,17876,-3576,25027,-17876,10726,-17876,3576,-25027,10726,-25027,3576,-17876,17876,-17876,25027,-25027,17876,-25027,25027,-10726,-10726,-10726,-3576,-3576,-10726,-3576,-3576,-10726,-17876,-10726,-25027,-3576,-17876,-3576,-25027,-17876,-10726,-17876,-3576,-25027,-10726,-25027,-3576,-17876,-17876,-17876,-25027,-25027,-17876,-25027,-25027,8886,8886,8886,12439,12439,8886,12439,12439,8886,5332,8886,1778,12439,5332,12439,1778,5332,8886,5332,12439,1778,8886,1778,12439,5332,5332,5332,1778,1778,5332,1778,1778,8886,19547,8886,15993,12439,19547,12439,15993,8886,23101,8886,26655,12439,23101,12439,26655,5332,19547,5332,15993,1778,19547,1778,15993,5332,23101,5332,26655,1778,23101,1778,26655,19547,8886,19547,12439,15993,8886,15993,12439,19547,5332,19547,1778,15993,5332,15993,1778,23101,8886,23101,12439,26655,8886,26655,12439,23101,5332,23101,1778,26655,5332,26655,1778,19547,19547,19547,15993,15993,19547,15993,15993,19547,23101,19547,26655,15993,23101,15993,26655,23101,19547,23101,15993,26655,19547,26655,15993,23101,23101,23101,26655,26655,23101,26655,26655,8886,-8886,8886,-12439,12439,-8886,12439,-12439,8886,-5332,8886,-1778,12439,-5332,12439,-1778,5332,-8886,5332,-12439,1778,-8886,1778,-12439,5332,-5332,5332,-1778,1778,-5332,1778,-1778,8886,-19547,8886,-15993,12439,-19547,12439,-15993,8886,-23101,8886,-26655,12439,-23101,12439,-26655,5332,-19547,5332,-15993,1778,-19547,1778,-15993,5332,-23101,5332,-26655,1778,-23101,1778,-26655,19547,-8886,19547,-12439,15993,-8886,15993,-12439,19547,-5332,19547,-1778,15993,-5332,15993,-1778,23101,-8886,23101,-12439,26655,-8886,26655,-12439,23101,-5332,23101,-1778,26655,-5332,26655,-1778,19547,-19547,19547,-15993,15993,-19547,15993,-15993,19547,-23101,19547,-26655,15993,-23101,15993,-26655,23101,-19547,23101,-15993,26655,-19547,26655,-15993,23101,-23101,23101,-26655,26655,-23101,26655,-26655,-8886,8886,-8886,12439,-12439,8886,-12439,12439,-8886,5332,-8886,1778,-12439,5332,-12439,1778,-5332,8886,-5332,12439,-1778,8886,-1778,12439,-5332,5332,-5332,1778,-1778,5332,-1778,1778,-8886,19547,-8886,15993,-12439,19547,-12439,15993,-8886,23101,-8886,26655,-12439,23101,-12439,26655,-5332,19547,-5332,15993,-1778,19547,-1778,15993,-5332,23101,-5332,26655,-1778,23101,-1778,26655,-19547,8886,-19547,12439,-15993,8886,-15993,12439,-19547,5332,-19547,1778,-15993,5332,-15993,1778,-23101,8886,-23101,12439,-26655,8886,-26655,12439,-23101,5332,-23101,1778,-26655,5332,-26655,1778,-19547,19547,-19547,15993,-15993,19547,-15993,15993,-19547,23101,-19547,26655,-15993,23101,-15993,26655,-23101,19547,-23101,15993,-26655,19547,-26655,15993,-23101,23101,-23101,26655,-26655,23101,-26655,26655,-8886,-8886,-8886,-12439,-12439,-8886,-12439,-12439,-8886,-5332,-8886,-1778,-12439,-5332,-12439,-1778,-5332,-8886,-5332,-12439,-1778,-8886,-1778,-12439,-5332,-5332,-5332,-1778,-1778,-5332,-1778,-1778,-8886,-19547,-8886,-15993,-12439,-19547,-12439,-15993,-8886,-23101,-8886,-26655,-12439,-23101,-12439,-26655,-5332,-19547,-5332,-15993,-1778,-19547,-1778,-15993,-5332,-23101,-5332,-26655,-1778,-23101,-1778,-26655,-19547,-8886,-19547,-12439,-15993,-8886,-15993,-12439,-19547,-5332,-19547,-1778,-15993,-5332,-15993,-1778,-23101,-8886,-23101,-12439,-26655,-8886,-26655,-12439,-23101,-5332,-23101,-1778,-26655,-5332,-26655,-1778,-19547,-19547,-19547,-15993,-15993,-19547,-15993,-15993,-19547,-23101,-19547,-26655,-15993,-23101,-15993,-26655,-23101,-19547,-23101,-15993,-26655,-19547,-26655,-15993,-23101,-23101,-23101,-26655,-26655,-23101,-26655,-26655}; diff --git a/openair1/PHY/NR_REFSIG/nr_refsig.h b/openair1/PHY/NR_REFSIG/nr_refsig.h index f6a8b5683b..cb165eed5e 100644 --- a/openair1/PHY/NR_REFSIG/nr_refsig.h +++ b/openair1/PHY/NR_REFSIG/nr_refsig.h @@ -37,6 +37,6 @@ void nr_init_pbch_dmrs(PHY_VARS_gNB* gNB); @param Nid is used for the initialization of x2, Physical cell Id by default or upper layer configured pdcch_scrambling_ID */ void nr_init_pdcch_dmrs(PHY_VARS_gNB* gNB, uint32_t Nid); - +void nr_init_pdsch_dmrs(PHY_VARS_gNB* gNB, uint32_t Nid); #endif diff --git a/openair1/PHY/NR_REFSIG/pss_nr.h b/openair1/PHY/NR_REFSIG/pss_nr.h index 2f39f26ee4..dc20ab24d7 100644 --- a/openair1/PHY/NR_REFSIG/pss_nr.h +++ b/openair1/PHY/NR_REFSIG/pss_nr.h @@ -137,7 +137,9 @@ int set_pss_nr(int ofdm_symbol_size); int pss_synchro_nr(PHY_VARS_NR_UE *PHY_vars_UE, int rate_change); int pss_search_time_nr(int **rxdata, ///rx data in time domain NR_DL_FRAME_PARMS *frame_parms, - int *eNB_id); + int fo_flag, + int *eNB_id, + int *f_off); #endif #undef EXTERN diff --git a/openair1/PHY/NR_REFSIG/refsig_defs_ue.h b/openair1/PHY/NR_REFSIG/refsig_defs_ue.h index a4cb2be7b8..a47e65d0b0 100644 --- a/openair1/PHY/NR_REFSIG/refsig_defs_ue.h +++ b/openair1/PHY/NR_REFSIG/refsig_defs_ue.h @@ -43,13 +43,12 @@ int nr_pdcch_dmrs_rx(PHY_VARS_NR_UE *ue, unsigned short nb_rb_corset); int nr_pdsch_dmrs_rx(PHY_VARS_NR_UE *ue, - uint8_t eNB_offset, unsigned int Ns, - unsigned int nr_gold_pdsch[2][20][2][21], + unsigned int *nr_gold_pdsch, int32_t *output, unsigned short p, - int length_dmrs, - unsigned short nb_rb_pdsch); + unsigned char lp, + unsigned short nb_pdsch_rb); void nr_gold_pbch(PHY_VARS_NR_UE* ue); diff --git a/openair1/PHY/NR_TRANSPORT/nr_dci.c b/openair1/PHY/NR_TRANSPORT/nr_dci.c index 12f313aee9..bd2abebdf0 100644 --- a/openair1/PHY/NR_TRANSPORT/nr_dci.c +++ b/openair1/PHY/NR_TRANSPORT/nr_dci.c @@ -20,7 +20,7 @@ */ /*! \file PHY/NR_TRANSPORT/nr_dci.c -* \brief Implements DCI encoding/decoding and PDCCH TX/RX procedures (38.212/38.213/38.214). V15.2.0 2018-06. +* \brief Implements DCI encoding and PDCCH TX procedures (38.212/38.213/38.214). V15.2.0 2018-06. * \author Guy De Souza * \date 2018 * \version 0.1 @@ -37,15 +37,15 @@ //#define DEBUG_CHANNEL_CODING #define PDCCH_TEST_POLAR_TEMP_FIX + extern short nr_mod_table[NR_MOD_TABLE_SIZE_SHORT]; uint16_t nr_get_dci_size(nfapi_nr_dci_format_e format, nfapi_nr_rnti_type_e rnti_type, - NR_BWP_PARMS* bwp, + uint16_t N_RB, nfapi_nr_config_request_t* config) { uint16_t size = 0; - uint16_t N_RB = bwp->N_RB; switch(format) { /*Only sizes for 0_0 and 1_0 are correct at the moment*/ @@ -53,7 +53,7 @@ uint16_t nr_get_dci_size(nfapi_nr_dci_format_e format, /// fixed: Format identifier 1, Hop flag 1, MCS 5, NDI 1, RV 2, HARQ PID 4, PUSCH TPC 2 Time Domain assgnmt 4 --20 size += 20; size += (uint8_t)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); // Freq domain assignment -- hopping scenario to be updated - size += nr_get_dci_size(NFAPI_NR_DL_DCI_FORMAT_1_0, rnti_type, bwp, config) - size; // Padding to match 1_0 size + size += nr_get_dci_size(NFAPI_NR_DL_DCI_FORMAT_1_0, rnti_type, N_RB, config) - size; // Padding to match 1_0 size // UL/SUL indicator assumed to be 0 break; @@ -149,11 +149,12 @@ void nr_pdcch_scrambling(uint32_t *in, s = lte_gold_generic(&x1, &x2, reset); reset = 0; if (i){ - in++; - out++; - } + in++; + out++; + } } (*out) ^= ((((*in)>>(i&0x1f))&1) ^ ((s>>(i&0x1f))&1))<<(i&0x1f); + // printf("nr_pdcch_scrambling: in %d => out %d\n",((*in)>>(i&0x1f))&1,((*out)>>(i&0x1f))&1); } } @@ -161,7 +162,7 @@ void nr_pdcch_scrambling(uint32_t *in, uint8_t nr_generate_dci_top(NR_gNB_PDCCH pdcch_vars, t_nrPolar_paramsPtr *nrPolar_params, uint32_t **gold_pdcch_dmrs, - int32_t** txdataF, + int32_t* txdataF, int16_t amp, NR_DL_FRAME_PARMS frame_parms, nfapi_nr_config_request_t config) @@ -169,7 +170,7 @@ uint8_t nr_generate_dci_top(NR_gNB_PDCCH pdcch_vars, int16_t mod_dmrs[NR_MAX_CSET_DURATION][NR_MAX_PDCCH_DMRS_LENGTH>>1]; // 3 for the max coreset duration uint8_t idx=0; - uint16_t a; + //uint16_t a; int k,l,k_prime,dci_idx, dmrs_idx; nr_cce_t cce; nr_reg_t reg; @@ -185,10 +186,12 @@ uint8_t nr_generate_dci_top(NR_gNB_PDCCH pdcch_vars, * in frequency: the first subcarrier is obtained by adding the first CRB overlapping the SSB and the rb_offset * in time: by its first slot and its first symbol*/ uint16_t cset_start_sc = frame_parms.first_carrier_offset + ((int)floor(frame_parms.ssb_start_subcarrier/NR_NB_SC_PER_RB)+pdcch_params.rb_offset)*NR_NB_SC_PER_RB; - uint8_t cset_start_symb = pdcch_params.first_slot*frame_parms.symbols_per_slot + pdcch_params.first_symbol; + // uint8_t cset_start_symb = pdcch_params.first_slot*frame_parms.symbols_per_slot + pdcch_params.first_symbol; + uint8_t cset_start_symb = pdcch_params.first_symbol; uint8_t cset_nsymb = pdcch_params.n_symb; dci_idx = 0; + /// DMRS QPSK modulation /*There is a need to shift from which index the pregenerated DMRS sequence is used * see 38211 r15.2.0 section 7.4.1.3.2: assumption is the reference point for k refers to the DMRS sequence*/ @@ -213,16 +216,16 @@ uint8_t nr_generate_dci_top(NR_gNB_PDCCH pdcch_vars, uint32_t encoder_output[NR_MAX_DCI_SIZE_DWORD]; uint16_t n_RNTI = (pdcch_params.search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_UE_SPECIFIC)? ((pdcch_params.scrambling_id)?pdcch_params.rnti:0) : 0; uint16_t Nid = (pdcch_params.search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_UE_SPECIFIC)? pdcch_params.scrambling_id : config.sch_config.physical_cell_id.value; -#ifdef PDCCH_TEST_POLAR_TEMP_FIX - t_nrPolar_paramsPtr currentPtr = NULL;//, polarParams = NULL; - nr_polar_init(¤tPtr, NR_POLAR_DCI_MESSAGE_TYPE, dci_alloc.size, dci_alloc.L); +//#ifdef PDCCH_TEST_POLAR_TEMP_FIX +// nr_polar_init(¤tPtr, NR_POLAR_DCI_MESSAGE_TYPE, dci_alloc.size, dci_alloc.L); // t_nrPolar_paramsPtr currentPtr = nr_polar_params(*nrPolar_params, NR_POLAR_DCI_MESSAGE_TYPE, dci_alloc.size, dci_alloc.L); -#else +//#else nr_polar_init(nrPolar_params, NR_POLAR_DCI_MESSAGE_TYPE, dci_alloc.size, dci_alloc.L); t_nrPolar_paramsPtr currentPtr = nr_polar_params(*nrPolar_params, NR_POLAR_DCI_MESSAGE_TYPE, dci_alloc.size, dci_alloc.L); -#endif +//#endif -polar_encoder_dci(dci_alloc.dci_pdu, encoder_output, currentPtr, pdcch_params.rnti); + //polar_encoder_dci(dci_alloc.dci_pdu, encoder_output, currentPtr, pdcch_params.rnti); + polar_encoder_fast(dci_alloc.dci_pdu, encoder_output, pdcch_params.rnti,currentPtr); #ifdef DEBUG_CHANNEL_CODING printf("polar rnti %d\n",pdcch_params.rnti); @@ -257,10 +260,7 @@ printf("scrambled output: [0]->0x%08x \t [1]->0x%08x \t [2]->0x%08x \t [3]->0x%0 } /// Resource mapping - a = (config.rf_config.tx_antenna_ports.value == 1) ? amp : (amp*ONE_OVER_SQRT2_Q15)>>15; - for (int aa = 0; aa < config.rf_config.tx_antenna_ports.value; aa++) - { if (cset_start_sc >= frame_parms.ofdm_symbol_size) cset_start_sc -= frame_parms.ofdm_symbol_size; @@ -277,20 +277,23 @@ printf("scrambled output: [0]->0x%08x \t [1]->0x%08x \t [2]->0x%08x \t [3]->0x%0 } } #ifdef DEBUG_DCI -printf("\n Ordered REG list:\n"); -for (int i=0; i> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = (a * mod_dmrs[l][(dmrs_idx<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms.ofdm_symbol_size + k)<<1] = (amp * mod_dmrs[l][dmrs_idx<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = (amp * mod_dmrs[l][(dmrs_idx<<1) + 1]) >> 15; +#ifdef DEBUG_PDCCH_DMRS + printf("symbol %d position %d => (%d,%d)\n",l,k,((int16_t*)txdataF[aa])[(l*frame_parms.ofdm_symbol_size + k)<<1] , ((int16_t*)txdataF[aa])[((l*frame_parms.ofdm_symbol_size + k)<<1)+1]); +#endif k+=4; if (k >= frame_parms.ofdm_symbol_size) k -= frame_parms.ofdm_symbol_size; @@ -312,15 +315,18 @@ printf("\n"); for (int m=0; m> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = (a * mod_dmrs[l][(dmrs_idx<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms.ofdm_symbol_size + k)<<1] = (amp * mod_dmrs[l][dmrs_idx<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = (amp * mod_dmrs[l][(dmrs_idx<<1) + 1]) >> 15; +#ifdef DEBUG_PDCCH_DMRS + printf("l %d position %d => (%d,%d)\n",l,k,((int16_t*)txdataF[aa])[(l*frame_parms.ofdm_symbol_size + k)<<1] , ((int16_t*)txdataF[aa])[((l*frame_parms.ofdm_symbol_size + k)<<1)+1]); +#endif k_prime++; dmrs_idx++; } } else { // DCI payload - ((int16_t*)txdataF[aa])[(l*frame_parms.ofdm_symbol_size + k)<<1] = (a * mod_dci[dci_idx<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = (a * mod_dci[(dci_idx<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms.ofdm_symbol_size + k)<<1] = (amp * mod_dci[dci_idx<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = (amp * mod_dci[(dci_idx<<1) + 1]) >> 15; //printf("dci output %d %d\n",(a * mod_dci[dci_idx<<1]) >> 15, (a * mod_dci[(dci_idx<<1) + 1]) >> 15); dci_idx++; } @@ -329,11 +335,6 @@ printf("\n"); k -= frame_parms.ofdm_symbol_size; } } - } - -#ifdef DEBUG_DCI - write_output("txdataF_dci.m", "txdataF_dci", txdataF[0], frame_parms.samples_per_frame_wCP>>1, 1, 1); -#endif return 0; } diff --git a/openair1/PHY/NR_TRANSPORT/nr_dci.h b/openair1/PHY/NR_TRANSPORT/nr_dci.h index 8d79311f35..4bf250dbff 100644 --- a/openair1/PHY/NR_TRANSPORT/nr_dci.h +++ b/openair1/PHY/NR_TRANSPORT/nr_dci.h @@ -29,13 +29,13 @@ typedef unsigned __int128 uint128_t; uint16_t nr_get_dci_size(nfapi_nr_dci_format_e format, nfapi_nr_rnti_type_e rnti_type, - NR_BWP_PARMS* bwp, + uint16_t N_RB, nfapi_nr_config_request_t* config); uint8_t nr_generate_dci_top(NR_gNB_PDCCH pdcch_vars, t_nrPolar_paramsPtr *nrPolar_params, uint32_t **gold_pdcch_dmrs, - int32_t** txdataF, + int32_t* txdataF, int16_t amp, NR_DL_FRAME_PARMS frame_parms, nfapi_nr_config_request_t config); @@ -51,7 +51,8 @@ void nr_fill_dci_and_dlsch(PHY_VARS_gNB *gNB, int subframe, gNB_L1_rxtx_proc_t *proc, NR_gNB_DCI_ALLOC_t *dci_alloc, - nfapi_nr_dl_config_request_pdu_t *pdu); + nfapi_nr_dl_config_dci_dl_pdu *pdu, + nfapi_nr_dl_config_dlsch_pdu *dlsch_pdu); void nr_fill_cce_list(NR_gNB_DCI_ALLOC_t* dci_alloc, uint16_t n_shift, uint8_t m); diff --git a/openair1/PHY/NR_TRANSPORT/nr_dci_tools.c b/openair1/PHY/NR_TRANSPORT/nr_dci_tools.c index c1a17e6da7..1d29b057fc 100644 --- a/openair1/PHY/NR_TRANSPORT/nr_dci_tools.c +++ b/openair1/PHY/NR_TRANSPORT/nr_dci_tools.c @@ -32,6 +32,10 @@ #include "nr_dci.h" +//#define DEBUG_FILL_DCI + +#include "nr_dlsch.h" + void nr_fill_cce_list(NR_gNB_DCI_ALLOC_t* dci_alloc, uint16_t n_shift, uint8_t m) { @@ -61,7 +65,7 @@ void nr_fill_cce_list(NR_gNB_DCI_ALLOC_t* dci_alloc, uint16_t n_shift, uint8_t m AssertFatal(cond==0, "CCE to REG interleaving: Invalid configuration leading to non integer C\n"); C = N_reg/(bsize*R); - tmp = L * (( Y + (uint16_t)(floor((m*N_cce)/(L*M_s_max))) + n_CI ) % ((uint16_t)floor(N_cce/L))); + tmp = L * (( Y + (m*N_cce)/(L*M_s_max) + n_CI ) % (N_cce/L)); LOG_I(PHY, "CCE list generation for candidate %d: bundle size %d ilv size %d tmp %d\n", m, bsize, R, tmp); for (uint8_t cce_idx=0; cce_idxframe_parms; - uint8_t n_shift; - uint32_t *dci_pdu = dci_alloc->dci_pdu; - memset((void*)dci_pdu,0,4*sizeof(uint32_t)); - nfapi_nr_dl_config_dci_dl_pdu_rel15_t *pdu_rel15 = &pdu->dci_dl_pdu.dci_dl_pdu_rel15; - nfapi_nr_dl_config_pdcch_parameters_rel15_t *params_rel15 = &pdu->dci_dl_pdu.pdcch_params_rel15; - nfapi_nr_config_request_t *cfg = &gNB->gNB_config; - uint16_t N_RB = fp->initial_bwp_dl.N_RB; + uint8_t n_shift; + + + uint64_t *dci_pdu = (uint64_t*)dci_alloc->dci_pdu; + memset((void*)dci_pdu,0,2*sizeof(uint64_t)); + nfapi_nr_dl_config_dci_dl_pdu_rel15_t *pdu_rel15 = &pdcch_pdu->dci_dl_pdu_rel15; + nfapi_nr_dl_config_pdcch_parameters_rel15_t *params_rel15 = &pdcch_pdu->pdcch_params_rel15; + + + nfapi_nr_config_request_t *cfg = &gNB->gNB_config; + NR_gNB_DLSCH_t *dlsch = gNB->dlsch[0][0]; + NR_DL_gNB_HARQ_t **harq = dlsch->harq_processes; + + uint16_t N_RB = params_rel15->n_RB_BWP; uint8_t fsize=0, pos=0, cand_idx=0; + + dci_alloc->L = 8; + memcpy((void*)&dci_alloc->pdcch_params, (void*)params_rel15, sizeof(nfapi_nr_dl_config_pdcch_parameters_rel15_t)); + dci_alloc->size = nr_get_dci_size(dci_alloc->pdcch_params.dci_format, + dci_alloc->pdcch_params.rnti_type, + N_RB, + cfg); + + AssertFatal(dci_alloc->size<=64, "DCI sizes above 64 bits not yet supported"); + n_shift = (dci_alloc->pdcch_params.config_type == NFAPI_NR_CSET_CONFIG_MIB_SIB1)? + cfg->sch_config.physical_cell_id.value : dci_alloc->pdcch_params.shift_index; + nr_fill_cce_list(dci_alloc, n_shift, cand_idx); + /// Payload generation switch(params_rel15->dci_format) { - case NFAPI_NR_DL_DCI_FORMAT_1_0: - switch(params_rel15->rnti_type) { - case NFAPI_NR_RNTI_RA: - // Freq domain assignment - fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); - for (int i=0; ifrequency_domain_assignment>>(fsize-i-1))&1)<time_domain_assignment>>(3-i))&1)<vrb_to_prb_mapping&1)<mcs>>(4-i))&1)<tb_scaling>>(1-i))&1)<rnti_type) { + case NFAPI_NR_RNTI_RA: + // printf("dci_pdu %llx\n",(unsigned long long int)*dci_pdu); + // Freq domain assignment + fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); + pos=fsize; + *dci_pdu |= ((pdu_rel15->frequency_domain_assignment&((1<size-pos)); +#ifdef DEBUG_FILL_DCI + printf("frequency-domain assignment %d (%d bits)=> %d (%llx)\n",pdu_rel15->frequency_domain_assignment,fsize,dci_alloc->size-pos,(unsigned long long int)*dci_pdu); +#endif + // Time domain assignment + pos+=4; + *dci_pdu |= ((pdu_rel15->time_domain_assignment&0xf) << (dci_alloc->size-pos)); +#ifdef DEBUG_FILL_DCI + printf("time-domain assignment %d (3 bits)=> %d (%llx)\n",pdu_rel15->time_domain_assignment,dci_alloc->size-pos,(unsigned long long int)*dci_pdu); +#endif + // VRB to PRB mapping + + pos++; + *dci_pdu |= (pdu_rel15->vrb_to_prb_mapping&0x1)<<(dci_alloc->size-pos); +#ifdef DEBUG_FILL_DCI + printf("vrb to prb mapping %d (1 bits)=> %d (%llx)\n",pdu_rel15->vrb_to_prb_mapping,dci_alloc->size-pos,(unsigned long long int)*dci_pdu); +#endif + // MCS + pos+=5; + *dci_pdu |= (pdu_rel15->mcs&0x1f)<<(dci_alloc->size-pos); +#ifdef DEBUG_FILL_DCI + printf("mcs %d (5 bits)=> %d (%llx)\n",pdu_rel15->mcs,dci_alloc->size-pos,(unsigned long long int)*dci_pdu); +#endif + // TB scaling + pos+=2; + *dci_pdu |= (pdu_rel15->tb_scaling&0x3)<<(dci_alloc->size-pos); +#ifdef DEBUG_FILL_DCI + printf("tb_scaling %d (2 bits)=> %d (%llx)\n",pdu_rel15->tb_scaling,dci_alloc->size-pos,(unsigned long long int)*dci_pdu); +#endif break; - case NFAPI_NR_UL_DCI_FORMAT_0_0: + case NFAPI_NR_RNTI_C: + // indicating a DL DCI format 1bit + pos++; + *dci_pdu |= (pdu_rel15->format_indicator&1)<<(dci_alloc->size-pos); + // Freq domain assignment (275rb >> fsize = 16) + fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); + pos+=fsize; + *dci_pdu |= ((pdu_rel15->frequency_domain_assignment&((1<size-pos)); + + if (((pdu_rel15->frequency_domain_assignment+1)&1) ==0) //fsize are all 1 38.212 p86 + { + // ra_preamble_index 6 bits + pos+=6; + *dci_pdu |= ((pdu_rel15->ra_preamble_index&0x3f)<<(dci_alloc->size-pos)); + + // UL/SUL indicator 1 bit + pos++; + *dci_pdu |= (pdu_rel15->ul_sul_indicator&1)<<(dci_alloc->size-pos); + + // SS/PBCH index 6 bits + pos+=6; + *dci_pdu |= ((pdu_rel15->ss_pbch_index&0x3f)<<(dci_alloc->size-pos)); + + // prach_mask_index 4 bits + pos+=4; + *dci_pdu |= ((pdu_rel15->prach_mask_index&0xf)<<(dci_alloc->size-pos)); + + } //end if + + else { + + // Time domain assignment 4bit + + pos+=4; + *dci_pdu |= ((pdu_rel15->time_domain_assignment&0xf) << (dci_alloc->size-pos)); + + + // VRB to PRB mapping 1bit + pos++; + *dci_pdu |= (pdu_rel15->vrb_to_prb_mapping&1)<<(dci_alloc->size-pos); + + // MCS 5bit //bit over 32, so dci_pdu ++ + pos+=5; + *dci_pdu |= (pdu_rel15->mcs&0x1f)<<(dci_alloc->size-pos); + + // New data indicator 1bit + pos++; + *dci_pdu |= (pdu_rel15->ndi&1)<<(dci_alloc->size-pos); + + // Redundancy version 2bit + pos+=2; + *dci_pdu |= (pdu_rel15->rv&0x3)<<(dci_alloc->size-pos); + + // HARQ process number 4bit + pos+=4; + *dci_pdu |= ((pdu_rel15->harq_pid&0xf)<<(dci_alloc->size-pos)); + + // Downlink assignment index 2bit + pos+=2; + *dci_pdu |= ((pdu_rel15->dai&3)<<(dci_alloc->size-pos)); + + // TPC command for scheduled PUCCH 2bit + pos+=2; + *dci_pdu |= ((pdu_rel15->tpc&3)<<(dci_alloc->size-pos)); + + // PUCCH resource indicator 3bit + pos+=3; + *dci_pdu |= ((pdu_rel15->pucch_resource_indicator&0x7)<<(dci_alloc->size-pos)); + + // PDSCH-to-HARQ_feedback timing indicator 3bit + pos+=3; + *dci_pdu |= ((pdu_rel15->pdsch_to_harq_feedback_timing_indicator&0x7)<<(dci_alloc->size-pos)); + + } //end else break; + + case NFAPI_NR_RNTI_P: + + // Short Messages Indicator – 2 bits + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->short_messages_indicator>>(1-i))&1)<<(dci_alloc->size-pos++); + // Short Messages – 8 bits + for (int i=0; i<8; i++) + *dci_pdu |= ((pdu_rel15->short_messages>>(7-i))&1)<<(dci_alloc->size-pos++); + // Freq domain assignment 0-16 bit + fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); + for (int i=0; ifrequency_domain_assignment>>(fsize-i-1))&1)<<(dci_alloc->size-pos++); + // Time domain assignment 4 bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->time_domain_assignment>>(3-i))&1)<<(dci_alloc->size-pos++); + // VRB to PRB mapping 1 bit + *dci_pdu |= (pdu_rel15->vrb_to_prb_mapping&1)<<(dci_alloc->size-pos++); + // MCS 5 bit + for (int i=0; i<5; i++) + *dci_pdu |= ((pdu_rel15->mcs>>(4-i))&1)<<(dci_alloc->size-pos++); + + // TB scaling 2 bit + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->tb_scaling>>(1-i))&1)<<(dci_alloc->size-pos++); + + + break; + + case NFAPI_NR_RNTI_SI: + // Freq domain assignment 0-16 bit + fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); + for (int i=0; ifrequency_domain_assignment>>(fsize-i-1))&1)<<(dci_alloc->size-pos++); + // Time domain assignment 4 bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->time_domain_assignment>>(3-i))&1)<<(dci_alloc->size-pos++); + // VRB to PRB mapping 1 bit + *dci_pdu |= (pdu_rel15->vrb_to_prb_mapping&1)<<(dci_alloc->size-pos++); + // MCS 5bit //bit over 32, so dci_pdu ++ + for (int i=0; i<5; i++) + *dci_pdu |= ((pdu_rel15->mcs>>(4-i))&1)<<(dci_alloc->size-pos++); + // Redundancy version 2bit + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->rv>>(1-i))&1)<<(dci_alloc->size-pos++); + + break; + + case NFAPI_NR_RNTI_TC: + // indicating a DL DCI format 1bit + *dci_pdu |= (pdu_rel15->format_indicator&1)<<(dci_alloc->size-pos++); + // Freq domain assignment 0-16 bit + fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); + for (int i=0; ifrequency_domain_assignment>>(fsize-i-1))&1)<<(dci_alloc->size-pos++); + // Time domain assignment 4 bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->time_domain_assignment>>(3-i))&1)<<(dci_alloc->size-pos++); + // VRB to PRB mapping 1 bit + *dci_pdu |= (pdu_rel15->vrb_to_prb_mapping&1)<<(dci_alloc->size-pos++); + // MCS 5bit //bit over 32, so dci_pdu ++ + for (int i=0; i<5; i++) + *dci_pdu |= ((pdu_rel15->mcs>>(4-i))&1)<<(dci_alloc->size-pos++); + // New data indicator 1bit + *dci_pdu |= (pdu_rel15->ndi&1)<<(dci_alloc->size-pos++); + // Redundancy version 2bit + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->rv>>(1-i))&1)<<(dci_alloc->size-pos++); + // HARQ process number 4bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->harq_pid>>(3-i))&1)<<(dci_alloc->size-pos++); + + // Downlink assignment index – 2 bits + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->dai>>(1-i))&1)<<(dci_alloc->size-pos++); + + // TPC command for scheduled PUCCH – 2 bits + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->tpc>>(1-i))&1)<<(dci_alloc->size-pos++); + + + // LOG_I(PHY, "DCI PDU: [0]->0x%08llx \t [1]->0x%08llx \t [2]->0x%08llx \t [3]->0x%08llx\n", + // dci_pdu[0], dci_pdu[1], dci_pdu[2], dci_pdu[3]); + + + // PDSCH-to-HARQ_feedback timing indicator – 3 bits + for (int i=0; i<3; i++) + *dci_pdu |= ((pdu_rel15->pdsch_to_harq_feedback_timing_indicator>>(2-i))&1)<<(dci_alloc->size-pos++); + + break; + } + + case NFAPI_NR_UL_DCI_FORMAT_0_0: + switch(params_rel15->rnti_type) + { + case NFAPI_NR_RNTI_C: + // indicating a DL DCI format 1bit + *dci_pdu |= (pdu_rel15->format_indicator&1)<<(dci_alloc->size-pos++); + // Freq domain assignment max 16 bit + fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); + for (int i=0; ifrequency_domain_assignment>>(fsize-i-1))&1)<<(dci_alloc->size-pos++); + // Time domain assignment 4bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->time_domain_assignment>>(3-i))&1)<<(dci_alloc->size-pos++); + // Frequency hopping flag – 1 bit + *dci_pdu |= (pdu_rel15->frequency_hopping_flag&1)<<(dci_alloc->size-pos++); + // MCS 5 bit + for (int i=0; i<5; i++) + *dci_pdu |= ((pdu_rel15->mcs>>(4-i))&1)<<(dci_alloc->size-pos++); + // New data indicator 1bit + *dci_pdu |= (pdu_rel15->ndi&1)<<(dci_alloc->size-pos++); + // Redundancy version 2bit + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->rv>>(1-i))&1)<<(dci_alloc->size-pos++); + // HARQ process number 4bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->harq_pid>>(3-i))&1)<<(dci_alloc->size-pos++); + + // TPC command for scheduled PUSCH – 2 bits + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->tpc>>(1-i))&1)<<(dci_alloc->size-pos++); + + // Padding bits + for(int a = pos;a<32;a++) + *dci_pdu |= (pdu_rel15->padding&1)<<(dci_alloc->size-pos++); + + // UL/SUL indicator – 1 bit + if (cfg->pucch_config.pucch_GroupHopping.value) + *dci_pdu |= (pdu_rel15->ul_sul_indicator&1)<<(dci_alloc->size-pos++); + + break; + + case NFAPI_NR_RNTI_TC: + + // indicating a DL DCI format 1bit + *dci_pdu |= (pdu_rel15->format_indicator&1)<<(dci_alloc->size-pos++); + // Freq domain assignment max 16 bit + fsize = (int)ceil( log2( (N_RB*(N_RB+1))>>1 ) ); + for (int i=0; ifrequency_domain_assignment>>(fsize-i-1))&1)<<(dci_alloc->size-pos++); + // Time domain assignment 4bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->time_domain_assignment>>(3-i))&1)<<(dci_alloc->size-pos++); + // Frequency hopping flag – 1 bit + *dci_pdu |= (pdu_rel15->frequency_hopping_flag&1)<<(dci_alloc->size-pos++); + // MCS 5 bit + for (int i=0; i<5; i++) + *dci_pdu |= ((pdu_rel15->mcs>>(4-i))&1)<<(dci_alloc->size-pos++); + // New data indicator 1bit + *dci_pdu |= (pdu_rel15->ndi&1)<<(dci_alloc->size-pos++); + // Redundancy version 2bit + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->rv>>(1-i))&1)<<(dci_alloc->size-pos++); + // HARQ process number 4bit + for (int i=0; i<4; i++) + *dci_pdu |= ((pdu_rel15->harq_pid>>(3-i))&1)<<(dci_alloc->size-pos++); + + // TPC command for scheduled PUSCH – 2 bits + for (int i=0; i<2; i++) + *dci_pdu |= ((pdu_rel15->tpc>>(1-i))&1)<<(dci_alloc->size-pos++); + + // Padding bits + for(int a = pos;a<32;a++) + *dci_pdu |= (pdu_rel15->padding&1)<<(dci_alloc->size-pos++); + + // UL/SUL indicator – 1 bit + if (cfg->pucch_config.pucch_GroupHopping.value) + *dci_pdu |= (pdu_rel15->ul_sul_indicator&1)<<(dci_alloc->size-pos++); + + break; + } + break; } - LOG_I(PHY, "DCI PDU: [0]->0x%08x \t [1]->0x%08x \t [2]->0x%08x \t [3]->0x%08x\n", - dci_pdu[0], dci_pdu[1], dci_pdu[2], dci_pdu[3]); - - /// rest of DCI alloc - dci_alloc->L = 8; - memcpy((void*)&dci_alloc->pdcch_params, (void*)params_rel15, sizeof(nfapi_nr_dl_config_pdcch_parameters_rel15_t)); - dci_alloc->size = nr_get_dci_size(dci_alloc->pdcch_params.dci_format, - dci_alloc->pdcch_params.rnti_type, - &fp->initial_bwp_dl, - cfg); - n_shift = (dci_alloc->pdcch_params.config_type == NFAPI_NR_CSET_CONFIG_MIB_SIB1)? - cfg->sch_config.physical_cell_id.value : dci_alloc->pdcch_params.shift_index; - nr_fill_cce_list(dci_alloc, n_shift, cand_idx); + LOG_I(PHY, "DCI PDU: [0]->0x%16llxx \t [1]->0x%16llx \n",(long long unsigned int)dci_pdu[0], (long long unsigned int)dci_pdu[1]); LOG_I(PHY, "DCI type %d payload (size %d) generated on candidate %d\n", dci_alloc->pdcch_params.dci_format, dci_alloc->size, cand_idx); + /// DLSCH struct + memcpy((void*)&harq[dci_alloc->harq_pid]->dlsch_pdu, (void*)dlsch_pdu, sizeof(nfapi_nr_dl_config_dlsch_pdu)); + } diff --git a/openair1/PHY/NR_TRANSPORT/nr_dlsch.c b/openair1/PHY/NR_TRANSPORT/nr_dlsch.c new file mode 100644 index 0000000000..b65b48c39e --- /dev/null +++ b/openair1/PHY/NR_TRANSPORT/nr_dlsch.c @@ -0,0 +1,374 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/NR_TRANSPORT/dlsch_decoding.c +* \brief Top-level routines for transmission of the PDSCH 38211 v 15.2.0 +* \author Guy De Souza +* \date 2018 +* \version 0.1 +* \company Eurecom +* \email: desouza@eurecom.fr +* \note +* \warning +*/ + +#include "nr_dlsch.h" +#include "nr_dci.h" +#include "nr_sch_dmrs.h" + +//#define DEBUG_DLSCH +//#define DEBUG_DLSCH_MAPPING + +uint8_t mod_order[5] = {1, 2, 4, 6, 8}; +uint16_t mod_offset[5] = {1,3,7,23,87}; + +void nr_pdsch_codeword_scrambling(uint8_t *in, + uint16_t size, + uint8_t q, + uint32_t Nid, + uint32_t n_RNTI, + uint32_t* out) { + + uint8_t reset, b_idx; + uint32_t x1, x2, s=0; + + reset = 1; + x2 = (n_RNTI<<15) + (q<<14) + Nid; + + for (int i=0; i>b_idx)&1))<>b_idx)&1)<<(order-j-1); + } + + out[i<<1] = nr_mod_table[(offset+idx)<<1]; + out[(i<<1)+1] = nr_mod_table[((offset+idx)<<1)+1]; + } +} + +void nr_pdsch_codeword_modulation(uint32_t *in, + uint8_t Qm, + uint32_t length, + int16_t *out) { + + uint16_t offset = (Qm==2)? NR_MOD_TABLE_QPSK_OFFSET : (Qm==4)? NR_MOD_TABLE_QAM16_OFFSET : \ + (Qm==6)? NR_MOD_TABLE_QAM64_OFFSET: (Qm==8)? NR_MOD_TABLE_QAM256_OFFSET : 0; + AssertFatal(offset, "Invalid modulation order %d\n", Qm); + + for (int i=0; i>b_idx)&1)<<(Qm-j-1); + } + + out[i<<1] = nr_mod_table[(offset+idx)<<1]; + out[(i<<1)+1] = nr_mod_table[((offset+idx)<<1)+1]; + } +} + +void nr_pdsch_layer_mapping(int16_t **mod_symbs, + uint8_t n_layers, + uint16_t n_symbs, + int16_t **tx_layers) { + + switch (n_layers) { + + case 1: + memcpy((void*)tx_layers[0], (void*)mod_symbs[0], (n_symbs<<1)*sizeof(int16_t)); + break; + + case 2: + case 3: + case 4: + for (int i=0; i>1; i++) + for (int l=0; l<2; l++) { + tx_layers[l][i<<1] = mod_symbs[0][((i<<1)+l)<<1]; + tx_layers[l][(i<<1)+1] = mod_symbs[0][(((i<<1)+l)<<1)+1]; + } + for (int i=0; i>2; i++) + for (int l=0; l<3; l++) { + tx_layers[l][i<<1] = mod_symbs[q][((i<<2)+l)<<1]; + tx_layers[l][(i<<1)+1] = mod_symbs[q][(((i<<2)+l)<<1)+1]; + } + break; + + default: + AssertFatal(0, "Invalid number of layers %d\n", n_layers); + } +} + +static inline uint16_t get_pdsch_dmrs_idx(uint8_t n, uint8_t k_prime, uint8_t delta, uint8_t dmrs_type) { + uint16_t dmrs_idx = (dmrs_type)? (6*n+k_prime+delta):((n<<2)+(k_prime<<1)+delta); + return dmrs_idx; +} + +uint8_t nr_generate_pdsch(NR_gNB_DLSCH_t dlsch, + NR_gNB_DCI_ALLOC_t dci_alloc, + uint32_t ***pdsch_dmrs, + int32_t** txdataF, + int16_t amp, + uint8_t slot, + NR_DL_FRAME_PARMS frame_parms, + nfapi_nr_config_request_t config) { + + NR_DL_gNB_HARQ_t *harq = dlsch.harq_processes[dci_alloc.harq_pid]; + nfapi_nr_dl_config_dlsch_pdu_rel15_t *rel15 = &harq->dlsch_pdu.dlsch_pdu_rel15; + nfapi_nr_dl_config_pdcch_parameters_rel15_t pdcch_params = dci_alloc.pdcch_params; + uint32_t scrambled_output[NR_MAX_NB_CODEWORDS][NR_MAX_PDSCH_ENCODED_LENGTH>>5]; + int16_t **mod_symbs = (int16_t**)dlsch.mod_symbs; + int16_t **tx_layers = (int16_t**)dlsch.txdataF; + int8_t Wf[2], Wt[2], l0, l_prime[2], delta; + uint16_t nb_symbols = rel15->nb_mod_symbols; + uint8_t Qm = rel15->modulation_order; + uint16_t encoded_length = nb_symbols*Qm; + + /// CRC, coding, interleaving and rate matching + nr_dlsch_encoding(harq->pdu, slot, &dlsch, &frame_parms); +#ifdef DEBUG_DLSCH +printf("PDSCH encoding:\nPayload:\n"); +for (int i=0; i>7; i++) { + for (int j=0; j<16; j++) + printf("0x%02x\t", harq->pdu[(i<<4)+j]); + printf("\n"); +} +printf("\nEncoded payload:\n"); +for (int i=0; i>3; i++) { + for (int j=0; j<8; j++) + printf("%d", harq->f[(i<<3)+j]); + printf("\t"); +} +printf("\n"); +#endif + + /// scrambling + for (int q=0; qnb_codewords; q++) + memset((void*)scrambled_output[q], 0, (encoded_length>>5)*sizeof(uint32_t)); + uint16_t n_RNTI = (pdcch_params.search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_UE_SPECIFIC)? \ + ((pdcch_params.scrambling_id)?pdcch_params.rnti:0) : 0; + uint16_t Nid = (pdcch_params.search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_UE_SPECIFIC)? \ + pdcch_params.scrambling_id : config.sch_config.physical_cell_id.value; + for (int q=0; qnb_codewords; q++) + nr_pdsch_codeword_scrambling(harq->f, + encoded_length, + q, + Nid, + n_RNTI, + scrambled_output[q]); +#ifdef DEBUG_DLSCH +printf("PDSCH scrambling:\n"); +for (int i=0; i>8; i++) { + for (int j=0; j<8; j++) + printf("0x%08x\t", scrambled_output[0][(i<<3)+j]); + printf("\n"); +} +#endif + + /// Modulation + for (int q=0; qnb_codewords; q++) + nr_pdsch_codeword_modulation(scrambled_output[q], + Qm, + encoded_length, + mod_symbs[q]); +#ifdef DEBUG_DLSCH +printf("PDSCH Modulation: Qm %d(%d)\n", Qm, nb_symbols); +for (int i=0; i>3; i++) { + for (int j=0; j<8; j++) { + printf("%d %d\t", mod_symbs[0][((i<<3)+j)<<1], mod_symbs[0][(((i<<3)+j)<<1)+1]); + } + printf("\n"); +} +#endif + + + /// Layer mapping + nr_pdsch_layer_mapping(mod_symbs, + rel15->nb_layers, + nb_symbols, + tx_layers); +#ifdef DEBUG_DLSCH +printf("Layer mapping (%d layers):\n", rel15->nb_layers); +for (int l=0; lnb_layers; l++) + for (int i=0; i<(nb_symbols/rel15->nb_layers)>>3; i++) { + for (int j=0; j<8; j++) { + printf("%d %d\t", tx_layers[l][((i<<3)+j)<<1], tx_layers[l][(((i<<3)+j)<<1)+1]); + } + printf("\n"); + } +#endif + + /// Antenna port mapping + //to be moved to init phase potentially, for now tx_layers 1-8 are mapped on antenna ports 1000-1007 + + /// DMRS QPSK modulation + uint16_t n_dmrs = (rel15->n_prb*rel15->nb_re_dmrs)<<1; + int16_t mod_dmrs[n_dmrs<<1]; + uint8_t dmrs_type = config.pdsch_config.dmrs_type.value; + l0 = get_l0(dmrs_type, 2);//config.pdsch_config.dmrs_typeA_position.value); + nr_modulation(pdsch_dmrs[l0][0], n_dmrs, MOD_QPSK, mod_dmrs); // currently only codeword 0 is modulated +#ifdef DEBUG_DLSCH +printf("DMRS modulation (single symbol %d, %d symbols, type %d):\n", l0, n_dmrs>>1, dmrs_type); +for (int i=0; i>4; i++) { + for (int j=0; j<8; j++) { + printf("%d %d\t", mod_dmrs[((i<<3)+j)<<1], mod_dmrs[(((i<<3)+j)<<1)+1]); + } + printf("\n"); +} +#endif + + /// Resource mapping + + + // Non interleaved VRB to PRB mapping + uint16_t start_sc = frame_parms.first_carrier_offset + frame_parms.ssb_start_subcarrier; + if (start_sc >= frame_parms.ofdm_symbol_size) + start_sc -= frame_parms.ofdm_symbol_size; + /*rel15->start_prb*NR_NB_SC_PER_RB + + ((pdcch_params.search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_COMMON) && (pdcch_params.dci_format == NFAPI_NR_DL_DCI_FORMAT_1_0))?\ + ((frame_parms.ssb_start_subcarrier/NR_NB_SC_PER_RB + pdcch_params.rb_offset)*NR_NB_SC_PER_RB) : 0;*/ + +#ifdef DEBUG_DLSCH_MAPPING +printf("PDSCH resource mapping started (start SC %d\tstart symbol %d\tN_PRB %d\tnb_symbols %d)\n", +start_sc, rel15->start_symbol, rel15->n_prb, rel15->nb_symbols); +#endif + + for (int ap=0; apnb_layers; ap++) { + + // DMRS params for this ap + get_Wt(Wt, ap, dmrs_type); + get_Wf(Wf, ap, dmrs_type); + delta = get_delta(ap, dmrs_type); + l_prime[0] = 0; // single symbol ap 0 + uint8_t dmrs_symbol = l0+l_prime[0]; +#ifdef DEBUG_DLSCH_MAPPING +printf("DMRS params for ap %d: Wt %d %d \t Wf %d %d \t delta %d \t l_prime %d \t l0 %d\tDMRS symbol %d\n", +ap, Wt[0], Wt[1], Wf[0], Wf[1], delta, l_prime[0], l0, dmrs_symbol); +#endif + uint8_t k_prime=0; + uint16_t m=0, n=0, dmrs_idx=0, k=0; + + for (int l=rel15->start_symbol; lstart_symbol+rel15->nb_symbols; l++) { + k = start_sc; + for (int i=0; in_prb*NR_NB_SC_PER_RB; i++) { +// Note the dmrs index modulo 2048 only works for that symbol size -- to be fixed soon +// Also the different amplitudes are for debug purposes and temporary + if ((l == dmrs_symbol) && (k == ((start_sc+get_pdsch_dmrs_idx(n, k_prime, delta, dmrs_type))&((1<<11)-1)))) { + ((int16_t*)txdataF[ap])[(l*frame_parms.ofdm_symbol_size + k)<<1] = (Wt[l_prime[0]]*Wf[k_prime]*(amp>>1)*mod_dmrs[dmrs_idx<<1]) >> 15; + ((int16_t*)txdataF[ap])[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = (Wt[l_prime[0]]*Wf[k_prime]*(amp>>1)*mod_dmrs[(dmrs_idx<<1) + 1]) >> 15; +#ifdef DEBUG_DLSCH_MAPPING +printf("dmrs_idx %d\t l %d \t k %d \t k_prime %d \t n %d \t txdataF: %d %d\n", +dmrs_idx, l, k, k_prime, n, ((int16_t*)txdataF[ap])[(l*frame_parms.ofdm_symbol_size + k)<<1], +((int16_t*)txdataF[ap])[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1]); +#endif + dmrs_idx++; + k_prime++; + k_prime&=1; + n+=(k_prime)?0:1; + } + + else { + + ((int16_t*)txdataF[ap])[(l*frame_parms.ofdm_symbol_size + k)<<1] = ((amp<<1) * tx_layers[ap][m<<1]) >> 15; + ((int16_t*)txdataF[ap])[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1] = ((amp<<1) * tx_layers[ap][(m<<1) + 1]) >> 15; +#ifdef DEBUG_DLSCH_MAPPING +printf("m %d\t l %d \t k %d \t txdataF: %d %d\n", +m, l, k, ((int16_t*)txdataF[ap])[(l*frame_parms.ofdm_symbol_size + k)<<1], +((int16_t*)txdataF[ap])[((l*frame_parms.ofdm_symbol_size + k)<<1) + 1]); +#endif + m++; + } + if (++k >= frame_parms.ofdm_symbol_size) + k -= frame_parms.ofdm_symbol_size; + } + } + } + return 0; +} diff --git a/openair1/PHY/NR_TRANSPORT/nr_dlsch.h b/openair1/PHY/NR_TRANSPORT/nr_dlsch.h new file mode 100644 index 0000000000..06969eded0 --- /dev/null +++ b/openair1/PHY/NR_TRANSPORT/nr_dlsch.h @@ -0,0 +1,105 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/LTE_TRANSPORT/defs.h +* \brief data structures for PDSCH/DLSCH/PUSCH/ULSCH physical and transport channel descriptors (TX/RX) +* \author R. Knopp +* \date 2011 +* \version 0.1 +* \company Eurecom +* \email: raymond.knopp@eurecom.fr, florian.kaltenberger@eurecom.fr, oscar.tonelli@yahoo.it +* \note +* \warning +*/ + +#include "PHY/defs_gNB.h" + +extern short nr_mod_table[NR_MOD_TABLE_SIZE_SHORT]; + +void nr_get_time_domain_allocation_type(nfapi_nr_config_request_t config, + nfapi_nr_dl_config_dci_dl_pdu dci_pdu, + nfapi_nr_dl_config_dlsch_pdu *dlsch_pdu); + +void nr_check_time_alloc(uint8_t S, uint8_t L, nfapi_nr_config_request_t config); + +uint16_t get_RIV(uint16_t rb_start, uint16_t L, uint16_t N_RB); + +uint16_t get_SLIV(uint8_t S, uint8_t L); + +uint8_t nr_get_S(uint8_t row_idx, uint8_t CP, uint8_t time_alloc_type, uint8_t dmrs_typeA_position); + +void nr_get_rbg_parms(NR_BWP_PARMS* bwp, uint8_t config_type); + +void nr_get_rbg_list(uint32_t bitmap, uint8_t n_rbg, uint8_t* rbg_list); + +void nr_get_PRG_parms(NR_BWP_PARMS* bwp, NR_gNB_DCI_ALLOC_t dci_alloc, uint8_t prb_bundling_type); + +uint8_t nr_get_Qm(uint8_t Imcs, uint8_t table_idx); + +uint32_t nr_get_code_rate(uint8_t Imcs, uint8_t table_idx); + +void nr_get_tbs(nfapi_nr_dl_config_dlsch_pdu *dlsch_pdu, + nfapi_nr_dl_config_dci_dl_pdu dci_pdu, + nfapi_nr_config_request_t config); + +void nr_pdsch_codeword_scrambling(uint8_t *in, + uint16_t size, + uint8_t q, + uint32_t Nid, + uint32_t n_RNTI, + uint32_t* out); + +void nr_pdsch_codeword_modulation(uint32_t *in, + uint8_t Qm, + uint32_t length, + int16_t *out); + +void nr_pdsch_layer_mapping(int16_t **mod_symbs, + uint8_t n_layers, + uint16_t n_symbs, + int16_t **tx_layers); + +uint8_t nr_generate_pdsch(NR_gNB_DLSCH_t dlsch, + NR_gNB_DCI_ALLOC_t dci_alloc, + uint32_t ***pdsch_dmrs, + int32_t** txdataF, + int16_t amp, + uint8_t subframe, + NR_DL_FRAME_PARMS frame_parms, + nfapi_nr_config_request_t config); + +/** \brief Computes available bits G. + @param nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs */ +uint32_t nr_get_G(uint16_t nb_rb, uint16_t nb_symb_sch,uint8_t nb_re_dmrs,uint16_t length_dmrs,uint8_t Qm, uint8_t Nl); + +uint32_t nr_get_E(uint32_t G, uint8_t C, uint8_t Qm, uint8_t Nl, uint8_t r); + +void free_gNB_dlsch(NR_gNB_DLSCH_t *dlsch); + +void clean_gNB_dlsch(NR_gNB_DLSCH_t *dlsch); + +int nr_dlsch_encoding(unsigned char *a, + uint8_t subframe, + NR_gNB_DLSCH_t *dlsch, + NR_DL_FRAME_PARMS* frame_parms); + +void nr_emulate_dlsch_payload(uint8_t* payload, uint16_t size); + diff --git a/openair1/PHY/NR_TRANSPORT/nr_dlsch_coding.c b/openair1/PHY/NR_TRANSPORT/nr_dlsch_coding.c new file mode 100644 index 0000000000..8a2877cc17 --- /dev/null +++ b/openair1/PHY/NR_TRANSPORT/nr_dlsch_coding.c @@ -0,0 +1,448 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.0 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/LTE_TRANSPORT/dlsch_coding.c +* \brief Top-level routines for implementing LDPC-coded (DLSCH) transport channels from 38-212, 15.2 +* \author H.Wang +* \date 2018 +* \version 0.1 +* \company Eurecom +* \email: +* \note +* \warning +*/ + +#include "PHY/defs_gNB.h" +#include "PHY/phy_extern.h" +#include "PHY/CODING/coding_extern.h" +#include "PHY/CODING/coding_defs.h" +#include "PHY/CODING/lte_interleaver_inline.h" +#include "PHY/CODING/nrLDPC_encoder/defs.h" +#include "PHY/NR_TRANSPORT/nr_transport.h" +#include "PHY/NR_TRANSPORT/nr_transport_common_proto.h" +#include "PHY/NR_TRANSPORT/nr_dlsch.h" +#include "SCHED_NR/sched_nr.h" +#include "defs.h" +#include "common/utils/LOG/vcd_signal_dumper.h" +#include "common/utils/LOG/log.h" +#include + +//#define DEBUG_DLSCH_CODING +//#define DEBUG_DLSCH_FREE 1 + +void free_gNB_dlsch(NR_gNB_DLSCH_t *dlsch) +{ + int i; + int r; + + if (dlsch) { +#ifdef DEBUG_DLSCH_FREE + printf("Freeing dlsch %p\n",dlsch); +#endif + + for (i=0; iMdlharq; i++) { +#ifdef DEBUG_DLSCH_FREE + printf("Freeing dlsch process %d\n",i); +#endif + + if (dlsch->harq_processes[i]) { +#ifdef DEBUG_DLSCH_FREE + printf("Freeing dlsch process %d (%p)\n",i,dlsch->harq_processes[i]); +#endif + + if (dlsch->harq_processes[i]->b) { + free16(dlsch->harq_processes[i]->b,MAX_DLSCH_PAYLOAD_BYTES); + dlsch->harq_processes[i]->b = NULL; +#ifdef DEBUG_DLSCH_FREE + printf("Freeing dlsch process %d b (%p)\n",i,dlsch->harq_processes[i]->b); +#endif + } + +#ifdef DEBUG_DLSCH_FREE + printf("Freeing dlsch process %d c (%p)\n",i,dlsch->harq_processes[i]->c); +#endif + + for (r=0; rharq_processes[i]->c[r]); +#endif + + if (dlsch->harq_processes[i]->c[r]) { + free16(dlsch->harq_processes[i]->c[r],1056); + dlsch->harq_processes[i]->c[r] = NULL; + } + if (dlsch->harq_processes[i]->d[r]) { + free16(dlsch->harq_processes[i]->d[r],3*8448); + dlsch->harq_processes[i]->d[r] = NULL; + } + + } + free16(dlsch->harq_processes[i],sizeof(NR_DL_gNB_HARQ_t)); + dlsch->harq_processes[i] = NULL; + } + } + + free16(dlsch,sizeof(NR_gNB_DLSCH_t)); + dlsch = NULL; + } + +} + +NR_gNB_DLSCH_t *new_gNB_dlsch(unsigned char Kmimo, + unsigned char Mdlharq, + uint32_t Nsoft, + uint8_t abstraction_flag, + NR_DL_FRAME_PARMS *frame_parms, + nfapi_nr_config_request_t *config) +{ + + NR_gNB_DLSCH_t *dlsch; + unsigned char exit_flag = 0,i,r,aa,layer; + int re; + unsigned char bw_scaling =1; + uint16_t N_RB = config->rf_config.dl_carrier_bandwidth.value; + + switch (N_RB) { + + case 106: + bw_scaling =2; + break; + + default: + bw_scaling =1; + break; + } + + dlsch = (NR_gNB_DLSCH_t *)malloc16(sizeof(NR_gNB_DLSCH_t)); + + if (dlsch) { + bzero(dlsch,sizeof(NR_gNB_DLSCH_t)); + dlsch->Kmimo = Kmimo; + dlsch->Mdlharq = Mdlharq; + dlsch->Mlimit = 4; + dlsch->Nsoft = Nsoft; + + for (layer=0; layerue_spec_bf_weights[layer] = (int32_t**)malloc16(64*sizeof(int32_t*)); + + for (aa=0; aa<64; aa++) { + dlsch->ue_spec_bf_weights[layer][aa] = (int32_t *)malloc16(OFDM_SYMBOL_SIZE_COMPLEX_SAMPLES*sizeof(int32_t)); + for (re=0;reue_spec_bf_weights[layer][aa][re] = 0x00007fff; + } + } + + dlsch->txdataF[layer] = (int32_t *)malloc16((NR_MAX_PDSCH_ENCODED_LENGTH>>1)*sizeof(int32_t*)); + } + + for (int q=0; qmod_symbs[q] = (int32_t *)malloc16((NR_MAX_PDSCH_ENCODED_LENGTH>>1)*sizeof(int32_t*)); + + dlsch->calib_dl_ch_estimates = (int32_t**)malloc16(64*sizeof(int32_t*)); + for (aa=0; aa<64; aa++) { + dlsch->calib_dl_ch_estimates[aa] = (int32_t *)malloc16(OFDM_SYMBOL_SIZE_COMPLEX_SAMPLES*sizeof(int32_t)); + + } + + for (i=0; i<10; i++) + dlsch->harq_ids[i] = 0; + + + for (i=0; iharq_processes[i] = (NR_DL_gNB_HARQ_t *)malloc16(sizeof(NR_DL_gNB_HARQ_t)); + LOG_T(PHY, "Required mem size %d (bw scaling %d), dlsch->harq_processes[%d] %p\n", + MAX_NR_DLSCH_PAYLOAD_BYTES/bw_scaling,bw_scaling, i,dlsch->harq_processes[i]); + + if (dlsch->harq_processes[i]) { + bzero(dlsch->harq_processes[i],sizeof(NR_DL_gNB_HARQ_t)); + // dlsch->harq_processes[i]->first_tx=1; + dlsch->harq_processes[i]->b = (unsigned char*)malloc16(MAX_NR_DLSCH_PAYLOAD_BYTES/bw_scaling); + dlsch->harq_processes[i]->pdu = (uint8_t*)malloc16(MAX_NR_DLSCH_PAYLOAD_BYTES/bw_scaling); + if (dlsch->harq_processes[i]->pdu) { + bzero(dlsch->harq_processes[i]->pdu,MAX_NR_DLSCH_PAYLOAD_BYTES/bw_scaling); + nr_emulate_dlsch_payload(dlsch->harq_processes[i]->pdu, (MAX_NR_DLSCH_PAYLOAD_BYTES/bw_scaling)>>3); + } else { + printf("Can't allocate PDU\n"); + exit_flag=1; + } + + if (dlsch->harq_processes[i]->b) { + bzero(dlsch->harq_processes[i]->b,MAX_NR_DLSCH_PAYLOAD_BYTES/bw_scaling); + } else { + printf("Can't get b\n"); + exit_flag=1; + } + + if (abstraction_flag==0) { + for (r=0; rharq_processes[i]->c[r] = (uint8_t*)malloc16(8448); + dlsch->harq_processes[i]->d[r] = (uint8_t*)malloc16(68*384); + if (dlsch->harq_processes[i]->c[r]) { + bzero(dlsch->harq_processes[i]->c[r],8448); + } else { + printf("Can't get c\n"); + exit_flag=2; + } + if (dlsch->harq_processes[i]->d[r]) { + bzero(dlsch->harq_processes[i]->d[r],(3*8448)); + } else { + printf("Can't get d\n"); + exit_flag=2; + } + } + } + } else { + printf("Can't get harq_p %d\n",i); + exit_flag=3; + } + } + + if (exit_flag==0) { + for (i=0; iharq_processes[i]->round=0; + } + + return(dlsch); + } + } + + LOG_D(PHY,"new_gNB_dlsch exit flag %d, size of %ld\n", + exit_flag, sizeof(NR_gNB_DLSCH_t)); + free_gNB_dlsch(dlsch); + return(NULL); + + +} + +void clean_gNB_dlsch(NR_gNB_DLSCH_t *dlsch) +{ + + unsigned char Mdlharq; + unsigned char i,j,r; + + if (dlsch) { + Mdlharq = dlsch->Mdlharq; + dlsch->rnti = 0; + dlsch->active = 0; + + for (i=0; i<10; i++) + dlsch->harq_ids[i] = Mdlharq; + + for (i=0; iharq_processes[i]) { + // dlsch->harq_processes[i]->Ndi = 0; + //dlsch->harq_processes[i]->status = 0; + dlsch->harq_processes[i]->round = 0; + + for (j=0; j<96; j++) + for (r=0; rharq_processes[i]->d[r]) + dlsch->harq_processes[i]->d[r][j] = NR_NULL; + + } + } + } +} + +int nr_dlsch_encoding(unsigned char *a, + uint8_t slot, + NR_gNB_DLSCH_t *dlsch, + NR_DL_FRAME_PARMS* frame_parms) +{ + + unsigned int G; + unsigned int crc=1; + + uint8_t harq_pid = dlsch->harq_ids[slot]; + nfapi_nr_dl_config_dlsch_pdu_rel15_t rel15 = dlsch->harq_processes[harq_pid]->dlsch_pdu.dlsch_pdu_rel15; + uint16_t nb_rb = rel15.n_prb; + uint8_t nb_symb_sch = rel15.nb_symbols; + uint32_t A, Z; + uint32_t *pz = &Z; + uint8_t mod_order = rel15.modulation_order; + uint16_t Kr=0,r,r_offset=0,Kr_bytes; + uint8_t *d_tmp[MAX_NUM_DLSCH_SEGMENTS]; + uint8_t kb,BG=1; + uint32_t E; + uint8_t Ilbrm = 0; + uint32_t Tbslbrm = 950984; //max tbs + uint8_t nb_re_dmrs = rel15.nb_re_dmrs; + uint16_t length_dmrs = 1; + + /* + uint8_t *channel_input[MAX_NUM_DLSCH_SEGMENTS]; //unsigned char + for(j=0;jsymbols_per_slot,0,0, rel15.nb_layers); + + // if (dlsch->harq_processes[harq_pid]->Ndi == 1) { // this is a new packet + if (dlsch->harq_processes[harq_pid]->round == 0) { // this is a new packet +#ifdef DEBUG_DLSCH_CODING + printf("encoding thinks this is a new packet \n"); +#endif + /* + int i; + printf("dlsch (tx): \n"); + for (i=0;i<(A>>3);i++) + printf("%02x.",a[i]); + printf("\n"); + */ + // Add 24-bit crc (polynomial A) to payload + crc = crc24a(a,A)>>8; + a[A>>3] = ((uint8_t*)&crc)[2]; + a[1+(A>>3)] = ((uint8_t*)&crc)[1]; + a[2+(A>>3)] = ((uint8_t*)&crc)[0]; + //printf("CRC %x (A %d)\n",crc,A); + //printf("a0 %d a1 %d a2 %d\n", a[A>>3], a[1+(A>>3)], a[2+(A>>3)]); + + dlsch->harq_processes[harq_pid]->B = A+24; + // dlsch->harq_processes[harq_pid]->b = a; + memcpy(dlsch->harq_processes[harq_pid]->b,a,(A/8)+4); + + nr_segmentation(dlsch->harq_processes[harq_pid]->b, + dlsch->harq_processes[harq_pid]->c, + dlsch->harq_processes[harq_pid]->B, + &dlsch->harq_processes[harq_pid]->C, + &dlsch->harq_processes[harq_pid]->K, + pz, + &dlsch->harq_processes[harq_pid]->F); + + kb = dlsch->harq_processes[harq_pid]->K/(*pz); + if ( kb==22){ + BG = 1; + } + else{ + BG = 2; + } + + Kr = dlsch->harq_processes[harq_pid]->K; + Kr_bytes = Kr>>3; + + //printf("segment Z %d kb %d k %d Kr %d BG %d\n", *pz,kb,dlsch->harq_processes[harq_pid]->K,Kr,BG); + + //start_meas(te_stats); + for (r=0; rharq_processes[harq_pid]->C; r++) { + d_tmp[r] = &dlsch->harq_processes[harq_pid]->d[r][0]; + //channel_input[r] = &dlsch->harq_processes[harq_pid]->d[r][0]; +#ifdef DEBUG_DLSCH_CODING + printf("Encoder: B %d F %d \n",dlsch->harq_processes[harq_pid]->B, dlsch->harq_processes[harq_pid]->F); + printf("start ldpc encoder segment %d/%d\n",r,dlsch->harq_processes[harq_pid]->C); + printf("input %d %d %d %d %d \n", dlsch->harq_processes[harq_pid]->c[r][0], dlsch->harq_processes[harq_pid]->c[r][1], dlsch->harq_processes[harq_pid]->c[r][2],dlsch->harq_processes[harq_pid]->c[r][3], dlsch->harq_processes[harq_pid]->c[r][4]); + for (int cnt =0 ; cnt < 22*(*pz)/8; cnt ++){ + printf("%d ", dlsch->harq_processes[harq_pid]->c[r][cnt]); + } + printf("\n"); + +#endif + //ldpc_encoder_orig((unsigned char*)dlsch->harq_processes[harq_pid]->c[r],dlsch->harq_processes[harq_pid]->d[r],Kr,BG,0); + //ldpc_encoder_optim((unsigned char*)dlsch->harq_processes[harq_pid]->c[r],(unsigned char*)&dlsch->harq_processes[harq_pid]->d[r][0],Kr,BG,NULL,NULL,NULL,NULL); + } + + //for (int i=0;i<68*384;i++) + // printf("channel_input[%d]=%d\n",i,channel_input[i]); + + + + /*printf("output %d %d %d %d %d \n", dlsch->harq_processes[harq_pid]->d[0][0], dlsch->harq_processes[harq_pid]->d[0][1], dlsch->harq_processes[harq_pid]->d[r][2],dlsch->harq_processes[harq_pid]->d[0][3], dlsch->harq_processes[harq_pid]->d[0][4]); + for (int cnt =0 ; cnt < 66*(*pz); cnt ++){ + printf("%d \n", dlsch->harq_processes[harq_pid]->d[0][cnt]); + } + printf("\n");*/ + + //ldpc_encoder_optim_8seg(dlsch->harq_processes[harq_pid]->c,d_tmp,Kr,BG,dlsch->harq_processes[harq_pid]->C,NULL,NULL,NULL,NULL); + ldpc_encoder_optim_8seg(dlsch->harq_processes[harq_pid]->c,dlsch->harq_processes[harq_pid]->d,Kr,BG,dlsch->harq_processes[harq_pid]->C,NULL,NULL,NULL,NULL); + + //stop_meas(te_stats); + //printf("end ldpc encoder -- output\n"); + +#ifdef DEBUG_DLSCH_CODING + write_output("enc_input0.m","enc_in0",&dlsch->harq_processes[harq_pid]->c[0][0],Kr_bytes,1,4); + write_output("enc_output0.m","enc0",&dlsch->harq_processes[harq_pid]->d[0][0],(3*8*Kr_bytes)+12,1,4); +#endif + + } + + for (r=0; rharq_processes[harq_pid]->C; r++) { +#ifdef DEBUG_DLSCH_CODING + printf("Rate Matching, Code segment %d (coded bits (G) %d,unpunctured/repeated bits per code segment %d,mod_order %d, nb_rb %d)...\n", + r, + G, + Kr*3, + mod_order,nb_rb); +#endif + + //start_meas(rm_stats); +#ifdef DEBUG_DLSCH_CODING + printf("rvidx in encoding = %d\n", rel15.redundancy_version); +#endif + + E = nr_get_E(G, dlsch->harq_processes[harq_pid]->C, mod_order, rel15.nb_layers, r); + + nr_rate_matching_ldpc(Ilbrm, + Tbslbrm, + BG, + *pz, + dlsch->harq_processes[harq_pid]->d[r], + dlsch->harq_processes[harq_pid]->e+r_offset, + dlsch->harq_processes[harq_pid]->C, + rel15.redundancy_version, + E); + +#ifdef DEBUG_DLSCH_CODING + for (int i =0; i<16; i++) + printf("output ratematching e[%d]= %d r_offset %d\n", i,dlsch->harq_processes[harq_pid]->e[i+r_offset], r_offset); +#endif + //stop_meas(rm_stats); + + //start_meas(i_stats); + nr_interleaving_ldpc(E, + mod_order, + dlsch->harq_processes[harq_pid]->e+r_offset, + dlsch->harq_processes[harq_pid]->f+r_offset); + //stop_meas(i_stats); + +#ifdef DEBUG_DLSCH_CODING + for (int i =0; i<16; i++) + printf("output interleaving f[%d]= %d r_offset %d\n", i,dlsch->harq_processes[harq_pid]->f[i+r_offset], r_offset); + + if (r==dlsch->harq_processes[harq_pid]->C-1) + write_output("enc_output.m","enc",dlsch->harq_processes[harq_pid]->f,G,1,4); +#endif + + r_offset += E; + } + + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_ENB_DLSCH_ENCODING, VCD_FUNCTION_OUT); + + return(0); +} diff --git a/openair1/PHY/NR_TRANSPORT/nr_dlsch_tools.c b/openair1/PHY/NR_TRANSPORT/nr_dlsch_tools.c new file mode 100644 index 0000000000..c1ac34075a --- /dev/null +++ b/openair1/PHY/NR_TRANSPORT/nr_dlsch_tools.c @@ -0,0 +1,250 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/NR_TRANSPORT/nr_dlsch_tools.c +* \brief Top-level routines for decoding DLSCH transport channel from 38-214, V15.2.0 2018-06 +* \author Guy De Souza +* \date 2018 +* \version 0.1 +* \company Eurecom +* \email: desouza@eurecom.fr +* \note +* \warning +*/ + +#include "nr_dlsch.h" + +extern void set_taus_seed(unsigned int seed_type); + +uint8_t nr_pdsch_default_time_alloc_A_S_nCP[23] = {2,3,2,3,2,3,2,3,2,3,9,10,4,6,5,5,9,12,1,1,2,4,8}; +uint8_t nr_pdsch_default_time_alloc_A_L_nCP[23] = {12,11,10,9,9,8,7,6,5,4,4,4,4,4,7,2,2,2,13,6,4,7,4}; +uint8_t nr_pdsch_default_time_alloc_A_S_eCP[23] = {2,3,2,3,2,3,2,3,2,3,6,8,4,6,5,5,9,10,1,1,2,4,8}; +uint8_t nr_pdsch_default_time_alloc_A_L_eCP[23] = {6,5,10,9,9,8,7,6,5,4,4,2,4,4,6,2,2,2,11,6,4,6,4}; +uint8_t nr_pdsch_default_time_alloc_B_S[16] = {2,4,6,8,10,2,4,2,4,6,8,10,2,2,3,2}; +uint8_t nr_pdsch_default_time_alloc_B_L[16] = {2,2,2,2,2,2,2,4,4,4,4,4,7,12,11,4}; +uint8_t nr_pdsch_default_time_alloc_C_S[15] = {2,4,6,8,10,2,4,6,8,10,2,2,3,0,2}; +uint8_t nr_pdsch_default_time_alloc_C_L[15] = {2,2,2,2,2,4,4,4,4,4,7,12,11,6,6}; + + + /// Time domain allocation routines + +void nr_get_time_domain_allocation_type(nfapi_nr_config_request_t config, + nfapi_nr_dl_config_dci_dl_pdu dci_pdu, + nfapi_nr_dl_config_dlsch_pdu *dlsch_pdu) { + + nfapi_nr_dl_config_pdcch_parameters_rel15_t params_rel15 = dci_pdu.pdcch_params_rel15; + uint8_t *alloc_type = &dlsch_pdu->dlsch_pdu_rel15.time_allocation_type; + uint8_t mux_pattern = params_rel15.mux_pattern; + uint8_t alloc_list_flag = dlsch_pdu->dlsch_pdu_rel15.time_alloc_list_flag; + + switch(params_rel15.rnti_type) { + + case NFAPI_NR_RNTI_SI: + AssertFatal(params_rel15.common_search_space_type == NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_0, + "Invalid common search space type %d for SI RNTI, expected %d\n", + params_rel15.common_search_space_type, NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_0); + + if (mux_pattern == NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1) + AssertFatal(config.subframe_config.dl_cyclic_prefix_type.value == NFAPI_CP_NORMAL, + "Invalid configuration CP extended for SI RNTI type 0 search space\n"); + + *alloc_type = (mux_pattern == NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1)?NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_A : + (mux_pattern == NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE2)?NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_B : + NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_C; + break; + + case NFAPI_NR_RNTI_RA: + case NFAPI_NR_RNTI_TC: + /*AssertFatal(dci_alloc.pdcch_params.common_search_space_type == NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_1, + "Invalid common search space type %d for RNTI %d, expected %d\n",dci_alloc.pdcch_params.common_search_space_type, + NFAPI_NR_COMMON_SEARCH_SPACE_TYPE_1, dci_alloc.rnti_type);*/ + *alloc_type = (alloc_list_flag) ? NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_ALLOC_LIST : NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_A; + break; + + case NFAPI_NR_RNTI_P: + break; + + case NFAPI_NR_RNTI_C: + case NFAPI_NR_RNTI_CS: + if (params_rel15.search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_COMMON) + *alloc_type = (alloc_list_flag)? NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_ALLOC_LIST : NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_A; + else + *alloc_type = NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_ALLOC_LIST; + + break; + + } +} + +uint16_t get_SLIV(uint8_t S, uint8_t L) { + return ( (uint16_t)(((L-1)<=7)? (14*(L-1)+S) : (14*(15-L)+(13-S))) ); +} + +static inline uint8_t get_K0(uint8_t row_idx, uint8_t time_alloc_type) { + return ( (time_alloc_type == NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_A)|| + (time_alloc_type == NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_C)? 0 : + ((row_idx==6)||(row_idx==7)||(row_idx==15))? 1 : 0); +} + +/*ideally combine the calculation of L in the same function once the right struct is defined*/ +uint8_t nr_get_S(uint8_t row_idx, uint8_t CP, uint8_t time_alloc_type, uint8_t dmrs_typeA_position) { + + uint8_t idx; + //uint8_t S; + + switch(time_alloc_type) { + case NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_A: + idx = (row_idx>7)? (row_idx+6) : (((row_idx-1)<<1)-1+((dmrs_typeA_position==2)?0:1)); + return ((CP==NFAPI_CP_NORMAL)?nr_pdsch_default_time_alloc_A_S_nCP[idx] : nr_pdsch_default_time_alloc_A_S_eCP[idx]); + break; + + case NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_B: + idx = (row_idx<14)? (row_idx-1) : (row_idx == 14)? row_idx-1+((dmrs_typeA_position==2)?0:1) : 15; + return (nr_pdsch_default_time_alloc_B_S[idx]); + break; + + case NFAPI_NR_PDSCH_TIME_DOMAIN_ALLOC_TYPE_DEFAULT_C: + AssertFatal((row_idx!=6)&&(row_idx!=7)&&(row_idx<17), "Invalid row index %d in %s %s\n", row_idx, __FUNCTION__, __FILE__); + idx = (row_idx<6)? (row_idx-1) : (row_idx<14)? (row_idx-3) : (row_idx == 14)? row_idx-3+((dmrs_typeA_position==2)?0:1) : (row_idx-2); + break; + + default: + AssertFatal(0, "Invalid Time domain allocation type %d in %s %s\n", time_alloc_type, __FUNCTION__, __FILE__); + } + return 0; // temp warning fix +} + +void nr_check_time_alloc(uint8_t S, uint8_t L, nfapi_nr_config_request_t config) { + + switch (config.subframe_config.dl_cyclic_prefix_type.value) { + case NFAPI_CP_NORMAL: + if (config.pdsch_config.mapping_type.value == NFAPI_NR_PDSCH_MAPPING_TYPE_A) { + AssertFatal(S<4, "Invalid value of S(%d) for mapping type A and normal CP\n", S); + + if (S==3) + AssertFatal(config.pdsch_config.mapping_type.value == 3, "Invalid S %d for dmrs_typeA_position %d\n", + S, config.pdsch_config.dmrs_typeA_position.value); + + AssertFatal((L>2)&&(L<15), "Invalid L %d for mapping type A and normal CP\n", L); + + AssertFatal(((S+L)>2)&&((S+L)<15), "Invalid S+L %d for mapping type A and normal CP\n", S+L); + } + else { + AssertFatal(S<13, "Invalid value of S(%d) for mapping type B and normal CP\n", S); + + AssertFatal((L>1)&&(L<8), "Invalid L %d for mapping type B and normal CP\n", L); + + AssertFatal(((S+L)>1)&&((S+L)<15), "Invalid S+L %d for mapping type B and normal CP\n", S+L); + } + break; + + case NFAPI_CP_EXTENDED: + if (config.pdsch_config.mapping_type.value == NFAPI_NR_PDSCH_MAPPING_TYPE_A) { + AssertFatal(S<4, "Invalid value of S(%d) for mapping type A and extended CP\n", S); + + if (S==3) + AssertFatal(config.pdsch_config.dmrs_typeA_position.value == 3, "Invalid S %d for dmrs_typeA_position %d\n", + S, config.pdsch_config.dmrs_typeA_position.value); + + AssertFatal((L>2)&&(L<13), "Invalid L %d for mapping type A and extended CP\n", L); + + AssertFatal(((S+L)>2)&&((S+L)<13), "Invalid S+L %d for mapping type A and extended CP\n", S+L); + } + else { + AssertFatal(S<11, "Invalid value of S(%d) for mapping type B and extended CP\n", S); + + AssertFatal((L>1)&&(L<7), "Invalid L %d for mapping type B and extended CP\n", L); + + AssertFatal(((S+L)>1)&&((S+L)<13), "Invalid S+L %d for mapping type B and extended CP\n", S+L); + } + break; + } +} + + + /// Frequency domain allocation routines + + // DL alloc type 0 +static inline uint8_t get_RBG_size_P(uint16_t n_RB, uint8_t RBG_config) { + if (RBG_config == NFAPI_NR_PDSCH_RBG_CONFIG_TYPE1) + return ((n_RB<37)?2:(n_RB<73)?4:(n_RB<145)?8:16); + else if (RBG_config == NFAPI_NR_PDSCH_RBG_CONFIG_TYPE2) + return ((n_RB<37)?4:(n_RB<73)?8:(n_RB<145)?16:16); + else + AssertFatal(0, "Invalid RBG config type (%d)\n", RBG_config); +} + +void nr_get_rbg_parms(NR_BWP_PARMS* bwp, uint8_t config_type) { + nr_rbg_parms_t* rbg_parms = &bwp->rbg_parms; + + rbg_parms->P = get_RBG_size_P(bwp->N_RB, config_type); + rbg_parms->start_size = rbg_parms->P - bwp->location%rbg_parms->P; + rbg_parms->end_size = ((bwp->location + bwp->N_RB)%rbg_parms->P)? ((bwp->location + bwp->N_RB)%rbg_parms->P) : rbg_parms->P; + rbg_parms->N_RBG = (uint8_t)ceil( (bwp->N_RB + (bwp->location%rbg_parms->P))/(float)rbg_parms->P); + LOG_I(PHY, "RBG parameters for BWP %d location %d N_RB %d:\n", bwp->bwp_id, bwp->location, bwp->N_RB); + LOG_I(PHY, "P %d\t start size %d\t endsize %d\t N_RBG %d\n", rbg_parms->P, rbg_parms->start_size, rbg_parms->end_size, rbg_parms->N_RBG); +} + +void nr_get_rbg_list(uint32_t bitmap, uint8_t n_rbg, uint8_t* rbg_list) { + uint8_t idx=0; + for (int i=0; i>(n_rbg-i-1))&1) + rbg_list[idx++]=i; +} + + // DL alloc type 1 + +uint16_t get_RIV(uint16_t rb_start, uint16_t L, uint16_t N_RB) { + if ((L-1)<=(N_RB>>1)) + return (N_RB*(L-1)+rb_start); + else + return (N_RB*(N_RB-L+1) + (N_RB-1-rb_start)); +} + + /// PRB bundling routines + +// Precoding granularity +static inline uint8_t nr_get_P_prime(uint8_t rnti_type, uint8_t dci_format, uint8_t prb_bundling_type) { + if (dci_format == NFAPI_NR_DL_DCI_FORMAT_1_0) + return (NFAPI_NR_PRG_GRANULARITY_2); + else // NFAPI_NR_DL_DCI_FORMAT_1_1 + return ((prb_bundling_type)?0:2);// incomplete for 1_1 +} + +void nr_get_PRG_parms(NR_BWP_PARMS* bwp, NR_gNB_DCI_ALLOC_t dci_alloc, uint8_t prb_bundling_type) { + nr_prg_parms_t* prg_parms = &bwp->prg_parms; + + prg_parms->P_prime = nr_get_P_prime(dci_alloc.pdcch_params.rnti_type, dci_alloc.pdcch_params.dci_format, prb_bundling_type); + prg_parms->start_size = prg_parms->P_prime - bwp->location%prg_parms->P_prime; + prg_parms->end_size = (bwp->location + bwp->N_RB)%prg_parms->P_prime; + if (!prg_parms->end_size) + prg_parms->end_size = prg_parms->P_prime; + prg_parms->N_PRG = ceil((float)bwp->N_RB/prg_parms->P_prime); + LOG_I(PHY, "PRG parameters for BWP %d location %d N_RB %d:\n", bwp->bwp_id, bwp->location, bwp->N_RB); + LOG_I(PHY, "P_prime %d\t start size %d\t endsize %d\t N_PRG %d\n", prg_parms->P_prime, prg_parms->start_size, prg_parms->end_size, prg_parms->N_PRG); +} + + /// Payload emulation +void nr_emulate_dlsch_payload(uint8_t* pdu, uint16_t size) { + set_taus_seed(0); + for (int i=0; isch_config.physical_cell_id.value &3; @@ -72,10 +72,7 @@ int nr_generate_pbch_dmrs(uint32_t *gold_pbch_dmrs, } /// Resource mapping - a = (config->rf_config.tx_antenna_ports.value == 1) ? amp : (amp*ONE_OVER_SQRT2_Q15)>>15; - for (int aa = 0; aa < config->rf_config.tx_antenna_ports.value; aa++) - { // PBCH DMRS are mapped within the SSB block on every fourth subcarrier starting from nushift of symbols 1, 2, 3 ///symbol 1 [0+nushift:4:236+nushift] -- 60 mod symbols @@ -86,12 +83,12 @@ int nr_generate_pbch_dmrs(uint32_t *gold_pbch_dmrs, #ifdef DEBUG_PBCH_DMRS printf("m %d at k %d of l %d\n", m, k, l); #endif - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1] = (a * mod_dmrs[m<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (a * mod_dmrs[(m<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1] = (amp * mod_dmrs[m<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (amp * mod_dmrs[(m<<1) + 1]) >> 15; #ifdef DEBUG_PBCH_DMRS printf("(%d,%d)\n", - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1], - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1)+1]); + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1], + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1)+1]); #endif k+=4; @@ -107,12 +104,12 @@ int nr_generate_pbch_dmrs(uint32_t *gold_pbch_dmrs, #ifdef DEBUG_PBCH_DMRS printf("m %d at k %d of l %d\n", m, k, l); #endif - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1] = (a * mod_dmrs[m<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (a * mod_dmrs[(m<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1] = (amp * mod_dmrs[m<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (amp * mod_dmrs[(m<<1) + 1]) >> 15; #ifdef DEBUG_PBCH_DMRS printf("(%d,%d)\n", - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1], - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1)+1]); + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1], + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1)+1]); #endif k+=(m==71)?148:4; // Jump from 44+nu to 192+nu @@ -128,12 +125,12 @@ int nr_generate_pbch_dmrs(uint32_t *gold_pbch_dmrs, #ifdef DEBUG_PBCH_DMRS printf("m %d at k %d of l %d\n", m, k, l); #endif - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1] = (a * mod_dmrs[m<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (a * mod_dmrs[(m<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1] = (amp * mod_dmrs[m<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (amp * mod_dmrs[(m<<1) + 1]) >> 15; #ifdef DEBUG_PBCH_DMRS printf("(%d,%d)\n", - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1], - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1)+1]); + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1], + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1)+1]); #endif k+=4; @@ -141,7 +138,6 @@ int nr_generate_pbch_dmrs(uint32_t *gold_pbch_dmrs, k-=frame_parms->ofdm_symbol_size; } - } #ifdef DEBUG_PBCH_DMRS @@ -225,13 +221,14 @@ void nr_init_pbch_interleaver(uint8_t *interleaver) { *(interleaver+i) = *(nr_pbch_payload_interleaving_pattern+j_hrf); else // Ssb bits:3 *(interleaver+i) = *(nr_pbch_payload_interleaving_pattern+j_ssb++); + } int nr_generate_pbch(NR_gNB_PBCH *pbch, t_nrPolar_paramsPtr polar_params, uint8_t *pbch_pdu, uint8_t *interleaver, - int32_t **txdataF, + int32_t *txdataF, int16_t amp, uint8_t ssb_start_symbol, uint8_t n_hf, @@ -243,12 +240,13 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, { int k,l,m; - int16_t a; + //int16_t a; int16_t mod_pbch_e[NR_POLAR_PBCH_E]; uint8_t idx=0; uint16_t M; uint8_t nushift; uint32_t unscrambling_mask; + uint64_t a_reversed=0; LOG_I(PHY, "PBCH generation started\n"); @@ -303,8 +301,13 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, printf("pbch_a_prime: 0x%08x\n", pbch->pbch_a_prime); #endif +// Encoder reversal + for (int i=0; ipbch_a_prime>>i)&1)<<(31-i); + + /// CRC, coding and rate matching - polar_encoder (&pbch->pbch_a_prime, pbch->pbch_e, polar_params); + polar_encoder_fast (&a_reversed, (uint32_t*)pbch->pbch_e, 0, polar_params); #ifdef DEBUG_PBCH_ENCODING printf("Channel coding:\n"); for (int i=0; isch_config.physical_cell_id.value &3; - a = (config->rf_config.tx_antenna_ports.value == 1) ? amp : (amp*ONE_OVER_SQRT2_Q15)>>15; - for (int aa = 0; aa < config->rf_config.tx_antenna_ports.value; aa++) - { // PBCH modulated symbols are mapped within the SSB block on symbols 1, 2, 3 excluding the subcarriers used for the PBCH DMRS ///symbol 1 [0:239] -- 180 mod symbols @@ -357,8 +357,8 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, #ifdef DEBUG_PBCH printf("m %d ssb_sc_idx %d at k %d of l %d\n", m, ssb_sc_idx, k, l); #endif - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1] = (a * mod_pbch_e[m<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (a * mod_pbch_e[(m<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1] = (amp * mod_pbch_e[m<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (amp * mod_pbch_e[(m<<1) + 1]) >> 15; k++; m++; } @@ -382,8 +382,8 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, #ifdef DEBUG_PBCH printf("m %d ssb_sc_idx %d at k %d of l %d\n", m, ssb_sc_idx, k, l); #endif - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1] = (a * mod_pbch_e[m<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (a * mod_pbch_e[(m<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1] = (amp * mod_pbch_e[m<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (amp * mod_pbch_e[(m<<1) + 1]) >> 15; k++; m++; } @@ -408,8 +408,8 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, #ifdef DEBUG_PBCH printf("m %d ssb_sc_idx %d at k %d of l %d\n", m, ssb_sc_idx, k, l); #endif - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1] = (a * mod_pbch_e[m<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (a * mod_pbch_e[(m<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1] = (amp * mod_pbch_e[m<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (amp * mod_pbch_e[(m<<1) + 1]) >> 15; k++; m++; } @@ -433,8 +433,8 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, #ifdef DEBUG_PBCH printf("m %d ssb_sc_idx %d at k %d of l %d\n", m, ssb_sc_idx, k, l); #endif - ((int16_t*)txdataF[aa])[(l*frame_parms->ofdm_symbol_size + k)<<1] = (a * mod_pbch_e[m<<1]) >> 15; - ((int16_t*)txdataF[aa])[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (a * mod_pbch_e[(m<<1) + 1]) >> 15; + ((int16_t*)txdataF)[(l*frame_parms->ofdm_symbol_size + k)<<1] = (amp * mod_pbch_e[m<<1]) >> 15; + ((int16_t*)txdataF)[((l*frame_parms->ofdm_symbol_size + k)<<1) + 1] = (amp * mod_pbch_e[(m<<1) + 1]) >> 15; k++; m++; } @@ -443,13 +443,6 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, k-=frame_parms->ofdm_symbol_size; } - } - - -#ifdef DEBUG_PBCH - write_output("txdataF_pbch.m", "txdataF_pbch", txdataF[0], frame_parms->samples_per_frame_wCP>>1, 1, 1); -#endif - return 0; } diff --git a/openair1/PHY/NR_TRANSPORT/nr_pss.c b/openair1/PHY/NR_TRANSPORT/nr_pss.c index f91f14f8a2..29dbe4ec99 100644 --- a/openair1/PHY/NR_TRANSPORT/nr_pss.c +++ b/openair1/PHY/NR_TRANSPORT/nr_pss.c @@ -25,7 +25,7 @@ //#define NR_PSS_DEBUG int nr_generate_pss( int16_t *d_pss, - int32_t **txdataF, + int32_t *txdataF, int16_t amp, uint8_t ssb_start_symbol, nfapi_nr_config_request_t* config, @@ -48,7 +48,7 @@ int nr_generate_pss( int16_t *d_pss, for (i=0; i < NR_PSS_LENGTH; i++) { m = (i + 43*Nid2)%(NR_PSS_LENGTH); - d_pss[i] = (1 - 2*x[m]) * 32767; + d_pss[i] = (1 - 2*x[m]) * 23170; } #ifdef NR_PSS_DEBUG @@ -57,10 +57,8 @@ int nr_generate_pss( int16_t *d_pss, #endif /// Resource mapping - a = (config->rf_config.tx_antenna_ports.value == 1) ? amp : (amp*ONE_OVER_SQRT2_Q15)>>15; + a = amp; - for (int aa = 0; aa < config->rf_config.tx_antenna_ports.value; aa++) - { // PSS occupies a predefined position (subcarriers 56-182, symbol 0) within the SSB block starting from k = frame_parms->first_carrier_offset + frame_parms->ssb_start_subcarrier + 56; //and @@ -70,13 +68,12 @@ int nr_generate_pss( int16_t *d_pss, for (m = 0; m < NR_PSS_LENGTH; m++) { // printf("pss: writing position k %d / %d\n",k,frame_parms->ofdm_symbol_size); - ((int16_t*)txdataF[aa])[2*(l*frame_parms->ofdm_symbol_size + k)] = (a * d_pss[m]) >> 15; + ((int16_t*)txdataF)[2*(l*frame_parms->ofdm_symbol_size + k)] = (a * d_pss[m]) >> 15; k++; if (k >= frame_parms->ofdm_symbol_size) k-=frame_parms->ofdm_symbol_size; } - } #ifdef NR_PSS_DEBUG LOG_M("pss_0.m", "pss_0", diff --git a/openair1/PHY/NR_TRANSPORT/nr_sch_dmrs.c b/openair1/PHY/NR_TRANSPORT/nr_sch_dmrs.c new file mode 100644 index 0000000000..4ec1083cfc --- /dev/null +++ b/openair1/PHY/NR_TRANSPORT/nr_sch_dmrs.c @@ -0,0 +1,89 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/NR_TRANSPORT/nr_sch_dmrs.c +* \brief +* \author +* \date +* \version +* \company Eurecom +* \email: +* \note +* \warning +*/ + +#include "nr_sch_dmrs.h" + + +/*Table 7.4.1.1.2-1 and 7.4.1.1.2-2 38211 Columns: ap - CDM group - Delta - Wf(0) - Wf(1) - Wt(0) - Wt(1)*/ +int8_t pdsch_dmrs_1[8][7] = {{0,0,0,1,1,1,1}, + {1,0,0,1,-1,1,1}, + {2,1,1,1,1,1,1}, + {3,1,1,1,-1,1,1}, + {4,0,0,1,1,1,-1}, + {5,0,0,1,-1,1,-1}, + {6,1,1,1,1,1,-1}, + {7,1,1,1,-1,1,-1}}; + +int8_t pdsch_dmrs_2[12][7] = {{0,0,0,1,1,1,1}, + {1,0,0,1,-1,1,1}, + {2,1,2,1,1,1,1}, + {3,1,2,1,-1,1,1}, + {4,2,4,1,1,1,1}, + {5,2,4,1,-1,1,1}, + {6,0,0,1,1,1,-1}, + {7,0,0,1,-1,1,-1}, + {8,1,2,1,1,1,-1}, + {9,1,2,1,-1,1,-1}, + {10,2,4,1,1,1,-1}, + {11,2,4,1,-1,1,-1}}; + +void get_l_prime(uint8_t *l_prime, uint8_t n_symbs) { + for (int i=0; irf_config.tx_antenna_ports.value == 1) ? amp : (amp*ONE_OVER_SQRT2_Q15)>>15; - - for (int aa = 0; aa < config->rf_config.tx_antenna_ports.value; aa++) - { // SSS occupies a predefined position (subcarriers 56-182, symbol 2) within the SSB block starting from k = frame_parms->first_carrier_offset + frame_parms->ssb_start_subcarrier + 56; //and l = ssb_start_symbol + 2; for (int m = 0; m < NR_SSS_LENGTH; m++) { - ((int16_t*)txdataF[aa])[2*(l*frame_parms->ofdm_symbol_size + k)] = (a * d_sss[m]) >> 15; + ((int16_t*)txdataF)[2*(l*frame_parms->ofdm_symbol_size + k)] = (amp * d_sss[m]) >> 15; k++; if (k >= frame_parms->ofdm_symbol_size) k-=frame_parms->ofdm_symbol_size; } - } #ifdef NR_SSS_DEBUG // write_output("sss_0.m", "sss_0", (void*)txdataF[0][l*frame_parms->ofdm_symbol_size], frame_parms->ofdm_symbol_size, 1, 1); #endif diff --git a/openair1/PHY/NR_TRANSPORT/nr_tbs_tools.c b/openair1/PHY/NR_TRANSPORT/nr_tbs_tools.c new file mode 100644 index 0000000000..f8b7454067 --- /dev/null +++ b/openair1/PHY/NR_TRANSPORT/nr_tbs_tools.c @@ -0,0 +1,193 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.0 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/LTE_TRANSPORT/dlsch_coding.c +* \brief Top-level routines for implementing LDPC-coded (DLSCH) transport channels from 38-212, 15.2 +* \author H.Wang +* \date 2018 +* \version 0.1 +* \company Eurecom +* \email: +* \note +* \warning +*/ + +#include "nr_dlsch.h" + +/// Target code rate tables indexed by Imcs +uint16_t nr_target_code_rate_table1[29] = {120, 157, 193, 251, 308, 379, 449, 526, 602, 679, 340, 378, 434, 490, 553, \ + 616, 658, 438, 466, 517, 567, 616, 666, 719, 772, 822, 873, 910, 948}; + // Imcs values 20 and 26 have been multiplied by 2 to avoid the floating point +uint16_t nr_target_code_rate_table2[28] = {120, 193, 308, 449, 602, 378, 434, 490, 553, 616, 658, 466, 517, 567, \ + 616, 666, 719, 772, 822, 873, 1365, 711, 754, 797, 841, 885, 1833, 948}; +uint16_t nr_target_code_rate_table3[29] = {30, 40, 50, 64, 78, 99, 120, 157, 193, 251, 308, 379, 449, 526, 602, 340, \ + 378, 434, 490, 553, 616, 438, 466, 517, 567, 616, 666, 719, 772}; +uint16_t nr_tbs_table[93] = {24, 32, 40, 48, 56, 64, 72, 80, 88, 96, 104, 112, 120, 128, 136, 144, 152, 160, 168, 176, 184, 192, 208, 224, 240, 256, 272, 288, 304, 320, \ + 336, 352, 368, 384, 408, 432, 456, 480, 504, 528, 552, 576, 608, 640, 672, 704, 736, 768, 808, 848, 888, 928, 984, 1032, 1064, 1128, 1160, 1192, 1224, 1256, \ + 1288, 1320, 1352, 1416, 1480, 1544, 1608, 1672, 1736, 1800, 1864, 1928, 2024, 2088, 2152, 2216, 2280, 2408, 2472, 2536, 2600, 2664, 2728, 2792, 2856, 2976, \ + 3104, 3240, 3368, 3496, 3624, 3752, 3824}; + +uint8_t nr_get_Qm(uint8_t Imcs, uint8_t table_idx) { + switch(table_idx) { + case 1: + return (((Imcs<10)||(Imcs==29))?2:((Imcs<17)||(Imcs==30))?4:((Imcs<29)||(Imcs==31))?6:-1); + break; + + case 2: + return (((Imcs<5)||(Imcs==28))?2:((Imcs<11)||(Imcs==29))?4:((Imcs<20)||(Imcs==30))?6:((Imcs<28)||(Imcs==31))?8:-1); + break; + + case 3: + return (((Imcs<15)||(Imcs==29))?2:((Imcs<21)||(Imcs==30))?4:((Imcs<29)||(Imcs==31))?6:-1); + break; + + default: + AssertFatal(0, "Invalid MCS table index %d (expected in range [1,3])\n", table_idx); + } +} + +uint32_t nr_get_code_rate(uint8_t Imcs, uint8_t table_idx) { + switch(table_idx) { + case 1: + return (nr_target_code_rate_table1[Imcs]); + break; + + case 2: + return (nr_target_code_rate_table2[Imcs]); + break; + + case 3: + return (nr_target_code_rate_table3[Imcs]); + break; + + default: + AssertFatal(0, "Invalid MCS table index %d (expected in range [1,3])\n", table_idx); + } +} + +static inline uint8_t is_codeword_disabled(uint8_t format, uint8_t Imcs, uint8_t rv) { + return ((format==NFAPI_NR_DL_DCI_FORMAT_1_1)&&(Imcs==26)&&(rv==1)); +} + +static inline uint8_t get_table_idx(uint8_t mcs_table, uint8_t dci_format, uint8_t rnti_type, uint8_t ss_type) { + if ((mcs_table == NFAPI_NR_MCS_TABLE_QAM256) && (dci_format == NFAPI_NR_DL_DCI_FORMAT_1_1) && ((rnti_type==NFAPI_NR_RNTI_C)||(rnti_type==NFAPI_NR_RNTI_CS))) + return 2; + else if ((mcs_table == NFAPI_NR_MCS_TABLE_QAM64_LOW_SE) && (rnti_type!=NFAPI_NR_RNTI_new) && (rnti_type==NFAPI_NR_RNTI_C) && (ss_type==NFAPI_NR_SEARCH_SPACE_TYPE_UE_SPECIFIC)) + return 3; + else if (rnti_type==NFAPI_NR_RNTI_new) + return 3; + else if ((mcs_table == NFAPI_NR_MCS_TABLE_QAM256) && (rnti_type==NFAPI_NR_RNTI_CS) && (dci_format == NFAPI_NR_DL_DCI_FORMAT_1_1)) + return 2; // Condition mcs_table not configured in sps_config necessary here but not yet implemented + /*else if((mcs_table == NFAPI_NR_MCS_TABLE_QAM64_LOW_SE) && (rnti_type==NFAPI_NR_RNTI_CS)) + * table_idx = 3; + * Note: the commented block refers to the case where the mcs_table is from sps_config*/ + else + return 1; +} + +void nr_get_tbs(nfapi_nr_dl_config_dlsch_pdu *dlsch_pdu, + nfapi_nr_dl_config_dci_dl_pdu dci_pdu, + nfapi_nr_config_request_t config) { + + LOG_I(MAC, "TBS calculation\n"); + + nfapi_nr_dl_config_pdcch_parameters_rel15_t params_rel15 = dci_pdu.pdcch_params_rel15; + nfapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_rel15 = &dlsch_pdu->dlsch_pdu_rel15; + uint8_t rnti_type = params_rel15.rnti_type; + uint8_t dci_format = params_rel15.dci_format; + uint8_t ss_type = params_rel15.search_space_type; + uint8_t N_PRB_oh = ((rnti_type==NFAPI_NR_RNTI_SI)||(rnti_type==NFAPI_NR_RNTI_RA)||(rnti_type==NFAPI_NR_RNTI_P))? 0 : \ + (config.pdsch_config.x_overhead.value); + uint8_t N_PRB_DMRS = (config.pdsch_config.dmrs_type.value == NFAPI_NR_DMRS_TYPE1)?6:4; //This only works for antenna port 1000 + uint8_t mcs_table = config.pdsch_config.mcs_table.value; + uint8_t N_sh_symb = dlsch_rel15->nb_symbols; + uint8_t Imcs = dlsch_rel15->mcs_idx; + uint16_t N_RE_prime = NR_NB_SC_PER_RB*N_sh_symb - N_PRB_DMRS - N_PRB_oh; + LOG_I(MAC, "N_RE_prime %d for %d symbols %d DMRS per PRB and %d overhead\n", N_RE_prime, N_sh_symb, N_PRB_DMRS, N_PRB_oh); + + uint16_t N_RE, Ninfo, Ninfo_prime, C, TBS=0, R; + uint8_t table_idx, Qm, n, scale; + + table_idx = get_table_idx(mcs_table, dci_format, rnti_type, ss_type); + scale = ((table_idx==2)&&((Imcs==20)||(Imcs==26)))?11:10; + + N_RE = min(156, N_RE_prime)*dlsch_rel15->n_prb; + R = nr_get_code_rate(Imcs, table_idx); + Qm = nr_get_Qm(Imcs, table_idx); + Ninfo = (N_RE*R*Qm*dlsch_rel15->nb_layers)>>scale; + + if (Ninfo <= 3824) { + n = max(3, (log2(Ninfo)-6)); + Ninfo_prime = max(24, (Ninfo>>n)<= Ninfo_prime) { + TBS = nr_tbs_table[i]; + break; + } + } + else { + n = log2(Ninfo-24)-5; + Ninfo_prime = max(3840, (ROUNDIDIV((Ninfo-24),(1<8424) { + C = CEILIDIV((Ninfo_prime+24),8424); + TBS = (C<<3)*CEILIDIV((Ninfo_prime+24),(C<<3)) - 24; + } + else + TBS = ((CEILIDIV((Ninfo_prime+24),8))<<3) - 24; + } + } + + dlsch_rel15->coding_rate = R; + dlsch_rel15->modulation_order = Qm; + dlsch_rel15->transport_block_size = TBS; + dlsch_rel15->nb_re_dmrs = N_PRB_DMRS; + dlsch_rel15->nb_mod_symbols = N_RE_prime*dlsch_rel15->n_prb*dlsch_rel15->nb_codewords; + + LOG_I(MAC, "TBS %d : N_RE %d N_PRB_DMRS %d N_sh_symb %d N_PRB_oh %d Ninfo %d Ninfo_prime %d R %d Qm %d table %d scale %d nb_symbols %d\n", + TBS, N_RE, N_PRB_DMRS, N_sh_symb, N_PRB_oh, Ninfo, Ninfo_prime, R, Qm, table_idx, scale, dlsch_rel15->nb_mod_symbols); +} + +uint32_t nr_get_G(uint16_t nb_rb, uint16_t nb_symb_sch,uint8_t nb_re_dmrs,uint16_t length_dmrs, uint8_t Qm, uint8_t Nl) { + uint32_t G; + G = ((NR_NB_SC_PER_RB*nb_symb_sch)-(nb_re_dmrs*length_dmrs))*nb_rb*Qm*Nl; + return(G); +} + +uint32_t nr_get_E(uint32_t G, uint8_t C, uint8_t Qm, uint8_t Nl, uint8_t r) { + uint32_t E; + uint8_t Cprime = C; //assume CBGTI not present + + AssertFatal(Nl>0,"Nl is 0\n"); + AssertFatal(Qm>0,"Qm is 0\n"); + + if (r <= Cprime - ((G/(Nl*Qm))%Cprime) - 1) + E = Nl*Qm*(G/(Nl*Qm*Cprime)); + else + E = Nl*Qm*((G/(Nl*Qm*Cprime))+1); + + return E; +} diff --git a/openair1/PHY/NR_TRANSPORT/nr_transport.h b/openair1/PHY/NR_TRANSPORT/nr_transport.h index 436ec10f79..c1a4052191 100644 --- a/openair1/PHY/NR_TRANSPORT/nr_transport.h +++ b/openair1/PHY/NR_TRANSPORT/nr_transport.h @@ -33,7 +33,7 @@ @returns 0 on success */ int nr_generate_pss( int16_t *d_pss, - int32_t **txdataF, + int32_t *txdataF, int16_t amp, uint8_t ssb_start_symbol, nfapi_nr_config_request_t* config, @@ -46,7 +46,7 @@ int nr_generate_pss( int16_t *d_pss, @returns 0 on success */ int nr_generate_sss( int16_t *d_sss, - int32_t **txdataF, + int32_t *txdataF, int16_t amp, uint8_t ssb_start_symbol, nfapi_nr_config_request_t* config, @@ -59,7 +59,7 @@ int nr_generate_sss( int16_t *d_sss, @returns 0 on success */ int nr_generate_pbch_dmrs(uint32_t *gold_pbch_dmrs, - int32_t **txdataF, + int32_t *txdataF, int16_t amp, uint8_t ssb_start_symbol, nfapi_nr_config_request_t* config, @@ -88,7 +88,7 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, t_nrPolar_paramsPtr polar_params, uint8_t *pbch_pdu, uint8_t *interleaver, - int32_t **txdataF, + int32_t *txdataF, int16_t amp, uint8_t ssb_start_symbol, uint8_t n_hf, @@ -106,4 +106,11 @@ int nr_generate_pbch(NR_gNB_PBCH *pbch, */ void nr_init_pbch_interleaver(uint8_t *interleaver); +NR_gNB_DLSCH_t *new_gNB_dlsch(unsigned char Kmimo, + unsigned char Mdlharq, + uint32_t Nsoft, + uint8_t abstraction_flag, + NR_DL_FRAME_PARMS *frame_parms, + nfapi_nr_config_request_t *config); + #endif /*__NR_TRANSPORT__H__*/ diff --git a/openair1/PHY/NR_TRANSPORT/nr_transport_common_proto.h b/openair1/PHY/NR_TRANSPORT/nr_transport_common_proto.h index b014748bca..b075da4e97 100644 --- a/openair1/PHY/NR_TRANSPORT/nr_transport_common_proto.h +++ b/openair1/PHY/NR_TRANSPORT/nr_transport_common_proto.h @@ -1,3 +1,4 @@ + /* * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more * contributor license agreements. See the NOTICE file distributed with @@ -33,6 +34,15 @@ #ifndef __NR_TRANSPORT_COMMON_PROTO__H__ #define __NR_TRANSPORT_COMMON_PROTO__H__ +#define MAX_NUM_NR_DLSCH_SEGMENTS 16 +#define MAX_NUM_NR_ULSCH_SEGMENTS MAX_NUM_NR_DLSCH_SEGMENTS + +#define MAX_NR_DLSCH_PAYLOAD_BYTES (MAX_NUM_NR_DLSCH_SEGMENTS*1056) +#define MAX_NR_ULSCH_PAYLOAD_BYTES (MAX_NUM_NR_DLSCH_SEGMENTS*1056) + +#define MAX_NUM_NR_CHANNEL_BITS (14*273*12*6) // 14 symbols, 273 RB +#define MAX_NUM_NR_RE (14*273*12) + // Functions below implement minor procedures from 38-214 /** \brief Computes Q based on I_MCS PDSCH and when 'MCS-Table-PDSCH' is set to "256QAM". Implements Table 5.1.3.1-2 from 38.214. diff --git a/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h b/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h index 52356a9438..f94b18fb27 100644 --- a/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h +++ b/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h @@ -90,7 +90,7 @@ short filt8_l0[8] = { 16384,8192,0,0,0,0,0,0}; short filt8_mr0[8] = { -0,8192,16384,8192,0,-8192,0,0}; +0,0,0,8192,16384,8192,0,-8192}; short filt8_r0[8] = { 0,8192,16384,24576,0,0,0,0}; @@ -98,6 +98,9 @@ short filt8_r0[8] = { short filt8_m0[8] = { 0,8192,16384,8192,0,0,0,0}; +short filt8_mm0[8]= { +0,0,0,8192,16384,8192,0,0}; + short filt8_l1[8] = { 24576,16384,0,0,0,0,0,0}; @@ -109,3 +112,6 @@ short filt8_r1[8] = { short filt8_m1[8] = { 0,0,8192,16384,8192,0,0,0}; + +short filt8_mm1[8]= { +0,0,0,0,8192,16384,8192,0}; diff --git a/openair1/PHY/NR_UE_ESTIMATION/nr_adjust_synch_ue.c b/openair1/PHY/NR_UE_ESTIMATION/nr_adjust_synch_ue.c new file mode 100644 index 0000000000..f1a3ef42ab --- /dev/null +++ b/openair1/PHY/NR_UE_ESTIMATION/nr_adjust_synch_ue.c @@ -0,0 +1,140 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +#include "PHY/types.h" +#include "PHY/defs_nr_UE.h" +#include "PHY/NR_UE_ESTIMATION/nr_estimation.h" +#include "PHY/impl_defs_top.h" +//#include "openair2/LAYER2/MAC/mac_proto.h" + +#include "common/utils/LOG/vcd_signal_dumper.h" + +#define DEBUG_PHY + +// Adjust location synchronization point to account for drift +// The adjustment is performed once per frame based on the +// last channel estimate of the receiver + +void nr_adjust_synch_ue(NR_DL_FRAME_PARMS *frame_parms, + PHY_VARS_NR_UE *ue, + module_id_t eNB_id, + uint8_t subframe, + unsigned char clear, + short coef) +{ + + static int max_pos_fil = 0; + static int count_max_pos_ok = 0; + static int first_time = 1; + int temp = 0, i, aa, max_val = 0, max_pos = 0; + int diff; + short Re,Im,ncoef; + + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_ADJUST_SYNCH, VCD_FUNCTION_IN); + + ncoef = 32767 - coef; + +//#ifdef DEBUG_PHY + LOG_D(PHY,"AbsSubframe %d.%d: rx_offset (before) = %d\n",ue->proc.proc_rxtx[0].frame_rx%1024,subframe,ue->rx_offset); +//#endif //DEBUG_PHY + + + // we only use channel estimates from tx antenna 0 here + for (i = 0; i < frame_parms->nb_prefix_samples; i++) { + temp = 0; + + for (aa=0; aanb_antennas_rx; aa++) { + Re = ((int16_t*)ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[subframe]].dl_ch_estimates_time[eNB_id][aa])[(i<<1)]; + Im = ((int16_t*)ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[subframe]].dl_ch_estimates_time[eNB_id][aa])[1+(i<<1)]; + temp += (Re*Re/2) + (Im*Im/2); + } + + if (temp > max_val) { + max_pos = i; + max_val = temp; + } + } + + // filter position to reduce jitter + if (clear == 1) + max_pos_fil = max_pos; + else + max_pos_fil = ((max_pos_fil * coef) + (max_pos * ncoef)) >> 15; + + // do not filter to have proactive timing adjustment + //max_pos_fil = max_pos; + + if(subframe == 1) + { + diff = max_pos_fil - (frame_parms->nb_prefix_samples>>3); + + if ( abs(diff) < SYNCH_HYST ) + ue->rx_offset = 0; + else + ue->rx_offset = diff; + + if(abs(diff)<5) + count_max_pos_ok ++; + else + count_max_pos_ok = 0; + + //printf("adjust sync count_max_pos_ok = %d\n",count_max_pos_ok); + + if(count_max_pos_ok > 10 && first_time == 1) + { + first_time = 0; + ue->time_sync_cell = 1; + if (ue->mac_enabled==1) { + LOG_I(PHY,"[UE%d] Sending synch status to higher layers\n",ue->Mod_id); + //mac_resynch(); + //dl_phy_sync_success(ue->Mod_id,ue->proc.proc_rxtx[0].frame_rx,0,1);//ue->common_vars.eNb_id); + ue->UE_mode[0] = PRACH; + } + else { + ue->UE_mode[0] = PUSCH; + } + } + + if ( ue->rx_offset < 0 ) + ue->rx_offset += frame_parms->samples_per_frame; + + if ( ue->rx_offset >= frame_parms->samples_per_frame ) + ue->rx_offset -= frame_parms->samples_per_frame; + + + + #ifdef DEBUG_PHY + LOG_D(PHY,"AbsSubframe %d.%d: ThreadId %d diff =%i rx_offset (final) = %i : clear %d,max_pos = %d,max_pos_fil = %d (peak %d) max_val %d target_pos %d \n", + ue->proc.proc_rxtx[ue->current_thread_id[subframe]].frame_rx, + subframe, + ue->current_thread_id[subframe], + diff, + ue->rx_offset, + clear, + max_pos, + max_pos_fil, + temp,max_val, + (frame_parms->nb_prefix_samples>>3)); + #endif //DEBUG_PHY + + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_ADJUST_SYNCH, VCD_FUNCTION_OUT); + } +} diff --git a/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c b/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c index ce4f8d4d01..145c779be9 100644 --- a/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c +++ b/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c @@ -28,7 +28,7 @@ #include "PHY/NR_REFSIG/refsig_defs_ue.h" #include "filt16a_32.h" #include "T.h" -//#define DEBUG_CH +//#define DEBUG_PDCCH int nr_pbch_channel_estimation(PHY_VARS_NR_UE *ue, @@ -48,7 +48,7 @@ int nr_pbch_channel_estimation(PHY_VARS_NR_UE *ue, //uint16_t Nid_cell = (eNB_offset == 0) ? ue->frame_parms.Nid_cell : ue->measurements.adj_cell_id[eNB_offset-1]; - uint8_t nushift,ssb_index=0, n_hf=0; + uint8_t nushift, ssb_index=0, n_hf=0; int **dl_ch_estimates =ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].dl_ch_estimates[eNB_offset]; int **rxdataF=ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].rxdataF; @@ -262,7 +262,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, unsigned short nb_rb_coreset) { int pilot[200] __attribute__((aligned(16))); - unsigned char aarx; + unsigned char aarx,p; unsigned short k; unsigned int pilot_cnt; int16_t ch[2],*pil,*rxF,*dl_ch,*fl,*fm,*fr; @@ -286,7 +286,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, k = coreset_start_subcarrier; -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("PDCCH Channel Estimation : ThreadId %d, eNB_offset %d ch_offset %d, OFDM size %d, Ncp=%d, l=%d, Ns=%d, k=%d symbol %d\n",ue->current_thread_id[Ns>>1], eNB_offset,ch_offset,ue->frame_parms.ofdm_symbol_size, ue->frame_parms.Ncp,l,Ns,k, symbol); #endif @@ -295,8 +295,10 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, fm = filt16a_m1; fr = filt16a_r1; - // generate pilot - nr_pdcch_dmrs_rx(ue,eNB_offset,Ns,ue->nr_gold_pdcch[eNB_offset][Ns][symbol], &pilot[0],2000,nb_rb_coreset); + + // generate pilot + nr_pdcch_dmrs_rx(ue,eNB_offset,Ns,ue->nr_gold_pdcch[eNB_offset][Ns>>1][symbol], &pilot[0],2000,nb_rb_coreset); + for (aarx=0; aarxframe_parms.nb_antennas_rx; aarx++) { @@ -309,18 +311,18 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, multadd_complex_vector_real_scalar(dl_ch-(ue->frame_parms.ofdm_symbol_size<<1), ue->ch_est_alpha,dl_ch-(ue->frame_parms.ofdm_symbol_size<<1), 1,ue->frame_parms.ofdm_symbol_size); - //#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("pdcch ch est pilot addr %p RB_DL %d\n",&pilot[0], ue->frame_parms.N_RB_DL); printf("k %d, first_carrier %d\n",k,ue->frame_parms.first_carrier_offset); printf("rxF addr %p\n", rxF); - printf("dl_ch addr %p\n",dl_ch); - //#endif - if ((ue->frame_parms.N_RB_DL&1)==0) { + printf("dl_ch addr %p\n",dl_ch); +#endif + // if ((ue->frame_parms.N_RB_DL&1)==0) { // Treat first 2 pilots specially (left edge) ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("ch 0 %d\n",((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])); printf("pilot 0 : rxF - > (%d,%d) addr %p ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],&rxF[0],ch[0],ch[1],pil[0],pil[1]); #endif @@ -335,7 +337,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("pilot 1 : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fm, @@ -348,7 +350,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("pilot 2 : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif @@ -357,7 +359,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, dl_ch, 16); -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH for (int m =0; m<12; m++) printf("data : dl_ch -> (%d,%d)\n",dl_ch[0+2*m],dl_ch[1+2*m]); #endif @@ -376,7 +378,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("pilot %d : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",pilot_cnt,rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fl, @@ -392,7 +394,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("pilot %d : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",pilot_cnt+1,rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fm, @@ -405,7 +407,7 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDCCH printf("pilot 1 : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif @@ -421,36 +423,84 @@ int nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, } - } + //} } + + void (*idft)(int16_t *,int16_t *, int); + + switch (ue->frame_parms.ofdm_symbol_size) { + case 128: + idft = idft128; + break; + + case 256: + idft = idft256; + break; + + case 512: + idft = idft512; + break; + + case 1024: + idft = idft1024; + break; + + case 1536: + idft = idft1536; + break; + + case 2048: + idft = idft2048; + break; + + default: + idft = idft512; + break; + } + + //if( (Ns== 2) && (l == 0)) + { + // do ifft of channel estimate + for (aarx=0; aarxframe_parms.nb_antennas_rx; aarx++) + for (p=0; pframe_parms.nb_antenna_ports_eNB; p++) { + if (ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].dl_ch_estimates[eNB_offset][(p<<1)+aarx]) + { + LOG_D(PHY,"Channel Impulse Computation Slot %d ThreadId %d Symbol %d \n", Ns, ue->current_thread_id[Ns>>1], l); + idft((int16_t*) &ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].dl_ch_estimates[eNB_offset][(p<<1)+aarx][0], + (int16_t*) ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].dl_ch_estimates_time[eNB_offset][(p<<1)+aarx],1); + } + } + } return(0); } int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, - uint8_t eNB_id, - uint8_t eNB_offset, - unsigned char Ns, - unsigned char l, - unsigned char symbol, - unsigned short bwp_start_subcarrier, - unsigned short nb_rb_pdsch) + uint8_t eNB_id, + uint8_t eNB_offset, + unsigned char Ns, + unsigned short p, + unsigned char l, + unsigned char symbol, + unsigned short bwp_start_subcarrier, + unsigned short nb_rb_pdsch) { - int pilot[200] __attribute__((aligned(16))); + int pilot[1320] __attribute__((aligned(16))); unsigned char aarx; unsigned short k; unsigned int pilot_cnt; - int16_t ch[2],*pil,*rxF,*dl_ch,*fl,*fm,*fr,*fml,*fmr; + int16_t ch[2],*pil,*rxF,*dl_ch,*fl,*fm,*fr,*fml,*fmr,*fmm; int ch_offset,symbol_offset; - //uint16_t Nid_cell = (eNB_offset == 0) ? ue->frame_parms.Nid_cell : ue->measurements.adj_cell_id[eNB_offset-1]; - uint8_t nushift=0; + uint8_t nushift; int **dl_ch_estimates =ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].dl_ch_estimates[eNB_offset]; int **rxdataF=ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[Ns>>1]].rxdataF; + nushift = (p>>1)&1; + ue->frame_parms.nushift = nushift; if (ue->high_speed_flag == 0) // use second channel estimate position for temporary storage ch_offset = ue->frame_parms.ofdm_symbol_size ; @@ -460,38 +510,41 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, symbol_offset = ue->frame_parms.ofdm_symbol_size*symbol; k = bwp_start_subcarrier; + int re_offset = k; #ifdef DEBUG_CH - printf("PBCH Channel Estimation : ThreadId %d, eNB_offset %d ch_offset %d, OFDM size %d, Ncp=%d, l=%d, Ns=%d, k=%d symbol %d\n",ue->current_thread_id[Ns>>1], eNB_offset,ch_offset,ue->frame_parms.ofdm_symbol_size, + printf("PDSCH Channel Estimation : ThreadId %d, eNB_offset %d ch_offset %d, symbol_offset %d OFDM size %d, Ncp=%d, l=%d, Ns=%d, k=%d symbol %d\n",ue->current_thread_id[Ns], eNB_offset,ch_offset,symbol_offset,ue->frame_parms.ofdm_symbol_size, ue->frame_parms.Ncp,l,Ns,k, symbol); #endif switch (nushift) { - case 0: - fl = filt8_l0; - fm = filt8_m0; - fr = filt8_r0; - fml = filt8_m0; - fmr = filt8_mr0; - break; + case 0: + fl = filt8_l0; + fm = filt8_m0; + fr = filt8_r0; + fmm = filt8_mm0; + fml = filt8_m0; + fmr = filt8_mr0; + break; - case 1: - fl = filt8_l1; - fm = filt8_m1; - fr = filt8_r1; - fml = filt8_ml1; - fmr = filt8_m1; - break; + case 1: + fl = filt8_l1; + fm = filt8_m1; + fr = filt8_r1; + fmm = filt8_mm1; + fml = filt8_ml1; + fmr = filt8_m1; + break; - default: - msg("pdsch_channel_estimation: nushift=%d -> ERROR\n",nushift); - return(-1); - break; - } + default: + msg("pdsch_channel_estimation: nushift=%d -> ERROR\n",nushift); + return(-1); + break; + } // generate pilot - nr_pdsch_dmrs_rx(ue,eNB_offset,Ns,ue->nr_gold_pdsch, &pilot[0],1000,1,nb_rb_pdsch); + nr_pdsch_dmrs_rx(ue,Ns,ue->nr_gold_pdsch[eNB_offset][Ns][0], &pilot[0],1000,0,nb_rb_pdsch); for (aarx=0; aarxframe_parms.nb_antennas_rx; aarx++) { @@ -504,33 +557,35 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, multadd_complex_vector_real_scalar(dl_ch-(ue->frame_parms.ofdm_symbol_size<<1), ue->ch_est_alpha,dl_ch-(ue->frame_parms.ofdm_symbol_size<<1), 1,ue->frame_parms.ofdm_symbol_size); -#ifdef DEBUG_CH +#ifdef DEBUG_PDSCH printf("ch est pilot addr %p RB_DL %d\n",&pilot[0], ue->frame_parms.N_RB_DL); printf("k %d, first_carrier %d\n",k,ue->frame_parms.first_carrier_offset); - printf("rxF addr %p\n", rxF); - printf("dl_ch addr %p\n",dl_ch); + printf("rxF addr %p p %d\n", rxF,p); + printf("dl_ch addr %p nushift %d\n",dl_ch,nushift); #endif - if ((ue->frame_parms.N_RB_DL&1)==0) { + //if ((ue->frame_parms.N_RB_DL&1)==0) { // Treat first 2 pilots specially (left edge) ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDSCH printf("ch 0 %d\n",((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])); printf("pilot 0 : rxF - > (%d,%d) addr %p ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],&rxF[0],ch[0],ch[1],pil[0],pil[1]); + printf("data 0 : rxF - > (%d,%d) addr %p ch -> (%d,%d), pil -> (%d,%d) \n",rxF[2],rxF[3],&rxF[2],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fl, ch, dl_ch, 8); pil+=2; - rxF+=4; + re_offset = (re_offset+2)&(ue->frame_parms.ofdm_symbol_size-1); + rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; //for (int i= 0; i<8; i++) //printf("dl_ch addr %p %d\n", dl_ch+i, *(dl_ch+i)); ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDSCH printf("pilot 1 : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fml, @@ -538,21 +593,36 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, dl_ch, 8); pil+=2; - rxF+=4; - dl_ch+=4; - k+=4; + re_offset = (re_offset+2)&(ue->frame_parms.ofdm_symbol_size-1); + rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; + //printf("dl_ch addr %p\n",dl_ch); + + ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); + ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); +#ifdef DEBUG_PDSCH + printf("pilot 2 : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); +#endif + multadd_real_vector_complex_scalar(fmm, + ch, + dl_ch, + 8); + + //for (int i= 0; i<16; i++) + //printf("dl_ch addr %p %d\n", dl_ch+i, *(dl_ch+i)); + + pil+=2; + re_offset = (re_offset+2)&(ue->frame_parms.ofdm_symbol_size-1); + rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; + dl_ch+=8; - for (pilot_cnt=2; pilot_cnt<(6*(nb_rb_pdsch-1)+4); pilot_cnt+=2) { - if ((pilot_cnt%6)==0) - dl_ch+=4; - - if (k >= ue->frame_parms.ofdm_symbol_size){ - k-=ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+k+nushift)];} + for (pilot_cnt=3; pilot_cnt<(6*nb_rb_pdsch-3); pilot_cnt+=2) { + //if ((pilot_cnt%6)==0) + //dl_ch+=4; + //printf("re_offset %d\n",re_offset); ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDSCH printf("pilot %d : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",pilot_cnt,rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fm, @@ -561,56 +631,69 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, 8); pil+=2; - rxF+=4; - dl_ch+=4; - + re_offset = (re_offset+2)&(ue->frame_parms.ofdm_symbol_size-1); + rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; + ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDSCH printf("pilot %d : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",pilot_cnt+1,rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif - multadd_real_vector_complex_scalar(fm, + multadd_real_vector_complex_scalar(fmm, ch, dl_ch, 8); pil+=2; - rxF+=4; - dl_ch+=4; - k+=4; + re_offset = (re_offset+2)&(ue->frame_parms.ofdm_symbol_size-1); + rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; + dl_ch+=8; } - + // Treat first 2 pilots specially (right edge) + ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); + ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); +#ifdef DEBUG_PDSCH + printf("pilot %d : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",pilot_cnt,rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); +#endif + multadd_real_vector_complex_scalar(fm, + ch, + dl_ch, + 8); + + //for (int i= 0; i<8; i++) + //printf("dl_ch addr %p %d\n", dl_ch+i, *(dl_ch+i)); + + pil+=2; + re_offset = (re_offset+2)&(ue->frame_parms.ofdm_symbol_size-1); + rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; + ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH +#ifdef DEBUG_PDSCH printf("ch 0 %d\n",((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])); - printf("pilot 0 : rxF - > (%d,%d) addr %p ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],&rxF[0],ch[0],ch[1],pil[0],pil[1]); + printf("pilot %d: rxF - > (%d,%d) addr %p ch -> (%d,%d), pil -> (%d,%d) \n",pilot_cnt+1,rxF[0],rxF[1],&rxF[0],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fmr, ch, dl_ch, 8); + pil+=2; - rxF+=4; - //for (int i= 0; i<8; i++) - //printf("dl_ch addr %p %d\n", dl_ch+i, *(dl_ch+i)); - + re_offset = (re_offset+2)&(ue->frame_parms.ofdm_symbol_size-1); + rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; + dl_ch+=8; + ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); -#ifdef DEBUG_CH - printf("pilot 1 : rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); +#ifdef DEBUG_PDSCH + printf("pilot %d: rxF - > (%d,%d) ch -> (%d,%d), pil -> (%d,%d) \n",pilot_cnt+2,rxF[0],rxF[1],ch[0],ch[1],pil[0],pil[1]); #endif multadd_real_vector_complex_scalar(fr, ch, dl_ch, 8); - pil+=2; - rxF+=4; - dl_ch+=4; - k+=4; - - } + //} } diff --git a/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h b/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h index 25d6755723..ae4295a99b 100644 --- a/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h +++ b/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h @@ -62,8 +62,17 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, uint8_t eNB_id, uint8_t eNB_offset, unsigned char Ns, + unsigned short p, unsigned char l, unsigned char symbol, unsigned short bwp_start_subcarrier, unsigned short nb_rb_pdsch); + +void nr_adjust_synch_ue(NR_DL_FRAME_PARMS *frame_parms, + PHY_VARS_NR_UE *ue, + module_id_t eNB_id, + uint8_t subframe, + unsigned char clear, + short coef); + #endif diff --git a/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c b/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c index b1a384d4a0..3d481a0468 100755 --- a/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c +++ b/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c @@ -49,62 +49,66 @@ //#define NR_LTE_PDCCH_DCI_SWITCH #define NR_PDCCH_DCI_RUN // activates new nr functions -#define NR_PDCCH_DCI_DEBUG // activates NR_PDCCH_DCI_DEBUG logs +//#define NR_PDCCH_DCI_DEBUG // activates NR_PDCCH_DCI_DEBUG logs #define NR_NBR_CORESET_ACT_BWP 3 // The number of CoreSets per BWP is limited to 3 (including initial CORESET: ControlResourceId 0) #define NR_NBR_SEARCHSPACE_ACT_BWP 10 // The number of SearSpaces per BWP is limited to 10 (including initial SEARCHSPACE: SearchSpaceId 0) #define PDCCH_TEST_POLAR_TEMP_FIX - +#ifdef LOG_I +#undef LOG_I +#define LOG_I(A,B...) printf(B) +#endif #ifdef NR_PDCCH_DCI_RUN -static const uint16_t conjugate[8]__attribute__((aligned(32))) = {-1,1,-1,1,-1,1,-1,1}; + +//static const int16_t conjugate[8]__attribute__((aligned(32))) = {-1,1,-1,1,-1,1,-1,1}; -void nr_pdcch_demapping_deinterleaving(uint16_t *llr, - uint16_t *z, +void nr_pdcch_demapping_deinterleaving(uint32_t *llr, + uint32_t *z, NR_DL_FRAME_PARMS *frame_parms, uint8_t coreset_time_dur, uint32_t coreset_nbr_rb, uint8_t reg_bundle_size_L, uint8_t coreset_interleaver_size_R, uint8_t n_shift){ -/* - * This function will do demapping and deinterleaving from llr containing demodulated symbols - * Demapping will regroup in REG and bundles - * Deinterleaving will order the bundles - * - * In the following example we can see the process. The llr contains the demodulated IQs, but they are not ordered from REG 0,1,2,.. - * In e_rx (z) we will order the REG ids and group them into bundles. - * Then we will put the bundles in the correct order as indicated in subclause 7.3.2.2 - * - llr --------------------------> e_rx (z) ----> e_rx (z) - | ... - | ... - | REG 26 -symbol 2 | ... - | ... - | REG 5 - | REG 2 + /* + * This function will do demapping and deinterleaving from llr containing demodulated symbols + * Demapping will regroup in REG and bundles + * Deinterleaving will order the bundles + * + * In the following example we can see the process. The llr contains the demodulated IQs, but they are not ordered from REG 0,1,2,.. + * In e_rx (z) we will order the REG ids and group them into bundles. + * Then we will put the bundles in the correct order as indicated in subclause 7.3.2.2 + * + llr --------------------------> e_rx (z) ----> e_rx (z) + | ... + | ... + | REG 26 + symbol 2 | ... + | ... + | REG 5 + | REG 2 - | ... - | ... - | REG 25 -symbol 1 | ... - | ... - | REG 4 - | REG 1 + | ... + | ... + | REG 25 + symbol 1 | ... + | ... + | REG 4 + | REG 1 - | ... - | ... ... ... - | REG 24 (bundle 7) ... ... -symbol 0 | ... bundle 3 bundle 6 - | ... bundle 2 bundle 1 - | REG 3 bundle 1 bundle 7 - | REG 0 (bundle 0) bundle 0 bundle 0 + | ... + | ... ... ... + | REG 24 (bundle 7) ... ... + symbol 0 | ... bundle 3 bundle 6 + | ... bundle 2 bundle 1 + | REG 3 bundle 1 bundle 7 + | REG 0 (bundle 0) bundle 0 bundle 0 -*/ + */ int c=0,r=0; uint16_t bundle_j=0, f_bundle_j=0,f_reg=0; uint32_t coreset_C=0; @@ -127,9 +131,10 @@ symbol 0 | ... bundle 3 bundle 6 bundle_j = (c*coreset_interleaver_size_R)+r; f_bundle_j = ((r*coreset_C)+c+n_shift)%((coreset_nbr_rb*coreset_time_dur)/reg_bundle_size_L); if (coreset_interleaved==0) f_bundle_j=bundle_j; - #ifndef NR_PDCCH_DCI_DEBUG - printf("\n\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_demapping_deinterleaving)-> [r=%d,c=%d] bundle_j(%d) interleaved at f_bundle_j(%d)\n",r,c,bundle_j,f_bundle_j); - #endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\n\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_demapping_deinterleaving)-> [r=%d,c=%d] bundle_j(%d) interleaved at f_bundle_j(%d)\n",r,c,bundle_j,f_bundle_j); +#endif } f_reg = (f_bundle_j*reg_bundle_size_L)+(reg%reg_bundle_size_L); //index_z = 9*reg; @@ -138,11 +143,12 @@ symbol 0 | ... bundle 3 bundle 6 index_llr = 9*((uint16_t)floor(f_reg/coreset_time_dur)+((f_reg%coreset_time_dur)*(coreset_nbr_rb))); for (int i=0; i<9; i++){ z[index_z + i] = llr[index_llr + i]; - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_demapping_deinterleaving)-> [reg=%d,bundle_j=%d] z[%d]=(%d,%d) <-> \t[f_reg=%d,fbundle_j=%d] llr[%d]=(%d,%d) \n", - reg,bundle_j,(index_z + i),*(char*) &z[index_z + i],*(1 + (char*) &z[index_z + i]), - f_reg,f_bundle_j,(index_llr + i),*(char*) &llr[index_llr + i], *(1 + (char*) &llr[index_llr + i])); - #endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_demapping_deinterleaving)-> [reg=%d,bundle_j=%d] z[%d]=(%d,%d) <-> \t[f_reg=%d,fbundle_j=%d] llr[%d]=(%d,%d) \n", + reg,bundle_j,(index_z + i),*(int16_t*) &z[index_z + i],*(1 + (int16_t*) &z[index_z + i]), + f_reg,f_bundle_j,(index_llr + i),*(int16_t*) &llr[index_llr + i], *(1 + (int16_t*) &llr[index_llr + i])); +#endif } if ((reg%reg_bundle_size_L) == 0) r++; } @@ -152,38 +158,39 @@ symbol 0 | ... bundle 3 bundle 6 #ifdef NR_PDCCH_DCI_RUN int32_t nr_pdcch_llr(NR_DL_FRAME_PARMS *frame_parms, int32_t **rxdataF_comp, - char *pdcch_llr, uint8_t symbol,uint32_t coreset_nbr_rb) { + int16_t *pdcch_llr, uint8_t symbol,uint32_t coreset_nbr_rb) { - int16_t *rxF = (int16_t*) &rxdataF_comp[0][(symbol * coreset_nbr_rb * 12)]; - int32_t i; - char *pdcch_llr8; + int16_t *rxF = (int16_t*) &rxdataF_comp[0][(symbol * coreset_nbr_rb * 12)]; + int32_t i; + int16_t *pdcch_llrp; - pdcch_llr8 = &pdcch_llr[2 * symbol * coreset_nbr_rb * 9]; + pdcch_llrp = &pdcch_llr[2 * symbol * coreset_nbr_rb * 9]; - if (!pdcch_llr8) { - printf("pdcch_qpsk_llr: llr is null, symbol %d\n", symbol); - return (-1); - } -#ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_llr)-> llr logs: pdcch qpsk llr for symbol %d (pos %d), llr offset %d\n",symbol,(symbol*frame_parms->N_RB_DL*12),pdcch_llr8-pdcch_llr); + + if (!pdcch_llrp) { + printf("pdcch_qpsk_llr: llr is null, symbol %d\n", symbol); + return (-1); + } +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_llr)-> llr logs: pdcch qpsk llr for symbol %d (pos %d), llr offset %d\n",symbol,(symbol*frame_parms->N_RB_DL*12),pdcch_llrp-pdcch_llr); #endif - //for (i = 0; i < (frame_parms->N_RB_DL * ((symbol == 0) ? 16 : 24)); i++) { - for (i = 0; i < (coreset_nbr_rb * ((symbol == 0) ? 18 : 18)); i++) { + //for (i = 0; i < (frame_parms->N_RB_DL * ((symbol == 0) ? 16 : 24)); i++) { + for (i = 0; i < (coreset_nbr_rb * ((symbol == 0) ? 18 : 18)); i++) { - if (*rxF > 31) - *pdcch_llr8 = 31; - else if (*rxF < -32) - *pdcch_llr8 = -32; - else - *pdcch_llr8 = (char) (*rxF); -#ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_llr)-> llr logs: rb=%d i=%d *rxF:%d => *pdcch_llr8:%d\n",i/18,i,*rxF,*pdcch_llr8); + if (*rxF > 31) + *pdcch_llrp = 31; + else if (*rxF < -32) + *pdcch_llrp = -32; + else + *pdcch_llrp = (*rxF); +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_llr)-> llr logs: rb=%d i=%d *rxF:%d => *pdcch_llrp:%d\n",i/18,i,*rxF,*pdcch_llrp); #endif - rxF++; - pdcch_llr8++; - } + rxF++; + pdcch_llrp++; + } - return (0); + return (0); } #endif @@ -237,7 +244,7 @@ void pdcch_channel_level(int32_t **dl_ch_estimates_ext, { int16_t rb; - uint8_t aatx,aarx; + uint8_t aarx; #if defined(__x86_64__) || defined(__i386__) __m128i *dl_ch128; __m128i avg128P; @@ -245,42 +252,42 @@ void pdcch_channel_level(int32_t **dl_ch_estimates_ext, int16x8_t *dl_ch128; int32x4_t *avg128P; #endif - for (aatx=0; aatxnb_antenna_ports_eNB; aatx++) - for (aarx=0; aarxnb_antennas_rx; aarx++) { - //clear average level + for (aarx=0; aarxnb_antennas_rx; aarx++) { + //clear average level #if defined(__x86_64__) || defined(__i386__) - avg128P = _mm_setzero_si128(); - dl_ch128=(__m128i *)&dl_ch_estimates_ext[(aatx<<1)+aarx][0]; + avg128P = _mm_setzero_si128(); + dl_ch128=(__m128i *)&dl_ch_estimates_ext[aarx][0]; #elif defined(__arm__) - + #endif - for (rb=0; rb>2; rb++) { + #if defined(__x86_64__) || defined(__i386__) - avg128P = _mm_add_epi32(avg128P,_mm_madd_epi16(dl_ch128[0],dl_ch128[0])); - avg128P = _mm_add_epi32(avg128P,_mm_madd_epi16(dl_ch128[1],dl_ch128[1])); - avg128P = _mm_add_epi32(avg128P,_mm_madd_epi16(dl_ch128[2],dl_ch128[2])); + avg128P = _mm_add_epi32(avg128P,_mm_madd_epi16(dl_ch128[0],dl_ch128[0])); + avg128P = _mm_add_epi32(avg128P,_mm_madd_epi16(dl_ch128[1],dl_ch128[1])); + avg128P = _mm_add_epi32(avg128P,_mm_madd_epi16(dl_ch128[2],dl_ch128[2])); #elif defined(__arm__) - + #endif - dl_ch128+=3; - /* - if (rb==0) { - print_shorts("dl_ch128",&dl_ch128[0]); - print_shorts("dl_ch128",&dl_ch128[1]); - print_shorts("dl_ch128",&dl_ch128[2]); - } - */ - } - - DevAssert( nb_rb ); - avg[(aatx<<1)+aarx] = (((int32_t*)&avg128P)[0] + - ((int32_t*)&avg128P)[1] + - ((int32_t*)&avg128P)[2] + - ((int32_t*)&avg128P)[3])/(nb_rb*12); - - // printf("Channel level : %d\n",avg[(aatx<<1)+aarx]); + // for (int i=0;i<24;i+=2) printf("pdcch channel re %d (%d,%d)\n",(rb*12)+(i>>1),((int16_t*)dl_ch128)[i],((int16_t*)dl_ch128)[i+1]); + dl_ch128+=3; + /* + if (rb==0) { + print_shorts("dl_ch128",&dl_ch128[0]); + print_shorts("dl_ch128",&dl_ch128[1]); + print_shorts("dl_ch128",&dl_ch128[2]); + } + */ } + + DevAssert( nb_rb ); + avg[aarx] = (((int32_t*)&avg128P)[0] + + ((int32_t*)&avg128P)[1] + + ((int32_t*)&avg128P)[2] + + ((int32_t*)&avg128P)[3])/(nb_rb*9); + + // printf("Channel level : %d\n",avg[(aatx<<1)+aarx]); + } #if defined(__x86_64__) || defined(__i386__) _mm_empty(); @@ -296,89 +303,6 @@ __m128i mmtmpPD0,mmtmpPD1,mmtmpPD2,mmtmpPD3; #endif -void pdcch_detection_mrc_i(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **rho, - int32_t **rho_i, - uint8_t symbol) -{ - - uint8_t aatx; - -#if defined(__x86_64__) || defined(__i386__) - __m128i *rxdataF_comp128_0,*rxdataF_comp128_1,*rxdataF_comp128_i0,*rxdataF_comp128_i1,*rho128_0,*rho128_1,*rho128_i0,*rho128_i1; -#elif defined(__arm__) - int16x8_t *rxdataF_comp128_0,*rxdataF_comp128_1,*rxdataF_comp128_i0,*rxdataF_comp128_i1,*rho128_0,*rho128_1,*rho128_i0,*rho128_i1; -#endif - int32_t i; - - if (frame_parms->nb_antennas_rx>1) { - for (aatx=0; aatxnb_antenna_ports_eNB; aatx++) { - //if (frame_parms->mode1_flag && (aatx>0)) break; - -#if defined(__x86_64__) || defined(__i386__) - rxdataF_comp128_0 = (__m128i *)&rxdataF_comp[(aatx<<1)][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_1 = (__m128i *)&rxdataF_comp[(aatx<<1)+1][symbol*frame_parms->N_RB_DL*12]; -#elif defined(__arm__) - rxdataF_comp128_0 = (int16x8_t *)&rxdataF_comp[(aatx<<1)][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_1 = (int16x8_t *)&rxdataF_comp[(aatx<<1)+1][symbol*frame_parms->N_RB_DL*12]; -#endif - // MRC on each re of rb on MF output - for (i=0; iN_RB_DL*3; i++) { -#if defined(__x86_64__) || defined(__i386__) - rxdataF_comp128_0[i] = _mm_adds_epi16(_mm_srai_epi16(rxdataF_comp128_0[i],1),_mm_srai_epi16(rxdataF_comp128_1[i],1)); -#elif defined(__arm__) - rxdataF_comp128_0[i] = vhaddq_s16(rxdataF_comp128_0[i],rxdataF_comp128_1[i]); -#endif - } - } - -#if defined(__x86_64__) || defined(__i386__) - rho128_0 = (__m128i *) &rho[0][symbol*frame_parms->N_RB_DL*12]; - rho128_1 = (__m128i *) &rho[1][symbol*frame_parms->N_RB_DL*12]; -#elif defined(__arm__) - rho128_0 = (int16x8_t *) &rho[0][symbol*frame_parms->N_RB_DL*12]; - rho128_1 = (int16x8_t *) &rho[1][symbol*frame_parms->N_RB_DL*12]; -#endif - for (i=0; iN_RB_DL*3; i++) { -#if defined(__x86_64__) || defined(__i386__) - rho128_0[i] = _mm_adds_epi16(_mm_srai_epi16(rho128_0[i],1),_mm_srai_epi16(rho128_1[i],1)); -#elif defined(__arm__) - rho128_0[i] = vhaddq_s16(rho128_0[i],rho128_1[i]); -#endif - } - -#if defined(__x86_64__) || defined(__i386__) - rho128_i0 = (__m128i *) &rho_i[0][symbol*frame_parms->N_RB_DL*12]; - rho128_i1 = (__m128i *) &rho_i[1][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_i0 = (__m128i *)&rxdataF_comp_i[0][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_i1 = (__m128i *)&rxdataF_comp_i[1][symbol*frame_parms->N_RB_DL*12]; -#elif defined(__arm__) - rho128_i0 = (int16x8_t*) &rho_i[0][symbol*frame_parms->N_RB_DL*12]; - rho128_i1 = (int16x8_t*) &rho_i[1][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_i0 = (int16x8_t *)&rxdataF_comp_i[0][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_i1 = (int16x8_t *)&rxdataF_comp_i[1][symbol*frame_parms->N_RB_DL*12]; - -#endif - // MRC on each re of rb on MF and rho - for (i=0; iN_RB_DL*3; i++) { -#if defined(__x86_64__) || defined(__i386__) - rxdataF_comp128_i0[i] = _mm_adds_epi16(_mm_srai_epi16(rxdataF_comp128_i0[i],1),_mm_srai_epi16(rxdataF_comp128_i1[i],1)); - rho128_i0[i] = _mm_adds_epi16(_mm_srai_epi16(rho128_i0[i],1),_mm_srai_epi16(rho128_i1[i],1)); -#elif defined(__arm__) - rxdataF_comp128_i0[i] = vhaddq_s16(rxdataF_comp128_i0[i],rxdataF_comp128_i1[i]); - rho128_i0[i] = vhaddq_s16(rho128_i0[i],rho128_i1[i]); - -#endif - } - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif -} #ifdef NR_PDCCH_DCI_RUN @@ -394,23 +318,23 @@ void nr_pdcch_extract_rbs_single(int32_t **rxdataF, uint32_t coreset_nbr_rb, uint32_t n_BWP_start) { -/* - * This function is demapping DM-RS PDCCH RE - * Implementing 38.211 Section 7.4.1.3.2 Mapping to physical resources - * PDCCH DM-RS signals are mapped on RE a_k_l where: - * k = 12*n + 4*kprime + 1 - * n=0,1,.. - * kprime=0,1,2 - * According to this equations, DM-RS PDCCH are mapped on k where k%12==1 || k%12==5 || k%12==9 - * - */ + /* + * This function is demapping DM-RS PDCCH RE + * Implementing 38.211 Section 7.4.1.3.2 Mapping to physical resources + * PDCCH DM-RS signals are mapped on RE a_k_l where: + * k = 12*n + 4*kprime + 1 + * n=0,1,.. + * kprime=0,1,2 + * According to this equations, DM-RS PDCCH are mapped on k where k%12==1 || k%12==5 || k%12==9 + * + */ // the bitmap coreset_frq_domain contains 45 bits - #define CORESET_FREQ_DOMAIN_BITMAP_SIZE 45 +#define CORESET_FREQ_DOMAIN_BITMAP_SIZE 45 // each bit is associated to 6 RBs - #define BIT_TO_NBR_RB_CORESET_FREQ_DOMAIN 6 - #define NBR_RE_PER_RB_WITH_DMRS 12 +#define BIT_TO_NBR_RB_CORESET_FREQ_DOMAIN 6 +#define NBR_RE_PER_RB_WITH_DMRS 12 // after removing the 3 DMRS RE, the RB contains 9 RE with PDCCH - #define NBR_RE_PER_RB_WITHOUT_DMRS 9 +#define NBR_RE_PER_RB_WITHOUT_DMRS 9 uint16_t c_rb, nb_rb = 0; // this variable will be incremented by 1 each time a bit set to '0' is found in coreset_freq_dom bitmap @@ -423,59 +347,59 @@ void nr_pdcch_extract_rbs_single(int32_t **rxdataF, symbol_mod = (symbol >= (7 - frame_parms->Ncp)) ? symbol - (7 - frame_parms->Ncp) : symbol; c_rb = n_BWP_start; // c_rb is the common resource block: RB within the BWP - #ifdef DEBUG_DCI_DECODING - LOG_I(PHY, "extract_rbs_single: symbol_mod %d\n",symbol_mod); - #endif +#ifdef DEBUG_DCI_DECODING + LOG_I(PHY, "extract_rbs_single: symbol_mod %d\n",symbol_mod); +#endif for (aarx = 0; aarx < frame_parms->nb_antennas_rx; aarx++) { if (high_speed_flag == 1){ dl_ch0 = &dl_ch_estimates[aarx][(symbol * (frame_parms->ofdm_symbol_size))]; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> dl_ch0 = &dl_ch_estimates[aarx = (%d)][ (symbol * (frame_parms->ofdm_symbol_size (%d))) = (%d)]\n", - aarx,frame_parms->ofdm_symbol_size,(symbol * (frame_parms->ofdm_symbol_size))); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> dl_ch0 = &dl_ch_estimates[aarx = (%d)][ (symbol * (frame_parms->ofdm_symbol_size (%d))) = (%d)]\n", + aarx,frame_parms->ofdm_symbol_size,(symbol * (frame_parms->ofdm_symbol_size))); +#endif } else { dl_ch0 = &dl_ch_estimates[aarx][0]; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> dl_ch0 = &dl_ch_estimates[aarx = (%d)][0]\n",aarx); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> dl_ch0 = &dl_ch_estimates[aarx = (%d)][0]\n",aarx); +#endif } dl_ch0_ext = &dl_ch_estimates_ext[aarx][symbol * (coreset_nbr_rb * NBR_RE_PER_RB_WITH_DMRS)]; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> dl_ch0_ext = &dl_ch_estimates_ext[aarx = (%d)][symbol * (frame_parms->N_RB_DL * 9) = (%d)]\n", - aarx,symbol * (coreset_nbr_rb * NBR_RE_PER_RB_WITH_DMRS)); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> dl_ch0_ext = &dl_ch_estimates_ext[aarx = (%d)][symbol * (frame_parms->N_RB_DL * 9) = (%d)]\n", + aarx,symbol * (coreset_nbr_rb * NBR_RE_PER_RB_WITH_DMRS)); +#endif rxF_ext = &rxdataF_ext[aarx][symbol * (coreset_nbr_rb * NBR_RE_PER_RB_WITH_DMRS)]; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> rxF_ext = &rxdataF_ext[aarx = (%d)][symbol * (frame_parms->N_RB_DL * 9) = (%d)]\n", - aarx,symbol * (coreset_nbr_rb * NBR_RE_PER_RB_WITH_DMRS)); - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> (for symbol=%d, aarx=%d), symbol_mod=%d, nushiftmod3=%d \n",symbol,aarx,symbol_mod,nushiftmod3); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> rxF_ext = &rxdataF_ext[aarx = (%d)][symbol * (frame_parms->N_RB_DL * 9) = (%d)]\n", + aarx,symbol * (coreset_nbr_rb * NBR_RE_PER_RB_WITH_DMRS)); + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> (for symbol=%d, aarx=%d), symbol_mod=%d, nushiftmod3=%d \n",symbol,aarx,symbol_mod,nushiftmod3); +#endif -/* - * The following for loop handles treatment of PDCCH contained in table rxdataF (in frequency domain) - * In NR the PDCCH IQ symbols are contained within RBs in the CORESET defined by higher layers which is located within the BWP - * Lets consider that the first RB to be considered as part of the CORESET and part of the PDCCH is n_BWP_start - * Several cases have to be handled differently as IQ symbols are situated in different parts of rxdataF: - * 1. Number of RBs in the system bandwidth is even - * 1.1 The RB is < than the N_RB_DL/2 -> IQ symbols are in the second half of the rxdataF (from first_carrier_offset) - * 1.2 The RB is >= than the N_RB_DL/2 -> IQ symbols are in the first half of the rxdataF (from element 0) - * 2. Number of RBs in the system bandwidth is odd - * (particular case when the RB with DC as it is treated differently: it is situated in symbol borders of rxdataF) - * 2.1 The RB is <= than the N_RB_DL/2 -> IQ symbols are in the second half of the rxdataF (from first_carrier_offset) - * 2.2 The RB is > than the N_RB_DL/2+1 -> IQ symbols are in the first half of the rxdataF (from element 0 + 2nd half RB containing DC) - * 2.3 The RB is == N_RB_DL/2+1 -> IQ symbols are in the lower border of the rxdataF for first 6 IQ element and the upper border of the rxdataF for the last 6 IQ elements - * If the first RB containing PDCCH within the UE BWP and within the CORESET is higher than half of the system bandwidth (N_RB_DL), - * then the IQ symbol is going to be found at the position 0+c_rb-N_RB_DL/2 in rxdataF and - * we have to point the pointer at (1+c_rb-N_RB_DL/2) in rxdataF - */ + /* + * The following for loop handles treatment of PDCCH contained in table rxdataF (in frequency domain) + * In NR the PDCCH IQ symbols are contained within RBs in the CORESET defined by higher layers which is located within the BWP + * Lets consider that the first RB to be considered as part of the CORESET and part of the PDCCH is n_BWP_start + * Several cases have to be handled differently as IQ symbols are situated in different parts of rxdataF: + * 1. Number of RBs in the system bandwidth is even + * 1.1 The RB is < than the N_RB_DL/2 -> IQ symbols are in the second half of the rxdataF (from first_carrier_offset) + * 1.2 The RB is >= than the N_RB_DL/2 -> IQ symbols are in the first half of the rxdataF (from element 0) + * 2. Number of RBs in the system bandwidth is odd + * (particular case when the RB with DC as it is treated differently: it is situated in symbol borders of rxdataF) + * 2.1 The RB is <= than the N_RB_DL/2 -> IQ symbols are in the second half of the rxdataF (from first_carrier_offset) + * 2.2 The RB is > than the N_RB_DL/2+1 -> IQ symbols are in the first half of the rxdataF (from element 0 + 2nd half RB containing DC) + * 2.3 The RB is == N_RB_DL/2+1 -> IQ symbols are in the lower border of the rxdataF for first 6 IQ element and the upper border of the rxdataF for the last 6 IQ elements + * If the first RB containing PDCCH within the UE BWP and within the CORESET is higher than half of the system bandwidth (N_RB_DL), + * then the IQ symbol is going to be found at the position 0+c_rb-N_RB_DL/2 in rxdataF and + * we have to point the pointer at (1+c_rb-N_RB_DL/2) in rxdataF + */ - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> n_BWP_start=%d, coreset_nbr_rb=%d\n",n_BWP_start,coreset_nbr_rb); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> n_BWP_start=%d, coreset_nbr_rb=%d\n",n_BWP_start,coreset_nbr_rb); +#endif for (c_rb = n_BWP_start; c_rb < (n_BWP_start + coreset_nbr_rb + (BIT_TO_NBR_RB_CORESET_FREQ_DOMAIN * offset_discontiguous)); c_rb++) { //c_rb_tmp = 0; @@ -487,33 +411,34 @@ void nr_pdcch_extract_rbs_single(int32_t **rxdataF, //c_rb_tmp = c_rb_tmp + 6; c_rb = c_rb + BIT_TO_NBR_RB_CORESET_FREQ_DOMAIN; offset_discontiguous ++; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> we entered here as coreset_freq_dom=%lx (bit %d) is 0, coreset_freq_domain is discontiguous\n",coreset_freq_dom,(46 - bitcnt_coreset_freq_dom)); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> we entered here as coreset_freq_dom=%lx (bit %d) is 0, coreset_freq_domain is discontiguous\n",coreset_freq_dom,(46 - bitcnt_coreset_freq_dom)); +#endif } } //c_rb = c_rb + c_rb_tmp; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> c_rb=%d\n",c_rb); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> c_rb=%d\n",c_rb); +#endif // first we set initial conditions for pointer to rxdataF depending on the situation of the first RB within the CORESET (c_rb = n_BWP_start) if ((c_rb < (frame_parms->N_RB_DL >> 1)) && ((frame_parms->N_RB_DL & 1) == 0)) { //if RB to be treated is lower than middle system bandwidth then rxdataF pointed at (offset + c_br + symbol * ofdm_symbol_size): even case rxF = &rxdataF[aarx][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))]; - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in even case c_rb (%d) is lower than half N_RB_DL -> rxF = &rxdataF[aarx = (%d)][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", - c_rb,aarx,(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))); - #endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in even case c_rb (%d) is lower than half N_RB_DL -> rxF = &rxdataF[aarx = (%d)][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", + c_rb,aarx,(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))); +#endif } if ((c_rb >= (frame_parms->N_RB_DL >> 1)) && ((frame_parms->N_RB_DL & 1) == 0)) { // number of RBs is even and c_rb is higher than half system bandwidth (we don't skip DC) // if these conditions are true the pointer has to be situated at the 1st part of the rxdataF rxF = &rxdataF[aarx][(12*(c_rb - (frame_parms->N_RB_DL>>1)) + (symbol * (frame_parms->ofdm_symbol_size)))]; // we point at the 1st part of the rxdataF in symbol - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in even case c_rb (%d) is higher than half N_RB_DL (not DC) -> rxF = &rxdataF[aarx = (%d)][(12*(c_rb - (frame_parms->N_RB_DL>>1)) + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in even case c_rb (%d) is higher than half N_RB_DL (not DC) -> rxF = &rxdataF[aarx = (%d)][(12*(c_rb - (frame_parms->N_RB_DL>>1)) + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", c_rb,aarx,(12*(c_rb - (frame_parms->N_RB_DL>>1)) + (symbol * (frame_parms->ofdm_symbol_size)))); - #endif +#endif //rxF = &rxdataF[aarx][(1 + 12*(c_rb - (frame_parms->N_RB_DL>>1)) + (symbol * (frame_parms->ofdm_symbol_size)))]; // we point at the 1st part of the rxdataF in symbol //#ifdef NR_PDCCH_DCI_DEBUG // printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in even case c_rb (%d) is higher than half N_RB_DL (not DC) -> rxF = &rxdataF[aarx = (%d)][(1 + 12*(c_rb - (frame_parms->N_RB_DL>>1)) + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", @@ -523,41 +448,43 @@ void nr_pdcch_extract_rbs_single(int32_t **rxdataF, if ((c_rb < (frame_parms->N_RB_DL >> 1)) && ((frame_parms->N_RB_DL & 1) != 0)){ //if RB to be treated is lower than middle system bandwidth then rxdataF pointed at (offset + c_br + symbol * ofdm_symbol_size): odd case rxF = &rxdataF[aarx][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))]; - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is lower or equal than half N_RB_DL -> rxF = &rxdataF[aarx = (%d)][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", - c_rb,aarx,(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))); - #endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is lower or equal than half N_RB_DL -> rxF = &rxdataF[aarx = (%d)][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", + c_rb,aarx,(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))); +#endif } if ((c_rb > (frame_parms->N_RB_DL >> 1)) && ((frame_parms->N_RB_DL & 1) != 0)){ // number of RBs is odd and c_rb is higher than half system bandwidth + 1 // if these conditions are true the pointer has to be situated at the 1st part of the rxdataF just after the first IQ symbols of the RB containing DC rxF = &rxdataF[aarx][(12*(c_rb - (frame_parms->N_RB_DL>>1)) - 6 + (symbol * (frame_parms->ofdm_symbol_size)))]; // we point at the 1st part of the rxdataF in symbol - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is higher than half N_RB_DL (not DC) -> rxF = &rxdataF[aarx = (%d)][(12*(c_rb - frame_parms->N_RB_DL) - 5 + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", - c_rb,aarx,(12*(c_rb - (frame_parms->N_RB_DL>>1)) - 6 + (symbol * (frame_parms->ofdm_symbol_size)))); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is higher than half N_RB_DL (not DC) -> rxF = &rxdataF[aarx = (%d)][(12*(c_rb - frame_parms->N_RB_DL) - 5 + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", + c_rb,aarx,(12*(c_rb - (frame_parms->N_RB_DL>>1)) - 6 + (symbol * (frame_parms->ofdm_symbol_size)))); +#endif } if ((c_rb == (frame_parms->N_RB_DL >> 1)) && ((frame_parms->N_RB_DL & 1) != 0)){ // treatment of RB containing the DC // if odd number RBs in system bandwidth and first RB to be treated is higher than middle system bandwidth (around DC) // we have to treat the RB in two parts: first part from i=0 to 5, the data is at the end of rxdataF (pointing at the end of the table) rxF = &rxdataF[aarx][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))]; - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is half N_RB_DL + 1 we treat DC case -> rxF = &rxdataF[aarx = (%d)][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", - c_rb,aarx,(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))); - #endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is half N_RB_DL + 1 we treat DC case -> rxF = &rxdataF[aarx = (%d)][(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size))) = (%d)]\n", + c_rb,aarx,(frame_parms->first_carrier_offset + 12 * c_rb + (symbol * (frame_parms->ofdm_symbol_size)))); +#endif /*if (symbol_mod > 300) { // this if is going to be removed as DM-RS signals are present in all symbols of PDCCH for (i = 0; i < 6; i++) { - dl_ch0_ext[i] = dl_ch0[i]; - rxF_ext[i] = rxF[i]; + dl_ch0_ext[i] = dl_ch0[i]; + rxF_ext[i] = rxF[i]; } rxF = &rxdataF[aarx][(symbol * (frame_parms->ofdm_symbol_size))]; // we point at the 1st part of the rxdataF in symbol #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is half N_RB_DL +1 we treat DC case -> rxF = &rxdataF[aarx = (%d)][(symbol * (frame_parms->ofdm_symbol_size)) = (%d)]\n", - c_rb,aarx,(symbol * (frame_parms->ofdm_symbol_size))); + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is half N_RB_DL +1 we treat DC case -> rxF = &rxdataF[aarx = (%d)][(symbol * (frame_parms->ofdm_symbol_size)) = (%d)]\n", + c_rb,aarx,(symbol * (frame_parms->ofdm_symbol_size))); #endif for (; i < 12; i++) { - dl_ch0_ext[i] = dl_ch0[i]; - rxF_ext[i] = rxF[(1 + i - 6)]; + dl_ch0_ext[i] = dl_ch0[i]; + rxF_ext[i] = rxF[(1 + i - 6)]; } nb_rb++; dl_ch0_ext += 12; @@ -576,10 +503,10 @@ void nr_pdcch_extract_rbs_single(int32_t **rxdataF, } // then we point at the begining of the symbol part of rxdataF do process second part of RB rxF = &rxdataF[aarx][((symbol * (frame_parms->ofdm_symbol_size)))]; // we point at the 1st part of the rxdataF in symbol - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is half N_RB_DL +1 we treat DC case -> rxF = &rxdataF[aarx = (%d)][(symbol * (frame_parms->ofdm_symbol_size)) = (%d)]\n", - c_rb,aarx,(symbol * (frame_parms->ofdm_symbol_size))); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> in odd case c_rb (%d) is half N_RB_DL +1 we treat DC case -> rxF = &rxdataF[aarx = (%d)][(symbol * (frame_parms->ofdm_symbol_size)) = (%d)]\n", + c_rb,aarx,(symbol * (frame_parms->ofdm_symbol_size))); +#endif for (; i < 12; i++) { if ((i != 9)) { dl_ch0_ext[j] = dl_ch0[i]; @@ -591,48 +518,49 @@ void nr_pdcch_extract_rbs_single(int32_t **rxdataF, dl_ch0_ext += NBR_RE_PER_RB_WITHOUT_DMRS; rxF_ext += NBR_RE_PER_RB_WITHOUT_DMRS; dl_ch0 += 12; - //rxF += 7; - //c_rb++; - //n_BWP_start++; // We have to increment this variable here to be consequent in the for loop afterwards + //rxF += 7; + //c_rb++; + //n_BWP_start++; // We have to increment this variable here to be consequent in the for loop afterwards //} } else { // treatment of any RB that does not contain the DC /*if (symbol_mod > 300) { memcpy(dl_ch0_ext, dl_ch0, 12 * sizeof(int32_t)); for (i = 0; i < 12; i++) { - rxF_ext[i] = rxF[i]; + rxF_ext[i] = rxF[i]; } nb_rb++; dl_ch0_ext += 12; rxF_ext += 12; dl_ch0 += 12; //rxF += 12; - } else {*/ + } else {*/ j = 0; for (i = 0; i < 12; i++) { if ((i != 1) && (i != 5) && (i != 9)) { rxF_ext[j] = rxF[i]; - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> RB[c_rb %d] \t RE[re %d] => rxF_ext[%d]=(%d,%d)\t rxF[%d]=(%d,%d)", - c_rb, i, j, *(short *) &rxF_ext[j],*(1 + (short*) &rxF_ext[j]), i, - *(short *) &rxF[i], *(1 + (short*) &rxF[i])); - #endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> RB[c_rb %d] \t RE[re %d] => rxF_ext[%d]=(%d,%d)\t rxF[%d]=(%d,%d)\n", + c_rb, i, j, *(short *) &rxF_ext[j],*(1 + (short*) &rxF_ext[j]), i, + *(short *) &rxF[i], *(1 + (short*) &rxF[i])); +#endif dl_ch0_ext[j] = dl_ch0[i]; //printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> ch %d => dl_ch0(%d,%d)\n", i, *(short *) &dl_ch0[i], *(1 + (short*) &dl_ch0[i])); //printf("\t-> dl_ch0[%d] => dl_ch0_ext[%d](%d,%d)\n", i,j, *(short *) &dl_ch0[i], *(1 + (short*) &dl_ch0[i])); j++; } else { - #ifndef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> RB[c_rb %d] \t RE[re %d] => rxF_ext[%d]=(%d,%d)\t rxF[%d]=(%d,%d) \t\t <==> DM-RS PDCCH, this is a pilot symbol\n", - c_rb, i, j, *(short *) &rxF_ext[j], *(1 + (short*) &rxF_ext[j]), i, - *(short *) &rxF[i], *(1 + (short*) &rxF[i])); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_extract_rbs_single)-> RB[c_rb %d] \t RE[re %d] => rxF_ext[%d]=(%d,%d)\t rxF[%d]=(%d,%d) \t\t <==> DM-RS PDCCH, this is a pilot symbol\n", + c_rb, i, j, *(short *) &rxF_ext[j], *(1 + (short*) &rxF_ext[j]), i, + *(short *) &rxF[i], *(1 + (short*) &rxF[i])); +#endif } } nb_rb++; dl_ch0_ext += NBR_RE_PER_RB_WITHOUT_DMRS; rxF_ext += NBR_RE_PER_RB_WITHOUT_DMRS; dl_ch0 += 12; - //rxF += 12; + //rxF += 12; //} } } @@ -654,12 +582,12 @@ void nr_pdcch_channel_compensation(int32_t **rxdataF_ext, { uint16_t rb; //,nb_rb=20; - uint8_t aatx,aarx; + uint8_t aarx; #if defined(__x86_64__) || defined(__i386__) -__m128i mmtmpP0,mmtmpP1,mmtmpP2,mmtmpP3; + __m128i mmtmpP0,mmtmpP1,mmtmpP2,mmtmpP3; #elif defined(__arm__) -int16x8_t mmtmpP0,mmtmpP1,mmtmpP2,mmtmpP3; + int16x8_t mmtmpP0,mmtmpP1,mmtmpP2,mmtmpP3; #endif #if defined(__x86_64__) || defined(__i386__) @@ -668,99 +596,100 @@ int16x8_t mmtmpP0,mmtmpP1,mmtmpP2,mmtmpP3; #endif - for (aatx=0; aatxnb_antenna_ports_eNB;aatx++) - for (aarx=0; aarxnb_antennas_rx; aarx++) { - + for (aarx=0; aarxnb_antennas_rx; aarx++) { + #if defined(__x86_64__) || defined(__i386__) - dl_ch128 = (__m128i *)&dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*coreset_nbr_rb*12]; - rxdataF128 = (__m128i *)&rxdataF_ext[aarx][symbol*coreset_nbr_rb*12]; - rxdataF_comp128 = (__m128i *)&rxdataF_comp[(aatx<<1)+aarx][symbol*coreset_nbr_rb*12]; - //printf("ch compensation dl_ch ext addr %p \n", &dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*20*12]); - //printf("rxdataf ext addr %p symbol %d\n", &rxdataF_ext[aarx][symbol*20*12], symbol); - //printf("rxdataf_comp addr %p\n",&rxdataF_comp[(aatx<<1)+aarx][symbol*20*12]); + dl_ch128 = (__m128i *)&dl_ch_estimates_ext[aarx][symbol*coreset_nbr_rb*12]; + rxdataF128 = (__m128i *)&rxdataF_ext[aarx][symbol*coreset_nbr_rb*12]; + rxdataF_comp128 = (__m128i *)&rxdataF_comp[aarx][symbol*coreset_nbr_rb*12]; + //printf("ch compensation dl_ch ext addr %p \n", &dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*20*12]); + //printf("rxdataf ext addr %p symbol %d\n", &rxdataF_ext[aarx][symbol*20*12], symbol); + //printf("rxdataf_comp addr %p\n",&rxdataF_comp[(aatx<<1)+aarx][symbol*20*12]); + +#elif defined(__arm__) + // to be filled in +#endif + + for (rb=0; rb<(coreset_nbr_rb*3)>>2; rb++) { + //printf("rb %d\n",rb); +#if defined(__x86_64__) || defined(__i386__) + // multiply by conjugated channel + mmtmpP0 = _mm_madd_epi16(dl_ch128[0],rxdataF128[0]); + // print_ints("re",&mmtmpP0); + // mmtmpP0 contains real part of 4 consecutive outputs (32-bit) + mmtmpP1 = _mm_shufflelo_epi16(dl_ch128[0],_MM_SHUFFLE(2,3,0,1)); + mmtmpP1 = _mm_shufflehi_epi16(mmtmpP1,_MM_SHUFFLE(2,3,0,1)); + mmtmpP1 = _mm_sign_epi16(mmtmpP1,*(__m128i*)&conjugate[0]); + // print_ints("im",&mmtmpP1); + mmtmpP1 = _mm_madd_epi16(mmtmpP1,rxdataF128[0]); + // mmtmpP1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpP0 = _mm_srai_epi32(mmtmpP0,output_shift); + // print_ints("re(shift)",&mmtmpP0); + mmtmpP1 = _mm_srai_epi32(mmtmpP1,output_shift); + // print_ints("im(shift)",&mmtmpP1); + mmtmpP2 = _mm_unpacklo_epi32(mmtmpP0,mmtmpP1); + mmtmpP3 = _mm_unpackhi_epi32(mmtmpP0,mmtmpP1); + // print_ints("c0",&mmtmpP2); + // print_ints("c1",&mmtmpP3); + rxdataF_comp128[0] = _mm_packs_epi32(mmtmpP2,mmtmpP3); + //print_shorts("rx:",rxdataF128); + //print_shorts("ch:",dl_ch128); + //print_shorts("pack:",rxdataF_comp128); + + // multiply by conjugated channel + mmtmpP0 = _mm_madd_epi16(dl_ch128[1],rxdataF128[1]); + // mmtmpP0 contains real part of 4 consecutive outputs (32-bit) + mmtmpP1 = _mm_shufflelo_epi16(dl_ch128[1],_MM_SHUFFLE(2,3,0,1)); + mmtmpP1 = _mm_shufflehi_epi16(mmtmpP1,_MM_SHUFFLE(2,3,0,1)); + mmtmpP1 = _mm_sign_epi16(mmtmpP1,*(__m128i*)&conjugate[0]); + mmtmpP1 = _mm_madd_epi16(mmtmpP1,rxdataF128[1]); + // mmtmpP1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpP0 = _mm_srai_epi32(mmtmpP0,output_shift); + mmtmpP1 = _mm_srai_epi32(mmtmpP1,output_shift); + mmtmpP2 = _mm_unpacklo_epi32(mmtmpP0,mmtmpP1); + mmtmpP3 = _mm_unpackhi_epi32(mmtmpP0,mmtmpP1); + rxdataF_comp128[1] = _mm_packs_epi32(mmtmpP2,mmtmpP3); + //print_shorts("rx:",rxdataF128+1); + //print_shorts("ch:",dl_ch128+1); + //print_shorts("pack:",rxdataF_comp128+1); + + // multiply by conjugated channel + mmtmpP0 = _mm_madd_epi16(dl_ch128[2],rxdataF128[2]); + // mmtmpP0 contains real part of 4 consecutive outputs (32-bit) + mmtmpP1 = _mm_shufflelo_epi16(dl_ch128[2],_MM_SHUFFLE(2,3,0,1)); + mmtmpP1 = _mm_shufflehi_epi16(mmtmpP1,_MM_SHUFFLE(2,3,0,1)); + mmtmpP1 = _mm_sign_epi16(mmtmpP1,*(__m128i*)&conjugate[0]); + mmtmpP1 = _mm_madd_epi16(mmtmpP1,rxdataF128[2]); + // mmtmpP1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpP0 = _mm_srai_epi32(mmtmpP0,output_shift); + mmtmpP1 = _mm_srai_epi32(mmtmpP1,output_shift); + mmtmpP2 = _mm_unpacklo_epi32(mmtmpP0,mmtmpP1); + mmtmpP3 = _mm_unpackhi_epi32(mmtmpP0,mmtmpP1); + rxdataF_comp128[2] = _mm_packs_epi32(mmtmpP2,mmtmpP3); + /////////////////////////////////////////////////////////////////////////////////////////////// + //print_shorts("rx:",rxdataF128+2); + //print_shorts("ch:",dl_ch128+2); + //print_shorts("pack:",rxdataF_comp128+2); + +#ifdef NR_PDCCH_DCI_DEBUG + for (int i=0; i<12 ; i++) + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_channel_compensation)-> rxdataF128[%d]=(%d,%d) X dlch[%d]=(%d,%d) rxdataF_comp128[%d]=(%d,%d)\n", + (rb*12)+i, ((short *)rxdataF128)[i<<1],((short*)rxdataF128)[1+(i<<1)], + (rb*12)+i, ((short *)dl_ch128)[i<<1],((short*)dl_ch128)[1+(i<<1)], + (rb*12)+i, ((short *)rxdataF_comp128)[i<<1],((short*)rxdataF_comp128)[1+(i<<1)]); + +#endif + + dl_ch128+=3; + rxdataF128+=3; + rxdataF_comp128+=3; + #elif defined(__arm__) -// to be filled in + // to be filled in #endif - - for (rb=0; rb rb=%d rxdataF128[%d]=(%d,%d) rxdataF_comp128[%d]=(%d,%d)\n", - rb, i, *(short *) &rxdataF128[i],*(1 + (short*) &rxdataF128[i]), - i,*(short *) &rxdataF_comp128[i], *(1 + (short*) &rxdataF_comp128[i])); - -#endif - - dl_ch128+=3; - rxdataF128+=3; - rxdataF_comp128+=3; - -#elif defined(__arm__) -// to be filled in -#endif - } } + } #if defined(__x86_64__) || defined(__i386__) _mm_empty(); _m_empty(); @@ -773,32 +702,29 @@ void pdcch_detection_mrc(NR_DL_FRAME_PARMS *frame_parms, uint8_t symbol) { - uint8_t aatx; #if defined(__x86_64__) || defined(__i386__) __m128i *rxdataF_comp128_0,*rxdataF_comp128_1; #elif defined(__arm__) - int16x8_t *rxdataF_comp128_0,*rxdataF_comp128_1; + int16x8_t *rxdataF_comp128_0,*rxdataF_comp128_1; #endif int32_t i; if (frame_parms->nb_antennas_rx>1) { - for (aatx=0; aatxnb_antenna_ports_eNB; aatx++) { #if defined(__x86_64__) || defined(__i386__) - rxdataF_comp128_0 = (__m128i *)&rxdataF_comp[(aatx<<1)][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_1 = (__m128i *)&rxdataF_comp[(aatx<<1)+1][symbol*frame_parms->N_RB_DL*12]; + rxdataF_comp128_0 = (__m128i *)&rxdataF_comp[0][symbol*frame_parms->N_RB_DL*12]; + rxdataF_comp128_1 = (__m128i *)&rxdataF_comp[1][symbol*frame_parms->N_RB_DL*12]; #elif defined(__arm__) - rxdataF_comp128_0 = (int16x8_t *)&rxdataF_comp[(aatx<<1)][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128_1 = (int16x8_t *)&rxdataF_comp[(aatx<<1)+1][symbol*frame_parms->N_RB_DL*12]; + rxdataF_comp128_0 = (int16x8_t *)&rxdataF_comp[0][symbol*frame_parms->N_RB_DL*12]; + rxdataF_comp128_1 = (int16x8_t *)&rxdataF_comp[1][symbol*frame_parms->N_RB_DL*12]; #endif - // MRC on each re of rb - for (i=0; iN_RB_DL*3; i++) { + // MRC on each re of rb + for (i=0; iN_RB_DL*3; i++) { #if defined(__x86_64__) || defined(__i386__) - rxdataF_comp128_0[i] = _mm_adds_epi16(_mm_srai_epi16(rxdataF_comp128_0[i],1),_mm_srai_epi16(rxdataF_comp128_1[i],1)); + rxdataF_comp128_0[i] = _mm_adds_epi16(_mm_srai_epi16(rxdataF_comp128_0[i],1),_mm_srai_epi16(rxdataF_comp128_1[i],1)); #elif defined(__arm__) - rxdataF_comp128_0[i] = vhaddq_s16(rxdataF_comp128_0[i],rxdataF_comp128_1[i]); + rxdataF_comp128_0[i] = vhaddq_s16(rxdataF_comp128_0[i],rxdataF_comp128_1[i]); #endif - } } } @@ -831,7 +757,6 @@ void pdcch_siso(NR_DL_FRAME_PARMS *frame_parms, } -int32_t avgP[4]; @@ -848,10 +773,6 @@ int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, uint16_t symbol_mon, NR_SEARCHSPACE_TYPE_t searchSpaceType) { - #ifdef MU_RECEIVER - uint8_t eNB_id_i=eNB_id+1; //add 1 to eNB_id to separate from wanted signal, chosen as the B/F'd pilots from the SeNB are shifted by 1 -#endif - NR_UE_COMMON *common_vars = &ue->common_vars; NR_DL_FRAME_PARMS *frame_parms = &ue->frame_parms; NR_UE_PDCCH **pdcch_vars = ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]]; @@ -860,8 +781,9 @@ int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, int do_common; if (searchSpaceType == common) do_common=1; if (searchSpaceType == ue_specific) do_common=0; - uint8_t log2_maxh, aatx, aarx; + uint8_t log2_maxh, aarx; int32_t avgs; + int32_t avgP[4]; // number of RB (1 symbol) or REG (12 RE) in one CORESET: higher-layer parameter CORESET-freq-dom // (bit map 45 bits: each bit indicates 6 RB in CORESET -> 1 bit MSB indicates PRB 0..6 are part of CORESET) @@ -883,7 +805,7 @@ int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, // For each BWP the number of CORESETs is limited to 3 (including initial CORESET Id=0 -> ControlResourceSetId (0..maxNrofControlReourceSets-1) (0..12-1) //uint32_t n_BWP_start = 0; //uint32_t n_rb_offset = 0; - uint32_t n_rb_offset = pdcch_vars2->coreset[nb_coreset_active].rb_offset+43; //to be removed 43 + uint32_t n_rb_offset = pdcch_vars2->coreset[nb_coreset_active].rb_offset+(int)floor(frame_parms->ssb_start_subcarrier/NR_NB_SC_PER_RB); // start time position for CORESET // parameter symbol_mon is a 14 bits bitmap indicating monitoring symbols within a slot uint8_t start_symbol = 0; @@ -899,8 +821,8 @@ int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, #ifdef NR_PDCCH_DCI_DEBUG printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> symbol_mon=(%d) and start_symbol=(%d)\n",symbol_mon,start_symbol); - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_freq_dom=(%ld) n_rb_offset=(%d) coreset_time_dur=(%d) n_shift=(%d) reg_bundle_size_L=(%d) coreset_interleaver_size_R=(%d) \n", - coreset_freq_dom,n_rb_offset,coreset_time_dur,n_shift,reg_bundle_size_L,coreset_interleaver_size_R); + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_freq_dom=(%ld) n_rb_offset=(%d) coreset_time_dur=(%d) n_shift=(%d) reg_bundle_size_L=(%d) coreset_interleaver_size_R=(%d) scrambling_ID=(%d) \n", + coreset_freq_dom,n_rb_offset,coreset_time_dur,n_shift,reg_bundle_size_L,coreset_interleaver_size_R,pdcch_DMRS_scrambling_id); #endif // @@ -922,34 +844,34 @@ int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, // for (int j=0; j < coreset_nbr_act; j++) { // for each active CORESET (max number of active CORESETs in a BWP is 3), // we calculate the number of RB for each CORESET bitmap - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_freq_dom=(%ld)\n",coreset_freq_dom); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_freq_dom=(%ld)\n",coreset_freq_dom); +#endif int i; //for each bit in the coreset_freq_dom bitmap for (i = 0; i < 45; i++) { // this loop counts each bit of the bit map coreset_freq_dom, and increments nbr_RB_coreset for each bit set to '1' if (((coreset_freq_dom & 0x1FFFFFFFFFFF) >> i) & 0x1) coreset_nbr_rb++; } coreset_nbr_rb = 6 * coreset_nbr_rb; // coreset_nbr_rb has to be multiplied by 6 to indicate the number of PRB or REG(=12 RE) within the CORESET - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_freq_dom=(%ld,%lx), coreset_nbr_rb=%d\n", coreset_freq_dom,coreset_freq_dom,coreset_nbr_rb); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_freq_dom=(%ld,%lx), coreset_nbr_rb=%d\n", coreset_freq_dom,coreset_freq_dom,coreset_nbr_rb); +#endif coreset_nbr_reg = coreset_time_dur * coreset_nbr_rb; coreset_C = (uint32_t)(coreset_nbr_reg / (reg_bundle_size_L * coreset_interleaver_size_R)); - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_nbr_rb=%d, coreset_nbr_reg=%d, coreset_C=(%d/(%d*%d))=%d\n", - coreset_nbr_rb, coreset_nbr_reg, coreset_nbr_reg, reg_bundle_size_L,coreset_interleaver_size_R, coreset_C); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> coreset_nbr_rb=%d, coreset_nbr_reg=%d, coreset_C=(%d/(%d*%d))=%d\n", + coreset_nbr_rb, coreset_nbr_reg, coreset_nbr_reg, reg_bundle_size_L,coreset_interleaver_size_R, coreset_C); +#endif for (int s = start_symbol; s < (start_symbol + coreset_time_dur); s++) { - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter process pdcch ofdm symbol s=%d where coreset_time_dur=%d\n",s,coreset_time_dur); - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_extract_rbs_single(is_secondary_ue=%d) to remove DM-RS PDCCH\n", - is_secondary_ue); - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in nr_pdcch_extract_rbs_single(rxdataF -> rxdataF_ext || dl_ch_estimates -> dl_ch_estimates_ext)\n"); - #endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_extract_rbs_single(is_secondary_ue=%d) to remove DM-RS PDCCH\n", + is_secondary_ue); + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in nr_pdcch_extract_rbs_single(rxdataF -> rxdataF_ext || dl_ch_estimates -> dl_ch_estimates_ext)\n"); +#endif nr_pdcch_extract_rbs_single(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id], pdcch_vars[eNB_id]->rxdataF_ext, @@ -961,37 +883,36 @@ int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, coreset_nbr_rb, n_rb_offset); - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter pdcch_channel_level(avgP=%d) => compute channel level based on ofdm symbol 0, pdcch_vars[eNB_id]->dl_ch_estimates_ext\n",avgP); - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in pdcch_channel_level(dl_ch_estimates_ext -> dl_ch_estimates_ext)\n"); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter pdcch_channel_level(avgP=%d) => compute channel level based on ofdm symbol 0, pdcch_vars[eNB_id]->dl_ch_estimates_ext\n",avgP); + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in pdcch_channel_level(dl_ch_estimates_ext -> dl_ch_estimates_ext)\n"); +#endif // compute channel level based on ofdm symbol 0 pdcch_channel_level(pdcch_vars[eNB_id]->dl_ch_estimates_ext, frame_parms, avgP, coreset_nbr_rb); avgs = 0; - for (aatx = 0; aatx < frame_parms->nb_antenna_ports_eNB; aatx++) - for (aarx = 0; aarx < frame_parms->nb_antennas_rx; aarx++) - avgs = cmax(avgs, avgP[(aarx << 1) + aatx]); + for (aarx = 0; aarx < frame_parms->nb_antennas_rx; aarx++) + avgs = cmax(avgs, avgP[aarx]); log2_maxh = (log2_approx(avgs) / 2) + 5; //+frame_parms->nb_antennas_rx; #ifdef UE_DEBUG_TRACE -LOG_D(PHY,"nr_tti_rx %d: pdcch log2_maxh = %d (%d,%d)\n",nr_tti_rx,log2_maxh,avgP[0],avgs); + LOG_D(PHY,"nr_tti_rx %d: pdcch log2_maxh = %d (%d,%d)\n",nr_tti_rx,log2_maxh,avgP[0],avgs); #endif #if T_TRACER -T(T_UE_PHY_PDCCH_ENERGY, T_INT(eNB_id), T_INT(0), T_INT(frame%1024), T_INT(nr_tti_rx), - T_INT(avgP[0]), T_INT(avgP[1]), T_INT(avgP[2]), T_INT(avgP[3])); + T(T_UE_PHY_PDCCH_ENERGY, T_INT(eNB_id), T_INT(0), T_INT(frame%1024), T_INT(nr_tti_rx), + T_INT(avgP[0]), T_INT(avgP[1]), T_INT(avgP[2]), T_INT(avgP[3])); +#endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_channel_compensation(log2_maxh=%d)\n",log2_maxh); + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in nr_pdcch_channel_compensation(rxdataF_ext x dl_ch_estimates_ext -> rxdataF_comp)\n"); #endif - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_channel_compensation(log2_maxh=%d)\n",log2_maxh); - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in nr_pdcch_channel_compensation(rxdataF_ext x dl_ch_estimates_ext -> rxdataF_comp)\n"); - #endif // compute LLRs for ofdm symbol 0 only nr_pdcch_channel_compensation(pdcch_vars[eNB_id]->rxdataF_ext, pdcch_vars[eNB_id]->dl_ch_estimates_ext, pdcch_vars[eNB_id]->rxdataF_comp, - (aatx > 1) ? pdcch_vars[eNB_id]->rho : NULL, + NULL, frame_parms, s, log2_maxh, @@ -1000,134 +921,72 @@ T(T_UE_PHY_PDCCH_ENERGY, T_INT(eNB_id), T_INT(0), T_INT(frame%1024), T_INT(nr_tt #ifdef DEBUG_PHY - if (nr_tti_rx==5) - write_output("rxF_comp_d.m","rxF_c_d",&pdcch_vars[eNB_id]->rxdataF_comp[0][s*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); + if (nr_tti_rx==5) + write_output("rxF_comp_d.m","rxF_c_d",&pdcch_vars[eNB_id]->rxdataF_comp[0][s*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); #endif -#ifdef MU_RECEIVER - - if (is_secondary_ue) { - //get MF output for interfering stream - pdcch_channel_compensation(pdcch_vars[eNB_id_i]->rxdataF_ext, - pdcch_vars[eNB_id_i]->dl_ch_estimates_ext, - pdcch_vars[eNB_id_i]->rxdataF_comp, - (aatx>1) ? pdcch_vars[eNB_id_i]->rho : NULL, - frame_parms, - 0, - log2_maxh);// log2_maxh+I0_shift -#ifdef DEBUG_PHY - write_output("rxF_comp_i.m","rxF_c_i",&pdcch_vars[eNB_id_i]->rxdataF_comp[0][s*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); -#endif - pdcch_dual_stream_correlation(frame_parms, - 0, - pdcch_vars[eNB_id]->dl_ch_estimates_ext, - pdcch_vars[eNB_id_i]->dl_ch_estimates_ext, - pdcch_vars[eNB_id]->dl_ch_rho_ext, - log2_maxh); - } - -#endif //MU_RECEIVER - if (frame_parms->nb_antennas_rx > 1) { -#ifdef MU_RECEIVER - - if (is_secondary_ue) { - pdcch_detection_mrc_i(frame_parms, - pdcch_vars[eNB_id]->rxdataF_comp, - pdcch_vars[eNB_id_i]->rxdataF_comp, - pdcch_vars[eNB_id]->rho, - pdcch_vars[eNB_id]->dl_ch_rho_ext, - 0); -#ifdef DEBUG_PHY - write_output("rxF_comp_d.m","rxF_c_d",&pdcch_vars[eNB_id]->rxdataF_comp[0][s*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); - write_output("rxF_comp_i.m","rxF_c_i",&pdcch_vars[eNB_id_i]->rxdataF_comp[0][s*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter pdcch_detection_mrc(frame_parms->nb_antennas_rx=%d)\n", + frame_parms->nb_antennas_rx); #endif - } else -#endif //MU_RECEIVER - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter pdcch_detection_mrc(frame_parms->nb_antennas_rx=%d)\n", - frame_parms->nb_antennas_rx); - #endif pdcch_detection_mrc(frame_parms, pdcch_vars[eNB_id]->rxdataF_comp,s); } -#ifdef MU_RECEIVER - - if (is_secondary_ue) { - pdcch_qpsk_qpsk_llr(frame_parms, - pdcch_vars[eNB_id]->rxdataF_comp, - pdcch_vars[eNB_id_i]->rxdataF_comp, - pdcch_vars[eNB_id]->dl_ch_rho_ext, - pdcch_vars[eNB_id]->llr16, //subsequent function require 16 bit llr, but output must be 8 bit (actually clipped to 4, because of the Viterbi decoder) - pdcch_vars[eNB_id]->llr, - 0); - /* - #ifdef DEBUG_PHY - if (subframe==5) { - write_output("llr8_seq.m","llr8",&pdcch_vars[eNB_id]->llr[s*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,4); - write_output("llr16_seq.m","llr16",&pdcch_vars[eNB_id]->llr16[s*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,4); - } - #endif*/ - } else { -#endif //MU_RECEIVER - - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_llr(for symbol %d), pdcch_vars[eNB_id]->rxdataF_comp ---> pdcch_vars[eNB_id]->llr \n",s); - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in nr_pdcch_llr(rxdataF_comp -> llr)\n"); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_llr(for symbol %d), pdcch_vars[eNB_id]->rxdataF_comp ---> pdcch_vars[eNB_id]->llr \n",s); + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> in nr_pdcch_llr(rxdataF_comp -> llr)\n"); +#endif nr_pdcch_llr(frame_parms, - pdcch_vars[eNB_id]->rxdataF_comp, - (char *) pdcch_vars[eNB_id]->llr, - s, - coreset_nbr_rb); - -#ifdef MU_RECEIVER -} -#endif //MU_RECEIVER + pdcch_vars[eNB_id]->rxdataF_comp, + pdcch_vars[eNB_id]->llr, + s, + coreset_nbr_rb); + #if T_TRACER -/* - T(T_UE_PHY_PDCCH_IQ, T_INT(frame_parms->N_RB_DL), T_INT(frame_parms->N_RB_DL), - T_INT(n_pdcch_symbols), - T_BUFFER(pdcch_vars[eNB_id]->rxdataF_comp, frame_parms->N_RB_DL*12*n_pdcch_symbols* 4)); -*/ + /* + T(T_UE_PHY_PDCCH_IQ, T_INT(frame_parms->N_RB_DL), T_INT(frame_parms->N_RB_DL), + T_INT(n_pdcch_symbols), + T_BUFFER(pdcch_vars[eNB_id]->rxdataF_comp, frame_parms->N_RB_DL*12*n_pdcch_symbols* 4)); + */ #endif #ifdef DEBUG_DCI_DECODING - printf("demapping: nr_tti_rx %d, mi %d, tdd_config %d\n",nr_tti_rx,get_mi(frame_parms,nr_tti_rx),frame_parms->tdd_config); + printf("demapping: nr_tti_rx %d, mi %d, tdd_config %d\n",nr_tti_rx,get_mi(frame_parms,nr_tti_rx),frame_parms->tdd_config); #endif } - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_demapping_deinterleaving()\n"); - #endif - -nr_pdcch_demapping_deinterleaving(pdcch_vars[eNB_id]->llr, - (uint16_t*) pdcch_vars[eNB_id]->e_rx, - frame_parms, - coreset_time_dur, - coreset_nbr_rb, - reg_bundle_size_L, - coreset_interleaver_size_R, - n_shift); - - +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we enter nr_pdcch_demapping_deinterleaving()\n"); +#endif + + nr_pdcch_demapping_deinterleaving((uint32_t*) pdcch_vars[eNB_id]->llr, + (uint32_t*) pdcch_vars[eNB_id]->e_rx, + frame_parms, + coreset_time_dur, + coreset_nbr_rb, + reg_bundle_size_L, + coreset_interleaver_size_R, + n_shift); + + nr_pdcch_unscrambling(pdcch_vars[eNB_id]->crnti, - frame_parms, - nr_tti_rx, - (uint16_t*) pdcch_vars[eNB_id]->e_rx, - coreset_time_dur*coreset_nbr_rb*9*2, - // get_nCCE(n_pdcch_symbols, frame_parms, mi) * 72, - pdcch_DMRS_scrambling_id, - do_common); - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we end nr_pdcch_unscrambling()\n"); - #endif - - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> Ending nr_rx_pdcch() function\n"); - #endif + frame_parms, + nr_tti_rx, + pdcch_vars[eNB_id]->e_rx, + coreset_time_dur*coreset_nbr_rb*9*2, + // get_nCCE(n_pdcch_symbols, frame_parms, mi) * 72, + pdcch_DMRS_scrambling_id, + do_common); +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> we end nr_pdcch_unscrambling()\n"); +#endif + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_rx_pdcch)-> Ending nr_rx_pdcch() function\n"); +#endif return (0); } #endif @@ -1135,26 +994,25 @@ nr_pdcch_demapping_deinterleaving(pdcch_vars[eNB_id]->llr, void pdcch_scrambling(NR_DL_FRAME_PARMS *frame_parms, - uint8_t nr_tti_rx, - uint8_t *e, - uint32_t length) -{ + uint8_t nr_tti_rx, + uint8_t *e, + uint32_t length) { int i; uint8_t reset; uint32_t x1, x2, s=0; - + reset = 1; // x1 is set in lte_gold_generic - + x2 = (nr_tti_rx<<9) + frame_parms->Nid_cell; //this is c_init in 36.211 Sec 6.8.2 - + for (i=0; i>(i&0x1f))&1)); if (e[i] != 2) // element is 2 e[i] = (e[i]&1) ^ ((s>>(i&0x1f))&1); @@ -1165,58 +1023,69 @@ void pdcch_scrambling(NR_DL_FRAME_PARMS *frame_parms, #ifdef NR_PDCCH_DCI_RUN void nr_pdcch_unscrambling(uint16_t crnti, NR_DL_FRAME_PARMS *frame_parms, uint8_t nr_tti_rx, - uint16_t *z, uint32_t length, uint16_t pdcch_DMRS_scrambling_id, int do_common) { - - int i; - uint8_t reset; - uint32_t x1, x2, s = 0; + int16_t *z, uint32_t length, uint16_t pdcch_DMRS_scrambling_id, int do_common) { + + int i; + uint8_t reset; + uint32_t x1, x2, s = 0; uint16_t n_id; //{0,1,...,65535} uint32_t n_rnti; - - reset = 1; - // x1 is set in first call to lte_gold_generic - //do_common=1; -if (do_common){ - n_id = frame_parms->Nid_cell; - n_rnti = 0; -} else { - n_id = pdcch_DMRS_scrambling_id; - n_rnti = (uint32_t)crnti; -} -//x2 = ((n_rnti * (1 << 16)) + n_id)%(1 << 31); -//uint32_t puissance_2_16 = ((1<<16)*n_rnti)+n_id; -//uint32_t puissance_2_31= (1<<30)*2; -//uint32_t calc_x2=puissance_2_16%puissance_2_31; - x2 = (((1<<16)*n_rnti)+n_id); //mod 2^31 is implicit //this is c_init in 38.211 v15.1.0 Section 7.3.2.3 -// x2 = (nr_tti_rx << 9) + frame_parms->Nid_cell; //this is c_init in 36.211 Sec 6.8.2 -#ifndef NR_PDCCH_DCI_DEBUG -printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> (c_init=%d, n_id=%d, n_rnti=%d, length=%d)\n",x2,n_id,n_rnti,length); + + reset = 1; + // x1 is set in first call to lte_gold_generic + //do_common=1; + if (do_common){ + n_id = frame_parms->Nid_cell; + n_rnti = 0; + } else { + n_id = pdcch_DMRS_scrambling_id; + n_rnti = (uint32_t)crnti; + } + //x2 = ((n_rnti * (1 << 16)) + n_id)%(1 << 31); + //uint32_t puissance_2_16 = ((1<<16)*n_rnti)+n_id; + //uint32_t puissance_2_31= (1<<30)*2; + //uint32_t calc_x2=puissance_2_16%puissance_2_31; + x2 = (((1<<16)*n_rnti)+n_id); //mod 2^31 is implicit //this is c_init in 38.211 v15.1.0 Section 7.3.2.3 + // x2 = (nr_tti_rx << 9) + frame_parms->Nid_cell; //this is c_init in 36.211 Sec 6.8.2 +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> (c_init=%d, n_id=%d, n_rnti=%d, length=%d)\n",x2,n_id,n_rnti,length); #endif - for (i = 0; i < length; i++) { - if ((i & 0x1f) == 0) { - s = lte_gold_generic(&x1, &x2, reset); - //printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> lte_gold[%d]=%x\n",i,s); - reset = 0; - } + for (i = 0; i < length; i++) { + if ((i & 0x1f) == 0) { + s = lte_gold_generic(&x1, &x2, reset); + //printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> lte_gold[%d]=%x\n",i,s); + reset = 0; + } - #ifndef NR_PDCCH_DCI_DEBUG - if (i%2 == 0) printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> unscrambling %d : scrambled_z=%d, => ", - i,*(char*) &z[(int)floor(i/2)]); - if (i%2 == 1) printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> unscrambling %d : scrambled_z=%d, => ", - i,*(1 + (char*) &z[(int)floor(i/2)])); - #endif - if (((s >> (i % 32)) & 1) == 1){ - if (i%2 == 0) *(char*) &z[(int)floor(i/2)] = -(*(char*) &z[(int)floor(i/2)]); - if (i%2 == 1) *(1 + (char*) &z[(int)floor(i/2)]) = -(*(1 + (char*) &z[(int)floor(i/2)])); - } - //llr[i] = -llr[i]; - //llr[i] = (-1)*llr[i]; - #ifndef NR_PDCCH_DCI_DEBUG - if (i%2 == 0) printf("unscrambled_z=%d\n",*(char*) &z[(int)floor(i/2)]); - if (i%2 == 1) printf("unscrambled_z=%d\n",*(1 + (char*) &z[(int)floor(i/2)])); - #endif - } + /* +#ifdef NR_PDCCH_DCI_DEBUG + if (i%2 == 0) printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> unscrambling %d : scrambled_z=%d, => ", + i,*(char*) &z[(int)floor(i/2)]); + if (i%2 == 1) printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> unscrambling %d : scrambled_z=%d, => ", + i,*(1 + (char*) &z[(int)floor(i/2)])); +#endif + if (((s >> (i % 32)) & 1) == 1){ + if (i%2 == 0) *(char*) &z[(int)floor(i/2)] = -(*(char*) &z[(int)floor(i/2)]); + if (i%2 == 1) *(1 + (char*) &z[(int)floor(i/2)]) = -(*(1 + (char*) &z[(int)floor(i/2)])); + } + //llr[i] = -llr[i]; + //llr[i] = (-1)*llr[i]; +#ifdef NR_PDCCH_DCI_DEBUG + if (i%2 == 0) printf("unscrambled_z=%d\n",*(char*) &z[(int)floor(i/2)]); + if (i%2 == 1) printf("unscrambled_z=%d\n",*(1 + (char*) &z[(int)floor(i/2)])); +#endif + */ +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> unscrambling %d : scrambled_z=%d, => ", + i,z[i]); +#endif + if (((s >> (i % 32)) & 1) == 1) z[i] = -z[i]; +#ifdef NR_PDCCH_DCI_DEBUG + printf("unscrambled_z=%d\n",z[i]); +#endif + } + } #endif @@ -1224,65 +1093,66 @@ printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_pdcch_unscrambling)-> (c_init=%d, n_id=%d, #ifdef NR_PDCCH_DCI_RUN void nr_dci_decoding_procedure0(int s, - int p, - int coreset_time_dur, - uint16_t coreset_nbr_rb, - NR_UE_PDCCH **pdcch_vars, - int do_common, - uint8_t nr_tti_rx, - NR_DCI_ALLOC_t *dci_alloc, - int16_t eNB_id, - uint8_t current_thread_id, - NR_DL_FRAME_PARMS *frame_parms, - t_nrPolar_paramsPtr *nrPolar_params, - //uint8_t mi, - uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES], - uint8_t L, - NR_UE_SEARCHSPACE_CSS_DCI_FORMAT_t format_css, - NR_UE_SEARCHSPACE_USS_DCI_FORMAT_t format_uss, - uint8_t sizeof_bits, - uint8_t sizeof_bytes, - uint8_t *dci_cnt, - crc_scrambled_t *crc_scrambled, - format_found_t *format_found, - uint16_t pdcch_DMRS_scrambling_id, - uint32_t *CCEmap0, - uint32_t *CCEmap1, - uint32_t *CCEmap2) { - + int p, + int coreset_time_dur, + uint16_t coreset_nbr_rb, + NR_UE_PDCCH **pdcch_vars, + int do_common, + uint8_t nr_tti_rx, + NR_DCI_ALLOC_t *dci_alloc, + int16_t eNB_id, + uint8_t current_thread_id, + NR_DL_FRAME_PARMS *frame_parms, + //uint8_t mi, + uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES], + uint8_t L, + NR_UE_SEARCHSPACE_CSS_DCI_FORMAT_t format_css, + NR_UE_SEARCHSPACE_USS_DCI_FORMAT_t format_uss, + uint8_t sizeof_bits, + uint8_t sizeof_bytes, + uint8_t *dci_cnt, + crc_scrambled_t *crc_scrambled, + format_found_t *format_found, + uint16_t pdcch_DMRS_scrambling_id, + uint32_t *CCEmap0, + uint32_t *CCEmap1, + uint32_t *CCEmap2) { + uint16_t crc, CCEind, nCCE[3]; uint32_t *CCEmap = NULL, CCEmap_mask = 0; uint8_t L2 = (1 << L); unsigned int Yk, nb_candidates = 0, i, m; unsigned int CCEmap_cand; - double *polar_input = malloc (sizeof(double) * 108*L2); - int8_t decoderState=0; + int8_t decoderState=0; + // A[p], p is the current active CORESET uint16_t A[3]={39827,39829,39839}; //Table 10.1-2: Maximum number of PDCCH candidates per slot and per serving cell as a function of the subcarrier spacing value 2^mu*15 KHz, mu {0,1,2,3} uint8_t m_max_slot_pdcch_Table10_1_2 [4] = {44,36,22,20}; //Table 10.1-3: Maximum number of non-overlapped CCEs per slot and per serving cell as a function of the subcarrier spacing value 2^mu*15 KHz, mu {0,1,2,3} - uint8_t cce_max_slot_pdcch_Table10_1_3 [4] = {56,56,48,32}; - + //uint8_t cce_max_slot_pdcch_Table10_1_3 [4] = {56,56,48,32}; + int coreset_nbr_cce_per_symbol=0; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> format_found is %d \n", *format_found); - #endif - + t_nrPolar_paramsPtr nrPolar_params = pdcch_vars[eNB_id]->nrPolar_params; + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> format_found is %d \n", *format_found); +#endif + //if (mode == NO_DCI) { // #ifdef NR_PDCCH_DCI_DEBUG // printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> skip DCI decoding: expect no DCIs at nr_tti_rx %d in current searchSpace\n", nr_tti_rx); // #endif // return; //} - - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> frequencyDomainResources=%lx, duration=%d\n", - pdcch_vars[eNB_id]->coreset[p].frequencyDomainResources, pdcch_vars[eNB_id]->coreset[p].duration); - #endif - + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> frequencyDomainResources=%lx, duration=%d\n", + pdcch_vars[eNB_id]->coreset[p].frequencyDomainResources, pdcch_vars[eNB_id]->coreset[p].duration); +#endif + // nCCE = get_nCCE(pdcch_vars[eNB_id]->num_pdcch_symbols, frame_parms, mi); for (int i = 0; i < 45; i++) { // this loop counts each bit of the bit map coreset_freq_dom, and increments nbr_RB_coreset for each bit set to '1' @@ -1294,59 +1164,59 @@ void nr_dci_decoding_procedure0(int s, // the number of symbols in the CORESET (pdcch_vars[eNB_id]->coreset[p].duration) // multiplied by the number of bits set to '1' in the frequencyDomainResources bitmap // (1 bit set to '1' corresponds to 6 RB and 1 CCE = 6 RB) - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> nCCE[%d]=%d\n",p,nCCE[p]); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> nCCE[%d]=%d\n",p,nCCE[p]); +#endif -/* if (nCCE > get_nCCE(3, frame_parms, 1)) { - LOG_D(PHY, - "skip DCI decoding: nCCE=%d > get_nCCE(3,frame_parms,1)=%d\n", - nCCE, get_nCCE(3, frame_parms, 1)); - return; + /* if (nCCE > get_nCCE(3, frame_parms, 1)) { + LOG_D(PHY, + "skip DCI decoding: nCCE=%d > get_nCCE(3,frame_parms,1)=%d\n", + nCCE, get_nCCE(3, frame_parms, 1)); + return; } - + if (nCCE < L2) { - LOG_D(PHY, "skip DCI decoding: nCCE=%d < L2=%d\n", nCCE, L2); - return; + LOG_D(PHY, "skip DCI decoding: nCCE=%d < L2=%d\n", nCCE, L2); + return; } - + if (mode == NO_DCI) { - LOG_D(PHY, "skip DCI decoding: expect no DCIs at nr_tti_rx %d\n", - nr_tti_rx); - return; + LOG_D(PHY, "skip DCI decoding: expect no DCIs at nr_tti_rx %d\n", + nr_tti_rx); + return; } -*/ + */ if (do_common == 1) { Yk = 0; if (pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.common_dci_formats == cformat2_0) { - // for dci_format_2_0, the nb_candidates is obtained from a different variable - switch (L2) { - case 1: - nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel1; - break; - case 2: - nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel2; - break; - case 4: - nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel4; - break; - case 8: - nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel8; - break; - case 16: - nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel16; - break; - default: - break; - } + // for dci_format_2_0, the nb_candidates is obtained from a different variable + switch (L2) { + case 1: + nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel1; + break; + case 2: + nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel2; + break; + case 4: + nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel4; + break; + case 8: + nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel8; + break; + case 16: + nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.sfi_nrofCandidates_aggrlevel16; + break; + default: + break; + } } else if (pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.common_dci_formats == cformat2_3) { - // for dci_format_2_3, the nb_candidates is obtained from a different variable - nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.srs_nrofCandidates; + // for dci_format_2_3, the nb_candidates is obtained from a different variable + nb_candidates = pdcch_vars[eNB_id]->searchSpace[s].searchSpaceType.srs_nrofCandidates; } else { nb_candidates = (L2 == 4) ? 4 : ((L2 == 8)? 2 : 1); // according to Table 10.1-1 (38.213 section 10.1) - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> we are in common searchSpace and nb_candidates=%d for L2=%d\n",nb_candidates,L2); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> we are in common searchSpace and nb_candidates=%d for L2=%d\n",nb_candidates,L2); +#endif } } else { switch (L2) { @@ -1368,7 +1238,7 @@ void nr_dci_decoding_procedure0(int s, default: break; } - + // Find first available in ue specific search space // according to procedure in Section 10.1 of 38.213 // compute Yk @@ -1376,358 +1246,353 @@ void nr_dci_decoding_procedure0(int s, for (i = 0; i <= nr_tti_rx; i++) Yk = (Yk * A[p%3]) % 65537; } - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> L2(%d) | nCCE[%d](%d) | Yk(%d) | nb_candidates(%d)\n",L2,p,nCCE[p],Yk,nb_candidates); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> L2(%d) | nCCE[%d](%d) | Yk(%d) | nb_candidates(%d)\n",L2,p,nCCE[p],Yk,nb_candidates); +#endif /* for (CCEind=0; - CCEind nCCE[p]) -// nb_candidates = nCCE[p] / L2; - -// In the next code line there is maybe a bug. The spec is not comparing Table 10.1-2 with nb_candidates, but with total number of candidates for all s and all p + CCEind nCCE[p]) + // nb_candidates = nCCE[p] / L2; + + // In the next code line there is maybe a bug. The spec is not comparing Table 10.1-2 with nb_candidates, but with total number of candidates for all s and all p int m_p_s_L_max = (m_max_slot_pdcch_Table10_1_2[1]<=nb_candidates ? m_max_slot_pdcch_Table10_1_2[1] : nb_candidates); if (L==4) m_p_s_L_max=1; // Table 10.1-2 is not defined for L=4 - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> m_max_slot_pdcch_Table10_1_2(%d)=%d\n",L,m_max_slot_pdcch_Table10_1_2[L]); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> m_max_slot_pdcch_Table10_1_2(%d)=%d\n",L,m_max_slot_pdcch_Table10_1_2[L]); +#endif for (m = 0; m < nb_candidates; m++) { int n_ci = 0; if (nCCE[p] < L2) return; + +#ifdef NR_PDCCH_DCI_DEBUG int debug1 = nCCE[p] / L2; int debug2 = L2*m_p_s_L_max; - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> debug1(%d)=nCCE[p]/L2 | nCCE[%d](%d) | L2(%d)\n",debug1,p,nCCE[p],L2); - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> debug2(%d)=L2*m_p_s_L_max | L2(%d) | m_p_s_L_max(%d)\n",debug2,L2,m_p_s_L_max); - #endif + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> debug1(%d)=nCCE[p]/L2 | nCCE[%d](%d) | L2(%d)\n",debug1,p,nCCE[p],L2); + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> debug2(%d)=L2*m_p_s_L_max | L2(%d) | m_p_s_L_max(%d)\n",debug2,L2,m_p_s_L_max); +#endif CCEind = (((Yk + (uint16_t)(floor((m*nCCE[p])/(L2*m_p_s_L_max))) + n_ci) % (uint16_t)(floor(nCCE[p] / L2))) * L2); - #ifdef NR_PDCCH_DCI_DEBUG - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> CCEind(%d) = (((Yk(%d) + ((m(%d)*nCCE[p](%d))/(L2(%d)*m_p_s_L_max(%d)))) % (nCCE[p] / L2)) * L2)\n", - CCEind,Yk,m,nCCE[p],L2,m_p_s_L_max); - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> n_candidate(m)=%d | CCEind=%d |",m,CCEind); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> CCEind(%d) = (((Yk(%d) + ((m(%d)*nCCE[p](%d))/(L2(%d)*m_p_s_L_max(%d)))) % (nCCE[p] / L2)) * L2)\n", + CCEind,Yk,m,nCCE[p],L2,m_p_s_L_max); + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> n_candidate(m)=%d | CCEind=%d |",m,CCEind); +#endif if (CCEind < 32) CCEmap = CCEmap0; else if (CCEind < 64) CCEmap = CCEmap1; else if (CCEind < 96) CCEmap = CCEmap2; - else { - LOG_E(PHY, "Illegal CCEind %d (Yk %d, m %d, nCCE %d, L2 %d\n",CCEind, Yk, m, nCCE, L2); - //mac_xface->macphy_exit("Illegal CCEind\n"); - return; // not reached - } - + else AssertFatal(1==0,"Illegal CCEind %d (Yk %d, m %d, nCCE %d, L2 %d\n",CCEind, Yk, m, nCCE[p], L2); + switch (L2) { - case 1: - CCEmap_mask = (1 << (CCEind & 0x1f)); - break; - case 2: - CCEmap_mask = (3 << (CCEind & 0x1f)); - break; - case 4: - CCEmap_mask = (0xf << (CCEind & 0x1f)); - break; - case 8: - CCEmap_mask = (0xff << (CCEind & 0x1f)); - break; - case 16: - CCEmap_mask = (0xfff << (CCEind & 0x1f)); - break; - default: - LOG_E(PHY, "Illegal L2 value %d\n", L2); - //mac_xface->macphy_exit("Illegal L2\n"); - return; // not reached + case 1: + CCEmap_mask = (1 << (CCEind & 0x1f)); + break; + case 2: + CCEmap_mask = (3 << (CCEind & 0x1f)); + break; + case 4: + CCEmap_mask = (0xf << (CCEind & 0x1f)); + break; + case 8: + CCEmap_mask = (0xff << (CCEind & 0x1f)); + break; + case 16: + CCEmap_mask = (0xfff << (CCEind & 0x1f)); + break; + default: + LOG_E(PHY, "Illegal L2 value %d\n", L2); + //mac_xface->macphy_exit("Illegal L2\n"); + return; // not reached } CCEmap_cand = (*CCEmap) & CCEmap_mask; // CCE is not allocated yet - #ifdef NR_PDCCH_DCI_DEBUG - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> CCEmap_cand=%d \n",CCEmap_cand); - #endif - +#ifdef NR_PDCCH_DCI_DEBUG + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> CCEmap_cand=%d \n",CCEmap_cand); +#endif + if (CCEmap_cand == 0) { - #ifdef DEBUG_DCI_DECODING - if (do_common == 1) - LOG_I(PHY,"[DCI search nPdcch %d - common] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x)\n", - pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask); - else - LOG_I(PHY,"[DCI search nPdcch %d - ue spec] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x) format %d\n", - pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask,format_c); - #endif - #ifndef NR_PDCCH_DCI_DEBUG - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we enter function dci_decoding(sizeof_bits=%d L=%d) -----\n",sizeof_bits,L); - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we have to replace this part of the code by polar decoding\n"); - #endif - -// for (int m=0; m < (nCCE[p]*6*9*2); m++) - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0: polar decoding)-> polar intput (with coreset_time_dur=%d, coreset_nbr_rb=%d, p=%d, CCEind=%d): \n", - coreset_time_dur,coreset_nbr_rb,p,CCEind); - #endif +#ifdef DEBUG_DCI_DECODING + if (do_common == 1) + LOG_I(PHY,"[DCI search nPdcch %d - common] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x)\n", + pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask); + else + LOG_I(PHY,"[DCI search nPdcch %d - ue spec] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x) format %d\n", + pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask,format_uss); +#endif +#ifdef NR_PDCCH_DCI_DEBUG + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we enter function dci_decoding(sizeof_bits=%d L=%d) -----\n",sizeof_bits,L); + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we have to replace this part of the code by polar decoding\n"); +#endif + + + // for (int m=0; m < (nCCE[p]*6*9*2); m++) +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0: polar decoding)-> polar intput (with coreset_time_dur=%d, coreset_nbr_rb=%d, p=%d, CCEind=%d): \n", + coreset_time_dur,coreset_nbr_rb,p,CCEind); +#endif + /* int reg_p=0,reg_e=0; for (int m=0; m < (L2*6); m++){ - reg_p = (((int)floor(m/coreset_time_dur))+((m%coreset_time_dur)*(L2*6/coreset_time_dur)))*9*2; - reg_e = m*9*2; - for (int i=0; i<9*2; i++){ - //polar_input[reg_p+i] = (pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]>0) ? (1.0):(-1.0); - polar_input[reg_e+i] = (pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]>0) ? (1/sqrt(2)):((-1)/sqrt(2)); - //printf("\t m=%d \tpolar_input[%d]=%lf <-> e_rx[%d]=%d\n",m,reg_e+i,polar_input[reg_e+i], - // ((CCEind*9*6*2) + reg_e + i),pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]); - //printf("\t m=%d \tpolar_input[%d]=%lf <-> e_rx[%d]=%d\n",m,reg_p+i,polar_input[reg_p+i], - // ((CCEind*9*6*2) + reg_e + i),pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]); - } - } + reg_p = (((int)floor(m/coreset_time_dur))+((m%coreset_time_dur)*(L2*6/coreset_time_dur)))*9*2; + reg_e = m*9*2; + for (int i=0; i<9*2; i++){ + //polar_input[reg_p+i] = (pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]>0) ? (1.0):(-1.0); + polar_input[reg_e+i] = (pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]>0) ? (1/sqrt(2)):((-1)/sqrt(2)); + //printf("\t m=%d \tpolar_input[%d]=%lf <-> e_rx[%d]=%d\n",m,reg_e+i,polar_input[reg_e+i], + // ((CCEind*9*6*2) + reg_e + i),pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]); + //printf("\t m=%d \tpolar_input[%d]=%lf <-> e_rx[%d]=%d\n",m,reg_p+i,polar_input[reg_p+i], + // ((CCEind*9*6*2) + reg_e + i),pdcch_vars[eNB_id]->e_rx[((CCEind*9*6*2) + reg_e + i)]); + } + } + +#ifdef NR_PDCCH_DCI_DEBUG + printf("\n"); + int j=0; + uint32_t polar_hex[27] = {0}; + for (int i=0; i polar_hex[%d]=%x\n",j,polar_hex[j]); +#endif + */ - #ifndef NR_PDCCH_DCI_DEBUG - printf("\n"); - int j=0; - uint32_t polar_hex[27] = {0}; - for (int i=0; i polar_hex[%d]=%x\n",j,polar_hex[j]); - #endif - -uint32_t dci_estimation[4]={0}; -uint16_t rnti=3; - #ifdef PDCCH_TEST_POLAR_TEMP_FIX - t_nrPolar_paramsPtr nrPolar_params = NULL; - nr_polar_init(&nrPolar_params, 1, 41, 8); - t_nrPolar_paramsPtr currentPtrDCI=nr_polar_params(nrPolar_params, 1, 41, 8); - decoderState = polar_decoder_dci(polar_input, - dci_estimation, - currentPtrDCI,//nrPolar_params, - 8,//NR_POLAR_DECODER_LISTSIZE, - 0,//NR_POLAR_DECODER_PATH_METRIC_APPROXIMATION, - rnti); //rnti - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0: \t\tpolar decoding)-> decoderState %d\n", decoderState); - #endif - //dci_estimation[0] = 0x01189400; - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0: \t\tdci_estimation: [0]->0x%08x \t [1]->0x%08x \t [2]->0x%08x \t [3]->0x%08x\n", - dci_estimation[0], dci_estimation[1], dci_estimation[2], dci_estimation[3]); + uint32_t dci_estimation[4]={0}; + + + nr_polar_init(&nrPolar_params, 1, sizeof_bits, L2); + + + t_nrPolar_paramsPtr currentPtrDCI=nr_polar_params(nrPolar_params, 1, sizeof_bits, L2); + decoderState = polar_decoder_int16(&pdcch_vars[eNB_id]->e_rx[CCEind*9*6*2], + (uint64_t*)dci_estimation, + currentPtrDCI); + crc = decoderState; + //crc = (crc16(&dci_decoded_output[current_thread_id][0], sizeof_bits) >> 16) ^ extract_crc(&dci_decoded_output[current_thread_id][0], sizeof_bits); +#ifdef NR_PDCCH_DCI_DEBUG + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we end function dci_decoding() with crc=%x\n",crc); + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we have to replace this part of the code by polar decoding\n"); +#endif +#ifdef DEBUG_DCI_DECODING + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0: crc =>%d\n",crc); +#endif //uint16_t tc_rnti, uint16_t int_rnti, uint16_t sfi_rnti, uint16_t tpc_pusch_rnti, uint16_t tpc_pucch_rnti, uint16_t tpc_srs__rnti +#ifdef NR_PDCCH_DCI_DEBUG + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> format_found=%d\n",*format_found); + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> crc_scrambled=%d\n",*crc_scrambled); +#endif - crc = rnti; - //crc = (crc16(&dci_decoded_output[current_thread_id][0], sizeof_bits) >> 16) ^ extract_crc(&dci_decoded_output[current_thread_id][0], sizeof_bits); - #ifndef NR_PDCCH_DCI_DEBUG - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we end function dci_decoding() with crc=%x\n",crc); - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> ... we have to replace this part of the code by polar decoding\n"); - #endif - #ifdef DEBUG_DCI_DECODING - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0: crc =>%d\n",crc); - #endif //uint16_t tc_rnti, uint16_t int_rnti, uint16_t sfi_rnti, uint16_t tpc_pusch_rnti, uint16_t tpc_pucch_rnti, uint16_t tpc_srs__rnti - #ifdef NR_PDCCH_DCI_DEBUG - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> format_found=%d\n",*format_found); - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> crc_scrambled=%d\n",*crc_scrambled); - #endif if (crc == crc_scrambled_values[_C_RNTI_]) { - *crc_scrambled =_c_rnti; - *format_found=1; + *crc_scrambled =_c_rnti; + *format_found=1; } if (crc == crc_scrambled_values[_CS_RNTI_]) { - *crc_scrambled =_cs_rnti; - *format_found=1; + *crc_scrambled =_cs_rnti; + *format_found=1; } if (crc == crc_scrambled_values[_NEW_RNTI_]) { - *crc_scrambled =_new_rnti; - *format_found=1; + *crc_scrambled =_new_rnti; + *format_found=1; } if (crc == crc_scrambled_values[_TC_RNTI_]) { - *crc_scrambled =_tc_rnti; - *format_found=_format_1_0_found; + *crc_scrambled =_tc_rnti; + *format_found=_format_1_0_found; } if (crc == crc_scrambled_values[_P_RNTI_]) { - *crc_scrambled =_p_rnti; - *format_found=_format_1_0_found; + *crc_scrambled =_p_rnti; + *format_found=_format_1_0_found; } if (crc == crc_scrambled_values[_SI_RNTI_]) { - *crc_scrambled =_si_rnti; - *format_found=_format_1_0_found; + *crc_scrambled =_si_rnti; + *format_found=_format_1_0_found; } if (crc == crc_scrambled_values[_RA_RNTI_]) { - *crc_scrambled =_ra_rnti; - *format_found=_format_1_0_found; + *crc_scrambled =_ra_rnti; + *format_found=_format_1_0_found; } if (crc == crc_scrambled_values[_SP_CSI_RNTI_]) { - *crc_scrambled =_sp_csi_rnti; - *format_found=_format_0_1_found; + *crc_scrambled =_sp_csi_rnti; + *format_found=_format_0_1_found; } if (crc == crc_scrambled_values[_SFI_RNTI_]) { - *crc_scrambled =_sfi_rnti; - *format_found=_format_2_0_found; + *crc_scrambled =_sfi_rnti; + *format_found=_format_2_0_found; } if (crc == crc_scrambled_values[_INT_RNTI_]) { - *crc_scrambled =_int_rnti; - *format_found=_format_2_1_found; + *crc_scrambled =_int_rnti; + *format_found=_format_2_1_found; } if (crc == crc_scrambled_values[_TPC_PUSCH_RNTI_]) { - *crc_scrambled =_tpc_pusch_rnti; - *format_found=_format_2_2_found; + *crc_scrambled =_tpc_pusch_rnti; + *format_found=_format_2_2_found; } if (crc == crc_scrambled_values[_TPC_PUCCH_RNTI_]) { - *crc_scrambled =_tpc_pucch_rnti; - *format_found=_format_2_2_found; + *crc_scrambled =_tpc_pucch_rnti; + *format_found=_format_2_2_found; } if (crc == crc_scrambled_values[_TPC_SRS_RNTI_]) { - *crc_scrambled =_tpc_srs_rnti; - *format_found=_format_2_3_found; + *crc_scrambled =_tpc_srs_rnti; + *format_found=_format_2_3_found; } #ifdef NR_PDCCH_DCI_DEBUG - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> format_found=%d\n",*format_found); - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> crc_scrambled=%d\n",*crc_scrambled); + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> format_found=%d\n",*format_found); + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> crc_scrambled=%d\n",*crc_scrambled); #endif if (*format_found!=255) { - dci_alloc[*dci_cnt].dci_length = sizeof_bits; - dci_alloc[*dci_cnt].rnti = crc; - dci_alloc[*dci_cnt].L = L; - dci_alloc[*dci_cnt].firstCCE = CCEind; - dci_alloc[*dci_cnt].dci_pdu[0] = dci_estimation[0]; - dci_alloc[*dci_cnt].dci_pdu[1] = dci_estimation[1]; - dci_alloc[*dci_cnt].dci_pdu[2] = dci_estimation[2]; - dci_alloc[*dci_cnt].dci_pdu[3] = dci_estimation[3]; - #ifdef NR_PDCCH_DCI_DEBUG - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> rnti matches -> DCI FOUND !!! crc =>%x, sizeof_bits %d, sizeof_bytes %d \n", - dci_alloc[*dci_cnt].rnti, dci_alloc[*dci_cnt].dci_length, sizeof_bytes); - printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> dci_cnt %d (format_css %d crc_scrambled %d) L %d, firstCCE %d pdu[0] %x pdu[1] %x pdu[2] %x pdu[3] %x \n", - *dci_cnt, format_css,*crc_scrambled,dci_alloc[*dci_cnt].L, dci_alloc[*dci_cnt].firstCCE,dci_alloc[*dci_cnt].dci_pdu[0],dci_alloc[*dci_cnt].dci_pdu[1],dci_alloc[*dci_cnt].dci_pdu[2],dci_alloc[*dci_cnt].dci_pdu[3]); - #endif - if ((format_css == cformat0_0_and_1_0) || (format_uss == uformat0_0_and_1_0)){ - if ((*crc_scrambled == _p_rnti) || (*crc_scrambled == _si_rnti) || (*crc_scrambled == _ra_rnti)){ - dci_alloc[*dci_cnt].format = format1_0; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_1_0_found; - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> a format1_0=%d and dci_cnt=%d\n",*format_found,*dci_cnt); - } else { - if ((dci_estimation[0]&1) == 0){ - dci_alloc[*dci_cnt].format = format0_0; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_0_0_found; - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> b format0_0=%d and dci_cnt=%d\n",*format_found,*dci_cnt); - } - if ((dci_estimation[0]&1) == 1){ - dci_alloc[*dci_cnt].format = format1_0; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_1_0_found; - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> c format1_0=%d and dci_cnt=%d\n",*format_found,*dci_cnt); - } - } - } - if (format_css == cformat2_0){ - dci_alloc[*dci_cnt].format = format2_0; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_2_0_found; - } - if (format_css == cformat2_1){ - dci_alloc[*dci_cnt].format = format2_1; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_2_1_found; - } - if (format_css == cformat2_2){ - dci_alloc[*dci_cnt].format = format2_2; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_2_2_found; - } - if (format_css == cformat2_3){ - dci_alloc[*dci_cnt].format = format2_3; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_2_3_found; - } - if (format_uss == uformat0_1_and_1_1){ - if ((dci_estimation[0]&1) == 0){ - dci_alloc[*dci_cnt].format = format0_1; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_0_1_found; - } - if ((dci_estimation[0]&1) == 1){ - dci_alloc[*dci_cnt].format = format1_1; - *dci_cnt = *dci_cnt + 1; - *format_found=_format_1_1_found; - } - } - // store first nCCE of group for PUCCH transmission of ACK/NAK - pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; -/* if (crc == si_rnti) { + dci_alloc[*dci_cnt].dci_length = sizeof_bits; + dci_alloc[*dci_cnt].rnti = crc; + dci_alloc[*dci_cnt].L = L; + dci_alloc[*dci_cnt].firstCCE = CCEind; + dci_alloc[*dci_cnt].dci_pdu[0] = dci_estimation[0]; + dci_alloc[*dci_cnt].dci_pdu[1] = dci_estimation[1]; + dci_alloc[*dci_cnt].dci_pdu[2] = dci_estimation[2]; + dci_alloc[*dci_cnt].dci_pdu[3] = dci_estimation[3]; +#ifdef NR_PDCCH_DCI_DEBUG + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> rnti matches -> DCI FOUND !!! crc =>%x, sizeof_bits %d, sizeof_bytes %d \n", + dci_alloc[*dci_cnt].rnti, dci_alloc[*dci_cnt].dci_length, sizeof_bytes); + printf ("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> dci_cnt %d (format_css %d crc_scrambled %d) L %d, firstCCE %d pdu[0] %x pdu[1] %x pdu[2] %x pdu[3] %x \n", + *dci_cnt, format_css,*crc_scrambled,dci_alloc[*dci_cnt].L, dci_alloc[*dci_cnt].firstCCE,dci_alloc[*dci_cnt].dci_pdu[0],dci_alloc[*dci_cnt].dci_pdu[1],dci_alloc[*dci_cnt].dci_pdu[2],dci_alloc[*dci_cnt].dci_pdu[3]); +#endif + if ((format_css == cformat0_0_and_1_0) || (format_uss == uformat0_0_and_1_0)){ + if ((*crc_scrambled == _p_rnti) || (*crc_scrambled == _si_rnti) || (*crc_scrambled == _ra_rnti)){ + dci_alloc[*dci_cnt].format = format1_0; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_1_0_found; + // printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> a format1_0=%d and dci_cnt=%d\n",*format_found,*dci_cnt); + } else { + if ((dci_estimation[0]&1) == 0){ + dci_alloc[*dci_cnt].format = format0_0; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_0_0_found; + // printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> b format0_0=%d and dci_cnt=%d\n",*format_found,*dci_cnt); + } + if ((dci_estimation[0]&1) == 1){ + dci_alloc[*dci_cnt].format = format1_0; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_1_0_found; + // printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> c format1_0=%d and dci_cnt=%d\n",*format_found,*dci_cnt); + } + } + } + if (format_css == cformat2_0){ + dci_alloc[*dci_cnt].format = format2_0; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_2_0_found; + } + if (format_css == cformat2_1){ + dci_alloc[*dci_cnt].format = format2_1; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_2_1_found; + } + if (format_css == cformat2_2){ + dci_alloc[*dci_cnt].format = format2_2; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_2_2_found; + } + if (format_css == cformat2_3){ + dci_alloc[*dci_cnt].format = format2_3; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_2_3_found; + } + if (format_uss == uformat0_1_and_1_1){ + if ((dci_estimation[0]&1) == 0){ + dci_alloc[*dci_cnt].format = format0_1; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_0_1_found; + } + if ((dci_estimation[0]&1) == 1){ + dci_alloc[*dci_cnt].format = format1_1; + *dci_cnt = *dci_cnt + 1; + *format_found=_format_1_1_found; + } + } + // store first nCCE of group for PUCCH transmission of ACK/NAK + pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; + /* if (crc == si_rnti) { dci_alloc[*dci_cnt].format = format_si; *dci_cnt = *dci_cnt + 1; - } else if (crc == p_rnti) { + } else if (crc == p_rnti) { dci_alloc[*dci_cnt].format = format_p; *dci_cnt = *dci_cnt + 1; - } else if (crc == ra_rnti) { + } else if (crc == ra_rnti) { dci_alloc[*dci_cnt].format = format_ra; // store first nCCE of group for PUCCH transmission of ACK/NAK pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; *dci_cnt = *dci_cnt + 1; - } else if (crc == pdcch_vars[eNB_id]->crnti) { + } else if (crc == pdcch_vars[eNB_id]->crnti) { if ((mode & UL_DCI) && (format_c == format0) - && ((dci_decoded_output[current_thread_id][0] & 0x80) - == 0)) { // check if pdu is format 0 or 1A - if (*format0_found == 0) { - dci_alloc[*dci_cnt].format = format0; - *format0_found = 1; - *dci_cnt = *dci_cnt + 1; - pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; - } - } else if (format_c == format0) { // this is a format 1A DCI - dci_alloc[*dci_cnt].format = format1A; - *dci_cnt = *dci_cnt + 1; - pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; - } else { - // store first nCCE of group for PUCCH transmission of ACK/NAK - if (*format_c_found == 0) { - dci_alloc[*dci_cnt].format = format_c; - *dci_cnt = *dci_cnt + 1; - *format_c_found = 1; - pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; - } + && ((dci_decoded_output[current_thread_id][0] & 0x80) + == 0)) { // check if pdu is format 0 or 1A + if (*format0_found == 0) { + dci_alloc[*dci_cnt].format = format0; + *format0_found = 1; + *dci_cnt = *dci_cnt + 1; + pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; } - }*/ - //LOG_I(PHY,"DCI decoding CRNTI [format: %d, nCCE[nr_tti_rx: %d]: %d ], AggregationLevel %d \n",format_c, nr_tti_rx, pdcch_vars[eNB_id]->nCCE[nr_tti_rx],L2); - // memcpy(&dci_alloc[*dci_cnt].dci_pdu[0],dci_decoded_output,sizeof_bytes); - switch (1 << L) { - case 1: - *CCEmap |= (1 << (CCEind & 0x1f)); - break; - case 2: - *CCEmap |= (1 << (CCEind & 0x1f)); - break; - case 4: - *CCEmap |= (1 << (CCEind & 0x1f)); - break; - case 8: - *CCEmap |= (1 << (CCEind & 0x1f)); - break; - case 16: - *CCEmap |= (1 << (CCEind & 0x1f)); - break; - } + } else if (format_c == format0) { // this is a format 1A DCI + dci_alloc[*dci_cnt].format = format1A; + *dci_cnt = *dci_cnt + 1; + pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; + } else { + // store first nCCE of group for PUCCH transmission of ACK/NAK + if (*format_c_found == 0) { + dci_alloc[*dci_cnt].format = format_c; + *dci_cnt = *dci_cnt + 1; + *format_c_found = 1; + pdcch_vars[eNB_id]->nCCE[nr_tti_rx] = CCEind; + } + } + }*/ + //LOG_I(PHY,"DCI decoding CRNTI [format: %d, nCCE[nr_tti_rx: %d]: %d ], AggregationLevel %d \n",format_c, nr_tti_rx, pdcch_vars[eNB_id]->nCCE[nr_tti_rx],L2); + // memcpy(&dci_alloc[*dci_cnt].dci_pdu[0],dci_decoded_output,sizeof_bytes); + switch (1 << L) { + case 1: + *CCEmap |= (1 << (CCEind & 0x1f)); + break; + case 2: + *CCEmap |= (1 << (CCEind & 0x1f)); + break; + case 4: + *CCEmap |= (1 << (CCEind & 0x1f)); + break; + case 8: + *CCEmap |= (1 << (CCEind & 0x1f)); + break; + case 16: + *CCEmap |= (1 << (CCEind & 0x1f)); + break; + } #ifdef DEBUG_DCI_DECODING - LOG_I(PHY,"[DCI search] Found DCI %d rnti %x Aggregation %d length %d format %s in CCE %d (CCEmap %x) candidate %d / %d \n", - *dci_cnt,crc,1< end candidate loop\n"); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure0)-> end candidate loop\n"); +#endif } #endif @@ -1737,86 +1602,86 @@ uint16_t rnti=3; /*void dci_decoding_procedure0(NR_UE_PDCCH **pdcch_vars, - int do_common, - dci_detect_mode_t mode, - uint8_t nr_tti_rx, - DCI_ALLOC_t *dci_alloc, - int16_t eNB_id, - uint8_t current_thread_id, - NR_DL_FRAME_PARMS *frame_parms, - uint8_t mi, - uint16_t si_rnti, - uint16_t ra_rnti, - uint16_t p_rnti, - uint8_t L, - uint8_t format_si, - uint8_t format_p, - uint8_t format_ra, - uint8_t format_c, - uint8_t sizeof_bits, - uint8_t sizeof_bytes, - uint8_t *dci_cnt, - uint8_t *format0_found, - uint8_t *format_c_found, - uint32_t *CCEmap0, - uint32_t *CCEmap1, - uint32_t *CCEmap2) -{ + int do_common, + dci_detect_mode_t mode, + uint8_t nr_tti_rx, + DCI_ALLOC_t *dci_alloc, + int16_t eNB_id, + uint8_t current_thread_id, + NR_DL_FRAME_PARMS *frame_parms, + uint8_t mi, + uint16_t si_rnti, + uint16_t ra_rnti, + uint16_t p_rnti, + uint8_t L, + uint8_t format_si, + uint8_t format_p, + uint8_t format_ra, + uint8_t format_c, + uint8_t sizeof_bits, + uint8_t sizeof_bytes, + uint8_t *dci_cnt, + uint8_t *format0_found, + uint8_t *format_c_found, + uint32_t *CCEmap0, + uint32_t *CCEmap1, + uint32_t *CCEmap2) + { uint16_t crc,CCEind,nCCE; uint32_t *CCEmap=NULL,CCEmap_mask=0; int L2=(1< \n"); -#endif + #ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (dci_decoding_procedure0)-> \n"); + #endif nCCE = get_nCCE(pdcch_vars[eNB_id]->num_pdcch_symbols,frame_parms,mi); if (nCCE > get_nCCE(3,frame_parms,1)) { - LOG_D(PHY,"skip DCI decoding: nCCE=%d > get_nCCE(3,frame_parms,1)=%d\n", nCCE, get_nCCE(3,frame_parms,1)); - return; + LOG_D(PHY,"skip DCI decoding: nCCE=%d > get_nCCE(3,frame_parms,1)=%d\n", nCCE, get_nCCE(3,frame_parms,1)); + return; } if (nCCEcrnti; + // Find first available in ue specific search space + // according to procedure in Section 9.1.1 of 36.213 (v. 8.6) + // compute Yk + Yk = (unsigned int)pdcch_vars[eNB_id]->crnti; - for (i=0; i<=nr_tti_rx; i++) - Yk = (Yk*39827)%65537; + for (i=0; i<=nr_tti_rx; i++) + Yk = (Yk*39827)%65537; - Yk = Yk % (nCCE/L2); + Yk = Yk % (nCCE/L2); - switch (L2) { - case 1: - case 2: - nb_candidates = 6; - break; + switch (L2) { + case 1: + case 2: + nb_candidates = 6; + break; - case 4: - case 8: - nb_candidates = 2; - break; + case 4: + case 8: + nb_candidates = 2; + break; - default: - DevParam(L2, do_common, eNB_id); - break; - } + default: + DevParam(L2, do_common, eNB_id); + break; + } } // for (CCEind=0; @@ -1824,196 +1689,196 @@ uint16_t rnti=3; // CCEind+=(1< nCCE) - nb_candidates = nCCE/L2; + nb_candidates = nCCE/L2; for (m=0; mmacphy_exit("Illegal CCEind\n"); - return; // not reached - } + if (CCEind<32) + CCEmap = CCEmap0; + else if (CCEind<64) + CCEmap = CCEmap1; + else if (CCEind<96) + CCEmap = CCEmap2; + else { + LOG_E(PHY,"Illegal CCEind %d (Yk %d, m %d, nCCE %d, L2 %d\n",CCEind,Yk,m,nCCE,L2); + mac_xface->macphy_exit("Illegal CCEind\n"); + return; // not reached + } - switch (L2) { - case 1: - CCEmap_mask = (1<<(CCEind&0x1f)); - break; + switch (L2) { + case 1: + CCEmap_mask = (1<<(CCEind&0x1f)); + break; - case 2: - CCEmap_mask = (3<<(CCEind&0x1f)); - break; + case 2: + CCEmap_mask = (3<<(CCEind&0x1f)); + break; - case 4: - CCEmap_mask = (0xf<<(CCEind&0x1f)); - break; + case 4: + CCEmap_mask = (0xf<<(CCEind&0x1f)); + break; - case 8: - CCEmap_mask = (0xff<<(CCEind&0x1f)); - break; + case 8: + CCEmap_mask = (0xff<<(CCEind&0x1f)); + break; - default: - LOG_E( PHY, "Illegal L2 value %d\n", L2 ); - mac_xface->macphy_exit( "Illegal L2\n" ); - return; // not reached - } + default: + LOG_E( PHY, "Illegal L2 value %d\n", L2 ); + mac_xface->macphy_exit( "Illegal L2\n" ); + return; // not reached + } - CCEmap_cand = (*CCEmap)&CCEmap_mask; + CCEmap_cand = (*CCEmap)&CCEmap_mask; - // CCE is not allocated yet + // CCE is not allocated yet - if (CCEmap_cand == 0) { -#ifdef DEBUG_DCI_DECODING + if (CCEmap_cand == 0) { + #ifdef DEBUG_DCI_DECODING - if (do_common == 1) - LOG_I(PHY,"[DCI search nPdcch %d - common] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x)\n", - pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask); - else - LOG_I(PHY,"[DCI search nPdcch %d - ue spec] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x) format %d\n", - pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask,format_c); + if (do_common == 1) + LOG_I(PHY,"[DCI search nPdcch %d - common] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x)\n", + pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask); + else + LOG_I(PHY,"[DCI search nPdcch %d - ue spec] Attempting candidate %d Aggregation Level %d DCI length %d at CCE %d/%d (CCEmap %x,CCEmap_cand %x) format %d\n", + pdcch_vars[eNB_id]->num_pdcch_symbols,m,L2,sizeof_bits,CCEind,nCCE,*CCEmap,CCEmap_mask,format_c); -#endif + #endif - dci_decoding(sizeof_bits, - L, - &pdcch_vars[eNB_id]->e_rx[CCEind*72], - &dci_decoded_output[current_thread_id][0]); + dci_decoding(sizeof_bits, + L, + &pdcch_vars[eNB_id]->e_rx[CCEind*72], + &dci_decoded_output[current_thread_id][0]); - // for (i=0;i<3+(sizeof_bits>>3);i++) - // printf("dci_decoded_output[%d] => %x\n",i,dci_decoded_output[i]); + // for (i=0;i<3+(sizeof_bits>>3);i++) + // printf("dci_decoded_output[%d] => %x\n",i,dci_decoded_output[i]); - crc = (crc16(&dci_decoded_output[current_thread_id][0],sizeof_bits)>>16) ^ extract_crc(&dci_decoded_output[current_thread_id][0],sizeof_bits); -#ifdef DEBUG_DCI_DECODING - printf("crc =>%x\n",crc); -#endif + crc = (crc16(&dci_decoded_output[current_thread_id][0],sizeof_bits)>>16) ^ extract_crc(&dci_decoded_output[current_thread_id][0],sizeof_bits); + #ifdef DEBUG_DCI_DECODING + printf("crc =>%x\n",crc); + #endif - if (((L>1) && ((crc == si_rnti)|| (crc == p_rnti)|| (crc == ra_rnti)))|| - (crc == pdcch_vars[eNB_id]->crnti)) { - dci_alloc[*dci_cnt].dci_length = sizeof_bits; - dci_alloc[*dci_cnt].rnti = crc; - dci_alloc[*dci_cnt].L = L; - dci_alloc[*dci_cnt].firstCCE = CCEind; + if (((L>1) && ((crc == si_rnti)|| (crc == p_rnti)|| (crc == ra_rnti)))|| + (crc == pdcch_vars[eNB_id]->crnti)) { + dci_alloc[*dci_cnt].dci_length = sizeof_bits; + dci_alloc[*dci_cnt].rnti = crc; + dci_alloc[*dci_cnt].L = L; + dci_alloc[*dci_cnt].firstCCE = CCEind; - //printf("DCI FOUND !!! crc =>%x, sizeof_bits %d, sizeof_bytes %d \n",crc, sizeof_bits, sizeof_bytes); - if (sizeof_bytes<=4) { - dci_alloc[*dci_cnt].dci_pdu[3] = dci_decoded_output[current_thread_id][0]; - dci_alloc[*dci_cnt].dci_pdu[2] = dci_decoded_output[current_thread_id][1]; - dci_alloc[*dci_cnt].dci_pdu[1] = dci_decoded_output[current_thread_id][2]; - dci_alloc[*dci_cnt].dci_pdu[0] = dci_decoded_output[current_thread_id][3]; -#ifdef DEBUG_DCI_DECODING - printf("DCI => %x,%x,%x,%x\n",dci_decoded_output[current_thread_id][0], - dci_decoded_output[current_thread_id][1], - dci_decoded_output[current_thread_id][2], - dci_decoded_output[current_thread_id][3]); -#endif - } else { - dci_alloc[*dci_cnt].dci_pdu[7] = dci_decoded_output[current_thread_id][0]; - dci_alloc[*dci_cnt].dci_pdu[6] = dci_decoded_output[current_thread_id][1]; - dci_alloc[*dci_cnt].dci_pdu[5] = dci_decoded_output[current_thread_id][2]; - dci_alloc[*dci_cnt].dci_pdu[4] = dci_decoded_output[current_thread_id][3]; - dci_alloc[*dci_cnt].dci_pdu[3] = dci_decoded_output[current_thread_id][4]; - dci_alloc[*dci_cnt].dci_pdu[2] = dci_decoded_output[current_thread_id][5]; - dci_alloc[*dci_cnt].dci_pdu[1] = dci_decoded_output[current_thread_id][6]; - dci_alloc[*dci_cnt].dci_pdu[0] = dci_decoded_output[current_thread_id][7]; -#ifdef DEBUG_DCI_DECODING - printf("DCI => %x,%x,%x,%x,%x,%x,%x,%x\n", - dci_decoded_output[current_thread_id][0],dci_decoded_output[current_thread_id][1],dci_decoded_output[current_thread_id][2],dci_decoded_output[current_thread_id][3], - dci_decoded_output[current_thread_id][4],dci_decoded_output[current_thread_id][5],dci_decoded_output[current_thread_id][6],dci_decoded_output[current_thread_id][7]); -#endif - } + //printf("DCI FOUND !!! crc =>%x, sizeof_bits %d, sizeof_bytes %d \n",crc, sizeof_bits, sizeof_bytes); + if (sizeof_bytes<=4) { + dci_alloc[*dci_cnt].dci_pdu[3] = dci_decoded_output[current_thread_id][0]; + dci_alloc[*dci_cnt].dci_pdu[2] = dci_decoded_output[current_thread_id][1]; + dci_alloc[*dci_cnt].dci_pdu[1] = dci_decoded_output[current_thread_id][2]; + dci_alloc[*dci_cnt].dci_pdu[0] = dci_decoded_output[current_thread_id][3]; + #ifdef DEBUG_DCI_DECODING + printf("DCI => %x,%x,%x,%x\n",dci_decoded_output[current_thread_id][0], + dci_decoded_output[current_thread_id][1], + dci_decoded_output[current_thread_id][2], + dci_decoded_output[current_thread_id][3]); + #endif + } else { + dci_alloc[*dci_cnt].dci_pdu[7] = dci_decoded_output[current_thread_id][0]; + dci_alloc[*dci_cnt].dci_pdu[6] = dci_decoded_output[current_thread_id][1]; + dci_alloc[*dci_cnt].dci_pdu[5] = dci_decoded_output[current_thread_id][2]; + dci_alloc[*dci_cnt].dci_pdu[4] = dci_decoded_output[current_thread_id][3]; + dci_alloc[*dci_cnt].dci_pdu[3] = dci_decoded_output[current_thread_id][4]; + dci_alloc[*dci_cnt].dci_pdu[2] = dci_decoded_output[current_thread_id][5]; + dci_alloc[*dci_cnt].dci_pdu[1] = dci_decoded_output[current_thread_id][6]; + dci_alloc[*dci_cnt].dci_pdu[0] = dci_decoded_output[current_thread_id][7]; + #ifdef DEBUG_DCI_DECODING + printf("DCI => %x,%x,%x,%x,%x,%x,%x,%x\n", + dci_decoded_output[current_thread_id][0],dci_decoded_output[current_thread_id][1],dci_decoded_output[current_thread_id][2],dci_decoded_output[current_thread_id][3], + dci_decoded_output[current_thread_id][4],dci_decoded_output[current_thread_id][5],dci_decoded_output[current_thread_id][6],dci_decoded_output[current_thread_id][7]); + #endif + } - if (crc==si_rnti) { - dci_alloc[*dci_cnt].format = format_si; - *dci_cnt = *dci_cnt+1; - } else if (crc==p_rnti) { - dci_alloc[*dci_cnt].format = format_p; - *dci_cnt = *dci_cnt+1; - } else if (crc==ra_rnti) { - dci_alloc[*dci_cnt].format = format_ra; - // store first nCCE of group for PUCCH transmission of ACK/NAK - pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; - *dci_cnt = *dci_cnt+1; - } else if (crc==pdcch_vars[eNB_id]->crnti) { + if (crc==si_rnti) { + dci_alloc[*dci_cnt].format = format_si; + *dci_cnt = *dci_cnt+1; + } else if (crc==p_rnti) { + dci_alloc[*dci_cnt].format = format_p; + *dci_cnt = *dci_cnt+1; + } else if (crc==ra_rnti) { + dci_alloc[*dci_cnt].format = format_ra; + // store first nCCE of group for PUCCH transmission of ACK/NAK + pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; + *dci_cnt = *dci_cnt+1; + } else if (crc==pdcch_vars[eNB_id]->crnti) { - if ((mode&UL_DCI)&&(format_c == format0)&&((dci_decoded_output[current_thread_id][0]&0x80)==0)) {// check if pdu is format 0 or 1A - if (*format0_found == 0) { - dci_alloc[*dci_cnt].format = format0; - *format0_found = 1; - *dci_cnt = *dci_cnt+1; - pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; - } - } else if (format_c == format0) { // this is a format 1A DCI - dci_alloc[*dci_cnt].format = format1A; - *dci_cnt = *dci_cnt+1; - pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; - } else { - // store first nCCE of group for PUCCH transmission of ACK/NAK - if (*format_c_found == 0) { - dci_alloc[*dci_cnt].format = format_c; - *dci_cnt = *dci_cnt+1; - *format_c_found = 1; - pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; - } - } - } + if ((mode&UL_DCI)&&(format_c == format0)&&((dci_decoded_output[current_thread_id][0]&0x80)==0)) {// check if pdu is format 0 or 1A + if (*format0_found == 0) { + dci_alloc[*dci_cnt].format = format0; + *format0_found = 1; + *dci_cnt = *dci_cnt+1; + pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; + } + } else if (format_c == format0) { // this is a format 1A DCI + dci_alloc[*dci_cnt].format = format1A; + *dci_cnt = *dci_cnt+1; + pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; + } else { + // store first nCCE of group for PUCCH transmission of ACK/NAK + if (*format_c_found == 0) { + dci_alloc[*dci_cnt].format = format_c; + *dci_cnt = *dci_cnt+1; + *format_c_found = 1; + pdcch_vars[eNB_id]->nCCE[nr_tti_rx]=CCEind; + } + } + } - //LOG_I(PHY,"DCI decoding CRNTI [format: %d, nCCE[nr_tti_rx: %d]: %d ], AggregationLevel %d \n",format_c, nr_tti_rx, pdcch_vars[eNB_id]->nCCE[nr_tti_rx],L2); - // memcpy(&dci_alloc[*dci_cnt].dci_pdu[0],dci_decoded_output,sizeof_bytes); + //LOG_I(PHY,"DCI decoding CRNTI [format: %d, nCCE[nr_tti_rx: %d]: %d ], AggregationLevel %d \n",format_c, nr_tti_rx, pdcch_vars[eNB_id]->nCCE[nr_tti_rx],L2); + // memcpy(&dci_alloc[*dci_cnt].dci_pdu[0],dci_decoded_output,sizeof_bytes); - switch (1<N_RB_DL) { case 6: - if (frame_type == TDD) { - format0_size_bits = sizeof_DCI0_1_5MHz_TDD_1_6_t; - format0_size_bytes = sizeof(DCI0_1_5MHz_TDD_1_6_t); - format1_size_bits = sizeof_DCI1_1_5MHz_TDD_t; - format1_size_bytes = sizeof(DCI1_1_5MHz_TDD_t); + if (frame_type == TDD) { + format0_size_bits = sizeof_DCI0_1_5MHz_TDD_1_6_t; + format0_size_bytes = sizeof(DCI0_1_5MHz_TDD_1_6_t); + format1_size_bits = sizeof_DCI1_1_5MHz_TDD_t; + format1_size_bytes = sizeof(DCI1_1_5MHz_TDD_t); - } else { - format0_size_bits = sizeof_DCI0_1_5MHz_FDD_t; - format0_size_bytes = sizeof(DCI0_1_5MHz_FDD_t); - format1_size_bits = sizeof_DCI1_1_5MHz_FDD_t; - format1_size_bytes = sizeof(DCI1_1_5MHz_FDD_t); - } + } else { + format0_size_bits = sizeof_DCI0_1_5MHz_FDD_t; + format0_size_bytes = sizeof(DCI0_1_5MHz_FDD_t); + format1_size_bits = sizeof_DCI1_1_5MHz_FDD_t; + format1_size_bytes = sizeof(DCI1_1_5MHz_FDD_t); + } - break; + break; case 25: default: - if (frame_type == TDD) { - format0_size_bits = sizeof_DCI0_5MHz_TDD_1_6_t; - format0_size_bytes = sizeof(DCI0_5MHz_TDD_1_6_t); - format1_size_bits = sizeof_DCI1_5MHz_TDD_t; - format1_size_bytes = sizeof(DCI1_5MHz_TDD_t); - } else { - format0_size_bits = sizeof_DCI0_5MHz_FDD_t; - format0_size_bytes = sizeof(DCI0_5MHz_FDD_t); - format1_size_bits = sizeof_DCI1_5MHz_FDD_t; - format1_size_bytes = sizeof(DCI1_5MHz_FDD_t); - } + if (frame_type == TDD) { + format0_size_bits = sizeof_DCI0_5MHz_TDD_1_6_t; + format0_size_bytes = sizeof(DCI0_5MHz_TDD_1_6_t); + format1_size_bits = sizeof_DCI1_5MHz_TDD_t; + format1_size_bytes = sizeof(DCI1_5MHz_TDD_t); + } else { + format0_size_bits = sizeof_DCI0_5MHz_FDD_t; + format0_size_bytes = sizeof(DCI0_5MHz_FDD_t); + format1_size_bits = sizeof_DCI1_5MHz_FDD_t; + format1_size_bytes = sizeof(DCI1_5MHz_FDD_t); + } - break; + break; case 50: - if (frame_type == TDD) { - format0_size_bits = sizeof_DCI0_10MHz_TDD_1_6_t; - format0_size_bytes = sizeof(DCI0_10MHz_TDD_1_6_t); - format1_size_bits = sizeof_DCI1_10MHz_TDD_t; - format1_size_bytes = sizeof(DCI1_10MHz_TDD_t); + if (frame_type == TDD) { + format0_size_bits = sizeof_DCI0_10MHz_TDD_1_6_t; + format0_size_bytes = sizeof(DCI0_10MHz_TDD_1_6_t); + format1_size_bits = sizeof_DCI1_10MHz_TDD_t; + format1_size_bytes = sizeof(DCI1_10MHz_TDD_t); - } else { - format0_size_bits = sizeof_DCI0_10MHz_FDD_t; - format0_size_bytes = sizeof(DCI0_10MHz_FDD_t); - format1_size_bits = sizeof_DCI1_10MHz_FDD_t; - format1_size_bytes = sizeof(DCI1_10MHz_FDD_t); - } + } else { + format0_size_bits = sizeof_DCI0_10MHz_FDD_t; + format0_size_bytes = sizeof(DCI0_10MHz_FDD_t); + format1_size_bits = sizeof_DCI1_10MHz_FDD_t; + format1_size_bytes = sizeof(DCI1_10MHz_FDD_t); + } - break; + break; case 100: - if (frame_type == TDD) { - format0_size_bits = sizeof_DCI0_20MHz_TDD_1_6_t; - format0_size_bytes = sizeof(DCI0_20MHz_TDD_1_6_t); - format1_size_bits = sizeof_DCI1_20MHz_TDD_t; - format1_size_bytes = sizeof(DCI1_20MHz_TDD_t); - } else { - format0_size_bits = sizeof_DCI0_20MHz_FDD_t; - format0_size_bytes = sizeof(DCI0_20MHz_FDD_t); - format1_size_bits = sizeof_DCI1_20MHz_FDD_t; - format1_size_bytes = sizeof(DCI1_20MHz_FDD_t); - } + if (frame_type == TDD) { + format0_size_bits = sizeof_DCI0_20MHz_TDD_1_6_t; + format0_size_bytes = sizeof(DCI0_20MHz_TDD_1_6_t); + format1_size_bits = sizeof_DCI1_20MHz_TDD_t; + format1_size_bytes = sizeof(DCI1_20MHz_TDD_t); + } else { + format0_size_bits = sizeof_DCI0_20MHz_FDD_t; + format0_size_bytes = sizeof(DCI0_20MHz_FDD_t); + format1_size_bits = sizeof_DCI1_20MHz_FDD_t; + format1_size_bytes = sizeof(DCI1_20MHz_FDD_t); + } - break; + break; } if (ue->prach_resources[eNB_id]) - ra_rnti = ue->prach_resources[eNB_id]->ra_RNTI; + ra_rnti = ue->prach_resources[eNB_id]->ra_RNTI; // Now check UE_SPEC format0/1A ue_spec search spaces at aggregation 8 dci_decoding_procedure0(pdcch_vars,0,mode, - nr_tti_rx, - dci_alloc, - eNB_id, - ue->current_thread_id[nr_tti_rx], - frame_parms, - mi, - ((ue->decode_SIB == 1) ? SI_RNTI : 0), - ra_rnti, - P_RNTI, - agregationLevel, - format1A, - format1A, - format1A, - format0, - format0_size_bits, - format0_size_bytes, - &dci_cnt, - &format0_found, - &format_c_found, - &CCEmap0, - &CCEmap1, - &CCEmap2); + nr_tti_rx, + dci_alloc, + eNB_id, + ue->current_thread_id[nr_tti_rx], + frame_parms, + mi, + ((ue->decode_SIB == 1) ? SI_RNTI : 0), + ra_rnti, + P_RNTI, + agregationLevel, + format1A, + format1A, + format1A, + format0, + format0_size_bits, + format0_size_bytes, + &dci_cnt, + &format0_found, + &format_c_found, + &CCEmap0, + &CCEmap1, + &CCEmap2); if ((CCEmap0==0xffff)|| - ((format0_found==1)&&(format_c_found==1))) - return(dci_cnt); + ((format0_found==1)&&(format_c_found==1))) + return(dci_cnt); if (DCIFormat == 1) { - if ((tmode < 3) || (tmode == 7)) { - //printf("Crnti decoding frame param agregation %d DCI %d \n",agregationLevel,DCIFormat); + if ((tmode < 3) || (tmode == 7)) { + //printf("Crnti decoding frame param agregation %d DCI %d \n",agregationLevel,DCIFormat); - // Now check UE_SPEC format 1 search spaces at aggregation 1 + // Now check UE_SPEC format 1 search spaces at aggregation 1 - //printf("[DCI search] Format 1/1A aggregation 1\n"); + //printf("[DCI search] Format 1/1A aggregation 1\n"); - old_dci_cnt=dci_cnt; - dci_decoding_procedure0(pdcch_vars,0,mode,nr_tti_rx, - dci_alloc, - eNB_id, - ue->current_thread_id[nr_tti_rx], - frame_parms, - mi, - ((ue->decode_SIB == 1) ? SI_RNTI : 0), - ra_rnti, - P_RNTI, - 0, - format1A, - format1A, - format1A, - format1, - format1_size_bits, - format1_size_bytes, - &dci_cnt, - &format0_found, - &format_c_found, - &CCEmap0, - &CCEmap1, - &CCEmap2); + old_dci_cnt=dci_cnt; + dci_decoding_procedure0(pdcch_vars,0,mode,nr_tti_rx, + dci_alloc, + eNB_id, + ue->current_thread_id[nr_tti_rx], + frame_parms, + mi, + ((ue->decode_SIB == 1) ? SI_RNTI : 0), + ra_rnti, + P_RNTI, + 0, + format1A, + format1A, + format1A, + format1, + format1_size_bits, + format1_size_bytes, + &dci_cnt, + &format0_found, + &format_c_found, + &CCEmap0, + &CCEmap1, + &CCEmap2); - if ((CCEmap0==0xffff) || - (format_c_found==1)) - return(dci_cnt); + if ((CCEmap0==0xffff) || + (format_c_found==1)) + return(dci_cnt); - if (dci_cnt>old_dci_cnt) - return(dci_cnt); + if (dci_cnt>old_dci_cnt) + return(dci_cnt); - //printf("Crnti 1 decoding frame param agregation %d DCI %d \n",agregationLevel,DCIFormat); + //printf("Crnti 1 decoding frame param agregation %d DCI %d \n",agregationLevel,DCIFormat); - } - else - { - AssertFatal(0,"Other Transmission mode not yet coded\n"); - } } else { - AssertFatal(0,"DCI format %d not yet implemented \n",DCIFormat); + AssertFatal(0,"Other Transmission mode not yet coded\n"); + } + } + else + { + AssertFatal(0,"DCI format %d not yet implemented \n",DCIFormat); } return(dci_cnt); -} + } */ #ifdef NR_PDCCH_DCI_RUN uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, - uint16_t eNB_id, - uint8_t nr_tti_rx, - int p, - crc_scrambled_t crc_scrambled, - uint16_t n_RB_ULBWP, - uint16_t n_RB_DLBWP, - uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS][NBR_NR_FORMATS], - uint8_t format){ + uint16_t eNB_id, + uint8_t nr_tti_rx, + int p, + crc_scrambled_t crc_scrambled, + uint16_t n_RB_ULBWP, + uint16_t n_RB_DLBWP, + uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS][NBR_NR_FORMATS], + uint8_t format){ #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_format_size)-> crc_scrambled=%d, n_RB_ULBWP=%d, n_RB_DLBWP=%d\n",crc_scrambled,n_RB_ULBWP,n_RB_DLBWP); + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_format_size)-> crc_scrambled=%d, n_RB_ULBWP=%d, n_RB_DLBWP=%d\n",crc_scrambled,n_RB_ULBWP,n_RB_DLBWP); #endif -/* - * function nr_dci_format_size calculates and returns the size in bits of a determined format - * it also returns an bi-dimensional array 'dci_fields_sizes' with x rows and y columns, where: - * x is the number of fields defined in TS 38.212 subclause 7.3.1 (Each field is mapped in the order in which it appears in the description in the specification) - * y is the number of formats - * e.g.: dci_fields_sizes[10][0] contains the size in bits of the field FREQ_DOM_RESOURCE_ASSIGNMENT_UL for format 0_0 - */ + /* + * function nr_dci_format_size calculates and returns the size in bits of a determined format + * it also returns an bi-dimensional array 'dci_fields_sizes' with x rows and y columns, where: + * x is the number of fields defined in TS 38.212 subclause 7.3.1 (Each field is mapped in the order in which it appears in the description in the specification) + * y is the number of formats + * e.g.: dci_fields_sizes[10][0] contains the size in bits of the field FREQ_DOM_RESOURCE_ASSIGNMENT_UL for format 0_0 + */ // pdsch_config contains the PDSCH-Config IE is used to configure the UE specific PDSCH parameters (TS 38.331) PDSCH_Config_t pdsch_config = ue->PDSCH_Config; @@ -2219,7 +2084,7 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, PDSCH_ServingCellConfig_t PDSCH_ServingCellConfig= ue->PDSCH_ServingCellConfig; NR_UE_PDCCH *pdcch_vars2 = ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]; -// 1 CARRIER_IN + // 1 CARRIER_IN // crossCarrierSchedulingConfig from higher layers, variable crossCarrierSchedulingConfig indicates if 'cross carrier scheduling' is enabled or not: // if No cross carrier scheduling: number of bits for CARRIER_IND is 0 // if Cross carrier scheduling: number of bits for CARRIER_IND is 3 @@ -2228,14 +2093,14 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, if (crossCarrierSchedulingConfig.schedulingCellInfo.other.cif_InSchedulingCell !=0 ) crossCarrierSchedulingConfig_ind=1; -// 2 SUL_IND_0_1, // 40 SRS_REQUEST, // 50 SUL_IND_0_0 + // 2 SUL_IND_0_1, // 40 SRS_REQUEST, // 50 SUL_IND_0_0 // UL/SUL indicator (TS 38.331, supplementary uplink is indicated in higher layer parameter ServCellAdd-SUL from IE ServingCellConfig and ServingCellConfigCommon): // 0 bit for UEs not configured with SUL in the cell or UEs configured with SUL in the cell but only PUCCH carrier in the cell is configured for PUSCH transmission // 1 bit for UEs configured with SUL in the cell as defined in Table 7.3.1.1.1-1 // sul_ind indicates whether SUL is configured in cell or not uint8_t sul_ind=ue->supplementaryUplink.supplementaryUplink; // this value will be 0 or 1 depending on higher layer parameter ServCellAdd-SUL. FIXME!!! -// 7 BANDWIDTH_PART_IND + // 7 BANDWIDTH_PART_IND // number of UL BWPs configured by higher layers uint8_t n_UL_BWP_RRC=1; // initialized to 1 but it has to be initialized by higher layers FIXME!!! n_UL_BWP_RRC = ((n_UL_BWP_RRC > 3)?n_UL_BWP_RRC:(n_UL_BWP_RRC+1)); @@ -2244,7 +2109,7 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, n_DL_BWP_RRC = ((n_DL_BWP_RRC > 3)?n_DL_BWP_RRC:(n_DL_BWP_RRC+1)); -// 10 FREQ_DOM_RESOURCE_ASSIGNMENT_UL + // 10 FREQ_DOM_RESOURCE_ASSIGNMENT_UL // if format0_0, only resource allocation type 1 is allowed // if format0_1, then resource allocation type 0 can be configured and N_RBG is defined in TS 38.214 subclause 6.1.2.2.1 @@ -2283,7 +2148,7 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, if ((ul_res_alloc_type_0 == 1) && (ul_res_alloc_type_1 == 1)) n_bits_freq_dom_res_assign_ul = ((n_bits_freq_dom_res_assign_ul>n_ul_RGB_tmp)?(n_bits_freq_dom_res_assign_ul+1):(n_ul_RGB_tmp+1)); -// 11 FREQ_DOM_RESOURCE_ASSIGNMENT_DL + // 11 FREQ_DOM_RESOURCE_ASSIGNMENT_DL // if format1_0, only resource allocation type 1 is allowed // if format1_1, then resource allocation type 0 can be configured and N_RBG is defined in TS 38.214 subclause 5.1.2.2.1 dl_resourceAllocation_t dl_resource_allocation_type = pdsch_config.dl_resourceAllocation; @@ -2308,61 +2173,61 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, } if (dl_res_alloc_type_1 == 1) n_bits_freq_dom_res_assign_dl = (uint8_t)(ceil(log2(n_RB_DLBWP*(n_RB_DLBWP+1)/2))); if ((dl_res_alloc_type_0 == 1) && (dl_res_alloc_type_1 == 1)) - n_bits_freq_dom_res_assign_dl = ((n_bits_freq_dom_res_assign_dl>n_dl_RGB_tmp)?(n_bits_freq_dom_res_assign_dl+1):(n_dl_RGB_tmp+1)); + n_bits_freq_dom_res_assign_dl = ((n_bits_freq_dom_res_assign_dl>n_dl_RGB_tmp)?(n_bits_freq_dom_res_assign_dl+1):(n_dl_RGB_tmp+1)); -// 12 TIME_DOM_RESOURCE_ASSIGNMENT + // 12 TIME_DOM_RESOURCE_ASSIGNMENT uint8_t pusch_alloc_list = pusch_config.n_push_alloc_list; uint8_t pdsch_alloc_list = pdsch_config.n_pdsh_alloc_list; -// 14 PRB_BUNDLING_SIZE_IND:0 bit if the higher layer parameter PRB_bundling is not configured or is set to 'static', or 1 bit if the higher layer parameter PRB_bundling is set to 'dynamic' according to Subclause 5.1.2.3 of [6, TS 38.214] + // 14 PRB_BUNDLING_SIZE_IND:0 bit if the higher layer parameter PRB_bundling is not configured or is set to 'static', or 1 bit if the higher layer parameter PRB_bundling is set to 'dynamic' according to Subclause 5.1.2.3 of [6, TS 38.214] static_bundleSize_t static_prb_BundlingType = pdsch_config.prbBundleType.staticBundling; bundleSizeSet1_t dynamic_prb_BundlingType1 = pdsch_config.prbBundleType.dynamicBundlig.bundleSizeSet1; bundleSizeSet2_t dynamic_prb_BundlingType2 = pdsch_config.prbBundleType.dynamicBundlig.bundleSizeSet2; uint8_t prb_BundlingType_size=0; if ((static_prb_BundlingType==st_n4)||(static_prb_BundlingType==st_wideband)) prb_BundlingType_size=0; if ((dynamic_prb_BundlingType1==dy_1_n4)||(dynamic_prb_BundlingType1==dy_1_wideband)||(dynamic_prb_BundlingType1==dy_1_n2_wideband)||(dynamic_prb_BundlingType1==dy_1_n4_wideband)|| - (dynamic_prb_BundlingType2==dy_2_n4)||(dynamic_prb_BundlingType2==dy_2_wideband)) prb_BundlingType_size=1; + (dynamic_prb_BundlingType2==dy_2_n4)||(dynamic_prb_BundlingType2==dy_2_wideband)) prb_BundlingType_size=1; -// 15 RATE_MATCHING_IND FIXME!!! + // 15 RATE_MATCHING_IND FIXME!!! // according to TS 38.212: Rate matching indicator – 0, 1, or 2 bits according to higher layer parameter rateMatchPattern uint8_t rateMatching_bits = pdsch_config.n_rateMatchPatterns; -// 16 ZP_CSI_RS_TRIGGER FIXME!!! + // 16 ZP_CSI_RS_TRIGGER FIXME!!! // 0, 1, or 2 bits as defined in Subclause 5.1.4.2 of [6, TS 38.214]. // is the number of ZP CSI-RS resource sets in the higher layer parameter zp-CSI-RS-Resource uint8_t n_zp_bits = pdsch_config.n_zp_CSI_RS_ResourceId; -// 17 FREQ_HOPPING_FLAG + // 17 FREQ_HOPPING_FLAG // freqHopping is defined by higher layer parameter frequencyHopping from IE PUSCH-Config. Values are ENUMERATED{mode1, mode2} frequencyHopping_t f_hopping = pusch_config.frequencyHopping; uint8_t freqHopping = 0; if ((f_hopping==f_hop_mode1)||(f_hopping==f_hop_mode2)) freqHopping = 1; -// 28 DAI + // 28 DAI pdsch_HARQ_ACK_Codebook_t pdsch_HARQ_ACK_Codebook = pdsch_config.pdsch_HARQ_ACK_Codebook; uint8_t n_dai = 0; uint8_t n_serving_cell_dl = 1; // this is hardcoded to 1 as we need to get this value from RRC higher layers parameters. FIXME!!! if ((pdsch_HARQ_ACK_Codebook == dynamic) && (n_serving_cell_dl == 1)) n_dai = 2; if ((pdsch_HARQ_ACK_Codebook == dynamic) && (n_serving_cell_dl > 1)) n_dai = 4; -// 29 FIRST_DAI + // 29 FIRST_DAI uint8_t codebook_HARQ_ACK = 0; // We need to get this value to calculate number of bits of fields 1st DAI and 2nd DAI. if (pdsch_HARQ_ACK_Codebook == semiStatic) codebook_HARQ_ACK = 1; if (pdsch_HARQ_ACK_Codebook == dynamic) codebook_HARQ_ACK = 2; -// 30 SECOND_DAI + // 30 SECOND_DAI uint8_t n_HARQ_ACK_sub_codebooks = 0; // We need to get this value to calculate number of bits of fields 1st DAI and 2nd DAI. FIXME!!! -// 35 PDSCH_TO_HARQ_FEEDBACK_TIME_IND + // 35 PDSCH_TO_HARQ_FEEDBACK_TIME_IND uint8_t pdsch_harq_t_ind = (uint8_t)ceil(log2(pucch_config_dedicated.dl_DataToUL_ACK[0])); -// 36 SRS_RESOURCE_IND + // 36 SRS_RESOURCE_IND // n_SRS is the number of configured SRS resources in the SRS resource set associated with the higher layer parameter usage of value 'codeBook' or 'nonCodeBook' // from SRS_ResourceSet_t type we should get the information of the usage parameter (with possible values beamManagement, codebook, nonCodebook, antennaSwitching) // at frame_parms->srs_nr->p_SRS_ResourceSetList[]->usage uint8_t n_SRS = ue->srs.number_srs_Resource_Set; -// 37 PRECOD_NBR_LAYERS -// 38 ANTENNA_PORTS + // 37 PRECOD_NBR_LAYERS + // 38 ANTENNA_PORTS txConfig_t txConfig = pusch_config.txConfig; transformPrecoder_t transformPrecoder = pusch_config.transformPrecoder; codebookSubset_t codebookSubset = pusch_config.codebookSubset; @@ -2374,24 +2239,24 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, if (txConfig == txConfig_codebook){ if (num_antenna_ports == 4) { if ((transformPrecoder == transformPrecoder_disabled) && ((maxRank == 2)||(maxRank == 3)||(maxRank == 4))) { // Table 7.3.1.1.2-2 - if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=6; - if (codebookSubset == codebookSubset_partialAndNonCoherent) precond_nbr_layers_bits=5; - if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=4; + if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=6; + if (codebookSubset == codebookSubset_partialAndNonCoherent) precond_nbr_layers_bits=5; + if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=4; } if (((transformPrecoder == transformPrecoder_enabled)||(transformPrecoder == transformPrecoder_disabled)) && (maxRank == 1)) { // Table 7.3.1.1.2-3 - if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=5; - if (codebookSubset == codebookSubset_partialAndNonCoherent) precond_nbr_layers_bits=4; - if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=2; + if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=5; + if (codebookSubset == codebookSubset_partialAndNonCoherent) precond_nbr_layers_bits=4; + if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=2; } } if (num_antenna_ports == 2) { if ((transformPrecoder == transformPrecoder_disabled) && (maxRank == 2)) { // Table 7.3.1.1.2-4 - if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=4; - if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=2; + if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=4; + if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=2; } if (((transformPrecoder == transformPrecoder_enabled)||(transformPrecoder == transformPrecoder_disabled)) && (maxRank == 1)) { // Table 7.3.1.1.2-5 - if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=3; - if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=1; + if (codebookSubset == codebookSubset_fullyAndPartialAndNonCoherent) precond_nbr_layers_bits=3; + if (codebookSubset == codebookSubset_nonCoherent) precond_nbr_layers_bits=1; } } } @@ -2415,16 +2280,16 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, if((dmrs_DownlinkConfig.pdsch_dmrs_type == pdsch_dmrs_type2) && (dmrs_DownlinkConfig.pdsch_maxLength == pdsch_len1)) antenna_ports_bits_dl = 5; // Table 7.3.1.2.2-3 if((dmrs_DownlinkConfig.pdsch_dmrs_type == pdsch_dmrs_type2) && (dmrs_DownlinkConfig.pdsch_maxLength == pdsch_len2)) antenna_ports_bits_dl = 6; // Table 7.3.1.2.2-4 -// 39 TCI + // 39 TCI uint8_t tci_bits=0; if (pdcch_vars2->coreset[p].tciPresentInDCI == tciPresentInDCI_enabled) tci_bits=3; -// 42 CSI_REQUEST + // 42 CSI_REQUEST // reportTriggerSize is defined in the CSI-MeasConfig IE (TS 38.331). // Size of CSI request field in DCI (bits). Corresponds to L1 parameter 'ReportTriggerSize' (see 38.214, section 5.2) uint8_t reportTriggerSize = csi_MeasConfig.reportTriggerSize; // value from 0..6 -// 43 CBGTI + // 43 CBGTI // for format 0_1 uint8_t maxCodeBlockGroupsPerTransportBlock = 0; if (PUSCH_ServingCellConfig.maxCodeBlockGroupsPerTransportBlock != 0) @@ -2432,17 +2297,17 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, // for format 1_1, as defined in Subclause 5.1.7 of [6, TS38.214] uint8_t maxCodeBlockGroupsPerTransportBlock_dl = 0; if (PDSCH_ServingCellConfig.maxCodeBlockGroupsPerTransportBlock_dl != 0) - maxCodeBlockGroupsPerTransportBlock_dl = pdsch_config.maxNrofCodeWordsScheduledByDCI; // FIXME!!! + maxCodeBlockGroupsPerTransportBlock_dl = pdsch_config.maxNrofCodeWordsScheduledByDCI; // FIXME!!! -// 44 CBGFI + // 44 CBGFI uint8_t cbgfi_bit = PDSCH_ServingCellConfig.codeBlockGroupFlushIndicator; -// 45 PTRS_DMRS + // 45 PTRS_DMRS // 0 bit if PTRS-UplinkConfig is not configured and transformPrecoder=disabled, or if transformPrecoder=enabled, or if maxRank=1 // 2 bits otherwise uint8_t ptrs_dmrs_bits=0; //FIXME!!! -// 46 BETA_OFFSET_IND + // 46 BETA_OFFSET_IND // at IE PUSCH-Config, beta_offset indicator – 0 if the higher layer parameter betaOffsets = semiStatic; otherwise 2 bits // uci-OnPUSCH // Selection between and configuration of dynamic and semi-static beta-offset. If the field is absent or released, the UE applies the value 'semiStatic' and the BetaOffsets @@ -2450,7 +2315,7 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, if (pusch_config.uci_onPusch.betaOffset_type == betaOffset_semiStatic); if (pusch_config.uci_onPusch.betaOffset_type == betaOffset_dynamic) betaOffsets = 2; -// 47 DMRS_SEQ_INI + // 47 DMRS_SEQ_INI uint8_t dmrs_seq_ini_bits_ul = 0; uint8_t dmrs_seq_ini_bits_dl = 0; //1 bit if both scramblingID0 and scramblingID1 are configured in DMRS-UplinkConfig @@ -2458,152 +2323,152 @@ uint16_t nr_dci_format_size (PHY_VARS_NR_UE *ue, //1 bit if both scramblingID0 and scramblingID1 are configured in DMRS-DownlinkConfig if ((dmrs_DownlinkConfig.scramblingID0 != 0) && (dmrs_DownlinkConfig.scramblingID0 != 0)) dmrs_seq_ini_bits_dl = 1; -/* - * For format 2_2 - * - * This format supports power control commands for semi-persistent scheduling. - * As we can already support power control commands dynamically with formats 0_0/0_1 (TPC PUSCH) and 1_0/1_1 (TPC PUCCH) - * - * This format will be implemented in the future FIXME!!! - * - */ -// 5 BLOCK_NUMBER: The parameter tpc-PUSCH or tpc-PUCCH provided by higher layers determines the index to the block number for an UL of a cell -// The following fields are defined for each block: Closed loop indicator and TPC command -// 6 CLOSE_LOOP_IND -// 41 TPC_CMD + /* + * For format 2_2 + * + * This format supports power control commands for semi-persistent scheduling. + * As we can already support power control commands dynamically with formats 0_0/0_1 (TPC PUSCH) and 1_0/1_1 (TPC PUCCH) + * + * This format will be implemented in the future FIXME!!! + * + */ + // 5 BLOCK_NUMBER: The parameter tpc-PUSCH or tpc-PUCCH provided by higher layers determines the index to the block number for an UL of a cell + // The following fields are defined for each block: Closed loop indicator and TPC command + // 6 CLOSE_LOOP_IND + // 41 TPC_CMD uint8_t tpc_cmd_bit_2_2 = 2; -/* - * For format 2_3 - * - * This format is used for power control of uplink sounding reference signals for devices which have not coupled SRS power control to the PUSCH power control - * either because independent control is desirable or because the device is configured without PUCCH and PUSCH - * - * This format will be implemented in the future FIXME!!! - * - */ -// 40 SRS_REQUEST -// 41 TPC_CMD + /* + * For format 2_3 + * + * This format is used for power control of uplink sounding reference signals for devices which have not coupled SRS power control to the PUSCH power control + * either because independent control is desirable or because the device is configured without PUCCH and PUSCH + * + * This format will be implemented in the future FIXME!!! + * + */ + // 40 SRS_REQUEST + // 41 TPC_CMD uint8_t tpc_cmd_bit_2_3 = 0; uint8_t dci_field_size_table [NBR_NR_DCI_FIELDS][NBR_NR_FORMATS] = { // This table contains the number of bits for each field (row) contained in each dci format (column). - // The values of the variables indicate field sizes in number of bits -//Format0_0 Format0_1 Format1_0 Format1_1 Formats2_0/1/2/3 -{1, 1, (((crc_scrambled == _p_rnti) || (crc_scrambled == _si_rnti) || (crc_scrambled == _ra_rnti)) ? 0:1), - 1, 0,0,0,0}, // 0 IDENTIFIER_DCI_FORMATS: -{0, ((crossCarrierSchedulingConfig_ind == 0) ? 0:3), - 0, ((crossCarrierSchedulingConfig_ind == 0) ? 0:3), - 0,0,0,0}, // 1 CARRIER_IND: 0 or 3 bits, as defined in Subclause x.x of [5, TS38.213] -{0, (sul_ind == 0)?0:1, 0, 0, 0,0,0,0}, // 2 SUL_IND_0_1: -{0, 0, 0, 0, 1,0,0,0}, // 3 SLOT_FORMAT_IND: size of DCI format 2_0 is configurable by higher layers up to 128 bits, according to Subclause 11.1.1 of [5, TS 38.213] -{0, 0, 0, 0, 0,1,0,0}, // 4 PRE_EMPTION_IND: size of DCI format 2_1 is configurable by higher layers up to 126 bits, according to Subclause 11.2 of [5, TS 38.213]. Each pre-emption indication is 14 bits -{0, 0, 0, 0, 0,0,0,0}, // 5 BLOCK_NUMBER: starting position of a block is determined by the parameter startingBitOfFormat2_3 -{0, 0, 0, 0, 0,0,1,0}, // 6 CLOSE_LOOP_IND -{0, (uint8_t)ceil(log2(n_UL_BWP_RRC)), - 0, (uint8_t)ceil(log2(n_DL_BWP_RRC)), - 0,0,0,0}, // 7 BANDWIDTH_PART_IND: -{0, 0, ((crc_scrambled == _p_rnti) ? 2:0), - 0, 0,0,0,0}, // 8 SHORT_MESSAGE_IND 2 bits if crc scrambled with P-RNTI -{0, 0, ((crc_scrambled == _p_rnti) ? 8:0), - 0, 0,0,0,0}, // 9 SHORT_MESSAGES 8 bit8 if crc scrambled with P-RNTI -{(uint8_t)(ceil(log2(n_RB_ULBWP*(n_RB_ULBWP+1)/2)))-n_UL_hopping, - n_bits_freq_dom_res_assign_ul, - 0, 0, 0,0,0,0}, // 10 FREQ_DOM_RESOURCE_ASSIGNMENT_UL: PUSCH hopping with resource allocation type 1 not considered - // (NOTE 1) If DCI format 0_0 is monitored in common search space - // and if the number of information bits in the DCI format 0_0 prior to padding - // is larger than the payload size of the DCI format 1_0 monitored in common search space - // the bitwidth of the frequency domain resource allocation field in the DCI format 0_0 - // is reduced such that the size of DCI format 0_0 equals to the size of the DCI format 1_0 -{0, 0, (uint8_t)ceil(log2(n_RB_DLBWP*(n_RB_DLBWP+1)/2)), - n_bits_freq_dom_res_assign_dl, - 0,0,0,0}, // 11 FREQ_DOM_RESOURCE_ASSIGNMENT_DL: -{4, (uint8_t)log2(pusch_alloc_list), - 4, (uint8_t)log2(pdsch_alloc_list), - 0,0,0,0}, // 12 TIME_DOM_RESOURCE_ASSIGNMENT: 0, 1, 2, 3, or 4 bits as defined in Subclause 6.1.2.1 of [6, TS 38.214]. The bitwidth for this field is determined as log2(I) bits, - // where I the number of entries in the higher layer parameter pusch-AllocationList -{0, 0, 1, (((dl_res_alloc_type_0==1)&&(dl_res_alloc_type_1==0))?0:1), - 0,0,0,0}, // 13 VRB_TO_PRB_MAPPING: 0 bit if only resource allocation type 0 -{0, 0, 0, prb_BundlingType_size, 0,0,0,0}, // 14 PRB_BUNDLING_SIZE_IND:0 bit if the higher layer parameter PRB_bundling is not configured or is set to 'static', or 1 bit if the higher layer parameter PRB_bundling is set to 'dynamic' according to Subclause 5.1.2.3 of [6, TS 38.214] -{0, 0, 0, rateMatching_bits, 0,0,0,0}, // 15 RATE_MATCHING_IND: 0, 1, or 2 bits according to higher layer parameter rate-match-PDSCH-resource-set -{0, 0, 0, n_zp_bits, 0,0,0,0}, // 16 ZP_CSI_RS_TRIGGER: -{1, (((ul_res_alloc_type_0==1)&&(ul_res_alloc_type_1==0))||(freqHopping == 0))?0:1, - 0, 0, 0,0,0,0}, // 17 FREQ_HOPPING_FLAG: 0 bit if only resource allocation type 0 -{0, 0, 0, 5, 0,0,0,0}, // 18 TB1_MCS: -{0, 0, 0, 1, 0,0,0,0}, // 19 TB1_NDI: -{0, 0, 0, 2, 0,0,0,0}, // 20 TB1_RV: -{0, 0, 0, 5, 0,0,0,0}, // 21 TB2_MCS: -{0, 0, 0, 1, 0,0,0,0}, // 22 TB2_NDI: -{0, 0, 0, 2, 0,0,0,0}, // 23 TB2_RV: -{5, 5, 5, 0, 0,0,0,0}, // 24 MCS: -{1, 1, (crc_scrambled == _c_rnti)?1:0,0, 0,0,0,0}, // 25 NDI: -{2, 2, (((crc_scrambled == _c_rnti) || (crc_scrambled == _si_rnti)) ? 2:0), - 0, 0,0,0,0}, // 26 RV: -{4, 4, (crc_scrambled == _c_rnti)?4:0,4, 0,0,0,0}, // 27 HARQ_PROCESS_NUMBER: -{0, 0, (crc_scrambled == _c_rnti)?2:0,n_dai, 0,0,0,0}, // 28 DAI: For format1_1: 4 if more than one serving cell are configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 MSB bits are the counter DAI and the 2 LSB bits are the total DAI - // 2 if one serving cell is configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 bits are the counter DAI - // 0 otherwise -{0, codebook_HARQ_ACK, 0, 0, 0,0,0,0}, // 29 FIRST_DAI: (1 or 2 bits) 1 bit for semi-static HARQ-ACK // 2 bits for dynamic HARQ-ACK codebook with single HARQ-ACK codebook -{0, (((codebook_HARQ_ACK == 2) && (n_HARQ_ACK_sub_codebooks==2))?2:0), - 0, 0, 0,0,0,0}, // 30 SECOND_DAI: (0 or 2 bits) 2 bits for dynamic HARQ-ACK codebook with two HARQ-ACK sub-codebooks // 0 bits otherwise -{0, 0, (((crc_scrambled == _p_rnti) || (crc_scrambled == _ra_rnti)) ? 2:0), - 0, 0,0,0,0}, // 31 TB_SCALING -{2, 2, 0, 0, 0,0,0,0}, // 32 TPC_PUSCH: -{0, 0, (crc_scrambled == _c_rnti)?2:0,2, 0,0,0,0}, // 33 TPC_PUCCH: -{0, 0, (crc_scrambled == _c_rnti)?3:0,3, 0,0,0,0}, // 34 PUCCH_RESOURCE_IND: -{0, 0, (crc_scrambled == _c_rnti)?3:0,pdsch_harq_t_ind, 0,0,0,0}, // 35 PDSCH_TO_HARQ_FEEDBACK_TIME_IND: -{0, (uint8_t)log2(n_SRS), 0, 0, 0,0,0,0}, // 36 SRS_RESOURCE_IND: -{0, precond_nbr_layers_bits, 0, 0, 0,0,0,0}, // 37 PRECOD_NBR_LAYERS: -{0, antenna_ports_bits_ul, 0, antenna_ports_bits_dl, 0,0,0,0}, // 38 ANTENNA_PORTS: -{0, 0, 0, tci_bits, 0,0,0,0}, // 39 TCI: 0 bit if higher layer parameter tci-PresentInDCI is not enabled; otherwise 3 bits -{0, (sul_ind == 0)?2:3, 0, (sul_ind == 0)?2:3, 0,0,0,2}, // 40 SRS_REQUEST: -{0, 0, 0, 0, 0,0,tpc_cmd_bit_2_2, - tpc_cmd_bit_2_3}, - // 41 TPC_CMD: -{0, reportTriggerSize, 0, 0, 0,0,0,0}, // 42 CSI_REQUEST: -{0, maxCodeBlockGroupsPerTransportBlock, - 0, maxCodeBlockGroupsPerTransportBlock_dl, - 0,0,0,0}, // 43 CBGTI: 0, 2, 4, 6, or 8 bits determined by higher layer parameter maxCodeBlockGroupsPerTransportBlock for the PDSCH -{0, 0, 0, cbgfi_bit, 0,0,0,0}, // 44 CBGFI: 0 or 1 bit determined by higher layer parameter codeBlockGroupFlushIndicator -{0, ptrs_dmrs_bits, 0, 0, 0,0,0,0}, // 45 PTRS_DMRS: -{0, betaOffsets, 0, 0, 0,0,0,0}, // 46 BETA_OFFSET_IND: -{0, dmrs_seq_ini_bits_ul, 0, dmrs_seq_ini_bits_dl, 0,0,0,0}, // 47 DMRS_SEQ_INI: 1 bit if the cell has two ULs and the number of bits for DCI format 1_0 before padding - // is larger than the number of bits for DCI format 0_0 before padding; 0 bit otherwise -{0, 1, 0, 0, 0,0,0,0}, // 48 UL_SCH_IND: value of "1" indicates UL-SCH shall be transmitted on the PUSCH and a value of "0" indicates UL-SCH shall not be transmitted on the PUSCH -{0, 0, 0, 0, 0,0,0,0}, // 49 PADDING_NR_DCI: - // (NOTE 2) If DCI format 0_0 is monitored in common search space - // and if the number of information bits in the DCI format 0_0 prior to padding - // is less than the payload size of the DCI format 1_0 monitored in common search space - // zeros shall be appended to the DCI format 0_0 - // until the payload size equals that of the DCI format 1_0 -{(sul_ind == 0)?0:1, 0, 0, 0, 0,0,0,0}, // 50 SUL_IND_0_0: -{0, 0, 0, 0, 0,0,0,0}, // 51 RA_PREAMBLE_INDEX (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) -{0, 0, 0, 0, 0,0,0,0}, // 52 SUL_IND_1_0 (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) -{0, 0, 0, 0, 0,0,0,0}, // 53 SS_PBCH_INDEX (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) -{0, 0, 0, 0, 0,0,0,0}, // 54 PRACH_MASK_INDEX (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) -{0, 0, ((crc_scrambled == _p_rnti)?6:(((crc_scrambled == _si_rnti) || (crc_scrambled == _ra_rnti))?16:0)), - 0, 0,0,0,0} // 55 RESERVED_NR_DCI -}; + // The values of the variables indicate field sizes in number of bits + //Format0_0 Format0_1 Format1_0 Format1_1 Formats2_0/1/2/3 + {1, 1, (((crc_scrambled == _p_rnti) || (crc_scrambled == _si_rnti) || (crc_scrambled == _ra_rnti)) ? 0:1), + 1, 0,0,0,0}, // 0 IDENTIFIER_DCI_FORMATS: + {0, ((crossCarrierSchedulingConfig_ind == 0) ? 0:3), + 0, ((crossCarrierSchedulingConfig_ind == 0) ? 0:3), + 0,0,0,0}, // 1 CARRIER_IND: 0 or 3 bits, as defined in Subclause x.x of [5, TS38.213] + {0, (sul_ind == 0)?0:1, 0, 0, 0,0,0,0}, // 2 SUL_IND_0_1: + {0, 0, 0, 0, 1,0,0,0}, // 3 SLOT_FORMAT_IND: size of DCI format 2_0 is configurable by higher layers up to 128 bits, according to Subclause 11.1.1 of [5, TS 38.213] + {0, 0, 0, 0, 0,1,0,0}, // 4 PRE_EMPTION_IND: size of DCI format 2_1 is configurable by higher layers up to 126 bits, according to Subclause 11.2 of [5, TS 38.213]. Each pre-emption indication is 14 bits + {0, 0, 0, 0, 0,0,0,0}, // 5 BLOCK_NUMBER: starting position of a block is determined by the parameter startingBitOfFormat2_3 + {0, 0, 0, 0, 0,0,1,0}, // 6 CLOSE_LOOP_IND + {0, (uint8_t)ceil(log2(n_UL_BWP_RRC)), + 0, (uint8_t)ceil(log2(n_DL_BWP_RRC)), + 0,0,0,0}, // 7 BANDWIDTH_PART_IND: + {0, 0, ((crc_scrambled == _p_rnti) ? 2:0), + 0, 0,0,0,0}, // 8 SHORT_MESSAGE_IND 2 bits if crc scrambled with P-RNTI + {0, 0, ((crc_scrambled == _p_rnti) ? 8:0), + 0, 0,0,0,0}, // 9 SHORT_MESSAGES 8 bit8 if crc scrambled with P-RNTI + {(uint8_t)(ceil(log2(n_RB_ULBWP*(n_RB_ULBWP+1)/2)))-n_UL_hopping, + n_bits_freq_dom_res_assign_ul, + 0, 0, 0,0,0,0}, // 10 FREQ_DOM_RESOURCE_ASSIGNMENT_UL: PUSCH hopping with resource allocation type 1 not considered + // (NOTE 1) If DCI format 0_0 is monitored in common search space + // and if the number of information bits in the DCI format 0_0 prior to padding + // is larger than the payload size of the DCI format 1_0 monitored in common search space + // the bitwidth of the frequency domain resource allocation field in the DCI format 0_0 + // is reduced such that the size of DCI format 0_0 equals to the size of the DCI format 1_0 + {0, 0, (uint8_t)ceil(log2(n_RB_DLBWP*(n_RB_DLBWP+1)/2)), + n_bits_freq_dom_res_assign_dl, + 0,0,0,0}, // 11 FREQ_DOM_RESOURCE_ASSIGNMENT_DL: + {4, (uint8_t)log2(pusch_alloc_list), + 4, (uint8_t)log2(pdsch_alloc_list), + 0,0,0,0}, // 12 TIME_DOM_RESOURCE_ASSIGNMENT: 0, 1, 2, 3, or 4 bits as defined in Subclause 6.1.2.1 of [6, TS 38.214]. The bitwidth for this field is determined as log2(I) bits, + // where I the number of entries in the higher layer parameter pusch-AllocationList + {0, 0, 1, (((dl_res_alloc_type_0==1)&&(dl_res_alloc_type_1==0))?0:1), + 0,0,0,0}, // 13 VRB_TO_PRB_MAPPING: 0 bit if only resource allocation type 0 + {0, 0, 0, prb_BundlingType_size, 0,0,0,0}, // 14 PRB_BUNDLING_SIZE_IND:0 bit if the higher layer parameter PRB_bundling is not configured or is set to 'static', or 1 bit if the higher layer parameter PRB_bundling is set to 'dynamic' according to Subclause 5.1.2.3 of [6, TS 38.214] + {0, 0, 0, rateMatching_bits, 0,0,0,0}, // 15 RATE_MATCHING_IND: 0, 1, or 2 bits according to higher layer parameter rate-match-PDSCH-resource-set + {0, 0, 0, n_zp_bits, 0,0,0,0}, // 16 ZP_CSI_RS_TRIGGER: + {1, (((ul_res_alloc_type_0==1)&&(ul_res_alloc_type_1==0))||(freqHopping == 0))?0:1, + 0, 0, 0,0,0,0}, // 17 FREQ_HOPPING_FLAG: 0 bit if only resource allocation type 0 + {0, 0, 0, 5, 0,0,0,0}, // 18 TB1_MCS: + {0, 0, 0, 1, 0,0,0,0}, // 19 TB1_NDI: + {0, 0, 0, 2, 0,0,0,0}, // 20 TB1_RV: + {0, 0, 0, 5, 0,0,0,0}, // 21 TB2_MCS: + {0, 0, 0, 1, 0,0,0,0}, // 22 TB2_NDI: + {0, 0, 0, 2, 0,0,0,0}, // 23 TB2_RV: + {5, 5, 5, 0, 0,0,0,0}, // 24 MCS: + {1, 1, (crc_scrambled == _c_rnti)?1:0,0, 0,0,0,0}, // 25 NDI: + {2, 2, (((crc_scrambled == _c_rnti) || (crc_scrambled == _si_rnti)) ? 2:0), + 0, 0,0,0,0}, // 26 RV: + {4, 4, (crc_scrambled == _c_rnti)?4:0,4, 0,0,0,0}, // 27 HARQ_PROCESS_NUMBER: + {0, 0, (crc_scrambled == _c_rnti)?2:0,n_dai, 0,0,0,0}, // 28 DAI: For format1_1: 4 if more than one serving cell are configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 MSB bits are the counter DAI and the 2 LSB bits are the total DAI + // 2 if one serving cell is configured in the DL and the higher layer parameter HARQ-ACK-codebook=dynamic, where the 2 bits are the counter DAI + // 0 otherwise + {0, codebook_HARQ_ACK, 0, 0, 0,0,0,0}, // 29 FIRST_DAI: (1 or 2 bits) 1 bit for semi-static HARQ-ACK // 2 bits for dynamic HARQ-ACK codebook with single HARQ-ACK codebook + {0, (((codebook_HARQ_ACK == 2) && (n_HARQ_ACK_sub_codebooks==2))?2:0), + 0, 0, 0,0,0,0}, // 30 SECOND_DAI: (0 or 2 bits) 2 bits for dynamic HARQ-ACK codebook with two HARQ-ACK sub-codebooks // 0 bits otherwise + {0, 0, (((crc_scrambled == _p_rnti) || (crc_scrambled == _ra_rnti)) ? 2:0), + 0, 0,0,0,0}, // 31 TB_SCALING + {2, 2, 0, 0, 0,0,0,0}, // 32 TPC_PUSCH: + {0, 0, (crc_scrambled == _c_rnti)?2:0,2, 0,0,0,0}, // 33 TPC_PUCCH: + {0, 0, (crc_scrambled == _c_rnti)?3:0,3, 0,0,0,0}, // 34 PUCCH_RESOURCE_IND: + {0, 0, (crc_scrambled == _c_rnti)?3:0,pdsch_harq_t_ind, 0,0,0,0}, // 35 PDSCH_TO_HARQ_FEEDBACK_TIME_IND: + {0, (uint8_t)log2(n_SRS), 0, 0, 0,0,0,0}, // 36 SRS_RESOURCE_IND: + {0, precond_nbr_layers_bits, 0, 0, 0,0,0,0}, // 37 PRECOD_NBR_LAYERS: + {0, antenna_ports_bits_ul, 0, antenna_ports_bits_dl, 0,0,0,0}, // 38 ANTENNA_PORTS: + {0, 0, 0, tci_bits, 0,0,0,0}, // 39 TCI: 0 bit if higher layer parameter tci-PresentInDCI is not enabled; otherwise 3 bits + {0, (sul_ind == 0)?2:3, 0, (sul_ind == 0)?2:3, 0,0,0,2}, // 40 SRS_REQUEST: + {0, 0, 0, 0, 0,0,tpc_cmd_bit_2_2, + tpc_cmd_bit_2_3}, + // 41 TPC_CMD: + {0, reportTriggerSize, 0, 0, 0,0,0,0}, // 42 CSI_REQUEST: + {0, maxCodeBlockGroupsPerTransportBlock, + 0, maxCodeBlockGroupsPerTransportBlock_dl, + 0,0,0,0}, // 43 CBGTI: 0, 2, 4, 6, or 8 bits determined by higher layer parameter maxCodeBlockGroupsPerTransportBlock for the PDSCH + {0, 0, 0, cbgfi_bit, 0,0,0,0}, // 44 CBGFI: 0 or 1 bit determined by higher layer parameter codeBlockGroupFlushIndicator + {0, ptrs_dmrs_bits, 0, 0, 0,0,0,0}, // 45 PTRS_DMRS: + {0, betaOffsets, 0, 0, 0,0,0,0}, // 46 BETA_OFFSET_IND: + {0, dmrs_seq_ini_bits_ul, 0, dmrs_seq_ini_bits_dl, 0,0,0,0}, // 47 DMRS_SEQ_INI: 1 bit if the cell has two ULs and the number of bits for DCI format 1_0 before padding + // is larger than the number of bits for DCI format 0_0 before padding; 0 bit otherwise + {0, 1, 0, 0, 0,0,0,0}, // 48 UL_SCH_IND: value of "1" indicates UL-SCH shall be transmitted on the PUSCH and a value of "0" indicates UL-SCH shall not be transmitted on the PUSCH + {0, 0, 0, 0, 0,0,0,0}, // 49 PADDING_NR_DCI: + // (NOTE 2) If DCI format 0_0 is monitored in common search space + // and if the number of information bits in the DCI format 0_0 prior to padding + // is less than the payload size of the DCI format 1_0 monitored in common search space + // zeros shall be appended to the DCI format 0_0 + // until the payload size equals that of the DCI format 1_0 + {(sul_ind == 0)?0:1, 0, 0, 0, 0,0,0,0}, // 50 SUL_IND_0_0: + {0, 0, 0, 0, 0,0,0,0}, // 51 RA_PREAMBLE_INDEX (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) + {0, 0, 0, 0, 0,0,0,0}, // 52 SUL_IND_1_0 (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) + {0, 0, 0, 0, 0,0,0,0}, // 53 SS_PBCH_INDEX (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) + {0, 0, 0, 0, 0,0,0,0}, // 54 PRACH_MASK_INDEX (random access procedure initiated by a PDCCH order not implemented, FIXME!!!) + {0, 0, ((crc_scrambled == _p_rnti)?6:(((crc_scrambled == _si_rnti) || (crc_scrambled == _ra_rnti))?16:0)), + 0, 0,0,0,0} // 55 RESERVED_NR_DCI + }; -// NOTE 1: adjustments in freq_dom_resource_assignment_UL to be done if necessary -// NOTE 2: adjustments in padding to be done if necessary + // NOTE 1: adjustments in freq_dom_resource_assignment_UL to be done if necessary + // NOTE 2: adjustments in padding to be done if necessary -uint8_t dci_size [8] = {0,0,0,0,0,0,0,0}; // will contain size for each format + uint8_t dci_size [8] = {0,0,0,0,0,0,0,0}; // will contain size for each format for (int i=0 ; i i=%d, j=%d\n", i, j); -//#endif + //#ifdef NR_PDCCH_DCI_DEBUG + // printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_format_size)-> i=%d, j=%d\n", i, j); + //#endif for (int j=0; j position in table is 2) dci_fields_sizes[j][i] = dci_field_size_table[j][i]; // dci_fields_sizes[j] contains the sizes of each field (j) for a determined format i //} } - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_format_size) dci_size[%d]=%d for n_RB_ULBWP=%d\n", - i,dci_size[i],n_RB_ULBWP); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_format_size) dci_size[%d]=%d for n_RB_ULBWP=%d\n", + i,dci_size[i],n_RB_ULBWP); +#endif } -#ifndef NR_PDCCH_DCI_DEBUG +#ifdef NR_PDCCH_DCI_DEBUG printf("\t\t<-NR_PDCCH_DCI_DEBUG (nr_dci_format_size) dci_fields_sizes[][] = { \n"); for (int j=0; jframe_parms; - t_nrPolar_paramsPtr *nrPolar_params = &ue->nrPolar_params; //uint8_t mi;// = get_mi(&ue->frame_parms, nr_tti_rx); //uint8_t tmode = ue->transmission_mode[eNB_id]; @@ -2766,25 +2632,25 @@ uint8_t nr_dci_decoding_procedure(int s, */ //dci_detect_mode_t mode = 3; //dci_detect_mode_select(&ue->frame_parms, nr_tti_rx); - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> searSpaceType=%d\n",do_common); - if (do_common==0) { - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> css_dci_format=%d\n",css_dci_format); - } else { - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> uss_dci_format=%d\n",uss_dci_format); - } - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> searSpaceType=%d\n",do_common); + if (do_common==0) { + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> css_dci_format=%d\n",css_dci_format); + } else { + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> uss_dci_format=%d\n",uss_dci_format); + } +#endif // A set of PDCCH candidates for a UE to monitor is defined in terms of PDCCH search spaces if (do_common==0) { // COMMON SearchSpaceType assigned to current SearchSpace/CORESET // Type0-PDCCH common search space for a DCI format with CRC scrambled by a SI-RNTI - // number of consecutive resource blocks and a number of consecutive symbols for - // the control resource set of the Type0-PDCCH common search space from - // the four most significant bits of RMSI-PDCCH-Config as described in Tables 13-1 through 13-10 - // and determines PDCCH monitoring occasions - // from the four least significant bits of RMSI-PDCCH-Config, - // included in MasterInformationBlock, as described in Tables 13-11 through 13-15 + // number of consecutive resource blocks and a number of consecutive symbols for + // the control resource set of the Type0-PDCCH common search space from + // the four most significant bits of RMSI-PDCCH-Config as described in Tables 13-1 through 13-10 + // and determines PDCCH monitoring occasions + // from the four least significant bits of RMSI-PDCCH-Config, + // included in MasterInformationBlock, as described in Tables 13-11 through 13-15 // Type0A-PDCCH common search space for a DCI format with CRC scrambled by a SI-RNTI // Type1-PDCCH common search space for a DCI format with CRC scrambled by a RA-RNTI, or a TC-RNTI, or a C-RNTI // Type2-PDCCH common search space for a DCI format with CRC scrambled by a P-RNTI @@ -2805,32 +2671,32 @@ uint8_t nr_dci_decoding_procedure(int s, // for format0_0 and format1_0, first we calculate dci pdu size format_0_0_1_0_size_bits = nr_dci_format_size(ue,eNB_id,nr_tti_rx,p,_c_rnti,n_RB_ULBWP,n_RB_DLBWP,dci_fields_sizes,0); format_0_0_1_0_size_bytes = (format_0_0_1_0_size_bits%8 == 0) ? (uint8_t)floor(format_0_0_1_0_size_bits/8) : (uint8_t)(floor(format_0_0_1_0_size_bits/8) + 1); - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for common searchSpaces with format css_dci_format=%d, format_0_0_1_0_size_bits=%d, format_0_0_1_0_size_bytes=%d\n", - css_dci_format,format_0_0_1_0_size_bits,format_0_0_1_0_size_bytes); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for common searchSpaces with format css_dci_format=%d, format_0_0_1_0_size_bits=%d, format_0_0_1_0_size_bytes=%d\n", + css_dci_format,format_0_0_1_0_size_bits,format_0_0_1_0_size_bytes); +#endif for (int aggregationLevel = 3; aggregationLevel<4 ; aggregationLevel++) { // We fix aggregationLevel to 3 for testing=> nbr of CCE=8 - //for (int aggregationLevel = 2; aggregationLevel<5 ; aggregationLevel++) { - // for aggregation level aggregationLevel. The number of candidates (for L2= 2^aggregationLevel) will be calculated in function nr_dci_decoding_procedure0 - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", - css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, nrPolar_params, - crc_scrambled_values, aggregationLevel, - cformat0_0_and_1_0, uformat0_0_and_1_0, - format_0_0_1_0_size_bits, format_0_0_1_0_size_bytes, &dci_cnt, - &crc_scrambled_, &format_found_, pdcch_DMRS_scrambling_id,&CCEmap0, &CCEmap1, &CCEmap2); - if (dci_cnt != old_dci_cnt){ - // we will exit the loop as we have found the DCI - aggregationLevel = 5; - format_0_0_1_0_size_bits = nr_dci_format_size(ue,eNB_id,nr_tti_rx,p,crc_scrambled_,n_RB_ULBWP,n_RB_DLBWP,dci_fields_sizes,0); // after decoding dci successfully we recalculate dci pdu size with correct crc scrambled to get the right field sizes - old_dci_cnt = dci_cnt; - for (int i=0; i common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", + css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, + crc_scrambled_values, aggregationLevel, + cformat0_0_and_1_0, uformat0_0_and_1_0, + format_0_0_1_0_size_bits, format_0_0_1_0_size_bytes, &dci_cnt, + &crc_scrambled_, &format_found_, pdcch_DMRS_scrambling_id,&CCEmap0, &CCEmap1, &CCEmap2); + if (dci_cnt != old_dci_cnt){ + // we will exit the loop as we have found the DCI + aggregationLevel = 5; + format_0_0_1_0_size_bits = nr_dci_format_size(ue,eNB_id,nr_tti_rx,p,crc_scrambled_,n_RB_ULBWP,n_RB_DLBWP,dci_fields_sizes,0); // after decoding dci successfully we recalculate dci pdu size with correct crc scrambled to get the right field sizes + old_dci_cnt = dci_cnt; + for (int i=0; i calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_0_size_bits=%d, format2_0_size_bytes=%d\n", - css_dci_format,format_2_0_size_bits,format_2_0_size_bytes); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_0_size_bits=%d, format2_0_size_bytes=%d\n", + css_dci_format,format_2_0_size_bits,format_2_0_size_bytes); +#endif for (int aggregationLevelSFI = 0; aggregationLevelSFI<5 ; aggregationLevelSFI++){ - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", - css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, nrPolar_params, - crc_scrambled_values, aggregationLevelSFI, - cformat2_0, uformat0_0_and_1_0, - format_2_0_size_bits, format_2_0_size_bytes, &dci_cnt, - &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); - if (dci_cnt != old_dci_cnt){ - // we will exit the loop as we have found the DCI - aggregationLevelSFI = 5; - old_dci_cnt = dci_cnt; - for (int i=0; i common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", + css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, + crc_scrambled_values, aggregationLevelSFI, + cformat2_0, uformat0_0_and_1_0, + format_2_0_size_bits, format_2_0_size_bytes, &dci_cnt, + &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); + if (dci_cnt != old_dci_cnt){ + // we will exit the loop as we have found the DCI + aggregationLevelSFI = 5; + old_dci_cnt = dci_cnt; + for (int i=0; i calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_1_size_bits=%d, format2_1_size_bytes=%d\n", - css_dci_format,format_2_1_size_bits,format_2_1_size_bytes); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_1_size_bits=%d, format2_1_size_bytes=%d\n", + css_dci_format,format_2_1_size_bits,format_2_1_size_bytes); +#endif for (int aggregationLevel = 0; aggregationLevel<5 ; aggregationLevel++){ - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", - css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, nrPolar_params, - crc_scrambled_values, aggregationLevel, - cformat2_1, uformat0_0_and_1_0, - format_2_1_size_bits, format_2_1_size_bytes, &dci_cnt, - &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); - if (dci_cnt != old_dci_cnt){ - // we will exit the loop as we have found the DCI - aggregationLevel = 5; - old_dci_cnt = dci_cnt; - for (int i=0; i common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", + css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, + crc_scrambled_values, aggregationLevel, + cformat2_1, uformat0_0_and_1_0, + format_2_1_size_bits, format_2_1_size_bytes, &dci_cnt, + &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); + if (dci_cnt != old_dci_cnt){ + // we will exit the loop as we have found the DCI + aggregationLevel = 5; + old_dci_cnt = dci_cnt; + for (int i=0; i calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_2_size_bits=%d, format2_2_size_bytes=%d\n", - css_dci_format,format_2_2_size_bits,format_2_2_size_bytes); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_2_size_bits=%d, format2_2_size_bytes=%d\n", + css_dci_format,format_2_2_size_bits,format_2_2_size_bytes); +#endif for (int aggregationLevel = 0; aggregationLevel<5 ; aggregationLevel++){ - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", - css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, nrPolar_params, - crc_scrambled_values, aggregationLevel, - cformat2_2, uformat0_0_and_1_0, - format_2_2_size_bits, format_2_2_size_bytes, &dci_cnt, - &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); - if (dci_cnt != old_dci_cnt){ - // we will exit the loop as we have found the DCI - aggregationLevel = 5; - old_dci_cnt = dci_cnt; - for (int i=0; i common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", + css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, + crc_scrambled_values, aggregationLevel, + cformat2_2, uformat0_0_and_1_0, + format_2_2_size_bits, format_2_2_size_bytes, &dci_cnt, + &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); + if (dci_cnt != old_dci_cnt){ + // we will exit the loop as we have found the DCI + aggregationLevel = 5; + old_dci_cnt = dci_cnt; + for (int i=0; i calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_3_size_bits=%d, format2_3_size_bytes=%d\n", - css_dci_format,format_2_3_size_bits,format_2_3_size_bytes); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for common searchSpaces with format css_dci_format=%d, format2_3_size_bits=%d, format2_3_size_bytes=%d\n", + css_dci_format,format_2_3_size_bits,format_2_3_size_bytes); +#endif for (int aggregationLevel = 0; aggregationLevel<5 ; aggregationLevel++){ - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", - css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, nrPolar_params, - crc_scrambled_values, aggregationLevel, - cformat2_3, uformat0_0_and_1_0, - format_2_3_size_bits, format_2_3_size_bytes, &dci_cnt, - &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); - if (dci_cnt != old_dci_cnt){ - // we will exit the loop as we have found the DCI - aggregationLevel = 5; - old_dci_cnt = dci_cnt; - for (int i=0; i common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", + css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, + crc_scrambled_values, aggregationLevel, + cformat2_3, uformat0_0_and_1_0, + format_2_3_size_bits, format_2_3_size_bytes, &dci_cnt, + &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); + if (dci_cnt != old_dci_cnt){ + // we will exit the loop as we have found the DCI + aggregationLevel = 5; + old_dci_cnt = dci_cnt; + for (int i=0; i calculating dci format size for UE-specific searchSpaces with format uss_dci_format=%d, format_0_0_1_0_size_bits=%d, format_0_0_1_0_size_bytes=%d\n", - css_dci_format,format_0_0_1_0_size_bits,format_0_0_1_0_size_bytes); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for UE-specific searchSpaces with format uss_dci_format=%d, format_0_0_1_0_size_bits=%d, format_0_0_1_0_size_bytes=%d\n", + css_dci_format,format_0_0_1_0_size_bits,format_0_0_1_0_size_bytes); +#endif for (int aggregationLevel = 0; aggregationLevel<5 ; aggregationLevel++) { // We fix aggregationLevel to 3 for testing=> nbr of CCE=8 - //for (int aggregationLevel = 2; aggregationLevel<5 ; aggregationLevel++) { - // for aggregation level aggregationLevel. The number of candidates (for L2= 2^aggregationLevel) will be calculated in function nr_dci_decoding_procedure0 - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", - css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms,nrPolar_params, - crc_scrambled_values, aggregationLevel, - cformat0_0_and_1_0, uformat0_0_and_1_0, - format_0_0_1_0_size_bits, format_0_0_1_0_size_bytes, &dci_cnt, - &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); - if (dci_cnt != old_dci_cnt){ - // we will exit the loop as we have found the DCI - aggregationLevel = 5; - old_dci_cnt = dci_cnt; - for (int i=0; i common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", + css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, + crc_scrambled_values, aggregationLevel, + cformat0_0_and_1_0, uformat0_0_and_1_0, + format_0_0_1_0_size_bits, format_0_0_1_0_size_bytes, &dci_cnt, + &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); + if (dci_cnt != old_dci_cnt){ + // we will exit the loop as we have found the DCI + aggregationLevel = 5; + old_dci_cnt = dci_cnt; + for (int i=0; i calculating dci format size for UE-specific searchSpaces with format uss_dci_format=%d, format_0_1_1_1_size_bits=%d, format_0_1_1_1_size_bytes=%d\n", - css_dci_format,format_0_1_1_1_size_bits,format_0_1_1_1_size_bytes); - #endif +#ifdef NR_PDCCH_DCI_DEBUG + printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> calculating dci format size for UE-specific searchSpaces with format uss_dci_format=%d, format_0_1_1_1_size_bits=%d, format_0_1_1_1_size_bytes=%d\n", + css_dci_format,format_0_1_1_1_size_bits,format_0_1_1_1_size_bytes); +#endif for (int aggregationLevel = 0; aggregationLevel<5 ; aggregationLevel++) { // We fix aggregationLevel to 3 for testing=> nbr of CCE=8 - //for (int aggregationLevel = 2; aggregationLevel<5 ; aggregationLevel++) { - // for aggregation level aggregationLevel. The number of candidates (for L2= 2^aggregationLevel) will be calculated in function nr_dci_decoding_procedure0 - #ifdef NR_PDCCH_DCI_DEBUG - printf("\t<-NR_PDCCH_DCI_DEBUG (nr_dci_decoding_procedure)-> common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", - css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms,nrPolar_params, - crc_scrambled_values, aggregationLevel, - cformat0_0_and_1_0, uformat0_1_and_1_1, - format_0_1_1_1_size_bits, format_0_1_1_1_size_bytes, &dci_cnt, - &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); - if (dci_cnt != old_dci_cnt){ - // we will exit the loop as we have found the DCI - aggregationLevel = 5; - old_dci_cnt = dci_cnt; - for (int i=0; i common searchSpaces with format css_dci_format=%d and aggregation_level=%d\n", + css_dci_format,(1<current_thread_id[nr_tti_rx], frame_parms, + crc_scrambled_values, aggregationLevel, + cformat0_0_and_1_0, uformat0_1_and_1_1, + format_0_1_1_1_size_bits, format_0_1_1_1_size_bytes, &dci_cnt, + &crc_scrambled_, &format_found_,pdcch_DMRS_scrambling_id, &CCEmap0, &CCEmap1, &CCEmap2); + if (dci_cnt != old_dci_cnt){ + // we will exit the loop as we have found the DCI + aggregationLevel = 5; + old_dci_cnt = dci_cnt; + for (int i=0; i at the end crc_scrambled=%d and format_found=%d\n",*crc_scrambled,*format_found); #endif diff --git a/openair1/PHY/NR_UE_TRANSPORT/dci_tools_nr.c b/openair1/PHY/NR_UE_TRANSPORT/dci_tools_nr.c index 61e40b82e4..1eafbfb7dc 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/dci_tools_nr.c +++ b/openair1/PHY/NR_UE_TRANSPORT/dci_tools_nr.c @@ -30,6 +30,7 @@ * \warning */ //#include "PHY/defs.h" +#include #include "PHY/defs_nr_UE.h" //#include "PHY/NR_TRANSPORT/nr_dci.h" //#include "PHY/NR_UE_TRANSPORT/nr_transport_ue.h" @@ -67,18 +68,28 @@ uint16_t nr_dci_field(uint32_t dci_pdu[4], uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS], uint8_t dci_field) { - uint16_t field_value = 0 ; + // uint16_t field_value = 0 ; // first_bit_position contains the position of the first bit of the corresponding field within the dci pdu payload - uint16_t first_bit_position = 0; + // last_bit_position contains the position of the last bit of the corresponding field within the dci pdu payload - uint16_t last_bit_position = 0; - uint8_t bit=0; + // uint16_t last_bit_position = 0; + // uint8_t bit=0; //printf("\tdci_field=%d, \tsize=%d \t|",dci_field,dci_fields_sizes[dci_field]); - for (int i=0; i 64 bits, not supported for now\n", + dci_size); + + uint16_t first_bit_position = dci_size; + + for (int i=0; i<=dci_field ; i++){ + first_bit_position = first_bit_position - dci_fields_sizes[i]; } - last_bit_position = first_bit_position + dci_fields_sizes[dci_field]; + // last_bit_position = first_bit_position + dci_fields_sizes[dci_field]; //printf("\tfirst_bit=%d,\tlast_bit=%d",first_bit_position,last_bit_position); + /* for (int i=0; i<4; i++) for (int j=0; j<32; j++){ if ((((i*32)+j) >= first_bit_position) && (((i*32)+j) < last_bit_position)){ @@ -87,7 +98,14 @@ uint16_t nr_dci_field(uint32_t dci_pdu[4], //printf(" bit(%d)=%d[%d] ",(i*32)+j,bit,field_value); } } - return field_value; + */ + + uint64_t *dci_pdu64=(uint64_t*)&dci_pdu[0]; + /* + printf("pdu %llx, field %d, pos %d, size %d => %u\n",(long long unsigned int)*dci_pdu64,dci_field,first_bit_position,dci_fields_sizes[dci_field], + (unsigned int)((*dci_pdu64>>first_bit_position)&((1<>first_bit_position)&((1<Mdlharq; i++) { + if (dlsch->harq_processes[i]) { + if (dlsch->harq_processes[i]->b) { + free16(dlsch->harq_processes[i]->b,MAX_DLSCH_PAYLOAD_BYTES); + dlsch->harq_processes[i]->b = NULL; + } + + for (r=0; rharq_processes[i]->c[r],1056); + dlsch->harq_processes[i]->c[r] = NULL; + } + + for (r=0; rharq_processes[i]->d[r]) { + free16(dlsch->harq_processes[i]->d[r],(3*8448)*sizeof(short)); + dlsch->harq_processes[i]->d[r] = NULL; + } + + for (r=0; r<(MAX_NUM_NR_DLSCH_SEGMENTS); r++) { + if (dlsch->harq_processes[i]->p_nrLDPC_procBuf[r]){ + nrLDPC_free_mem(dlsch->harq_processes[i]->p_nrLDPC_procBuf[r]); + dlsch->harq_processes[i]->p_nrLDPC_procBuf[r] = NULL; + } + } + + free16(dlsch->harq_processes[i],sizeof(NR_DL_UE_HARQ_t)); + dlsch->harq_processes[i] = NULL; + } + } + + free16(dlsch,sizeof(NR_UE_DLSCH_t)); + dlsch = NULL; + } +} + +NR_UE_DLSCH_t *new_nr_ue_dlsch(uint8_t Kmimo,uint8_t Mdlharq,uint32_t Nsoft,uint8_t max_ldpc_iterations,uint8_t N_RB_DL, uint8_t abstraction_flag) +{ + + NR_UE_DLSCH_t *dlsch; + uint8_t exit_flag = 0,i,r; + + unsigned char bw_scaling =1; + + switch (N_RB_DL) { + case 6: + bw_scaling =16; + break; + + case 25: + bw_scaling =4; + break; + + case 50: + bw_scaling =2; + break; + + default: + bw_scaling =1; + break; + } + + dlsch = (NR_UE_DLSCH_t *)malloc16(sizeof(NR_UE_DLSCH_t)); + + if (dlsch) { + memset(dlsch,0,sizeof(NR_UE_DLSCH_t)); + dlsch->Kmimo = Kmimo; + dlsch->Mdlharq = Mdlharq; + dlsch->Nsoft = Nsoft; + dlsch->max_ldpc_iterations = max_ldpc_iterations; + + for (i=0; iharq_processes[i] = (NR_DL_UE_HARQ_t *)malloc16(sizeof(NR_DL_UE_HARQ_t)); + + if (dlsch->harq_processes[i]) { + memset(dlsch->harq_processes[i],0,sizeof(NR_DL_UE_HARQ_t)); + dlsch->harq_processes[i]->first_tx=1; + dlsch->harq_processes[i]->b = (uint8_t*)malloc16(MAX_DLSCH_PAYLOAD_BYTES/bw_scaling); + + if (dlsch->harq_processes[i]->b) + memset(dlsch->harq_processes[i]->b,0,MAX_DLSCH_PAYLOAD_BYTES/bw_scaling); + else + exit_flag=3; + + if (abstraction_flag == 0) { + for (r=0; rharq_processes[i]->p_nrLDPC_procBuf[r] = nrLDPC_init_mem(); + dlsch->harq_processes[i]->c[r] = (uint8_t*)malloc16(1056); + + if (dlsch->harq_processes[i]->c[r]) + memset(dlsch->harq_processes[i]->c[r],0,1056); + else + exit_flag=2; + + dlsch->harq_processes[i]->d[r] = (short*)malloc16((3*8448)*sizeof(short)); + + if (dlsch->harq_processes[i]->d[r]) + memset(dlsch->harq_processes[i]->d[r],0,(3*8448)*sizeof(short)); + else + exit_flag=2; + } + } + } else { + exit_flag=1; + } + } + + if (exit_flag==0) + return(dlsch); + } + + printf("new_ue_dlsch with size %zu: exit_flag = %u\n",sizeof(NR_DL_UE_HARQ_t), exit_flag); + free_nr_ue_dlsch(dlsch); + + return(NULL); +} + +void nr_dlsch_unscrambling(int16_t* llr, + uint32_t size, + uint8_t q, + uint32_t Nid, + uint32_t n_RNTI) { + + uint8_t reset; + uint32_t x1, x2, s=0; + + reset = 1; + x2 = (n_RNTI<<15) + (q<<14) + Nid; + + for (int i=0; i>(i&0x1f))&1)==1) + llr[i] = -llr[i]; + } + +} + +uint32_t nr_dlsch_decoding(PHY_VARS_NR_UE *phy_vars_ue, + short *dlsch_llr, + NR_DL_FRAME_PARMS *frame_parms, + NR_UE_DLSCH_t *dlsch, + NR_DL_UE_HARQ_t *harq_process, + uint32_t frame, + uint16_t nb_symb_sch, + uint8_t nr_tti_rx, + uint8_t harq_pid, + uint8_t is_crnti, + uint8_t llr8_flag) +{ + +#if UE_TIMING_TRACE + time_stats_t *dlsch_rate_unmatching_stats=&phy_vars_ue->dlsch_rate_unmatching_stats; + time_stats_t *dlsch_turbo_decoding_stats=&phy_vars_ue->dlsch_turbo_decoding_stats; + time_stats_t *dlsch_deinterleaving_stats=&phy_vars_ue->dlsch_deinterleaving_stats; +#endif + uint32_t A,E; + uint32_t G; + uint32_t ret,offset; + int32_t no_iteration_ldpc; + //short dummy_w[MAX_NUM_DLSCH_SEGMENTS][3*(8448+64)]; + uint32_t r,r_offset=0,Kr=8424,Kr_bytes,K_bytes_F,err_flag=0; + uint8_t crc_type; + int8_t llrProcBuf[OAI_LDPC_MAX_NUM_LLR] __attribute__ ((aligned(32))); + t_nrLDPC_dec_params decParams; + t_nrLDPC_dec_params* p_decParams = &decParams; + t_nrLDPC_time_stats procTime; + t_nrLDPC_time_stats* p_procTime =&procTime ; + t_nrLDPC_procBuf** p_nrLDPC_procBuf = harq_process->p_nrLDPC_procBuf; + + int16_t z [68*384]; + int8_t l [68*384]; + //__m128i l; + int16_t inv_d [68*384]; +// int16_t *p_invd =&inv_d; + uint8_t kb, kc; + uint8_t Ilbrm = 0; + uint32_t Tbslbrm = 950984; + uint16_t nb_rb = 30; //to update + //uint16_t nb_symb_sch = 12; + uint8_t nb_re_dmrs = 6; + uint16_t length_dmrs = 1; + + uint32_t i,j; +// uint32_t k; + + __m128i *pv = (__m128i*)&z; + __m128i *pl = (__m128i*)&l; + + //NR_DL_UE_HARQ_t *harq_process = dlsch->harq_processes[0]; + + if (!dlsch_llr) { + printf("dlsch_decoding.c: NULL dlsch_llr pointer\n"); + return(dlsch->max_ldpc_iterations); + } + + if (!harq_process) { + printf("dlsch_decoding.c: NULL harq_process pointer\n"); + return(dlsch->max_ldpc_iterations); + } + + if (!frame_parms) { + printf("dlsch_decoding.c: NULL frame_parms pointer\n"); + return(dlsch->max_ldpc_iterations); + } + + /*if (nr_tti_rx> (10*frame_parms->ttis_per_subframe-1)) { + printf("dlsch_decoding.c: Illegal subframe index %d\n",nr_tti_rx); + return(dlsch->max_ldpc_iterations); + }*/ + + /*if (harq_process->harq_ack.ack != 2) { + LOG_D(PHY, "[UE %d] DLSCH @ SF%d : ACK bit is %d instead of DTX even before PDSCH is decoded!\n", + phy_vars_ue->Mod_id, nr_tti_rx, harq_process->harq_ack.ack); + }*/ + + // nb_rb = dlsch->nb_rb; + + /* + if (nb_rb > frame_parms->N_RB_DL) { + printf("dlsch_decoding.c: Illegal nb_rb %d\n",nb_rb); + return(max_ldpc_iterations); + }*/ + + /*harq_pid = dlsch->current_harq_pid[phy_vars_ue->current_thread_id[subframe]]; + if (harq_pid >= 8) { + printf("dlsch_decoding.c: Illegal harq_pid %d\n",harq_pid); + return(max_ldpc_iterations); + } + */ + + nb_rb = harq_process->nb_rb; + + harq_process->trials[harq_process->round]++; + + harq_process->TBS = nr_compute_tbs(harq_process->mcs,nb_rb,nb_symb_sch,nb_re_dmrs,length_dmrs, harq_process->Nl); + + A = harq_process->TBS; + ret = dlsch->max_ldpc_iterations; + + harq_process->G = nr_get_G(nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs, harq_process->Qm,harq_process->Nl); + G = harq_process->G; + + LOG_I(PHY,"DLSCH Decoding, harq_pid %d TBS %d G %d mcs %d Nl %d nb_symb_sch %d nb_rb %d\n",harq_pid,A,G, harq_process->mcs, harq_process->Nl, nb_symb_sch,nb_rb); + + if (harq_process->round == 0) { + // This is a new packet, so compute quantities regarding segmentation + harq_process->B = A+24; + nr_segmentation(NULL, + NULL, + harq_process->B, + &harq_process->C, + &harq_process->K, + &harq_process->Z, + &harq_process->F); + p_decParams->Z = harq_process->Z; + +#ifdef DEBUG_DLSCH_DECODING + printf("dlsch decoding nr segmentation Z %d\n", p_decParams->Z); + if (!frame%100) + printf("K %d C %d Z %d nl %d \n", harq_process->K, harq_process->C, p_decParams->Z, harq_process->Nl); +#endif + } + + kb = harq_process->K/harq_process->Z; + if ( kb==22){ + p_decParams->BG = 1; + p_decParams->R = 13; + kc = 68; + } + else{ + p_decParams->BG = 2; + p_decParams->R = 13; + kc = 52; + } + + p_decParams->numMaxIter = dlsch->max_ldpc_iterations; + Kr = p_decParams->Z*kb; + p_decParams->outMode= 0; + + + err_flag = 0; + r_offset = 0; + + unsigned char bw_scaling =1; + + switch (frame_parms->N_RB_DL) { + + case 106: + bw_scaling =2; + break; + + default: + bw_scaling =1; + break; + } + + if (harq_process->C > MAX_NUM_NR_DLSCH_SEGMENTS/bw_scaling) { + LOG_E(PHY,"Illegal harq_process->C %d > %d\n",harq_process->C,MAX_NUM_NR_DLSCH_SEGMENTS/bw_scaling); + return((1+dlsch->max_ldpc_iterations)); + } +#ifdef DEBUG_DLSCH_DECODING + printf("Segmentation: C %d, K %d\n",harq_process->C,harq_process->K); +#endif + + opp_enabled=1; + + Kr = harq_process->K; + Kr_bytes = Kr>>3; + + K_bytes_F = Kr_bytes-(harq_process->F>>3); + + Tbslbrm = nr_compute_tbs(28,nb_rb,frame_parms->symbols_per_slot,0,0, harq_process->Nl); + + for (r=0; rC; r++) { + + //printf("start rx segment %d\n",r); + E = nr_get_E(G, harq_process->C, harq_process->Qm, harq_process->Nl, r); + +#if UE_TIMING_TRACE + start_meas(dlsch_deinterleaving_stats); +#endif + nr_deinterleaving_ldpc(E, + harq_process->Qm, + harq_process->w[r], + dlsch_llr+r_offset); + + //for (int i =0; i<16; i++) + // printf("rx output deinterleaving w[%d]= %d r_offset %d\n", i,harq_process->w[r][i], r_offset); + +#if UE_TIMING_TRACE + stop_meas(dlsch_deinterleaving_stats); +#endif + +#if UE_TIMING_TRACE + start_meas(dlsch_rate_unmatching_stats); +#endif + +#ifdef DEBUG_DLSCH_DECODING + LOG_D(PHY,"HARQ_PID %d Rate Matching Segment %d (coded bits %d,unpunctured/repeated bits %d, TBS %d, mod_order %d, nb_rb %d, Nl %d, rv %d, round %d)...\n", + harq_pid,r, G, + Kr*3, + harq_process->TBS, + harq_process->Qm, + harq_process->nb_rb, + harq_process->Nl, + harq_process->rvidx, + harq_process->round); +#endif + + if (nr_rate_matching_ldpc_rx(Ilbrm, + Tbslbrm, + p_decParams->BG, + p_decParams->Z, + harq_process->d[r], + harq_process->w[r], + harq_process->C, + harq_process->rvidx, + (harq_process->round==0)?1:0, + E)==-1) { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + LOG_E(PHY,"dlsch_decoding.c: Problem in rate_matching\n"); + return(dlsch->max_ldpc_iterations); + } else + { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + } + + //for (int i =0; i<16; i++) + // printf("rx output ratematching d[%d]= %d r_offset %d\n", i,harq_process->d[r][i], r_offset); + + r_offset += E; + +#ifdef DEBUG_DLSCH_DECODING + if (r==0) { + write_output("decoder_llr.m","decllr",dlsch_llr,G,1,0); + write_output("decoder_in.m","dec",&harq_process->d[0][0],(3*8*Kr_bytes)+12,1,0); + } + + printf("decoder input(segment %d) :",r); + int i; for (i=0;i<(3*8*Kr_bytes)+12;i++) + printf("%d : %d\n",i,harq_process->d[r][i]); + printf("\n"); +#endif + + + // printf("Clearing c, %p\n",harq_process->c[r]); + memset(harq_process->c[r],0,Kr_bytes); + + // printf("done\n"); + if (harq_process->C == 1) + crc_type = CRC24_A; + else + crc_type = CRC24_B; + + if (err_flag == 0) { +/* + LOG_I(PHY, "turbo algo Kr=%d cb_cnt=%d C=%d nbRB=%d crc_type %d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d maxIter %d\n", + Kr,r,harq_process->C,harq_process->nb_rb,crc_type,A,harq_process->TBS, + harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round,dlsch->max_ldpc_iterations); +*/ + +#if UE_TIMING_TRACE + start_meas(dlsch_turbo_decoding_stats); +#endif + + //LOG_E(PHY,"AbsSubframe %d.%d Start turbo segment %d/%d A %d ",frame%1024,nr_tti_rx,r,harq_process->C-1, A); + + //printf("harq process dr iteration %d\n", p_decParams->numMaxIter); + + for (int cnt =0; cnt < (kc-2)*p_decParams->Z; cnt++){ + inv_d[cnt] = (1)*harq_process->d[r][cnt]; + } + + /*for (int cnt =0; cnt < 16; cnt++){ + printf("dr %d inv_d %d \n", harq_process->d[r][cnt], inv_d[cnt]); + } + + printf(" \n"); + + printf("end dr \n"); + for (int cnt =(50*p_decParams->Z-16) ; cnt < 50*p_decParams->Z; cnt++){ + printf("%d ", harq_process->d[r][cnt]); + } + printf(" \n");*/ + + memset(pv,0,2*harq_process->Z*sizeof(int16_t)); + //memset(pl,0,2*p_decParams->Z*sizeof(int8_t)); + memset((pv+K_bytes_F),127,harq_process->F*sizeof(int16_t)); + + for (i=((2*p_decParams->Z)>>3), j = 0; i < K_bytes_F+((2*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=Kr_bytes+((2*p_decParams->Z)>>3),j=Kr_bytes; i < ((kc*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=0, j=0; j < ((kc*p_decParams->Z)>>4); i+=2, j++) + { + + pl[j] = _mm_packs_epi16(pv[i],pv[i+1]); + + } + + no_iteration_ldpc = nrLDPC_decoder(p_decParams, + (int8_t*)&pl[0], + llrProcBuf, + p_nrLDPC_procBuf[r], + p_procTime); + + if (check_crc(llrProcBuf,harq_process->B,harq_process->F,crc_type)) { + printf("CRC OK\n"); + ret = 2; + } + else { + printf("CRC NOK\n"); + ret = 1+dlsch->max_ldpc_iterations; + } + + + nb_total_decod++; + if (no_iteration_ldpc > dlsch->max_ldpc_iterations){ + nb_error_decod++; + } + + ret=no_iteration_ldpc; + + //if (!nb_total_decod%10000){ + //printf("Error number of iteration LPDC %d %ld/%ld \n", no_iteration_ldpc, nb_error_decod,nb_total_decod);fflush(stdout); + //} + //else + //printf("OK number of iteration LPDC %d\n", no_iteration_ldpc); + + for (int m=0; m < Kr>>3; m ++) + { + harq_process->c[r][m]= (uint8_t) llrProcBuf[m]; + } + +#ifdef DEBUG_DLSCH_DECODING + //printf("output decoder %d %d %d %d %d \n", harq_process->c[r][0], harq_process->c[r][1], harq_process->c[r][2],harq_process->c[r][3], harq_process->c[r][4]); + for (int k=0;k>3;k++) + printf("output decoder [%d] = 0x%02x \n", k, harq_process->c[r][k]); + printf("no_iterations_ldpc %d (ret %d)\n",no_iteration_ldpc,ret); + //write_output("dec_output.m","dec0",harq_process->c[0],Kr_bytes,1,4); +#endif + + +#if UE_TIMING_TRACE + stop_meas(dlsch_turbo_decoding_stats); +#endif + } + + /*printf("Segmentation: C %d r %d, dlsch_rate_unmatching_stats %5.3f dlsch_deinterleaving_stats %5.3f dlsch_turbo_decoding_stats %5.3f \n", + harq_process->C, + r, + dlsch_rate_unmatching_stats->p_time/(cpuf*1000.0), + dlsch_deinterleaving_stats->p_time/(cpuf*1000.0), + dlsch_turbo_decoding_stats->p_time/(cpuf*1000.0));*/ + + + if ((err_flag == 0) && (ret>=(1+dlsch->max_ldpc_iterations))) {// a Code segment is in error so break; + LOG_D(PHY,"AbsSubframe %d.%d CRC failed, segment %d/%d \n",frame%1024,nr_tti_rx,r,harq_process->C-1); + err_flag = 1; + } + } + + int32_t frame_rx_prev = frame; + int32_t tti_rx_prev = nr_tti_rx - 1; + if (tti_rx_prev < 0) { + frame_rx_prev--; + tti_rx_prev += 10*frame_parms->ttis_per_subframe; + } + frame_rx_prev = frame_rx_prev%1024; + + if (err_flag == 1) { +#if UE_DEBUG_TRACE + LOG_I(PHY,"[UE %d] DLSCH: Setting NAK for SFN/SF %d/%d (pid %d, status %d, round %d, TBS %d, mcs %d) Kr %d r %d harq_process->round %d\n", + phy_vars_ue->Mod_id, frame, nr_tti_rx, harq_pid,harq_process->status, harq_process->round,harq_process->TBS,harq_process->mcs,Kr,r,harq_process->round); +#endif + harq_process->harq_ack.ack = 0; + harq_process->harq_ack.harq_id = harq_pid; + harq_process->harq_ack.send_harq_status = 1; + harq_process->errors[harq_process->round]++; + harq_process->round++; + + + // printf("Rate: [UE %d] DLSCH: Setting NACK for subframe %d (pid %d, round %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round); + if (harq_process->round >= dlsch->Mdlharq) { + harq_process->status = SCH_IDLE; + harq_process->round = 0; + } + if(is_crnti) + { + LOG_D(PHY,"[UE %d] DLSCH: Setting NACK for nr_tti_rx %d (pid %d, pid status %d, round %d/Max %d, TBS %d)\n", + phy_vars_ue->Mod_id,nr_tti_rx,harq_pid,harq_process->status,harq_process->round,dlsch->Mdlharq,harq_process->TBS); + } + + return((1+dlsch->max_ldpc_iterations)); + } else { +#if UE_DEBUG_TRACE + LOG_I(PHY,"[UE %d] DLSCH: Setting ACK for nr_tti_rx %d TBS %d mcs %d nb_rb %d harq_process->round %d\n", + phy_vars_ue->Mod_id,nr_tti_rx,harq_process->TBS,harq_process->mcs,harq_process->nb_rb, harq_process->round); +#endif + + harq_process->status = SCH_IDLE; + harq_process->round = 0; + harq_process->harq_ack.ack = 1; + harq_process->harq_ack.harq_id = harq_pid; + harq_process->harq_ack.send_harq_status = 1; + //LOG_I(PHY,"[UE %d] DLSCH: Setting ACK for SFN/SF %d/%d (pid %d, status %d, round %d, TBS %d, mcs %d)\n", + // phy_vars_ue->Mod_id, frame, subframe, harq_pid, harq_process->status, harq_process->round,harq_process->TBS,harq_process->mcs); + + if(is_crnti) + { + LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for nr_tti_rx %d (pid %d, round %d, TBS %d)\n",phy_vars_ue->Mod_id,nr_tti_rx,harq_pid,harq_process->round,harq_process->TBS); + } + //LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for subframe %d (pid %d, round %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round); + + } + + // Reassembly of Transport block here + offset = 0; + Kr = harq_process->K; + Kr_bytes = Kr>>3; + + /* + printf("harq_pid %d\n",harq_pid); + printf("F %d, Fbytes %d\n",harq_process->F,harq_process->F>>3); + printf("C %d\n",harq_process->C); + */ + for (r=0; rC; r++) { + + memcpy(harq_process->b+offset, + harq_process->c[r], + Kr_bytes- - (harq_process->F>>3) -((harq_process->C>1)?3:0)); + offset += (Kr_bytes - (harq_process->F>>3) - ((harq_process->C>1)?3:0)); + +#ifdef DEBUG_DLSCH_DECODING + printf("Segment %d : Kr= %d bytes\n",r,Kr_bytes); + printf("copied %d bytes to b sequence (harq_pid %d)\n", + (Kr_bytes - (harq_process->F>>3)-((harq_process->C>1)?3:0)),harq_pid); + printf("b[0] = %x,c[%d] = %x\n", + harq_process->b[offset], + harq_process->F>>3, + harq_process->c[r]); +#endif + + } + + dlsch->last_iteration_cnt = ret; + + return(ret); +} + +#ifdef UE_DLSCH_PARALLELISATION +uint32_t nr_dlsch_decoding_mthread(PHY_VARS_NR_UE *phy_vars_ue, + UE_nr_rxtx_proc_t *proc, + int eNB_id, + short *dlsch_llr, + NR_DL_FRAME_PARMS *frame_parms, + NR_UE_DLSCH_t *dlsch, + NR_DL_UE_HARQ_t *harq_process, + uint32_t frame, + uint16_t nb_symb_sch, + uint8_t nr_tti_rx, + uint8_t harq_pid, + uint8_t is_crnti, + uint8_t llr8_flag) +{ + +#if UE_TIMING_TRACE + time_stats_t *dlsch_rate_unmatching_stats=&phy_vars_ue->dlsch_rate_unmatching_stats; + time_stats_t *dlsch_turbo_decoding_stats=&phy_vars_ue->dlsch_turbo_decoding_stats; + time_stats_t *dlsch_deinterleaving_stats=&phy_vars_ue->dlsch_deinterleaving_stats; +#endif + uint32_t A,E; + uint32_t G; + uint32_t ret,offset; + //short dummy_w[MAX_NUM_DLSCH_SEGMENTS][3*(8448+64)]; + uint32_t r,r_offset=0,Kr,Kr_bytes,err_flag=0,K_bytes_F; + uint8_t crc_type; + //UE_rxtx_proc_t *proc = &phy_vars_ue->proc; + int32_t no_iteration_ldpc; + int Cby2; + /*uint8_t C; + uint8_t Qm; + uint8_t Nl; + uint8_t r_thread; + uint32_t Er, Gp,GpmodC;*/ + t_nrLDPC_dec_params decParams; + t_nrLDPC_dec_params* p_decParams = &decParams; + t_nrLDPC_time_stats procTime; + t_nrLDPC_time_stats* p_procTime =&procTime ; + int8_t llrProcBuf[OAI_LDPC_MAX_NUM_LLR] __attribute__ ((aligned(32))); + t_nrLDPC_procBuf* p_nrLDPC_procBuf = harq_process->p_nrLDPC_procBuf[0]; + + int16_t z [68*384]; + int8_t l [68*384]; + //__m128i l; + int16_t inv_d [68*384]; + //int16_t *p_invd =&inv_d; + uint8_t kb, kc; + uint8_t Ilbrm = 0; + uint32_t Tbslbrm = 950984; + uint16_t nb_rb = 30; //to update + //uint16_t nb_symb_sch = 12; + uint8_t nb_re_dmrs = 6; + uint16_t length_dmrs = 1; + + uint32_t i,j; + //uint32_t k; + + __m128i *pv = (__m128i*)&z; + __m128i *pl = (__m128i*)&l; + + + if (!dlsch_llr) { + printf("dlsch_decoding.c: NULL dlsch_llr pointer\n"); + return(dlsch->max_ldpc_iterations); + } + + if (!harq_process) { + printf("dlsch_decoding.c: NULL harq_process pointer\n"); + return(dlsch->max_ldpc_iterations); + } + + if (!frame_parms) { + printf("dlsch_decoding.c: NULL frame_parms pointer\n"); + return(dlsch->max_ldpc_iterations); + } + + /* if (nr_tti_rx> (10*frame_parms->ttis_per_subframe-1)) { + printf("dlsch_decoding.c: Illegal subframe index %d\n",nr_tti_rx); + return(dlsch->max_ldpc_iterations); + } + + if (dlsch->harq_ack[nr_tti_rx].ack != 2) { + LOG_D(PHY, "[UE %d] DLSCH @ SF%d : ACK bit is %d instead of DTX even before PDSCH is decoded!\n", + phy_vars_ue->Mod_id, nr_tti_rx, dlsch->harq_ack[nr_tti_rx].ack); + }*/ + + /* + if (nb_rb > frame_parms->N_RB_DL) { + printf("dlsch_decoding.c: Illegal nb_rb %d\n",nb_rb); + return(max_ldpc_iterations); + }*/ + + /*harq_pid = dlsch->current_harq_pid[phy_vars_ue->current_thread_id[subframe]]; + if (harq_pid >= 8) { + printf("dlsch_decoding.c: Illegal harq_pid %d\n",harq_pid); + return(max_ldpc_iterations); + } + */ + + nb_rb = harq_process->nb_rb; + harq_process->trials[harq_process->round]++; + + harq_process->TBS = nr_compute_tbs(harq_process->mcs,nb_rb,nb_symb_sch,nb_re_dmrs,length_dmrs, harq_process->Nl); + + A = harq_process->TBS; //2072 for QPSK 1/3 + + ret = dlsch->max_ldpc_iterations; + + harq_process->G = nr_get_G(nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs, harq_process->Qm,harq_process->Nl); + + G = harq_process->G; + + proc->decoder_main_available = 1; + proc->decoder_thread_available = 0; + proc->decoder_thread_available1 = 0; + //get_G(frame_parms,nb_rb,dlsch->rb_alloc,mod_order,num_pdcch_symbols,phy_vars_ue->frame,subframe); + + // printf("DLSCH Decoding, harq_pid %d Ndi %d\n",harq_pid,harq_process->Ndi); + + if (harq_process->round == 0) { + // This is a new packet, so compute quantities regarding segmentation + harq_process->B = A+24; + nr_segmentation(NULL, + NULL, + harq_process->B, + &harq_process->C, + &harq_process->K, + &harq_process->Z, + &harq_process->F); + p_decParams->Z = harq_process->Z; + + } + + kb = harq_process->K/harq_process->Z; + if ( kb==22){ + p_decParams->BG = 1; + p_decParams->R = 13; + kc = 68; + } + else{ + p_decParams->BG = 2; + p_decParams->R = 13; + kc = 52; + } + + p_decParams->numMaxIter = 2; + Kr = p_decParams->Z*kb; + p_decParams->outMode= 0; + + /* + else { + printf("dlsch_decoding.c: Ndi>0 not checked yet!!\n"); + return(max_ldpc_iterations); + } + */ + err_flag = 0; + r_offset = 0; + + unsigned char bw_scaling =1; + + switch (frame_parms->N_RB_DL) { + case 106: + bw_scaling =2; + break; + + default: + bw_scaling =1; + break; + } + + if (harq_process->C > MAX_NUM_DLSCH_SEGMENTS/bw_scaling) { + LOG_E(PHY,"Illegal harq_process->C %d > %d\n",harq_process->C,MAX_NUM_DLSCH_SEGMENTS/bw_scaling); + return((1+dlsch->max_ldpc_iterations)); + } +#ifdef DEBUG_DLSCH_DECODING + printf("Segmentation: C %d, K %d\n",harq_process->C,harq_process->K); +#endif + + opp_enabled=1; +if (harq_process->C>1) { // wakeup worker if more than 1 segment + if (pthread_mutex_lock(&proc->mutex_dlsch_td) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] error locking mutex for UE dlsch td\n",phy_vars_ue->Mod_id ); + exit_fun("nothing to add"); + } + + /*Qm= harq_process->Qm; + Nl=harq_process->Nl; + r_thread = harq_process->C/2-1; + C= harq_process->C; + + Gp = G/Nl/Qm; + GpmodC = Gp%C; + + + + if (r_thread < (C-(GpmodC))) + Er = Nl*Qm * (Gp/C); + else + Er = Nl*Qm * ((GpmodC==0?0:1) + (Gp/C)); + printf("mthread Er %d\n", Er); + + printf("mthread instance_cnt_dlsch_td %d\n", proc->instance_cnt_dlsch_td);*/ + + proc->instance_cnt_dlsch_td++; + proc->eNB_id = eNB_id; + proc->harq_pid = harq_pid; + proc->llr8_flag = llr8_flag; + //proc->r[0] = 1; + + if (proc->instance_cnt_dlsch_td == 0) + { + LOG_D(PHY,"unblock dlsch td processing thread blocked on instance_cnt_dlsch_td : %d \n", proc->instance_cnt_dlsch_td ); + if (pthread_cond_signal(&proc->cond_dlsch_td) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] ERROR pthread_cond_signal for UE dlsch td\n", phy_vars_ue->Mod_id); + exit_fun("nothing to add"); + } + if (pthread_mutex_unlock(&proc->mutex_dlsch_td) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] error unlocking mutex for UE dlsch td \n",phy_vars_ue->Mod_id ); + exit_fun("nothing to add"); + } + + } else + { + LOG_E( PHY, "[SCHED][UE %d] UE dlsch td thread busy (IC %d)!!\n", phy_vars_ue->Mod_id, proc->instance_cnt_dlsch_td); + if (proc->instance_cnt_dlsch_td > 4) + exit_fun("instance_cnt_dlsch_td > 4"); + } + //AssertFatal(pthread_cond_signal(&proc->cond_slot1_dl_processing) ==0 ,""); + AssertFatal(pthread_mutex_unlock(&proc->mutex_dlsch_td) ==0,""); + + if (harq_process->C>2) { + if (pthread_mutex_lock(&proc->mutex_dlsch_td1) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] error locking mutex for UE dlsch td\n",phy_vars_ue->Mod_id ); + exit_fun("nothing to add"); + } + + proc->instance_cnt_dlsch_td1++; + proc->eNB_id = eNB_id; + proc->harq_pid = harq_pid; + proc->llr8_flag = llr8_flag; + // proc->Er = Er; + + if (proc->instance_cnt_dlsch_td1 == 0) + { + LOG_D(PHY,"unblock slot1 dl processing thread blocked on instance_cnt_dlsch_td : %d \n", proc->instance_cnt_dlsch_td1 ); + if (pthread_cond_signal(&proc->cond_dlsch_td1) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] ERROR pthread_cond_signal for UE dlsch td\n", phy_vars_ue->Mod_id); + exit_fun("nothing to add"); + } + if (pthread_mutex_unlock(&proc->mutex_dlsch_td1) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] error unlocking mutex for UE dlsch td \n",phy_vars_ue->Mod_id ); + exit_fun("nothing to add"); + } + + } else + { + LOG_E( PHY, "[SCHED][UE %d] UE dlsch td thread 1 busy (IC %d)!!\n", phy_vars_ue->Mod_id, proc->instance_cnt_dlsch_td1); + if (proc->instance_cnt_dlsch_td1 > 4) + exit_fun("instance_cnt_dlsch_td1 > 4"); + } + + AssertFatal(pthread_mutex_unlock(&proc->mutex_dlsch_td1) ==0,""); + + } + /* + if (pthread_mutex_timedlock(&proc->mutex_td,&wait) != 0) { + printf("[eNB] ERROR pthread_mutex_lock for TD thread (IC %d)\n", proc->instance_cnt_td); + exit_fun( "error locking mutex_fep" ); + return -1; + } + + if (proc->instance_cnt_td==0) { + printf("[UE] TD thread busy\n"); + exit_fun("TD thread busy"); + pthread_mutex_unlock( &proc->mutex_td ); + return -1; + } + + ++proc->instance_cnt_td; + + proc->tdp.UE = phy_vars_ue; + proc->tdp.eNB_id = eNB_id; + proc->tdp.harq_pid = harq_pid; + proc->tdp.llr8_flag = llr8_flag; + + printf("----- 2thread llr flag %d tdp flag %d\n",llr8_flag, proc->tdp.llr8_flag); + + + // wakeup worker to do second half segments + if (pthread_cond_signal(&proc->cond_td) != 0) { + printf("[UE] ERROR pthread_cond_signal for td thread exit\n"); + exit_fun( "ERROR pthread_cond_signal" ); + return (1+dlsch->last_iteration_cnt); + } + + pthread_mutex_unlock( &proc->mutex_td );*/ + + + Cby2 = 1; //harq_process->C/2; + //proc->decoder_main_available = 1; + } + else { + Cby2 = 1; + } + + //for (r=0; rK; + Kr_bytes = Kr>>3; + K_bytes_F = Kr_bytes-(harq_process->F>>3); + + Tbslbrm = nr_compute_tbs(28,nb_rb,frame_parms->symbols_per_slot,0,0, harq_process->Nl); + + E = nr_get_E(G, harq_process->C, harq_process->Qm, harq_process->Nl, r); + + /* + printf("Subblock deinterleaving, dlsch_llr %p, w %p\n", + dlsch_llr+r_offset, + &harq_process->w[r]); + */ +#if UE_TIMING_TRACE + start_meas(dlsch_deinterleaving_stats); +#endif + nr_deinterleaving_ldpc(E, + harq_process->Qm, + harq_process->w[r], + dlsch_llr+r_offset); + +#ifdef DEBUG_DLSCH_DECODING + for (int i =0; i<16; i++) + printf("rx output deinterleaving w[%d]= %d r_offset %d\n", i,harq_process->w[r][i], r_offset); +#endif + +#if UE_TIMING_TRACE + stop_meas(dlsch_deinterleaving_stats); +#endif + +#if UE_TIMING_TRACE + start_meas(dlsch_rate_unmatching_stats); +#endif + +#ifdef DEBUG_DLSCH_DECODING + LOG_D(PHY,"HARQ_PID %d Rate Matching Segment %d (coded bits %d,unpunctured/repeated bits %d, TBS %d, mod_order %d, nb_rb %d, Nl %d, rv %d, round %d)...\n", + harq_pid,r, G, + Kr*3, + harq_process->TBS, + harq_process->Qm, + harq_process->nb_rb, + harq_process->Nl, + harq_process->rvidx, + harq_process->round); +#endif + + if (nr_rate_matching_ldpc_rx(Ilbrm, + Tbslbrm, + p_decParams->BG, + p_decParams->Z, + harq_process->d[r], + harq_process->w[r], + harq_process->C, + harq_process->rvidx, + (harq_process->round==0)?1:0, + E)==-1) { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + LOG_E(PHY,"dlsch_decoding.c: Problem in rate_matching\n"); + return(dlsch->max_ldpc_iterations); + } else + { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + } + + //for (int i =0; i<16; i++) + // printf("rx output ratematching d[%d]= %d r_offset %d\n", i,harq_process->d[r][i], r_offset); + + //r_offset += E; + //printf("main thread r_offset %d\n",r_offset); + +#ifdef DEBUG_DLSCH_DECODING + for (int i =0; i<16; i++) + printf("rx output ratematching d[%d]= %d r_offset %d\n", i,harq_process->d[r][i], r_offset); +#endif + +#ifdef DEBUG_DLSCH_DECODING + + if (r==0) { + write_output("decoder_llr.m","decllr",dlsch_llr,G,1,0); + write_output("decoder_in.m","dec",&harq_process->d[0][96],(3*8*Kr_bytes)+12,1,0); + } + + printf("decoder input(segment %d) :",r); + for (int i=0;i<(3*8*Kr_bytes);i++) + printf("%d : %d\n",i,harq_process->d[r][i]); + printf("\n"); +#endif + + + // printf("Clearing c, %p\n",harq_process->c[r]); + memset(harq_process->c[r],0,Kr_bytes); + + // printf("done\n"); + if (harq_process->C == 1) + crc_type = CRC24_A; + else + crc_type = CRC24_B; + + //#ifndef __AVX2__ + + if (err_flag == 0) { +/* + LOG_I(PHY, "turbo algo Kr=%d cb_cnt=%d C=%d nbRB=%d crc_type %d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d maxIter %d\n", + Kr,r,harq_process->C,harq_process->nb_rb,crc_type,A,harq_process->TBS, + harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round,dlsch->max_ldpc_iterations); +*/ + if (llr8_flag) { + AssertFatal (Kr >= 256, "turbo algo issue Kr=%d cb_cnt=%d C=%d nbRB=%d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d\n", + Kr,r,harq_process->C,harq_process->nb_rb,A,harq_process->TBS,harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round); + } +#if UE_TIMING_TRACE + start_meas(dlsch_turbo_decoding_stats); +#endif + LOG_D(PHY,"mthread AbsSubframe %d.%d Start turbo segment %d/%d \n",frame%1024,nr_tti_rx,r,harq_process->C-1); + + for (int cnt =0; cnt < (kc-2)*p_decParams->Z; cnt++){ + inv_d[cnt] = (1)*harq_process->d[r][cnt]; + } + + memset(pv,0,2*p_decParams->Z*sizeof(int16_t)); + //memset(pl,0,2*p_decParams->Z*sizeof(int8_t)); + memset((pv+K_bytes_F),127,harq_process->F*sizeof(int16_t)); + + + for (i=((2*p_decParams->Z)>>3), j = 0; i < K_bytes_F+((2*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=Kr_bytes+((2*p_decParams->Z)>>3),j=Kr_bytes; i < ((kc*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=0, j=0; j < ((kc*p_decParams->Z)>>4); i+=2, j++) + { + + pl[j] = _mm_packs_epi16(pv[i],pv[i+1]); + + } + + no_iteration_ldpc = nrLDPC_decoder(p_decParams, + (int8_t*)&pl[0], + llrProcBuf, + p_nrLDPC_procBuf, + p_procTime); + + nb_total_decod++; + if (no_iteration_ldpc > 10){ + nb_error_decod++; + ret = 1+dlsch->max_ldpc_iterations; + } + else { + ret=2; + } + + if (check_crc(llrProcBuf,harq_process->B,harq_process->F,crc_type)) { + printf("CRC OK\n"); + ret = 2; + } + else { + printf("CRC NOK\n"); + ret = 1+dlsch->max_ldpc_iterations; + } + + //if (!nb_total_decod%10000){ + printf("Error number of iteration LPDC %d %ld/%ld \n", no_iteration_ldpc, nb_error_decod,nb_total_decod);fflush(stdout); + //} + + //else + //printf("OK number of iteration LPDC %d\n", no_iteration_ldpc); + + for (int m=0; m < Kr>>3; m ++) + { + harq_process->c[r][m]= (uint8_t) llrProcBuf[m]; + } + + /*for (int u=0; u < Kr>>3; u ++) + { + ullrProcBuf[u]= (uint8_t) llrProcBuf[u]; + } + + + printf("output unsigned ullrProcBuf \n"); + + for (int j=0; j < Kr>>3; j ++) + { + + printf(" %d \n", ullrProcBuf[j]); + + } + printf(" \n");*/ + //printf("output channel decoder %d %d %d %d %d \n", harq_process->c[r][0], harq_process->c[r][1], harq_process->c[r][2],harq_process->c[r][3], harq_process->c[r][4]); + + //printf("output decoder %d %d %d %d %d \n", harq_process->c[r][0], harq_process->c[r][1], harq_process->c[r][2],harq_process->c[r][3], harq_process->c[r][4]); +#ifdef DEBUG_DLSCH_DECODING + for (int k=0;k<32;k++) + printf("output decoder [%d] = 0x%02x \n", k, harq_process->c[r][k]); +#endif + +#if UE_TIMING_TRACE + stop_meas(dlsch_turbo_decoding_stats); +#endif + } + + + if ((err_flag == 0) && (ret>=(1+dlsch->max_ldpc_iterations))) {// a Code segment is in error so break; + LOG_D(PHY,"AbsSubframe %d.%d CRC failed, segment %d/%d \n",frame%1024,nr_tti_rx,r,harq_process->C-1); + err_flag = 1; + } + //} //loop r + + int32_t frame_rx_prev = frame; + int32_t tti_rx_prev = nr_tti_rx - 1; + if (tti_rx_prev < 0) { + frame_rx_prev--; + tti_rx_prev += 10*frame_parms->ttis_per_subframe; + } + frame_rx_prev = frame_rx_prev%1024; + + if (err_flag == 1) { +#if UE_DEBUG_TRACE + LOG_I(PHY,"[UE %d] DLSCH: Setting NAK for SFN/SF %d/%d (pid %d, status %d, round %d, TBS %d, mcs %d) Kr %d r %d harq_process->round %d\n", + phy_vars_ue->Mod_id, frame, nr_tti_rx, harq_pid,harq_process->status, harq_process->round,harq_process->TBS,harq_process->mcs,Kr,r,harq_process->round); +#endif + harq_process->harq_ack.ack = 0; + harq_process->harq_ack.harq_id = harq_pid; + harq_process->harq_ack.send_harq_status = 1; + harq_process->errors[harq_process->round]++; + harq_process->round++; + + + // printf("Rate: [UE %d] DLSCH: Setting NACK for subframe %d (pid %d, round %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round); + if (harq_process->round >= dlsch->Mdlharq) { + harq_process->status = SCH_IDLE; + harq_process->round = 0; + } + if(is_crnti) + { + LOG_D(PHY,"[UE %d] DLSCH: Setting NACK for nr_tti_rx %d (pid %d, pid status %d, round %d/Max %d, TBS %d)\n", + phy_vars_ue->Mod_id,nr_tti_rx,harq_pid,harq_process->status,harq_process->round,dlsch->Mdlharq,harq_process->TBS); + } + + return((1+dlsch->max_ldpc_iterations)); + } else { +#if UE_DEBUG_TRACE + LOG_I(PHY,"[UE %d] DLSCH: Setting ACK for nr_tti_rx %d TBS %d mcs %d nb_rb %d\n", + phy_vars_ue->Mod_id,nr_tti_rx,harq_process->TBS,harq_process->mcs,harq_process->nb_rb); +#endif + + harq_process->status = SCH_IDLE; + harq_process->round = 0; + harq_process->harq_ack.ack = 1; + harq_process->harq_ack.harq_id = harq_pid; + harq_process->harq_ack.send_harq_status = 1; + //LOG_I(PHY,"[UE %d] DLSCH: Setting ACK for SFN/SF %d/%d (pid %d, status %d, round %d, TBS %d, mcs %d)\n", + // phy_vars_ue->Mod_id, frame, subframe, harq_pid, harq_process->status, harq_process->round,harq_process->TBS,harq_process->mcs); + + if(is_crnti) + { + LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for nr_tti_rx %d (pid %d, round %d, TBS %d)\n",phy_vars_ue->Mod_id,nr_tti_rx,harq_pid,harq_process->round,harq_process->TBS); + } + //LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for subframe %d (pid %d, round %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round); + + } + + // Reassembly of Transport block here + offset = 0; + + /* + printf("harq_pid %d\n",harq_pid); + printf("F %d, Fbytes %d\n",harq_process->F,harq_process->F>>3); + printf("C %d\n",harq_process->C); + */ + uint32_t wait = 0; + if (harq_process->C==2){ + while((proc->decoder_thread_available == 0) ) + { + usleep(1); + wait++; + } + } + else if ((harq_process->C==3) ){ + while((proc->decoder_thread_available == 0) || (proc->decoder_thread_available1 == 0)) + { + usleep(1); + wait++; + } + } + + proc->decoder_main_available = 0; + + for (r=0; rC; r++) { + + Kr = harq_process->K; //to check if same K in all segments + Kr_bytes = Kr>>3; + + memcpy(harq_process->b+offset, + harq_process->c[r], + Kr_bytes- - (harq_process->F>>3) -((harq_process->C>1)?3:0)); + offset += (Kr_bytes - (harq_process->F>>3) - ((harq_process->C>1)?3:0)); + +#ifdef DEBUG_DLSCH_DECODING + printf("Segment %d : Kr= %d bytes\n",r,Kr_bytes); + printf("copied %d bytes to b sequence (harq_pid %d)\n", + (Kr_bytes - (harq_process->F>>3)-((harq_process->C>1)?3:0)),harq_pid); + printf("b[0] = %x,c[%d] = %x\n", + harq_process->b[offset], + harq_process->F>>3, + harq_process->c[r]); +#endif + } + + dlsch->last_iteration_cnt = ret; + //proc->decoder_thread_available = 0; + //proc->decoder_main_available = 0; + + //wait for worker to finish + //wait_on_busy_condition(&proc->mutex_td,&proc->cond_td,&proc->instance_cnt_dlsch td,"dlsch td thread"); + + //return( (ret>proc->tdp.ret) ? ret : proc->tdp.ret ); + + return(ret); +} +#endif + +#ifdef UE_DLSCH_PARALLELISATION +#define FIFO_PRIORITY 39 +void *nr_dlsch_decoding_2thread0(void *arg) +{ + static __thread int UE_dlsch_td_retval; + struct nr_rxtx_thread_data *rtd = arg; + UE_nr_rxtx_proc_t *proc = rtd->proc; + PHY_VARS_NR_UE *phy_vars_ue = rtd->UE; + NR_DL_FRAME_PARMS *frame_parms = &phy_vars_ue->frame_parms; + int llr8_flag1; + int32_t no_iteration_ldpc; + t_nrLDPC_dec_params decParams; + t_nrLDPC_dec_params* p_decParams = &decParams; + t_nrLDPC_time_stats procTime; + t_nrLDPC_time_stats* p_procTime =&procTime ; + int8_t llrProcBuf[OAI_LDPC_MAX_NUM_LLR] __attribute__ ((aligned(32))); + t_nrLDPC_procBuf* p_nrLDPC_procBuf; + int16_t z [68*384]; + int8_t l [68*384]; + //__m128i l; + int16_t inv_d [68*384]; + int16_t *p_invd =&inv_d; + uint8_t kb, kc; + uint8_t Ilbrm = 0; + uint32_t Tbslbrm = 950984; + uint16_t nb_rb = 30; //to update + uint16_t nb_symb_sch = 12; + uint8_t nb_re_dmrs = 6; + uint16_t length_dmrs = 1; + + uint32_t i,j; + uint32_t k; + + __m128i *pv = (__m128i*)&z; + __m128i *pl = (__m128i*)&l; + + proc->instance_cnt_dlsch_td=-1; + proc->nr_tti_rx=proc->sub_frame_start; + + proc->decoder_thread_available = 0; + + char threadname[256]; + sprintf(threadname,"UE_thread_dlsch_td_%d", proc->sub_frame_start); + + cpu_set_t cpuset; + CPU_ZERO(&cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 0 && threads.dlsch_td_one != -1 ) + CPU_SET(threads.dlsch_td_one, &cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 1 && threads.dlsch_td_two != -1 ) + CPU_SET(threads.dlsch_td_two, &cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 2 && threads.dlsch_td_three != -1 ) + CPU_SET(threads.dlsch_td_three, &cpuset); + + + +#if UE_TIMING_TRACE + time_stats_t *dlsch_rate_unmatching_stats=&phy_vars_ue->dlsch_rate_unmatching_stats; + time_stats_t *dlsch_turbo_decoding_stats=&phy_vars_ue->dlsch_turbo_decoding_stats; + time_stats_t *dlsch_deinterleaving_stats=&phy_vars_ue->dlsch_deinterleaving_stats; +#endif + uint32_t A,E; + uint32_t G; + uint32_t ret,offset; +// short dummy_w[MAX_NUM_DLSCH_SEGMENTS][3*(8448+64)]; + uint32_t r,r_offset=0,Kr,Kr_bytes,err_flag=0,K_bytes_F; + uint8_t crc_type; + uint8_t C,Cprime; + uint8_t Qm; + uint8_t Nl; + //uint32_t Er; + + init_thread(900000,1000000 , FIFO_PRIORITY-1, &cpuset, threadname); + while (!oai_exit) { + + //proc->decoder_thread_available = 1; + + if (pthread_mutex_lock(&proc->mutex_dlsch_td) != 0) { + LOG_E( PHY, "[SCHED][UE] error locking mutex for UE dlsch td\n" ); + exit_fun("nothing to add"); + } + while (proc->instance_cnt_dlsch_td < 0) { + // most of the time, the thread is waiting here + pthread_cond_wait( &proc->cond_dlsch_td, &proc->mutex_dlsch_td ); + } + if (pthread_mutex_unlock(&proc->mutex_dlsch_td) != 0) { + LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE dlsch_td \n" ); + exit_fun("nothing to add"); + } + + uint32_t wait = 0; + while(proc->decoder_main_available == 0) + { + usleep(1); + wait++; + } + + //proc->decoder_thread_available = 0; + //PHY_VARS_NR_UE *phy_vars_ue = tdp->UE; + int eNB_id = proc->eNB_id; + int harq_pid = proc->harq_pid; + llr8_flag1 = proc->llr8_flag; + //r_offset = proc->Er; + //UE_rxtx_proc_t *proc = tdp->proc; + int frame = proc->frame_rx; + int subframe = proc->nr_tti_rx; + NR_UE_DLSCH_t *dlsch = phy_vars_ue->dlsch[phy_vars_ue->current_thread_id[subframe]][eNB_id][0]; + NR_DL_UE_HARQ_t *harq_process = dlsch->harq_processes[harq_pid]; + short *dlsch_llr = phy_vars_ue->pdsch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->llr[0]; + //printf("2thread0 llr flag %d tdp flag %d\n",llr8_flag1, tdp->llr8_flag); + p_nrLDPC_procBuf = harq_process->p_nrLDPC_procBuf[1]; + + /* + if (nb_rb > frame_parms->N_RB_DL) { + printf("dlsch_decoding.c: Illegal nb_rb %d\n",nb_rb); + return(max_ldpc_iterations); + }*/ + + /*harq_pid = dlsch->current_harq_pid[phy_vars_ue->current_thread_id[subframe]]; + if (harq_pid >= 8) { + printf("dlsch_decoding.c: Illegal harq_pid %d\n",harq_pid); + return(max_ldpc_iterations); + } + */ + + nb_rb = harq_process->nb_rb; + + harq_process->trials[harq_process->round]++; + + harq_process->TBS = nr_compute_tbs(harq_process->mcs,nb_rb,nb_symb_sch,nb_re_dmrs,length_dmrs, harq_process->Nl); + + A = harq_process->TBS; //2072 for QPSK 1/3 + + ret = dlsch->max_ldpc_iterations; + + harq_process->G = nr_get_G(nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs, harq_process->Qm,harq_process->Nl); + G = harq_process->G; + + if (harq_process->round == 0) { + // This is a new packet, so compute quantities regarding segmentation + harq_process->B = A+24; + + nr_segmentation(NULL, + NULL, + harq_process->B, + &harq_process->C, + &harq_process->K, + &harq_process->Z, + &harq_process->F); + p_decParams->Z = harq_process->Z; + + } + + kb = harq_process->K/harq_process->Z; + if ( kb==22){ + p_decParams->BG = 1; + p_decParams->R = 13; + kc = 68; + } + else{ + p_decParams->BG = 2; + p_decParams->R = 13; + kc = 52; + } + + p_decParams->numMaxIter = 2; + Kr = p_decParams->Z*kb; + p_decParams->outMode= 0; + + /* + else { + printf("dlsch_decoding.c: Ndi>0 not checked yet!!\n"); + return(max_ldpc_iterations); + } + */ + err_flag = 0; + //r_offset = 0; + + /* + unsigned char bw_scaling =1; + + switch (frame_parms->N_RB_DL) { + case 106: + bw_scaling =2; + break; + + default: + bw_scaling =1; + break; + } + + if (harq_process->C > MAX_NUM_DLSCH_SEGMENTS/bw_scaling) { + LOG_E(PHY,"Illegal harq_process->C %d > %d\n",harq_process->C,MAX_NUM_DLSCH_SEGMENTS/bw_scaling); + return((1+dlsch->max_ldpc_iterations)); + }*/ +#ifdef DEBUG_DLSCH_DECODING + printf("Segmentation: C %d, Cminus %d, Kminus %d, Kplus %d\n",harq_process->C,harq_process->Cminus,harq_process->Kminus,harq_process->Kplus); +#endif + + opp_enabled=1; + + Qm= harq_process->Qm; + Nl=harq_process->Nl; + //r_thread = harq_process->C/2-1; + C= harq_process->C; + + Cprime = C; //assume CBGTI not present + + if (r <= Cprime - ((G/(Nl*Qm))%Cprime) - 1) + r_offset = Nl*Qm*(G/(Nl*Qm*Cprime)); + else + r_offset = Nl*Qm*((G/(Nl*Qm*Cprime))+1); + + // printf("thread0 r_offset %d\n",r_offset); + + //for (r=(harq_process->C/2); rC; r++) { + r=1; //(harq_process->C/2); + + + Kr = harq_process->K; + Kr_bytes = Kr>>3; + K_bytes_F = Kr_bytes-(harq_process->F>>3); + + Tbslbrm = nr_compute_tbs(28,nb_rb,frame_parms->symbols_per_slot,0,0, harq_process->Nl); + + E = nr_get_E(G, harq_process->C, harq_process->Qm, harq_process->Nl, r); + +#if UE_TIMING_TRACE + start_meas(dlsch_deinterleaving_stats); +#endif + nr_deinterleaving_ldpc(E, + harq_process->Qm, + harq_process->w[r], + dlsch_llr+r_offset); + + //for (int i =0; i<16; i++) + // printf("rx output deinterleaving w[%d]= %d r_offset %d\n", i,harq_process->w[r][i], r_offset); + +#if UE_TIMING_TRACE + stop_meas(dlsch_deinterleaving_stats); +#endif + +#if UE_TIMING_TRACE + start_meas(dlsch_rate_unmatching_stats); +#endif + +#ifdef DEBUG_DLSCH_DECODING + LOG_D(PHY,"HARQ_PID %d Rate Matching Segment %d (coded bits %d,unpunctured/repeated bits %d, TBS %d, mod_order %d, nb_rb %d, Nl %d, rv %d, round %d)...\n", + harq_pid,r, G, + Kr*3, + harq_process->TBS, + harq_process->Qm, + harq_process->nb_rb, + harq_process->Nl, + harq_process->rvidx, + harq_process->round); +#endif + + if (nr_rate_matching_ldpc_rx(Ilbrm, + Tbslbrm, + p_decParams->BG, + p_decParams->Z, + harq_process->d[r], + harq_process->w[r], + harq_process->C, + harq_process->rvidx, + (harq_process->round==0)?1:0, + E)==-1) { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + LOG_E(PHY,"dlsch_decoding.c: Problem in rate_matching\n"); + //return(dlsch->max_ldpc_iterations); + } else + { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + } + + //for (int i =0; i<16; i++) + // printf("rx output ratematching d[%d]= %d r_offset %d\n", i,harq_process->d[r][i], r_offset); + + //r_offset += E; + +#ifdef DEBUG_DLSCH_DECODING + if (r==0) { + write_output("decoder_llr.m","decllr",dlsch_llr,G,1,0); + write_output("decoder_in.m","dec",&harq_process->d[0][0],(3*8*Kr_bytes)+12,1,0); + } + + printf("decoder input(segment %d) :",r); + int i; for (i=0;i<(3*8*Kr_bytes)+12;i++) + printf("%d : %d\n",i,harq_process->d[r][i]); + printf("\n"); +#endif + + + // printf("Clearing c, %p\n",harq_process->c[r]); + memset(harq_process->c[r],0,Kr_bytes); + + // printf("done\n"); + if (harq_process->C == 1) + crc_type = CRC24_A; + else + crc_type = CRC24_B; + + +#if 1 + if (err_flag == 0) { +/* + LOG_I(PHY, "turbo algo Kr=%d cb_cnt=%d C=%d nbRB=%d crc_type %d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d maxIter %d\n", + Kr,r,harq_process->C,harq_process->nb_rb,crc_type,A,harq_process->TBS, + harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round,dlsch->max_ldpc_iterations); +*/ + if (llr8_flag1) { + AssertFatal (Kr >= 256, "turbo algo issue Kr=%d cb_cnt=%d C=%d nbRB=%d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d\n", + Kr,r,harq_process->C,harq_process->nb_rb,A,harq_process->TBS,harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round); + } +#if UE_TIMING_TRACE + start_meas(dlsch_turbo_decoding_stats); +#endif +// LOG_D(PHY,"AbsSubframe %d.%d Start turbo segment %d/%d \n",frame%1024,subframe,r,harq_process->C-1); + + for (int cnt =0; cnt < (kc-2)*p_decParams->Z; cnt++){ + inv_d[cnt] = (1)*harq_process->d[r][cnt]; + } + + memset(pv,0,2*p_decParams->Z*sizeof(int16_t)); + //memset(pl,0,2*p_decParams->Z*sizeof(int8_t)); + memset((pv+K_bytes_F),127,harq_process->F*sizeof(int16_t)); + + for (i=((2*p_decParams->Z)>>3), j = 0; i < K_bytes_F+((2*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=Kr_bytes+((2*p_decParams->Z)>>3),j=Kr_bytes; i < ((kc*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=0, j=0; j < ((kc*p_decParams->Z)>>4); i+=2, j++) + { + + pl[j] = _mm_packs_epi16(pv[i],pv[i+1]); + + } + + no_iteration_ldpc = nrLDPC_decoder(p_decParams, + (int8_t*)&pl[0], + llrProcBuf, + p_nrLDPC_procBuf, + p_procTime); + + if (no_iteration_ldpc > 10) + printf("Error number of iteration LPDC %d\n", no_iteration_ldpc); + //else + //printf("OK number of iteration LPDC %d\n", no_iteration_ldpc); + + for (int m=0; m < Kr>>3; m ++) + { + harq_process->c[r][m]= (uint8_t) llrProcBuf[m]; + } + + /*for (int u=0; u < Kr>>3; u ++) + { + ullrProcBuf[u]= (uint8_t) llrProcBuf[u]; + } + + + printf("output unsigned ullrProcBuf \n"); + + for (int j=0; j < Kr>>3; j ++) + { + + printf(" %d \n", ullrProcBuf[j]); + + } + printf(" \n");*/ +#ifdef DEBUG_DLSCH_DECODING + for (int k=0;k<2;k++) + printf("segment 1 output decoder [%d] = 0x%02x \n", k, harq_process->c[r][k]); +#endif + +#if UE_TIMING_TRACE + stop_meas(dlsch_turbo_decoding_stats); +#endif + } + + if ((err_flag == 0) && (ret>=(1+dlsch->max_ldpc_iterations))) {// a Code segment is in error so break; +// LOG_D(PHY,"AbsSubframe %d.%d CRC failed, segment %d/%d \n",frame%1024,subframe,r,harq_process->C-1); + err_flag = 1; + } + //} + + /*int32_t frame_rx_prev = frame; + int32_t subframe_rx_prev = subframe - 1; + if (subframe_rx_prev < 0) { + frame_rx_prev--; + subframe_rx_prev += 10; + } + frame_rx_prev = frame_rx_prev%1024;*/ +#if 0 + if (err_flag == 1) { +//#if UE_DEBUG_TRACE + LOG_I(PHY,"[UE %d] THREAD 0 DLSCH: Setting NAK for SFN/SF %d/%d (pid %d, status %d, round %d, TBS %d, mcs %d) Kr %d r %d harq_process->round %d\n", + phy_vars_ue->Mod_id, frame, subframe, harq_pid,harq_process->status, harq_process->round,harq_process->TBS,harq_process->mcs,Kr,r,harq_process->round); +//#endif + dlsch->harq_ack[subframe].ack = 0; + dlsch->harq_ack[subframe].harq_id = harq_pid; + dlsch->harq_ack[subframe].send_harq_status = 1; + harq_process->errors[harq_process->round]++; + harq_process->round++; + + + // printf("Rate: [UE %d] DLSCH: Setting NACK for subframe %d (pid %d, round %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round); + if (harq_process->round >= dlsch->Mdlharq) { + harq_process->status = SCH_IDLE; + harq_process->round = 0; + } +/* if(is_crnti) + { + LOG_D(PHY,"[UE %d] DLSCH: Setting NACK for subframe %d (pid %d, pid status %d, round %d/Max %d, TBS %d)\n", + phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->status,harq_process->round,dlsch->Mdlharq,harq_process->TBS); + }*/ + + + //return((1+dlsch->max_ldpc_iterations)); + } else { +#if UE_DEBUG_TRACE + LOG_I(PHY,"[UE %d] THREAD 0 DLSCH: Setting ACK for subframe %d TBS %d mcs %d nb_rb %d\n", + phy_vars_ue->Mod_id,subframe,harq_process->TBS,harq_process->mcs,harq_process->nb_rb); +#endif + + harq_process->status = SCH_IDLE; + harq_process->round = 0; + dlsch->harq_ack[subframe].ack = 1; + dlsch->harq_ack[subframe].harq_id = harq_pid; + dlsch->harq_ack[subframe].send_harq_status = 1; + //LOG_I(PHY,"[UE %d] DLSCH: Setting ACK for SFN/SF %d/%d (pid %d, status %d, round %d, TBS %d, mcs %d)\n", + // phy_vars_ue->Mod_id, frame, subframe, harq_pid, harq_process->status, harq_process->round,harq_process->TBS,harq_process->mcs); + +/* if(is_crnti) + { + LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for subframe %d (pid %d, round %d, TBS %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round,harq_process->TBS); + } + LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for subframe %d (pid %d, round %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round); + + }*/ + + // Reassembly of Transport block here + offset = 0; + + /* + printf("harq_pid %d\n",harq_pid); + printf("F %d, Fbytes %d\n",harq_process->F,harq_process->F>>3); + printf("C %d\n",harq_process->C); + */ + for (r=0; rC; r++) { + if (rCminus) + Kr = harq_process->Kminus; + else + Kr = harq_process->Kplus; + + Kr_bytes = Kr>>3; + + // printf("Segment %d : Kr= %d bytes\n",r,Kr_bytes); + if (r==0) { + memcpy(harq_process->b, + &harq_process->c[0][(harq_process->F>>3)], + Kr_bytes - (harq_process->F>>3)- ((harq_process->C>1)?3:0)); + offset = Kr_bytes - (harq_process->F>>3) - ((harq_process->C>1)?3:0); + // printf("copied %d bytes to b sequence (harq_pid %d)\n", + // Kr_bytes - (harq_process->F>>3),harq_pid); + // printf("b[0] = %x,c[%d] = %x\n", + // harq_process->b[0], + // harq_process->F>>3, + // harq_process->c[0][(harq_process->F>>3)]); + } else { + memcpy(harq_process->b+offset, + harq_process->c[r], + Kr_bytes- ((harq_process->C>1)?3:0)); + offset += (Kr_bytes - ((harq_process->C>1)?3:0)); + } + } + + dlsch->last_iteration_cnt = ret; + + //return(ret); + } +#endif + + proc->decoder_thread_available = 1; + //proc->decoder_main_available = 0; + + if (pthread_mutex_lock(&proc->mutex_dlsch_td) != 0) { + LOG_E( PHY, "[SCHED][UE] error locking mutex for UE RXTX\n" ); + exit_fun("noting to add"); + } + proc->instance_cnt_dlsch_td--; + if (pthread_mutex_unlock(&proc->mutex_dlsch_td) != 0) { + LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE td1\n" ); + exit_fun("noting to add"); + } + } + + // thread finished + free(arg); + return &UE_dlsch_td_retval; +} +#endif + +#ifdef UE_DLSCH_PARALLELISATION +#define FIFO_PRIORITY 39 +void *nr_dlsch_decoding_2thread1(void *arg) +{ + static __thread int UE_dlsch_td_retval1; + struct nr_rxtx_thread_data *rtd = arg; + UE_nr_rxtx_proc_t *proc = rtd->proc; + PHY_VARS_NR_UE *phy_vars_ue = rtd->UE; + NR_DL_FRAME_PARMS *frame_parms = &phy_vars_ue->frame_parms; + + int llr8_flag1; + int32_t no_iteration_ldpc; + t_nrLDPC_dec_params decParams; + t_nrLDPC_dec_params* p_decParams = &decParams; + t_nrLDPC_time_stats procTime; + t_nrLDPC_time_stats* p_procTime =&procTime ; + t_nrLDPC_procBuf* p_nrLDPC_procBuf; + int8_t llrProcBuf[OAI_LDPC_MAX_NUM_LLR] __attribute__ ((aligned(32))); + int16_t z [68*384]; + int8_t l [68*384]; + //__m128i l; + int16_t inv_d [68*384]; + //int16_t *p_invd =&inv_d; + uint8_t kb, kc; + uint8_t Ilbrm = 0; + uint32_t Tbslbrm = 950984; + uint16_t nb_rb = 30; //to update + uint16_t nb_symb_sch = 12; + uint8_t nb_re_dmrs = 6; + uint16_t length_dmrs = 1; + + uint32_t i,j; + //uint32_t k; + + __m128i *pv = (__m128i*)&z; + __m128i *pl = (__m128i*)&l; + + proc->instance_cnt_dlsch_td1=-1; + proc->nr_tti_rx=proc->sub_frame_start; + + printf("start thread 1\n"); + proc->decoder_thread_available1 = 0; + + char threadname[256]; + sprintf(threadname,"UE_thread_dlsch_td1_%d", proc->sub_frame_start); + + cpu_set_t cpuset; + CPU_ZERO(&cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 0 && threads.dlsch_td_one != -1 ) + CPU_SET(threads.dlsch_td_one, &cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 1 && threads.dlsch_td_two != -1 ) + CPU_SET(threads.dlsch_td_two, &cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 2 && threads.dlsch_td_three != -1 ) + CPU_SET(threads.dlsch_td_three, &cpuset); + + + +#if UE_TIMING_TRACE + time_stats_t *dlsch_rate_unmatching_stats=&phy_vars_ue->dlsch_rate_unmatching_stats; + time_stats_t *dlsch_turbo_decoding_stats=&phy_vars_ue->dlsch_turbo_decoding_stats; + time_stats_t *dlsch_deinterleaving_stats=&phy_vars_ue->dlsch_deinterleaving_stats; +#endif + uint32_t A,E; + uint32_t G; + uint32_t ret,offset; + uint32_t r,r_offset=0,Kr,Kr_bytes,err_flag=0,K_bytes_F; + uint8_t crc_type; + uint8_t C,Cprime; + uint8_t Qm; + uint8_t Nl; + //uint32_t Er; + + init_thread(900000,1000000 , FIFO_PRIORITY-1, &cpuset, threadname); + printf("2thread1 oai_exit %d\n", oai_exit); + while (!oai_exit) { + if (pthread_mutex_lock(&proc->mutex_dlsch_td1) != 0) { + LOG_E( PHY, "[SCHED][UE] error locking mutex for UE dlsch td\n" ); + exit_fun("nothing to add"); + } + while (proc->instance_cnt_dlsch_td1 < 0) { + // most of the time, the thread is waiting here + pthread_cond_wait( &proc->cond_dlsch_td1, &proc->mutex_dlsch_td1 ); + } + if (pthread_mutex_unlock(&proc->mutex_dlsch_td1) != 0) { + LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE dlsch_td \n" ); + exit_fun("nothing to add"); + } + + //printf("2thread1 main available %d\n", proc->decoder_main_available); + + uint32_t wait = 0; + while(proc->decoder_main_available == 0) + { + usleep(1); + wait++; + } + + //proc->decoder_thread_available1 = 0; + //PHY_VARS_NR_UE *phy_vars_ue = tdp->UE; + int eNB_id = proc->eNB_id; + int harq_pid = proc->harq_pid; + llr8_flag1 = proc->llr8_flag; + //r_offset = proc->Er; + //UE_rxtx_proc_t *proc = tdp->proc; + int frame = proc->frame_rx; + int subframe = proc->nr_tti_rx; + NR_UE_DLSCH_t *dlsch = phy_vars_ue->dlsch[phy_vars_ue->current_thread_id[subframe]][eNB_id][0]; + NR_DL_UE_HARQ_t *harq_process = dlsch->harq_processes[harq_pid]; + short *dlsch_llr = phy_vars_ue->pdsch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->llr[0]; + //printf("2thread0 llr flag %d tdp flag %d\n",llr8_flag1, tdp->llr8_flag); + //printf("2thread1 nr_tti_tx %d subframe %d SF thread id %d r_offset %d\n", proc->nr_tti_rx, subframe, phy_vars_ue->current_thread_id[subframe], r_offset); + p_nrLDPC_procBuf = harq_process->p_nrLDPC_procBuf[2]; + + /* + if (nb_rb > frame_parms->N_RB_DL) { + printf("dlsch_decoding.c: Illegal nb_rb %d\n",nb_rb); + return(max_ldpc_iterations); + }*/ + + /*harq_pid = dlsch->current_harq_pid[phy_vars_ue->current_thread_id[subframe]]; + if (harq_pid >= 8) { + printf("dlsch_decoding.c: Illegal harq_pid %d\n",harq_pid); + return(max_ldpc_iterations); + } + */ + + nb_rb = harq_process->nb_rb; + + harq_process->trials[harq_process->round]++; + harq_process->TBS = nr_compute_tbs(harq_process->mcs,nb_rb,nb_symb_sch,nb_re_dmrs,length_dmrs, harq_process->Nl); + + A = harq_process->TBS; //2072 for QPSK 1/3 + + ret = dlsch->max_ldpc_iterations; + + harq_process->G = nr_get_G(nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs, harq_process->Qm,harq_process->Nl); + + G = harq_process->G; + //get_G(frame_parms,nb_rb,dlsch->rb_alloc,mod_order,num_pdcch_symbols,phy_vars_ue->frame,subframe); + + //printf("DLSCH Decoding, A %d harq_pid %d G %d\n",A, harq_pid,harq_process->G); + + if (harq_process->round == 0) { + // This is a new packet, so compute quantities regarding segmentation + harq_process->B = A+24; + nr_segmentation(NULL, + NULL, + harq_process->B, + &harq_process->C, + &harq_process->K, + &harq_process->Z, + &harq_process->F); + p_decParams->Z = harq_process->Z; + + } + + kb = harq_process->K/harq_process->Z; + if ( kb==22){ + p_decParams->BG = 1; + p_decParams->R = 89; + kc = 68; + } + else{ + p_decParams->BG = 2; + p_decParams->R = 13; + kc = 52; + } + + p_decParams->numMaxIter = 2; + Kr = p_decParams->Z*kb; + p_decParams->outMode= 0; + /* + else { + printf("dlsch_decoding.c: Ndi>0 not checked yet!!\n"); + return(max_ldpc_iterations); + } + */ + err_flag = 0; + //r_offset = 0; + + /* + unsigned char bw_scaling =1; + + switch (frame_parms->N_RB_DL) { + case 106: + bw_scaling =2; + break; + + default: + bw_scaling =1; + break; + } + + if (harq_process->C > MAX_NUM_DLSCH_SEGMENTS/bw_scaling) { + LOG_E(PHY,"Illegal harq_process->C %d > %d\n",harq_process->C,MAX_NUM_DLSCH_SEGMENTS/bw_scaling); + return((1+dlsch->max_ldpc_iterations)); + }*/ +#ifdef DEBUG_DLSCH_DECODING + printf("Segmentation: C %d, Cminus %d, Kminus %d, Kplus %d\n",harq_process->C,harq_process->Cminus,harq_process->Kminus,harq_process->Kplus); +#endif + + opp_enabled=1; + + Qm= harq_process->Qm; + Nl=harq_process->Nl; + //r_thread = harq_process->C/2-1; + C= harq_process->C; + + Cprime = C; //assume CBGTI not present + + if (r <= Cprime - ((G/(Nl*Qm))%Cprime) - 1) + r_offset = Nl*Qm*(G/(Nl*Qm*Cprime)); + else + r_offset = Nl*Qm*((G/(Nl*Qm*Cprime))+1); + + //printf("sub thread r_offset %d\n", r_offset); + + //for (r=(harq_process->C/2); rC; r++) { + r=2; //(harq_process->C/2); + r_offset = r*r_offset; + //printf("thread1 r=%d r_offset %d \n",r, r_offset); + + + Kr = harq_process->K; + Kr_bytes = Kr>>3; + K_bytes_F = Kr_bytes-(harq_process->F>>3); + + Tbslbrm = nr_compute_tbs(28,nb_rb,frame_parms->symbols_per_slot,0,0, harq_process->Nl); + + E = nr_get_E(G, harq_process->C, harq_process->Qm, harq_process->Nl, r); + + /* + printf("Subblock deinterleaving, d %p w %p\n", + harq_process->d[r], + harq_process->w); + */ +#if UE_TIMING_TRACE + start_meas(dlsch_deinterleaving_stats); +#endif + nr_deinterleaving_ldpc(E, + harq_process->Qm, + harq_process->w[r], + dlsch_llr+r_offset); + + //for (int i =0; i<16; i++) + // printf("rx output deinterleaving w[%d]= %d r_offset %d\n", i,harq_process->w[r][i], r_offset); + +#if UE_TIMING_TRACE + stop_meas(dlsch_deinterleaving_stats); +#endif + +#if UE_TIMING_TRACE + start_meas(dlsch_rate_unmatching_stats); +#endif + +#ifdef DEBUG_DLSCH_DECODING + LOG_D(PHY,"HARQ_PID %d Rate Matching Segment %d (coded bits %d,unpunctured/repeated bits %d, TBS %d, mod_order %d, nb_rb %d, Nl %d, rv %d, round %d)...\n", + harq_pid,r, G, + Kr*3, + harq_process->TBS, + harq_process->Qm, + harq_process->nb_rb, + harq_process->Nl, + harq_process->rvidx, + harq_process->round); +#endif + + if (nr_rate_matching_ldpc_rx(Ilbrm, + Tbslbrm, + p_decParams->BG, + p_decParams->Z, + harq_process->d[r], + harq_process->w[r], + harq_process->C, + harq_process->rvidx, + (harq_process->round==0)?1:0, + E)==-1) { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + LOG_E(PHY,"dlsch_decoding.c: Problem in rate_matching\n"); + //return(dlsch->max_ldpc_iterations); + } else + { +#if UE_TIMING_TRACE + stop_meas(dlsch_rate_unmatching_stats); +#endif + } + + //for (int i =0; i<16; i++) + // printf("rx output ratematching d[%d]= %d r_offset %d\n", i,harq_process->d[r][i], r_offset); + + //r_offset += E; + +#ifdef DEBUG_DLSCH_DECODING + if (r==0) { + write_output("decoder_llr.m","decllr",dlsch_llr,G,1,0); + write_output("decoder_in.m","dec",&harq_process->d[0][0],(3*8*Kr_bytes)+12,1,0); + } + + printf("decoder input(segment %d) :",r); + int i; for (i=0;i<(3*8*Kr_bytes)+12;i++) + printf("%d : %d\n",i,harq_process->d[r][i]); + printf("\n"); +#endif + + + // printf("Clearing c, %p\n",harq_process->c[r]); + memset(harq_process->c[r],0,Kr_bytes); + + // printf("done\n"); + if (harq_process->C == 1) + crc_type = CRC24_A; + else + crc_type = CRC24_B; + + if (err_flag == 0) { +/* + LOG_I(PHY, "turbo algo Kr=%d cb_cnt=%d C=%d nbRB=%d crc_type %d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d maxIter %d\n", + Kr,r,harq_process->C,harq_process->nb_rb,crc_type,A,harq_process->TBS, + harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round,dlsch->max_ldpc_iterations); +*/ + if (llr8_flag1) { + AssertFatal (Kr >= 256, "turbo algo issue Kr=%d cb_cnt=%d C=%d nbRB=%d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d\n", + Kr,r,harq_process->C,harq_process->nb_rb,A,harq_process->TBS,harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round); + } +#if UE_TIMING_TRACE + start_meas(dlsch_turbo_decoding_stats); +#endif +// LOG_D(PHY,"AbsSubframe %d.%d Start turbo segment %d/%d \n",frame%1024,subframe,r,harq_process->C-1); + + memset(pv,0,2*p_decParams->Z*sizeof(int16_t)); + //memset(pl,0,2*p_decParams->Z*sizeof(int8_t)); + memset((pv+K_bytes_F),127,harq_process->F*sizeof(int16_t)); + + for (i=((2*p_decParams->Z)>>3), j = 0; i < K_bytes_F+((2*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=Kr_bytes+((2*p_decParams->Z)>>3),j=Kr_bytes; i < ((kc*p_decParams->Z)>>3); i++, j++) + { + pv[i]= _mm_loadu_si128((__m128i*)(&inv_d[8*j])); + } + + for (i=0, j=0; j < ((kc*p_decParams->Z)>>4); i+=2, j++) + { + + pl[j] = _mm_packs_epi16(pv[i],pv[i+1]); + + } + + no_iteration_ldpc = nrLDPC_decoder(p_decParams, + (int8_t*)&pl[0], + llrProcBuf, + p_nrLDPC_procBuf, + p_procTime); + + if (no_iteration_ldpc > 10) + printf("Error number of iteration LPDC %d\n", no_iteration_ldpc); + //else + // printf("OK number of iteration LPDC %d\n", no_iteration_ldpc); + + for (int m=0; m < Kr>>3; m ++) + { + harq_process->c[r][m]= (uint8_t) llrProcBuf[m]; + } + + /*for (int u=0; u < Kr>>3; u ++) + { + ullrProcBuf[u]= (uint8_t) llrProcBuf[u]; + } + + + printf("output unsigned ullrProcBuf \n"); + + for (int j=0; j < Kr>>3; j ++) + { + + printf(" %d \n", ullrProcBuf[j]); + + } + printf(" \n");*/ +#endif + //printf("output channel decoder %d %d %d %d %d \n", harq_process->c[r][0], harq_process->c[r][1], harq_process->c[r][2],harq_process->c[r][3], harq_process->c[r][4]); + + //printf("output decoder %d %d %d %d %d \n", harq_process->c[r][0], harq_process->c[r][1], harq_process->c[r][2],harq_process->c[r][3], harq_process->c[r][4]); + +#if UE_TIMING_TRACE + stop_meas(dlsch_turbo_decoding_stats); +#endif + } + + if ((err_flag == 0) && (ret>=(1+dlsch->max_ldpc_iterations))) {// a Code segment is in error so break; +// LOG_D(PHY,"AbsSubframe %d.%d CRC failed, segment %d/%d \n",frame%1024,subframe,r,harq_process->C-1); + err_flag = 1; + } + //} + + /*int32_t frame_rx_prev = frame; + int32_t subframe_rx_prev = subframe - 1; + if (subframe_rx_prev < 0) { + frame_rx_prev--; + subframe_rx_prev += 10; + } + frame_rx_prev = frame_rx_prev%1024;*/ + + + proc->decoder_thread_available1 = 1; + //proc->decoder_main_available = 0; + + //printf("2thread1 proc->instance_cnt_dlsch_td1 %d\n", proc->instance_cnt_dlsch_td1); + + if (pthread_mutex_lock(&proc->mutex_dlsch_td1) != 0) { + LOG_E( PHY, "[SCHED][UE] error locking mutex for UE RXTX\n" ); + exit_fun("noting to add"); + } + proc->instance_cnt_dlsch_td1--; + if (pthread_mutex_unlock(&proc->mutex_dlsch_td1) != 0) { + LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE td1\n" ); + exit_fun("noting to add"); + } + //printf("end 2thread1 proc->instance_cnt_dlsch_td1 %d\n", proc->instance_cnt_dlsch_td1); + } + + //printf("after 2thread1 after oai exit proc->instance_cnt_dlsch_td %d\n", proc->instance_cnt_dlsch_td1); + // thread finished + free(arg); + return &UE_dlsch_td_retval1; +} +#endif + diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c b/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c new file mode 100644 index 0000000000..a9f6b02567 --- /dev/null +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c @@ -0,0 +1,2236 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.0 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/LTE_TRANSPORT/dlsch_demodulation.c + * \brief Top-level routines for demodulating the PDSCH physical channel from 38-211, V15.2 2018-06 + * \author H.Wang + * \date 2018 + * \version 0.1 + * \company Eurecom + * \note + * \warning + */ +#include "PHY/defs_nr_UE.h" +#include "PHY/phy_extern_nr_ue.h" +#include "nr_transport_proto_ue.h" +//#include "SCHED/defs.h" +//#include "PHY/defs.h" +//#include "extern.h" +#include "PHY/sse_intrin.h" +#include "T.h" + +#ifndef USER_MODE +#define NOCYGWIN_STATIC static +#else +#define NOCYGWIN_STATIC +#endif + +/* dynamic shift for LLR computation for TM3/4 + * set as command line argument, see lte-softmodem.c + * default value: 0 + */ +int16_t nr_dlsch_demod_shift = 0; +//int16_t interf_unaw_shift = 13; + +//#define DEBUG_HARQ + +//#define DEBUG_PHY 1 +//#define DEBUG_DLSCH_DEMOD 1 + + + +// [MCS][i_mod (0,1,2) = (2,4,6)] +//unsigned char offset_mumimo_llr_drange_fix=0; +//inferference-free case +/*unsigned char interf_unaw_shift_tm4_mcs[29]={5, 3, 4, 3, 3, 2, 1, 1, 2, 0, 1, 1, 1, 1, 0, 0, + 1, 1, 1, 1, 0, 2, 1, 0, 1, 0, 1, 0, 0} ;*/ + +//unsigned char interf_unaw_shift_tm1_mcs[29]={5, 5, 4, 3, 3, 3, 2, 2, 4, 4, 2, 3, 3, 3, 1, 1, +// 0, 1, 1, 2, 5, 4, 4, 6, 5, 1, 0, 5, 6} ; // mcs 21, 26, 28 seem to be errorneous + +/* +unsigned char offset_mumimo_llr_drange[29][3]={{8,8,8},{7,7,7},{7,7,7},{7,7,7},{6,6,6},{6,6,6},{6,6,6},{5,5,5},{4,4,4},{1,2,4}, // QPSK +{5,5,4},{5,5,5},{5,5,5},{3,3,3},{2,2,2},{2,2,2},{2,2,2}, // 16-QAM +{2,2,1},{3,3,3},{3,3,3},{3,3,1},{2,2,2},{2,2,2},{0,0,0},{0,0,0},{0,0,0},{0,0,0},{0,0,0},{0,0,0}}; //64-QAM +*/ + /* + //first optimization try + unsigned char offset_mumimo_llr_drange[29][3]={{7, 8, 7},{6, 6, 7},{6, 6, 7},{6, 6, 6},{5, 6, 6},{5, 5, 6},{5, 5, 6},{4, 5, 4},{4, 3, 4},{3, 2, 2},{6, 5, 5},{5, 4, 4},{5, 5, 4},{3, 3, 2},{2, 2, 1},{2, 1, 1},{2, 2, 2},{3, 3, 3},{3, 3, 2},{3, 3, 2},{3, 2, 1},{2, 2, 2},{2, 2, 2},{0, 0, 0},{0, 0, 0},{0, 0, 0},{0, 0, 0},{0, 0, 0}}; + */ + //second optimization try + /* + unsigned char offset_mumimo_llr_drange[29][3]={{5, 8, 7},{4, 6, 8},{3, 6, 7},{7, 7, 6},{4, 7, 8},{4, 7, 4},{6, 6, 6},{3, 6, 6},{3, 6, 6},{1, 3, 4},{1, 1, 0},{3, 3, 2},{3, 4, 1},{4, 0, 1},{4, 2, 2},{3, 1, 2},{2, 1, 0},{2, 1, 1},{1, 0, 1},{1, 0, 1},{0, 0, 0},{1, 0, 0},{0, 0, 0},{0, 1, 0},{1, 0, 0},{0, 0, 0},{0, 0, 0},{0, 0, 0},{0, 0, 0}}; w + */ +//unsigned char offset_mumimo_llr_drange[29][3]= {{0, 6, 5},{0, 4, 5},{0, 4, 5},{0, 5, 4},{0, 5, 6},{0, 5, 3},{0, 4, 4},{0, 4, 4},{0, 3, 3},{0, 1, 2},{1, 1, 0},{1, 3, 2},{3, 4, 1},{2, 0, 0},{2, 2, 2},{1, 1, 1},{2, 1, 0},{2, 1, 1},{1, 0, 1},{1, 0, 1},{0, 0, 0},{1, 0, 0},{0, 0, 0},{0, 1, 0},{1, 0, 0},{0, 0, 0},{0, 0, 0},{0, 0, 0},{0, 0, 0}}; + + +extern void print_shorts(char *s,int16_t *x); + + +int nr_rx_pdsch(PHY_VARS_NR_UE *ue, + PDSCH_t type, + unsigned char eNB_id, + unsigned char eNB_id_i, //if this == ue->n_connected_eNB, we assume MU interference + uint32_t frame, + uint8_t nr_tti_rx, + unsigned char symbol, + unsigned char first_symbol_flag, + RX_type_t rx_type, + unsigned char i_mod, + unsigned char harq_pid) +{ + + NR_UE_COMMON *common_vars = &ue->common_vars; + NR_UE_PDSCH **pdsch_vars; + NR_DL_FRAME_PARMS *frame_parms = &ue->frame_parms; + PHY_NR_MEASUREMENTS *measurements = &ue->measurements; + NR_UE_DLSCH_t **dlsch; + + int avg[4]; +// int avg_0[2]; +// int avg_1[2]; + +#if UE_TIMING_TRACE + uint8_t slot = 0; +#endif + + unsigned char aatx,aarx; + + unsigned short nb_rb = 0, nb_re =0, round; + int avgs = 0;// rb; + NR_DL_UE_HARQ_t *dlsch0_harq,*dlsch1_harq = 0; + + uint8_t beamforming_mode; + uint32_t *rballoc; + + int32_t **rxdataF_comp_ptr; + int32_t **dl_ch_mag_ptr; + int32_t codeword_TB0 = -1; + int32_t codeword_TB1 = -1; + + //to be updated higher layer + unsigned short start_rb = 0; + unsigned short nb_rb_pdsch = 50; + int16_t *pllr_symbol_cw0; + int16_t *pllr_symbol_cw1; + //int16_t *pllr_symbol_cw0_deint; + //int16_t *pllr_symbol_cw1_deint; + uint32_t llr_offset_symbol; + //uint16_t bundle_L = 2; + uint8_t l0 =2; + + switch (type) { + case SI_PDSCH: + pdsch_vars = &ue->pdsch_vars_SI[eNB_id]; + dlsch = &ue->dlsch_SI[eNB_id]; + dlsch0_harq = dlsch[0]->harq_processes[harq_pid]; + beamforming_mode = 0; + break; + + case RA_PDSCH: + pdsch_vars = &ue->pdsch_vars_ra[eNB_id]; + dlsch = &ue->dlsch_ra[eNB_id]; + dlsch0_harq = dlsch[0]->harq_processes[harq_pid]; + beamforming_mode = 0; + break; + + case PDSCH: + pdsch_vars = ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]]; + dlsch = ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id]; + + + //set active for testing -> to be removed + dlsch[0]->harq_processes[harq_pid]->status = ACTIVE; + dlsch[0]->harq_processes[harq_pid]->Qm = 2; + dlsch[0]->harq_processes[harq_pid]->mcs = 9; + dlsch[0]->harq_processes[harq_pid]->Nl=1; + dlsch[0]->harq_processes[harq_pid]->round=0; + dlsch[0]->harq_processes[harq_pid]->nb_rb = nb_rb_pdsch; + frame_parms->nushift = 0; + + //printf("status TB0 = %d, status TB1 = %d \n", dlsch[0]->harq_processes[harq_pid]->status, dlsch[1]->harq_processes[harq_pid]->status); + LOG_D(PHY,"AbsSubframe %d.%d / Sym %d harq_pid %d, harq status %d.%d \n", + frame,nr_tti_rx,symbol,harq_pid, + dlsch[0]->harq_processes[harq_pid]->status, + dlsch[1]->harq_processes[harq_pid]->status); + + if ((dlsch[0]->harq_processes[harq_pid]->status == ACTIVE) && + (dlsch[1]->harq_processes[harq_pid]->status == ACTIVE)){ + codeword_TB0 = dlsch[0]->harq_processes[harq_pid]->codeword; + codeword_TB1 = dlsch[1]->harq_processes[harq_pid]->codeword; + dlsch0_harq = dlsch[codeword_TB0]->harq_processes[harq_pid]; + dlsch1_harq = dlsch[codeword_TB1]->harq_processes[harq_pid]; +#ifdef DEBUG_HARQ + printf("[DEMOD] I am assuming both TBs are active\n"); +#endif + } + else if ((dlsch[0]->harq_processes[harq_pid]->status == ACTIVE) && + (dlsch[1]->harq_processes[harq_pid]->status != ACTIVE) ) { + codeword_TB0 = dlsch[0]->harq_processes[harq_pid]->codeword; + dlsch0_harq = dlsch[0]->harq_processes[harq_pid]; + dlsch1_harq = NULL; + codeword_TB1 = -1; +#ifdef DEBUG_HARQ + printf("[DEMOD] I am assuming only TB0 is active\n"); +#endif + } + else if ((dlsch[0]->harq_processes[harq_pid]->status != ACTIVE) && + (dlsch[1]->harq_processes[harq_pid]->status == ACTIVE) ){ + codeword_TB1 = dlsch[1]->harq_processes[harq_pid]->codeword; + dlsch0_harq = dlsch[1]->harq_processes[harq_pid]; + dlsch1_harq = NULL; + codeword_TB0 = -1; +#ifdef DEBUG_HARQ + printf("[DEMOD] I am assuming only TB1 is active, it is in cw %d\n", dlsch0_harq->codeword); +#endif + } + else { + LOG_E(PHY,"[UE][FATAL] Frame %d nr_tti_rx %d: no active DLSCH\n",ue->proc.proc_rxtx[0].frame_rx,nr_tti_rx); + return(-1); + } + beamforming_mode = ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id]; + break; + + default: + LOG_E(PHY,"[UE][FATAL] Frame %d nr_tti_rx %d: Unknown PDSCH format %d\n",ue->proc.proc_rxtx[0].frame_rx,nr_tti_rx,type); + return(-1); + break; + } +#ifdef DEBUG_HARQ + printf("[DEMOD] MIMO mode = %d\n", dlsch0_harq->mimo_mode); + printf("[DEMOD] cw for TB0 = %d, cw for TB1 = %d\n", codeword_TB0, codeword_TB1); +#endif + + DevAssert(dlsch0_harq); + round = dlsch0_harq->round; + //printf("round = %d\n", round); + + if (eNB_id > 2) { + LOG_W(PHY,"dlsch_demodulation.c: Illegal eNB_id %d\n",eNB_id); + return(-1); + } + + if (!common_vars) { + LOG_W(PHY,"dlsch_demodulation.c: Null common_vars\n"); + return(-1); + } + + if (!dlsch[0]) { + LOG_W(PHY,"dlsch_demodulation.c: Null dlsch_ue pointer\n"); + return(-1); + } + + if (!pdsch_vars) { + LOG_W(PHY,"dlsch_demodulation.c: Null pdsch_vars pointer\n"); + return(-1); + } + + if (!frame_parms) { + LOG_W(PHY,"dlsch_demodulation.c: Null frame_parms\n"); + return(-1); + } + + if (((frame_parms->Ncp == NORMAL) && (symbol>=7)) || + ((frame_parms->Ncp == EXTENDED) && (symbol>=6))) + rballoc = dlsch0_harq->rb_alloc_odd; + else + rballoc = dlsch0_harq->rb_alloc_even; + + + if (dlsch0_harq->mimo_mode>DUALSTREAM_PUSCH_PRECODING) { + LOG_E(PHY,"This transmission mode is not yet supported!\n"); + return(-1); + } + + + if ((dlsch0_harq->mimo_mode==LARGE_CDD) || ((dlsch0_harq->mimo_mode>=DUALSTREAM_UNIFORM_PRECODING1) && (dlsch0_harq->mimo_mode<=DUALSTREAM_PUSCH_PRECODING))) { + DevAssert(dlsch1_harq); + if (eNB_id!=eNB_id_i) { + LOG_E(PHY,"TM3/TM4 requires to set eNB_id==eNB_id_i!\n"); + return(-1); + } + } + +#if UE_TIMING_TRACE + if(symbol > ue->frame_parms.symbols_per_tti>>1) + { + slot = 1; + } +#endif + +#ifdef DEBUG_HARQ + printf("Demod dlsch0_harq->pmi_alloc %d\n", dlsch0_harq->pmi_alloc); +#endif +#if 0 + if (frame_parms->nb_antenna_ports_eNB>1 && beamforming_mode==0) { +#ifdef DEBUG_DLSCH_MOD + LOG_I(PHY,"dlsch: using pmi %x (%p), rb_alloc %x\n",pmi2hex_2Ar1(dlsch0_harq->pmi_alloc),dlsch[0],dlsch0_harq->rb_alloc_even[0]); +#endif + +#if UE_TIMING_TRACE + start_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#endif + nb_rb = dlsch_extract_rbs_dual(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, + common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id], + pdsch_vars[eNB_id]->rxdataF_ext, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + dlsch0_harq->pmi_alloc, + pdsch_vars[eNB_id]->pmi_ext, + rballoc, + symbol, + nr_tti_rx, + ue->high_speed_flag, + frame_parms, + dlsch0_harq->mimo_mode); +#ifdef DEBUG_DLSCH_MOD + printf("dlsch: using pmi %lx, rb_alloc %x, pmi_ext ",pmi2hex_2Ar1(dlsch0_harq->pmi_alloc),*rballoc); + for (rb=0;rbpmi_ext[rb]); + printf("\n"); +#endif + + + if (rx_type >= rx_IC_single_stream) { + if (eNB_id_in_connected_eNB) // we are in TM5 + nb_rb = dlsch_extract_rbs_dual(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, + common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id_i], + pdsch_vars[eNB_id_i]->rxdataF_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + dlsch0_harq->pmi_alloc, + pdsch_vars[eNB_id_i]->pmi_ext, + rballoc, + symbol, + nr_tti_rx, + ue->high_speed_flag, + frame_parms, + dlsch0_harq->mimo_mode); + else + nb_rb = dlsch_extract_rbs_dual(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, + common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id], + pdsch_vars[eNB_id_i]->rxdataF_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + dlsch0_harq->pmi_alloc, + pdsch_vars[eNB_id_i]->pmi_ext, + rballoc, + symbol, + nr_tti_rx, + ue->high_speed_flag, + frame_parms, + dlsch0_harq->mimo_mode); + } + } else + +#endif + if (beamforming_mode==0) { //else if nb_antennas_ports_eNB==1 && beamforming_mode == 0 + //printf("start nr dlsch extract nr_tti_rx %d thread id %d \n", nr_tti_rx, ue->current_thread_id[nr_tti_rx]); + nb_rb = nr_dlsch_extract_rbs_single(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, + common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id], + pdsch_vars[eNB_id]->rxdataF_ext, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + dlsch0_harq->pmi_alloc, + pdsch_vars[eNB_id]->pmi_ext, + rballoc, + symbol, + start_rb, + nb_rb_pdsch, + nr_tti_rx, + ue->high_speed_flag, + frame_parms); + + +/* + if (rx_type==rx_IC_single_stream) { + if (eNB_id_in_connected_eNB) + nb_rb = dlsch_extract_rbs_single(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, + common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id_i], + pdsch_vars[eNB_id_i]->rxdataF_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + dlsch0_harq->pmi_alloc, + pdsch_vars[eNB_id_i]->pmi_ext, + rballoc, + symbol, + nr_tti_rx, + ue->high_speed_flag, + frame_parms); + else + nb_rb = dlsch_extract_rbs_single(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, + common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id], + pdsch_vars[eNB_id_i]->rxdataF_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + dlsch0_harq->pmi_alloc, + pdsch_vars[eNB_id_i]->pmi_ext, + rballoc, + symbol, + nr_tti_rx, + ue->high_speed_flag, + frame_parms); + }*/ + } /*else if (beamforming_mode==7) { //else if beamforming_mode == 7 + nb_rb = dlsch_extract_rbs_TM7(common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF, + pdsch_vars[eNB_id]->dl_bf_ch_estimates, + pdsch_vars[eNB_id]->rxdataF_ext, + pdsch_vars[eNB_id]->dl_bf_ch_estimates_ext, + rballoc, + symbol, + nr_tti_rx, + ue->high_speed_flag, + frame_parms); + + } else if(beamforming_mode>7) { + LOG_W(PHY,"dlsch_demodulation: beamforming mode not supported yet.\n"); + }*/ + + //printf("nb_rb = %d, eNB_id %d\n",nb_rb,eNB_id); + if (nb_rb==0) { + LOG_D(PHY,"dlsch_demodulation.c: nb_rb=0\n"); + return(-1); + } + +#if UE_TIMING_TRACE + stop_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot%d Symbol %d Flag %d type %d: Pilot/Data extraction %5.2f \n",frame,nr_tti_rx,slot, + symbol,ue->high_speed_flag,type,symbol,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#else + LOG_I(PHY, "[AbsSFN %d.%d] Slot%d Symbol %d Flag %d type %d: Pilot/Data extraction %5.2f \n",frame,nr_tti_rx,slot,symbol, + ue->high_speed_flag,type,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#endif +#endif + +#if UE_TIMING_TRACE + start_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#endif + aatx = frame_parms->nb_antenna_ports_eNB; + aarx = frame_parms->nb_antennas_rx; + + nr_dlsch_scale_channel(pdsch_vars[eNB_id]->dl_ch_estimates_ext, + frame_parms, + dlsch, + symbol, + nb_rb); + + if ((dlsch0_harq->mimo_moden_connected_eNB) && + (dlsch0_harq->dl_power_off==0) + ) // TM5 two-user + { + nr_dlsch_scale_channel(pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + frame_parms, + dlsch, + symbol, + nb_rb); + } + +#if UE_TIMING_TRACE + stop_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot%d Symbol %d: Channel Scale %5.2f \n",frame,nr_tti_rx,slot,symbol,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#else + LOG_I(PHY, "[AbsSFN %d.%d] Slot%d Symbol %d: Channel Scale %5.2f \n",frame,nr_tti_rx,slot,symbol,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#endif +#endif + +#if UE_TIMING_TRACE + start_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#endif + if (first_symbol_flag==1) { + if (beamforming_mode==0){ + if (dlsch0_harq->mimo_modedl_ch_estimates_ext, + frame_parms, + avg, + symbol, + nb_rb); + avgs = 0; + for (aatx=0;aatxnb_antenna_ports_eNB;aatx++) + for (aarx=0;aarxnb_antennas_rx;aarx++) + avgs = cmax(avgs,avg[(aatx<<1)+aarx]); + + pdsch_vars[eNB_id]->log2_maxh = (log2_approx(avgs)/2)+1; + }/* + else if ((dlsch0_harq->mimo_mode == LARGE_CDD) || + ((dlsch0_harq->mimo_mode >=DUALSTREAM_UNIFORM_PRECODING1) && + (dlsch0_harq->mimo_mode <=DUALSTREAM_PUSCH_PRECODING))) + { + dlsch_channel_level_TM34(pdsch_vars[eNB_id]->dl_ch_estimates_ext, + frame_parms, + pdsch_vars[eNB_id]->pmi_ext, + avg_0, + avg_1, + symbol, + nb_rb, + dlsch0_harq->mimo_mode); + + LOG_D(PHY,"Channel Level TM34 avg_0 %d, avg_1 %d, rx_type %d, rx_standard %d, dlsch_demod_shift %d \n", avg_0[0], + avg_1[0], rx_type, rx_standard, dlsch_demod_shift); + if (rx_type>rx_standard) { + avg_0[0] = (log2_approx(avg_0[0])/2) + dlsch_demod_shift;// + 2 ;//+ 4; + avg_1[0] = (log2_approx(avg_1[0])/2) + dlsch_demod_shift;// + 2 ;//+ 4; + pdsch_vars[eNB_id]->log2_maxh0 = cmax(avg_0[0],0); + pdsch_vars[eNB_id]->log2_maxh1 = cmax(avg_1[0],0); + // printf("dlsch_demod_shift %d\n", dlsch_demod_shift); + } + else { + avg_0[0] = (log2_approx(avg_0[0])/2) - 13 + interf_unaw_shift; + avg_1[0] = (log2_approx(avg_1[0])/2) - 13 + interf_unaw_shift; + pdsch_vars[eNB_id]->log2_maxh0 = cmax(avg_0[0],0); + pdsch_vars[eNB_id]->log2_maxh1 = cmax(avg_1[0],0); + } + } + else if (dlsch0_harq->mimo_moden_connected_eNB) && (dlsch0_harq->dl_power_off==0)) { + dlsch_channel_level_TM56(pdsch_vars[eNB_id]->dl_ch_estimates_ext, + frame_parms, + pdsch_vars[eNB_id]->pmi_ext, + avg, + symbol, + nb_rb); + avg[0] = log2_approx(avg[0]) - 13 + offset_mumimo_llr_drange[dlsch0_harq->mcs][(i_mod>>1)-1]; + pdsch_vars[eNB_id]->log2_maxh = cmax(avg[0],0); + + } + else if (dlsch0_harq->dl_power_off==1) { //TM6 + + nr_dlsch_channel_level(pdsch_vars[eNB_id]->dl_ch_estimates_ext, + frame_parms, + avg, + symbol, + nb_rb); + + avgs = 0; + for (aatx=0;aatxnb_antenna_ports_eNB;aatx++) + for (aarx=0;aarxnb_antennas_rx;aarx++) + avgs = cmax(avgs,avg[(aatx<<1)+aarx]); + + pdsch_vars[eNB_id]->log2_maxh = (log2_approx(avgs)/2) + 1; + pdsch_vars[eNB_id]->log2_maxh++; + + } + } + + } + else if (beamforming_mode==7) + dlsch_channel_level_TM7(pdsch_vars[eNB_id]->dl_bf_ch_estimates_ext, + frame_parms, + avg, + symbol, + nb_rb);*/ + } +//#ifdef UE_DEBUG_TRACE + LOG_I(PHY,"[DLSCH] AbsSubframe %d.%d log2_maxh = %d [log2_maxh0 %d log2_maxh1 %d] (%d,%d)\n", + frame%1024,nr_tti_rx, pdsch_vars[eNB_id]->log2_maxh, + pdsch_vars[eNB_id]->log2_maxh0, + pdsch_vars[eNB_id]->log2_maxh1, + avg[0],avgs); + //LOG_D(PHY,"[DLSCH] mimo_mode = %d\n", dlsch0_harq->mimo_mode); +//#endif + + //wait until pdcch is decoded + //proc->channel_level = 1; + } + + /* + uint32_t wait = 0; + while(proc->channel_level == 0) + { + usleep(1); + wait++; + } + */ + +#if T_TRACER + if (type == PDSCH) + { + T(T_UE_PHY_PDSCH_ENERGY, T_INT(eNB_id), T_INT(0), T_INT(frame%1024), T_INT(nr_tti_rx), + T_INT(avg[0]), T_INT(avg[1]), T_INT(avg[2]), T_INT(avg[3])); + } +#endif + +#if UE_TIMING_TRACE + stop_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot%d Symbol %d first_symbol_flag %d: Channel Level %5.2f \n",frame,nr_tti_rx,slot,symbol,first_symbol_flag,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#else + LOG_I(PHY, "[AbsSFN %d.%d] Slot%d Symbol %d first_symbol_flag %d: Channel Level %5.2f \n",frame,nr_tti_rx,slot,symbol,first_symbol_flag,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#endif +#endif + + +#if UE_TIMING_TRACE + start_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#endif +// Now channel compensation + if (dlsch0_harq->mimo_moderxdataF_ext, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + pdsch_vars[eNB_id]->rxdataF_comp0, + (aatx>1) ? pdsch_vars[eNB_id]->rho : NULL, + frame_parms, + symbol, + first_symbol_flag, + dlsch0_harq->Qm, + nb_rb, + pdsch_vars[eNB_id]->log2_maxh, + measurements); // log2_maxh+I0_shift + /*if (symbol == 5) { + write_output("rxF_comp_d.m","rxF_c_d",&pdsch_vars[eNB_id]->rxdataF_comp0[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); + } */ + + /*if ((rx_type==rx_IC_single_stream) && + (eNB_id_in_connected_eNB)) { + nr_dlsch_channel_compensation(pdsch_vars[eNB_id_i]->rxdataF_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + pdsch_vars[eNB_id_i]->dl_ch_mag0, + pdsch_vars[eNB_id_i]->dl_ch_magb0, + pdsch_vars[eNB_id_i]->rxdataF_comp0, + (aatx>1) ? pdsch_vars[eNB_id_i]->rho : NULL, + frame_parms, + symbol, + first_symbol_flag, + i_mod, + nb_rb, + pdsch_vars[eNB_id]->log2_maxh, + measurements); // log2_maxh+I0_shift +#ifdef DEBUG_PHY + if (symbol == 5) { + write_output("rxF_comp_d.m","rxF_c_d",&pdsch_vars[eNB_id]->rxdataF_comp0[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); + write_output("rxF_comp_i.m","rxF_c_i",&pdsch_vars[eNB_id_i]->rxdataF_comp0[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); + } +#endif + + dlsch_dual_stream_correlation(frame_parms, + symbol, + nb_rb, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->log2_maxh); + }*/ + } + +#if 0 + else if ((dlsch0_harq->mimo_mode == LARGE_CDD) || ((dlsch0_harq->mimo_mode >=DUALSTREAM_UNIFORM_PRECODING1) && + (dlsch0_harq->mimo_mode <=DUALSTREAM_PUSCH_PRECODING))){ + dlsch_channel_compensation_TM34(frame_parms, + pdsch_vars[eNB_id], + measurements, + eNB_id, + symbol, + dlsch0_harq->Qm, + dlsch1_harq->Qm, + harq_pid, + dlsch0_harq->round, + dlsch0_harq->mimo_mode, + nb_rb, + pdsch_vars[eNB_id]->log2_maxh0, + pdsch_vars[eNB_id]->log2_maxh1); + /* if (symbol == 5) { + write_output("rxF_comp_d00.m","rxF_c_d00",&pdsch_vars[eNB_id]->rxdataF_comp0[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);// should be QAM + write_output("rxF_comp_d01.m","rxF_c_d01",&pdsch_vars[eNB_id]->rxdataF_comp0[1][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);//should be almost 0 + write_output("rxF_comp_d10.m","rxF_c_d10",&pdsch_vars[eNB_id]->rxdataF_comp1[harq_pid][round][0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);//should be almost 0 + write_output("rxF_comp_d11.m","rxF_c_d11",&pdsch_vars[eNB_id]->rxdataF_comp1[harq_pid][round][1][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);//should be QAM + } */ + // compute correlation between signal and interference channels (rho12 and rho21) + dlsch_dual_stream_correlation(frame_parms, // this is doing h11'*h12 and h21'*h22 + symbol, + nb_rb, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + &(pdsch_vars[eNB_id]->dl_ch_estimates_ext[2]), + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->log2_maxh0); + //printf("rho stream1 =%d\n", &pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round] ); + //to be optimized (just take complex conjugate) + dlsch_dual_stream_correlation(frame_parms, // this is doing h12'*h11 and h22'*h21 + symbol, + nb_rb, + &(pdsch_vars[eNB_id]->dl_ch_estimates_ext[2]), + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->log2_maxh1); + // printf("rho stream2 =%d\n",&pdsch_vars[eNB_id]->dl_ch_rho2_ext ); + //printf("TM3 log2_maxh : %d\n",pdsch_vars[eNB_id]->log2_maxh); + /* if (symbol == 5) { + write_output("rho0_0.m","rho0_0",&pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round][0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);// should be QAM + write_output("rho2_0.m","rho2_0",&pdsch_vars[eNB_id]->dl_ch_rho2_ext[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);//should be almost 0 + write_output("rho0_1.m.m","rho0_1",&pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round][1][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);//should be almost 0 + write_output("rho2_1.m","rho2_1",&pdsch_vars[eNB_id]->dl_ch_rho2_ext[1][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);//should be QAM + } */ + + } else if (dlsch0_harq->mimo_moden_connected_eNB) && (dlsch0_harq->dl_power_off==0)) { + dlsch_channel_compensation_TM56(pdsch_vars[eNB_id]->rxdataF_ext, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id]->pmi_ext, + frame_parms, + measurements, + eNB_id, + symbol, + dlsch0_harq->Qm, + nb_rb, + pdsch_vars[eNB_id]->log2_maxh, + dlsch0_harq->dl_power_off); + + for (rb=0; rbpmi_ext[rb]) { + case 0: + pdsch_vars[eNB_id_i]->pmi_ext[rb]=1; + break; + case 1: + pdsch_vars[eNB_id_i]->pmi_ext[rb]=0; + break; + case 2: + pdsch_vars[eNB_id_i]->pmi_ext[rb]=3; + break; + case 3: + pdsch_vars[eNB_id_i]->pmi_ext[rb]=2; + break; + } + // if (rb==0) + // printf("pmi %d, pmi_i %d\n",pdsch_vars[eNB_id]->pmi_ext[rb],pdsch_vars[eNB_id_i]->pmi_ext[rb]); + } + dlsch_channel_compensation_TM56(pdsch_vars[eNB_id_i]->rxdataF_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + pdsch_vars[eNB_id_i]->dl_ch_mag0, + pdsch_vars[eNB_id_i]->dl_ch_magb0, + pdsch_vars[eNB_id_i]->rxdataF_comp0, + pdsch_vars[eNB_id_i]->pmi_ext, + frame_parms, + measurements, + eNB_id_i, + symbol, + i_mod, + nb_rb, + pdsch_vars[eNB_id]->log2_maxh, + dlsch0_harq->dl_power_off); +#ifdef DEBUG_PHY + if (symbol==5) { + write_output("rxF_comp_d.m","rxF_c_d",&pdsch_vars[eNB_id]->rxdataF_comp0[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); + write_output("rxF_comp_i.m","rxF_c_i",&pdsch_vars[eNB_id_i]->rxdataF_comp0[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); + } +#endif + dlsch_dual_stream_correlation(frame_parms, + symbol, + nb_rb, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + pdsch_vars[eNB_id_i]->dl_ch_estimates_ext, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->log2_maxh); + } else if (dlsch0_harq->dl_power_off==1) { + dlsch_channel_compensation_TM56(pdsch_vars[eNB_id]->rxdataF_ext, + pdsch_vars[eNB_id]->dl_ch_estimates_ext, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id]->pmi_ext, + frame_parms, + measurements, + eNB_id, + symbol, + dlsch0_harq->Qm, + nb_rb, + pdsch_vars[eNB_id]->log2_maxh, + 1); + + } + + + } else if (dlsch0_harq->mimo_mode==TM7) { //TM7 + + nr_dlsch_channel_compensation(pdsch_vars[eNB_id]->rxdataF_ext, + pdsch_vars[eNB_id]->dl_bf_ch_estimates_ext, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + pdsch_vars[eNB_id]->rxdataF_comp0, + (aatx>1) ? pdsch_vars[eNB_id]->rho : NULL, + frame_parms, + symbol, + first_symbol_flag, + get_Qm(dlsch0_harq->mcs), + nb_rb, + //9, + pdsch_vars[eNB_id]->log2_maxh, + measurements); // log2_maxh+I0_shift + } +#endif + +#if UE_TIMING_TRACE + stop_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot%d Symbol %d log2_maxh %d channel_level %d: Channel Comp %5.2f \n",frame,nr_tti_rx,slot,symbol,pdsch_vars[eNB_id]->log2_maxh,proc->channel_level,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#else + LOG_I(PHY, "[AbsSFN %d.%d] Slot%d Symbol %d log2_maxh %d channel_level %d: Channel Comp %5.2f \n",frame,nr_tti_rx,slot,symbol,pdsch_vars[eNB_id]->log2_maxh,proc->channel_level,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#endif +#endif +// MRC +#if UE_TIMING_TRACE + start_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#endif + +#if 0 + if (frame_parms->nb_antennas_rx > 1) { + if ((dlsch0_harq->mimo_mode == LARGE_CDD) || + ((dlsch0_harq->mimo_mode >=DUALSTREAM_UNIFORM_PRECODING1) && + (dlsch0_harq->mimo_mode <=DUALSTREAM_PUSCH_PRECODING))){ // TM3 or TM4 + if (frame_parms->nb_antenna_ports_eNB == 2) { + dlsch_detection_mrc_TM34(frame_parms, + pdsch_vars[eNB_id], + harq_pid, + dlsch0_harq->round, + symbol, + nb_rb, + 1); + /* if (symbol == 5) { + write_output("rho0_mrc.m","rho0_0",&pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round][0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);// should be QAM + write_output("rho2_mrc.m","rho2_0",&pdsch_vars[eNB_id]->dl_ch_rho2_ext[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1);//should be almost 0 + } */ + } + } else { + dlsch_detection_mrc(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id_i]->rxdataF_comp0, + pdsch_vars[eNB_id]->rho, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + pdsch_vars[eNB_id_i]->dl_ch_mag0, + pdsch_vars[eNB_id_i]->dl_ch_magb0, + symbol, + nb_rb, + rx_type==rx_IC_single_stream); + } + } + // printf("Combining"); + if ((dlsch0_harq->mimo_mode == SISO) || + ((dlsch0_harq->mimo_mode >= UNIFORM_PRECODING11) && + (dlsch0_harq->mimo_mode <= PUSCH_PRECODING0)) || + (dlsch0_harq->mimo_mode == TM7)) { + /* + dlsch_siso(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp, + pdsch_vars[eNB_id_i]->rxdataF_comp, + symbol, + nb_rb); + */ + } else if (dlsch0_harq->mimo_mode == ALAMOUTI) { + dlsch_alamouti(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + symbol, + nb_rb); + } +#endif + + //printf("start compute LLR\n"); + if ((dlsch0_harq->mimo_mode == LARGE_CDD) || + ((dlsch0_harq->mimo_mode >=DUALSTREAM_UNIFORM_PRECODING1) && + (dlsch0_harq->mimo_mode <=DUALSTREAM_PUSCH_PRECODING))) { + rxdataF_comp_ptr = pdsch_vars[eNB_id]->rxdataF_comp1[harq_pid][round]; + dl_ch_mag_ptr = pdsch_vars[eNB_id]->dl_ch_mag1[harq_pid][round]; + } + else { + rxdataF_comp_ptr = pdsch_vars[eNB_id_i]->rxdataF_comp0; + dl_ch_mag_ptr = pdsch_vars[eNB_id_i]->dl_ch_mag0; + //i_mod should have been passed as a parameter + } + +#if UE_TIMING_TRACE + stop_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot%d Symbol %d: Channel Combine %5.2f \n",frame,nr_tti_rx,slot,symbol,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#else + LOG_I(PHY, "[AbsSFN %d.%d] Slot%d Symbol %d: Channel Combine %5.2f \n",frame,nr_tti_rx,slot,symbol,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#endif +#endif + +#if UE_TIMING_TRACE + start_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#endif + //printf("LLR dlsch0_harq->Qm %d rx_type %d cw0 %d cw1 %d symbol %d \n",dlsch0_harq->Qm,rx_type,codeword_TB0,codeword_TB1,symbol); + // compute LLRs + // -> // compute @pointer where llrs should filled for this ofdm-symbol + + pdsch_vars[eNB_id]->llr_offset[l0-1] = 0; + llr_offset_symbol = pdsch_vars[eNB_id]->llr_offset[symbol-1]; + //pllr_symbol_cw0_deint = (int8_t*)pdsch_vars[eNB_id]->llr[0]; + //pllr_symbol_cw1_deint = (int8_t*)pdsch_vars[eNB_id]->llr[1]; + pllr_symbol_cw0 = pdsch_vars[eNB_id]->llr[0]; + pllr_symbol_cw1 = pdsch_vars[eNB_id]->llr[1]; + pllr_symbol_cw0 += llr_offset_symbol; + pllr_symbol_cw1 += llr_offset_symbol; + + nb_re= (symbol==l0)? (nb_rb*6):(nb_rb*12); + pdsch_vars[eNB_id]->llr_offset[symbol] = nb_re*dlsch0_harq->Qm + llr_offset_symbol; + + /*LOG_I(PHY,"compute LLRs [symbol %d] NbRB %d Qm %d LLRs-Length %d LLR-Offset %d @LLR Buff %x @LLR Buff(symb) %x\n", + symbol, + nb_rb,dlsch0_harq->Qm, + pdsch_vars[eNB_id]->llr_length[symbol], + pdsch_vars[eNB_id]->llr_offset[symbol], + (int16_t*)pdsch_vars[eNB_id]->llr[0], + pllr_symbol_cw0);*/ + + /*printf("compute LLRs [symbol %d] NbRB %d Qm %d LLRs-Length %d LLR-Offset %d @LLR Buff %p @LLR Buff(symb) %p\n", + symbol, + nb_rb,dlsch0_harq->Qm, + pdsch_vars[eNB_id]->llr_length[symbol], + pdsch_vars[eNB_id]->llr_offset[symbol], + pdsch_vars[eNB_id]->llr[0], + pllr_symbol_cw0);*/ + + switch (dlsch0_harq->Qm) { + case 2 : + if ((rx_type==rx_standard) || (codeword_TB1 == -1)) { + nr_dlsch_qpsk_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pllr_symbol_cw0, + symbol, + first_symbol_flag, + nb_rb, + beamforming_mode); + + } else if (codeword_TB0 == -1){ + + nr_dlsch_qpsk_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pllr_symbol_cw1, + symbol, + first_symbol_flag, + nb_rb, + beamforming_mode); + } + else if (rx_type >= rx_IC_single_stream) { + if (dlsch1_harq->Qm == 2) { + nr_dlsch_qpsk_qpsk_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,2,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_qpsk_qpsk_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,2,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + else if (dlsch1_harq->Qm == 4) { + nr_dlsch_qpsk_16qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + dl_ch_mag_ptr,//i + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,2,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_16qam_qpsk_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0,//i + dl_ch_mag_ptr, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,4,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + else { + nr_dlsch_qpsk_64qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + dl_ch_mag_ptr,//i + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,2,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_64qam_qpsk_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0,//i + dl_ch_mag_ptr, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,6,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + } + break; + case 4 : + if ((rx_type==rx_standard ) || (codeword_TB1 == -1)) { + nr_dlsch_16qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id]->llr[0], + pdsch_vars[eNB_id]->dl_ch_mag0, + symbol,first_symbol_flag,nb_rb, + pdsch_vars[eNB_id]->llr128, + beamforming_mode); + } else if (codeword_TB0 == -1){ + nr_dlsch_16qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id]->llr[1], + pdsch_vars[eNB_id]->dl_ch_mag0, + symbol,first_symbol_flag,nb_rb, + pdsch_vars[eNB_id]->llr128_2ndstream, + beamforming_mode); + } + else if (rx_type >= rx_IC_single_stream) { + if (dlsch1_harq->Qm == 2) { + nr_dlsch_16qam_qpsk_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,4,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_qpsk_16qam_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0,//i + pdsch_vars[eNB_id]->dl_ch_mag0,//i + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,2,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + else if (dlsch1_harq->Qm == 4) { + nr_dlsch_16qam_16qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + pdsch_vars[eNB_id]->dl_ch_mag0, + dl_ch_mag_ptr,//i + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,4,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_16qam_16qam_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0,//i + dl_ch_mag_ptr, + pdsch_vars[eNB_id]->dl_ch_mag0,//i + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,4,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + else { + nr_dlsch_16qam_64qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + pdsch_vars[eNB_id]->dl_ch_mag0, + dl_ch_mag_ptr,//i + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,4,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_64qam_16qam_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0, + dl_ch_mag_ptr, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,6,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + } + break; + case 6 : + if ((rx_type==rx_standard) || (codeword_TB1 == -1)) { + nr_dlsch_64qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + (int16_t*)pllr_symbol_cw0, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + symbol,first_symbol_flag,nb_rb, + pdsch_vars[eNB_id]->llr_offset[symbol], + beamforming_mode); + } else if (codeword_TB0 == -1){ + nr_dlsch_64qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pllr_symbol_cw1, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + symbol,first_symbol_flag,nb_rb, + pdsch_vars[eNB_id]->llr_offset[symbol], + beamforming_mode); + } + else if (rx_type >= rx_IC_single_stream) { + if (dlsch1_harq->Qm == 2) { + nr_dlsch_64qam_qpsk_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,6,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_qpsk_64qam_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0,//i + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,2,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + else if (dlsch1_harq->Qm == 4) { + nr_dlsch_64qam_16qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + pdsch_vars[eNB_id]->dl_ch_mag0, + dl_ch_mag_ptr,//i + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + pdsch_vars[eNB_id]->llr[0], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,6,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_16qam_64qam_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0,//i + dl_ch_mag_ptr, + pdsch_vars[eNB_id]->dl_ch_mag0,//i + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pdsch_vars[eNB_id]->llr[1], + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,4,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr128_2ndstream); + } + } + else { + nr_dlsch_64qam_64qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + rxdataF_comp_ptr,//i + pdsch_vars[eNB_id]->dl_ch_mag0, + dl_ch_mag_ptr,//i + pdsch_vars[eNB_id]->dl_ch_rho2_ext, + (int16_t*)pllr_symbol_cw0, + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch0_harq->rb_alloc_even,6,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr_offset[symbol]); + if (rx_type==rx_IC_dual_stream) { + nr_dlsch_64qam_64qam_llr(frame_parms, + rxdataF_comp_ptr, + pdsch_vars[eNB_id]->rxdataF_comp0,//i + dl_ch_mag_ptr, + pdsch_vars[eNB_id]->dl_ch_mag0,//i + pdsch_vars[eNB_id]->dl_ch_rho_ext[harq_pid][round], + pllr_symbol_cw1, + symbol,first_symbol_flag,nb_rb, + adjust_G2(frame_parms,dlsch1_harq->rb_alloc_even,6,nr_tti_rx,symbol), + pdsch_vars[eNB_id]->llr_offset[symbol]); + } + } + } + break; + default: + LOG_W(PHY,"rx_dlsch.c : Unknown mod_order!!!!\n"); + return(-1); + break; + } + if (dlsch1_harq) { + switch (get_Qm(dlsch1_harq->mcs)) { + case 2 : + if (rx_type==rx_standard) { + nr_dlsch_qpsk_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pllr_symbol_cw0, + symbol,first_symbol_flag,nb_rb, + beamforming_mode); + } + break; + case 4: + if (rx_type==rx_standard) { + nr_dlsch_16qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pdsch_vars[eNB_id]->llr[0], + pdsch_vars[eNB_id]->dl_ch_mag0, + symbol,first_symbol_flag,nb_rb, + pdsch_vars[eNB_id]->llr128, + beamforming_mode); + } + break; + case 6 : + if (rx_type==rx_standard) { + nr_dlsch_64qam_llr(frame_parms, + pdsch_vars[eNB_id]->rxdataF_comp0, + pllr_symbol_cw0, + pdsch_vars[eNB_id]->dl_ch_mag0, + pdsch_vars[eNB_id]->dl_ch_magb0, + symbol,first_symbol_flag,nb_rb, + pdsch_vars[eNB_id]->llr_offset[symbol], + beamforming_mode); + } + break; + default: + LOG_W(PHY,"rx_dlsch.c : Unknown mod_order!!!!\n"); + return(-1); + break; + } + } + + //nr_dlsch_deinterleaving(symbol,bundle_L,(int16_t*)pllr_symbol_cw0,(int16_t*)pllr_symbol_cw0_deint, nb_rb_pdsch); + +#if UE_TIMING_TRACE + stop_meas(&ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot%d Symbol %d: LLR Computation %5.2f \n",frame,nr_tti_rx,slot,symbol,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#else + LOG_I(PHY, "[AbsSFN %d.%d] Slot%d Symbol %d: LLR Computation %5.2f \n",frame,nr_tti_rx,slot,symbol,ue->generic_stat_bis[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); +#endif +#endif +// Please keep it: useful for debugging +#if 0 + if( (symbol == 13) && (nr_tti_rx==0) && (dlsch0_harq->Qm == 6) /*&& (nb_rb==25)*/) + { + LOG_E(PHY,"Dump Phy Chan Est \n"); + if(1) + { +#if 1 + write_output("rxdataF0.m" , "rxdataF0", &common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF[0][0],14*frame_parms->ofdm_symbol_size,1,1); + //write_output("rxdataF1.m" , "rxdataF1", &common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF[0][0],14*frame_parms->ofdm_symbol_size,1,1); + write_output("dl_ch_estimates00.m", "dl_ch_estimates00", &common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id][0][0],14*frame_parms->ofdm_symbol_size,1,1); + //write_output("dl_ch_estimates01.m", "dl_ch_estimates01", &common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id][1][0],14*frame_parms->ofdm_symbol_size,1,1); + //write_output("dl_ch_estimates10.m", "dl_ch_estimates10", &common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id][2][0],14*frame_parms->ofdm_symbol_size,1,1); + //write_output("dl_ch_estimates11.m", "dl_ch_estimates11", &common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].dl_ch_estimates[eNB_id][3][0],14*frame_parms->ofdm_symbol_size,1,1); + + + //write_output("rxdataF_ext00.m" , "rxdataF_ext00", &pdsch_vars[eNB_id]->rxdataF_ext[0][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("rxdataF_ext01.m" , "rxdataF_ext01", &pdsch_vars[eNB_id]->rxdataF_ext[1][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("rxdataF_ext10.m" , "rxdataF_ext10", &pdsch_vars[eNB_id]->rxdataF_ext[2][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("rxdataF_ext11.m" , "rxdataF_ext11", &pdsch_vars[eNB_id]->rxdataF_ext[3][0],14*frame_parms->N_RB_DL*12,1,1); + write_output("dl_ch_estimates_ext00.m", "dl_ch_estimates_ext00", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[0][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("dl_ch_estimates_ext01.m", "dl_ch_estimates_ext01", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[1][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("dl_ch_estimates_ext10.m", "dl_ch_estimates_ext10", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[2][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("dl_ch_estimates_ext11.m", "dl_ch_estimates_ext11", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[3][0],14*frame_parms->N_RB_DL*12,1,1); + write_output("rxdataF_comp00.m","rxdataF_comp00", &pdsch_vars[eNB_id]->rxdataF_comp0[0][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("rxdataF_comp01.m","rxdataF_comp01", &pdsch_vars[eNB_id]->rxdataF_comp0[1][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("rxdataF_comp10.m","rxdataF_comp10", &pdsch_vars[eNB_id]->rxdataF_comp1[harq_pid][round][0][0],14*frame_parms->N_RB_DL*12,1,1); + //write_output("rxdataF_comp11.m","rxdataF_comp11", &pdsch_vars[eNB_id]->rxdataF_comp1[harq_pid][round][1][0],14*frame_parms->N_RB_DL*12,1,1); +#endif + write_output("llr0.m","llr0", &pdsch_vars[eNB_id]->llr[0][0],(14*nb_rb*12*dlsch1_harq->Qm) - 4*(nb_rb*4*dlsch1_harq->Qm),1,0); + //write_output("llr1.m","llr1", &pdsch_vars[eNB_id]->llr[1][0],(14*nb_rb*12*dlsch1_harq->Qm) - 4*(nb_rb*4*dlsch1_harq->Qm),1,0); + + + AssertFatal(0," "); + } + + } +#endif + +#if T_TRACER + T(T_UE_PHY_PDSCH_IQ, T_INT(eNB_id), T_INT(ue->Mod_id), T_INT(frame%1024), + T_INT(nr_tti_rx), T_INT(nb_rb), + T_INT(frame_parms->N_RB_UL), T_INT(frame_parms->symbols_per_tti), + T_BUFFER(&pdsch_vars[eNB_id]->rxdataF_comp0[eNB_id][0], + 2 * /* ulsch[UE_id]->harq_processes[harq_pid]->nb_rb */ frame_parms->N_RB_UL *12*frame_parms->symbols_per_tti*2)); +#endif + return(0); + +} + +void nr_dlsch_deinterleaving(uint8_t symbol, + uint16_t L, + uint16_t *llr, + uint16_t *llr_deint, + uint16_t nb_rb_pdsch) +{ + + uint32_t bundle_idx, N_bundle, R, C, r,c; + int32_t m,k; + uint8_t nb_re; + + R=2; + N_bundle = nb_rb_pdsch/L; + C=N_bundle/R; + + uint32_t *bundle_deint = malloc(N_bundle*sizeof(uint32_t)); + + printf("N_bundle %d L %d nb_rb_pdsch %d\n",N_bundle, L,nb_rb_pdsch); + + if (symbol==2) + nb_re = 6; + else + nb_re = 12; + + + AssertFatal(llr!=NULL,"nr_dlsch_deinterleaving: FATAL llr is Null\n"); + + + for (c =0; c< C; c++){ + for (r=0; rnb_antenna_ports_eNB; aatx++) { + if (mod_order == 4) { + QAM_amp128 = _mm_set1_epi16(QAM16_n1); // 2/sqrt(10) + QAM_amp128b = _mm_setzero_si128(); + } else if (mod_order == 6) { + QAM_amp128 = _mm_set1_epi16(QAM64_n1); // + QAM_amp128b = _mm_set1_epi16(QAM64_n2); + } + + // printf("comp: rxdataF_comp %p, symbol %d\n",rxdataF_comp[0],symbol); + + for (aarx=0; aarxnb_antennas_rx; aarx++) { + + dl_ch128 = (__m128i *)&dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*nb_rb*12]; + dl_ch_mag128 = (__m128i *)&dl_ch_mag[(aatx<<1)+aarx][symbol*nb_rb*12]; + dl_ch_mag128b = (__m128i *)&dl_ch_magb[(aatx<<1)+aarx][symbol*nb_rb*12]; + rxdataF128 = (__m128i *)&rxdataF_ext[aarx][symbol*nb_rb*12]; + rxdataF_comp128 = (__m128i *)&rxdataF_comp[(aatx<<1)+aarx][symbol*nb_rb*12]; + + + for (rb=0; rb2) { + // get channel amplitude if not QPSK + + mmtmpD0 = _mm_madd_epi16(dl_ch128[0],dl_ch128[0]); + mmtmpD0 = _mm_srai_epi32(mmtmpD0,(output_shift-2)); + + mmtmpD1 = _mm_madd_epi16(dl_ch128[1],dl_ch128[1]); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,(output_shift-2)); + + mmtmpD0 = _mm_packs_epi32(mmtmpD0,mmtmpD1); + + // store channel magnitude here in a new field of dlsch + + dl_ch_mag128[0] = _mm_unpacklo_epi16(mmtmpD0,mmtmpD0); + dl_ch_mag128b[0] = dl_ch_mag128[0]; + dl_ch_mag128[0] = _mm_mulhi_epi16(dl_ch_mag128[0],QAM_amp128); + dl_ch_mag128[0] = _mm_slli_epi16(dl_ch_mag128[0],1); + //print_ints("Re(ch):",(int16_t*)&mmtmpD0); + //print_shorts("QAM_amp:",(int16_t*)&QAM_amp128); + //print_shorts("mag:",(int16_t*)&dl_ch_mag128[0]); + dl_ch_mag128[1] = _mm_unpackhi_epi16(mmtmpD0,mmtmpD0); + dl_ch_mag128b[1] = dl_ch_mag128[1]; + dl_ch_mag128[1] = _mm_mulhi_epi16(dl_ch_mag128[1],QAM_amp128); + dl_ch_mag128[1] = _mm_slli_epi16(dl_ch_mag128[1],1); + + if (pilots==0) { + mmtmpD0 = _mm_madd_epi16(dl_ch128[2],dl_ch128[2]); + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + mmtmpD1 = _mm_packs_epi32(mmtmpD0,mmtmpD0); + + dl_ch_mag128[2] = _mm_unpacklo_epi16(mmtmpD1,mmtmpD1); + dl_ch_mag128b[2] = dl_ch_mag128[2]; + + dl_ch_mag128[2] = _mm_mulhi_epi16(dl_ch_mag128[2],QAM_amp128); + dl_ch_mag128[2] = _mm_slli_epi16(dl_ch_mag128[2],1); + } + + dl_ch_mag128b[0] = _mm_mulhi_epi16(dl_ch_mag128b[0],QAM_amp128b); + dl_ch_mag128b[0] = _mm_slli_epi16(dl_ch_mag128b[0],1); + + + dl_ch_mag128b[1] = _mm_mulhi_epi16(dl_ch_mag128b[1],QAM_amp128b); + dl_ch_mag128b[1] = _mm_slli_epi16(dl_ch_mag128b[1],1); + + if (pilots==0) { + dl_ch_mag128b[2] = _mm_mulhi_epi16(dl_ch_mag128b[2],QAM_amp128b); + dl_ch_mag128b[2] = _mm_slli_epi16(dl_ch_mag128b[2],1); + } + } + + // multiply by conjugated channel + mmtmpD0 = _mm_madd_epi16(dl_ch128[0],rxdataF128[0]); + // print_ints("re",&mmtmpD0); + + // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) + mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[0],_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)&conjugate[0]); + // print_ints("im",&mmtmpD1); + mmtmpD1 = _mm_madd_epi16(mmtmpD1,rxdataF128[0]); + // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + // print_ints("re(shift)",&mmtmpD0); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + // print_ints("im(shift)",&mmtmpD1); + mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); + mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); + // print_ints("c0",&mmtmpD2); + // print_ints("c1",&mmtmpD3); + rxdataF_comp128[0] = _mm_packs_epi32(mmtmpD2,mmtmpD3); + // print_shorts("rx:",rxdataF128); + // print_shorts("ch:",dl_ch128); + // print_shorts("pack:",rxdataF_comp128); + + // multiply by conjugated channel + mmtmpD0 = _mm_madd_epi16(dl_ch128[1],rxdataF128[1]); + // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) + mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[1],_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)conjugate); + mmtmpD1 = _mm_madd_epi16(mmtmpD1,rxdataF128[1]); + // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); + mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); + + rxdataF_comp128[1] = _mm_packs_epi32(mmtmpD2,mmtmpD3); + // print_shorts("rx:",rxdataF128+1); + // print_shorts("ch:",dl_ch128+1); + // print_shorts("pack:",rxdataF_comp128+1); + + if (pilots==0) { + // multiply by conjugated channel + mmtmpD0 = _mm_madd_epi16(dl_ch128[2],rxdataF128[2]); + // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) + mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[2],_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)conjugate); + mmtmpD1 = _mm_madd_epi16(mmtmpD1,rxdataF128[2]); + // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); + mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); + + rxdataF_comp128[2] = _mm_packs_epi32(mmtmpD2,mmtmpD3); + // print_shorts("rx:",rxdataF128+2); + // print_shorts("ch:",dl_ch128+2); + // print_shorts("pack:",rxdataF_comp128+2); + + dl_ch128+=3; + dl_ch_mag128+=3; + dl_ch_mag128b+=3; + rxdataF128+=3; + rxdataF_comp128+=3; + } else { // we have a smaller PDSCH in symbols with pilots so skip last group of 4 REs and increment less + dl_ch128+=2; + dl_ch_mag128+=2; + dl_ch_mag128b+=2; + rxdataF128+=2; + rxdataF_comp128+=2; + } + + } + } + } + + if (rho) { + + + for (aarx=0; aarxnb_antennas_rx; aarx++) { + rho128 = (__m128i *)&rho[aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch128 = (__m128i *)&dl_ch_estimates_ext[aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch128_2 = (__m128i *)&dl_ch_estimates_ext[2+aarx][symbol*frame_parms->N_RB_DL*12]; + + for (rb=0; rbrx_correlation[0][aarx] = signal_energy(&rho[aarx][symbol*nb_rb*12],rb*12); + } + } + } + + _mm_empty(); + _m_empty(); + +#elif defined(__arm__) + + + unsigned short rb; + unsigned char aatx,aarx,symbol_mod,pilots=0; + + int16x4_t *dl_ch128,*dl_ch128_2,*rxdataF128; + int32x4_t mmtmpD0,mmtmpD1,mmtmpD0b,mmtmpD1b; + int16x8_t *dl_ch_mag128,*dl_ch_mag128b,mmtmpD2,mmtmpD3,mmtmpD4; + int16x8_t QAM_amp128,QAM_amp128b; + int16x4x2_t *rxdataF_comp128,*rho128; + + int16_t conj[4]__attribute__((aligned(16))) = {1,-1,1,-1}; + int32x4_t output_shift128 = vmovq_n_s32(-(int32_t)output_shift); + + symbol_mod = (symbol>=(7-frame_parms->Ncp)) ? symbol-(7-frame_parms->Ncp) : symbol; + + if ((symbol_mod == 0) || (symbol_mod == (4-frame_parms->Ncp))) { + if (frame_parms->nb_antenna_ports_eNB==1) { // 10 out of 12 so don't reduce size + nb_rb=1+(5*nb_rb/6); + } + else { + pilots=1; + } + } + + for (aatx=0; aatxnb_antenna_ports_eNB; aatx++) { + if (mod_order == 4) { + QAM_amp128 = vmovq_n_s16(QAM16_n1); // 2/sqrt(10) + QAM_amp128b = vmovq_n_s16(0); + } else if (mod_order == 6) { + QAM_amp128 = vmovq_n_s16(QAM64_n1); // + QAM_amp128b = vmovq_n_s16(QAM64_n2); + } + // printf("comp: rxdataF_comp %p, symbol %d\n",rxdataF_comp[0],symbol); + + for (aarx=0; aarxnb_antennas_rx; aarx++) { + dl_ch128 = (int16x4_t*)&dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch_mag128 = (int16x8_t*)&dl_ch_mag[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch_mag128b = (int16x8_t*)&dl_ch_magb[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; + rxdataF128 = (int16x4_t*)&rxdataF_ext[aarx][symbol*frame_parms->N_RB_DL*12]; + rxdataF_comp128 = (int16x4x2_t*)&rxdataF_comp[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; + + for (rb=0; rb2) { + // get channel amplitude if not QPSK + mmtmpD0 = vmull_s16(dl_ch128[0], dl_ch128[0]); + // mmtmpD0 = [ch0*ch0,ch1*ch1,ch2*ch2,ch3*ch3]; + mmtmpD0 = vqshlq_s32(vqaddq_s32(mmtmpD0,vrev64q_s32(mmtmpD0)),output_shift128); + // mmtmpD0 = [ch0*ch0 + ch1*ch1,ch0*ch0 + ch1*ch1,ch2*ch2 + ch3*ch3,ch2*ch2 + ch3*ch3]>>output_shift128 on 32-bits + mmtmpD1 = vmull_s16(dl_ch128[1], dl_ch128[1]); + mmtmpD1 = vqshlq_s32(vqaddq_s32(mmtmpD1,vrev64q_s32(mmtmpD1)),output_shift128); + mmtmpD2 = vcombine_s16(vmovn_s32(mmtmpD0),vmovn_s32(mmtmpD1)); + // mmtmpD2 = [ch0*ch0 + ch1*ch1,ch0*ch0 + ch1*ch1,ch2*ch2 + ch3*ch3,ch2*ch2 + ch3*ch3,ch4*ch4 + ch5*ch5,ch4*ch4 + ch5*ch5,ch6*ch6 + ch7*ch7,ch6*ch6 + ch7*ch7]>>output_shift128 on 16-bits + mmtmpD0 = vmull_s16(dl_ch128[2], dl_ch128[2]); + mmtmpD0 = vqshlq_s32(vqaddq_s32(mmtmpD0,vrev64q_s32(mmtmpD0)),output_shift128); + mmtmpD1 = vmull_s16(dl_ch128[3], dl_ch128[3]); + mmtmpD1 = vqshlq_s32(vqaddq_s32(mmtmpD1,vrev64q_s32(mmtmpD1)),output_shift128); + mmtmpD3 = vcombine_s16(vmovn_s32(mmtmpD0),vmovn_s32(mmtmpD1)); + if (pilots==0) { + mmtmpD0 = vmull_s16(dl_ch128[4], dl_ch128[4]); + mmtmpD0 = vqshlq_s32(vqaddq_s32(mmtmpD0,vrev64q_s32(mmtmpD0)),output_shift128); + mmtmpD1 = vmull_s16(dl_ch128[5], dl_ch128[5]); + mmtmpD1 = vqshlq_s32(vqaddq_s32(mmtmpD1,vrev64q_s32(mmtmpD1)),output_shift128); + mmtmpD4 = vcombine_s16(vmovn_s32(mmtmpD0),vmovn_s32(mmtmpD1)); + } + + dl_ch_mag128b[0] = vqdmulhq_s16(mmtmpD2,QAM_amp128b); + dl_ch_mag128b[1] = vqdmulhq_s16(mmtmpD3,QAM_amp128b); + dl_ch_mag128[0] = vqdmulhq_s16(mmtmpD2,QAM_amp128); + dl_ch_mag128[1] = vqdmulhq_s16(mmtmpD3,QAM_amp128); + + if (pilots==0) { + dl_ch_mag128b[2] = vqdmulhq_s16(mmtmpD4,QAM_amp128b); + dl_ch_mag128[2] = vqdmulhq_s16(mmtmpD4,QAM_amp128); + } + } + + mmtmpD0 = vmull_s16(dl_ch128[0], rxdataF128[0]); + //mmtmpD0 = [Re(ch[0])Re(rx[0]) Im(ch[0])Im(ch[0]) Re(ch[1])Re(rx[1]) Im(ch[1])Im(ch[1])] + mmtmpD1 = vmull_s16(dl_ch128[1], rxdataF128[1]); + //mmtmpD1 = [Re(ch[2])Re(rx[2]) Im(ch[2])Im(ch[2]) Re(ch[3])Re(rx[3]) Im(ch[3])Im(ch[3])] + mmtmpD0 = vcombine_s32(vpadd_s32(vget_low_s32(mmtmpD0),vget_high_s32(mmtmpD0)), + vpadd_s32(vget_low_s32(mmtmpD1),vget_high_s32(mmtmpD1))); + //mmtmpD0 = [Re(ch[0])Re(rx[0])+Im(ch[0])Im(ch[0]) Re(ch[1])Re(rx[1])+Im(ch[1])Im(ch[1]) Re(ch[2])Re(rx[2])+Im(ch[2])Im(ch[2]) Re(ch[3])Re(rx[3])+Im(ch[3])Im(ch[3])] + + mmtmpD0b = vmull_s16(vrev32_s16(vmul_s16(dl_ch128[0],*(int16x4_t*)conj)), rxdataF128[0]); + //mmtmpD0 = [-Im(ch[0])Re(rx[0]) Re(ch[0])Im(rx[0]) -Im(ch[1])Re(rx[1]) Re(ch[1])Im(rx[1])] + mmtmpD1b = vmull_s16(vrev32_s16(vmul_s16(dl_ch128[1],*(int16x4_t*)conj)), rxdataF128[1]); + //mmtmpD0 = [-Im(ch[2])Re(rx[2]) Re(ch[2])Im(rx[2]) -Im(ch[3])Re(rx[3]) Re(ch[3])Im(rx[3])] + mmtmpD1 = vcombine_s32(vpadd_s32(vget_low_s32(mmtmpD0b),vget_high_s32(mmtmpD0b)), + vpadd_s32(vget_low_s32(mmtmpD1b),vget_high_s32(mmtmpD1b))); + //mmtmpD1 = [-Im(ch[0])Re(rx[0])+Re(ch[0])Im(rx[0]) -Im(ch[1])Re(rx[1])+Re(ch[1])Im(rx[1]) -Im(ch[2])Re(rx[2])+Re(ch[2])Im(rx[2]) -Im(ch[3])Re(rx[3])+Re(ch[3])Im(rx[3])] + + mmtmpD0 = vqshlq_s32(mmtmpD0,output_shift128); + mmtmpD1 = vqshlq_s32(mmtmpD1,output_shift128); + rxdataF_comp128[0] = vzip_s16(vmovn_s32(mmtmpD0),vmovn_s32(mmtmpD1)); + mmtmpD0 = vmull_s16(dl_ch128[2], rxdataF128[2]); + mmtmpD1 = vmull_s16(dl_ch128[3], rxdataF128[3]); + mmtmpD0 = vcombine_s32(vpadd_s32(vget_low_s32(mmtmpD0),vget_high_s32(mmtmpD0)), + vpadd_s32(vget_low_s32(mmtmpD1),vget_high_s32(mmtmpD1))); + mmtmpD0b = vmull_s16(vrev32_s16(vmul_s16(dl_ch128[2],*(int16x4_t*)conj)), rxdataF128[2]); + mmtmpD1b = vmull_s16(vrev32_s16(vmul_s16(dl_ch128[3],*(int16x4_t*)conj)), rxdataF128[3]); + mmtmpD1 = vcombine_s32(vpadd_s32(vget_low_s32(mmtmpD0b),vget_high_s32(mmtmpD0b)), + vpadd_s32(vget_low_s32(mmtmpD1b),vget_high_s32(mmtmpD1b))); + mmtmpD0 = vqshlq_s32(mmtmpD0,output_shift128); + mmtmpD1 = vqshlq_s32(mmtmpD1,output_shift128); + rxdataF_comp128[1] = vzip_s16(vmovn_s32(mmtmpD0),vmovn_s32(mmtmpD1)); + + if (pilots==0) { + mmtmpD0 = vmull_s16(dl_ch128[4], rxdataF128[4]); + mmtmpD1 = vmull_s16(dl_ch128[5], rxdataF128[5]); + mmtmpD0 = vcombine_s32(vpadd_s32(vget_low_s32(mmtmpD0),vget_high_s32(mmtmpD0)), + vpadd_s32(vget_low_s32(mmtmpD1),vget_high_s32(mmtmpD1))); + + mmtmpD0b = vmull_s16(vrev32_s16(vmul_s16(dl_ch128[4],*(int16x4_t*)conj)), rxdataF128[4]); + mmtmpD1b = vmull_s16(vrev32_s16(vmul_s16(dl_ch128[5],*(int16x4_t*)conj)), rxdataF128[5]); + mmtmpD1 = vcombine_s32(vpadd_s32(vget_low_s32(mmtmpD0b),vget_high_s32(mmtmpD0b)), + vpadd_s32(vget_low_s32(mmtmpD1b),vget_high_s32(mmtmpD1b))); + + + mmtmpD0 = vqshlq_s32(mmtmpD0,output_shift128); + mmtmpD1 = vqshlq_s32(mmtmpD1,output_shift128); + rxdataF_comp128[2] = vzip_s16(vmovn_s32(mmtmpD0),vmovn_s32(mmtmpD1)); + + + dl_ch128+=6; + dl_ch_mag128+=3; + dl_ch_mag128b+=3; + rxdataF128+=6; + rxdataF_comp128+=3; + + } else { // we have a smaller PDSCH in symbols with pilots so skip last group of 4 REs and increment less + dl_ch128+=4; + dl_ch_mag128+=2; + dl_ch_mag128b+=2; + rxdataF128+=4; + rxdataF_comp128+=2; + } + } + } + } + + if (rho) { + for (aarx=0; aarxnb_antennas_rx; aarx++) { + rho128 = (int16x4x2_t*)&rho[aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch128 = (int16x4_t*)&dl_ch_estimates_ext[aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch128_2 = (int16x4_t*)&dl_ch_estimates_ext[2+aarx][symbol*frame_parms->N_RB_DL*12]; + for (rb=0; rbrx_correlation[0][aarx] = signal_energy(&rho[aarx][symbol*frame_parms->N_RB_DL*12],rb*12); + } + } + } +#endif +} + + +void nr_dlsch_scale_channel(int **dl_ch_estimates_ext, + NR_DL_FRAME_PARMS *frame_parms, + NR_UE_DLSCH_t **dlsch_ue, + uint8_t symbol, + unsigned short nb_rb) +{ + +#if defined(__x86_64__)||defined(__i386__) + + short rb, ch_amp; + unsigned char aatx,aarx,pilots=0,symbol_mod; + __m128i *dl_ch128, ch_amp128; + + + if (symbol==2){ + nb_rb = nb_rb>>1; + pilots=1; + } + + // Determine scaling amplitude based the symbol + +ch_amp = 1024*8; //((pilots) ? (dlsch_ue[0]->sqrt_rho_b) : (dlsch_ue[0]->sqrt_rho_a)); + + LOG_D(PHY,"Scaling PDSCH Chest in OFDM symbol %d by %d, pilots %d nb_rb %d NCP %d symbol %d\n",symbol_mod,ch_amp,pilots,nb_rb,frame_parms->Ncp,symbol); + // printf("Scaling PDSCH Chest in OFDM symbol %d by %d\n",symbol_mod,ch_amp); + + ch_amp128 = _mm_set1_epi16(ch_amp); // Q3.13 + + for (aatx=0; aatxnb_antenna_ports_eNB; aatx++) { + for (aarx=0; aarxnb_antennas_rx; aarx++) { + + dl_ch128=(__m128i *)&dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*nb_rb*12]; + + for (rb=0;rb>x; + //printf("nb_rb*nre = %d = %d * 2^(%d)\n",nb_rb*nre,y,x); + + for (aatx=0; aatxnb_antenna_ports_eNB; aatx++) + for (aarx=0; aarxnb_antennas_rx; aarx++) { + //clear average level + avg128D = _mm_setzero_si128(); + // 5 is always a symbol with no pilots for both normal and extended prefix + + dl_ch128=(__m128i *)&dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*nb_rb*12]; + + for (rb=0;rbNcp-1)))&&(frame_parms->nb_antenna_ports_eNB!=1)) { + dl_ch128+=2; + } + else {*/ + avg128D = _mm_add_epi32(avg128D,_mm_srai_epi16(_mm_madd_epi16(dl_ch128[2],dl_ch128[2]),x)); + //avg128D = _mm_add_epi32(avg128D,_mm_madd_epi16(dl_ch128[2],_mm_srai_epi16(_mm_mulhi_epi16(dl_ch128[2], coeff128),15))); + dl_ch128+=3; + // } + /* + if (rb==0) { + print_shorts("dl_ch128",&dl_ch128[0]); + print_shorts("dl_ch128",&dl_ch128[1]); + print_shorts("dl_ch128",&dl_ch128[2]); + } + */ + } + + avg[(aatx<<1)+aarx] =(((int32_t*)&avg128D)[0] + + ((int32_t*)&avg128D)[1] + + ((int32_t*)&avg128D)[2] + + ((int32_t*)&avg128D)[3])/y; + // printf("Channel level : %d\n",avg[(aatx<<1)+aarx]); + } + + _mm_empty(); + _m_empty(); + +#elif defined(__arm__) + + short rb; + unsigned char aatx,aarx,nre=12,symbol_mod; + int32x4_t avg128D; + int16x4_t *dl_ch128; + + symbol_mod = (symbol>=(7-frame_parms->Ncp)) ? symbol-(7-frame_parms->Ncp) : symbol; + + for (aatx=0; aatxnb_antenna_ports_eNB; aatx++) + for (aarx=0; aarxnb_antennas_rx; aarx++) { + //clear average level + avg128D = vdupq_n_s32(0); + // 5 is always a symbol with no pilots for both normal and extended prefix + + dl_ch128=(int16x4_t *)&dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; + + for (rb=0; rbNcp-1)))&&(frame_parms->nb_antenna_ports_eNB!=1)) { + dl_ch128+=4; + } else { + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[4],dl_ch128[4])); + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[5],dl_ch128[5])); + dl_ch128+=6; + } + + /* + if (rb==0) { + print_shorts("dl_ch128",&dl_ch128[0]); + print_shorts("dl_ch128",&dl_ch128[1]); + print_shorts("dl_ch128",&dl_ch128[2]); + } + */ + } + + if (symbol==2) //assume start symbol 2 + nre=6; + else + nre=12; + + avg[(aatx<<1)+aarx] = (((int32_t*)&avg128D)[0] + + ((int32_t*)&avg128D)[1] + + ((int32_t*)&avg128D)[2] + + ((int32_t*)&avg128D)[3])/(nb_rb*nre); + + // printf("Channel level : %d\n",avg[(aatx<<1)+aarx]); + } + + +#endif +} + +//============================================================================================== +// Extraction functions +//============================================================================================== + +unsigned short nr_dlsch_extract_rbs_single(int **rxdataF, + int **dl_ch_estimates, + int **rxdataF_ext, + int **dl_ch_estimates_ext, + unsigned short pmi, + unsigned char *pmi_ext, + unsigned int *rb_alloc, + unsigned char symbol, + unsigned short start_rb, + unsigned short nb_rb_pdsch, + unsigned char nr_tti_rx, + uint32_t high_speed_flag, + NR_DL_FRAME_PARMS *frame_parms) { + + + + unsigned short k,rb; + unsigned char i,aarx; //,nsymb,sss_symb,pss_symb=0,l; + int *dl_ch0,*dl_ch0_ext,*rxF,*rxF_ext; + + + + unsigned char pilots=0,j=0; + + //symbol_mod = (symbol>=(7-frame_parms->Ncp)) ? symbol-(7-frame_parms->Ncp) : symbol; + pilots = (symbol==2) ? 1 : 0; //to updated from config + //l=symbol; + //nsymb = (frame_parms->Ncp==NORMAL) ? 14:12; + k = frame_parms->first_carrier_offset + 516; //0 + + /* if (frame_parms->frame_type == TDD) { // TDD + sss_symb = nsymb-1; + pss_symb = 2; + } else { + sss_symb = (nsymb>>1)-2; + pss_symb = (nsymb>>1)-1; + }*/ + + + for (aarx=0; aarxnb_antennas_rx; aarx++) { + + if (high_speed_flag == 1) + dl_ch0 = &dl_ch_estimates[aarx][(2*(frame_parms->ofdm_symbol_size))]; + else + dl_ch0 = &dl_ch_estimates[aarx][0]; + + dl_ch0_ext = &dl_ch_estimates_ext[aarx][symbol*(nb_rb_pdsch*12)]; + + rxF_ext = &rxdataF_ext[aarx][symbol*(nb_rb_pdsch*12)]; + rxF = &rxdataF[aarx][(k+(symbol*(frame_parms->ofdm_symbol_size)))]; + + //printf("extract single addr %p pilots %d symbol %d, l %d RB_DL %d nushift %d k %d\n", rxF, pilots, symbol, l, frame_parms->N_RB_DL,frame_parms->nushift,k); + + //if ((frame_parms->N_RB_DL&1) == 0){ // even number of RBs + + for (rb=start_rb;rb=frame_parms->ofdm_symbol_size) { + rxF = &rxdataF[aarx][(symbol*(frame_parms->ofdm_symbol_size))]; + k=k-(frame_parms->ofdm_symbol_size); + } + + //*pmi_ext = (pmi>>((rb>>2)<<1))&3; + //memcpy(dl_ch0_ext,dl_ch0,12*sizeof(int)); + + /* + printf("rb %d\n",rb); + for (i=0;i<12;i++) + printf("(%d %d)",((short *)dl_ch0)[i<<1],((short*)dl_ch0)[1+(i<<1)]); + printf("\n"); + */ + if (pilots==0) { + for (i=0; i<12; i++) { + rxF_ext[i]=rxF[i]; + dl_ch0_ext[i]=dl_ch0[i]; + /*printf("rb %ld %p: (%d,%d)\n",(rxF+i-&rxdataF[aarx][( (symbol*(frame_parms->ofdm_symbol_size)))]),rxF, + ((short*)&rxF[i])[0],((short*)&rxF[i])[1]); + printf("rb %ld %p: dl_ch (%d,%d)\n",(rxF+i-&rxdataF[aarx][( (symbol*(frame_parms->ofdm_symbol_size)))]),rxF, + ((short*)&dl_ch0[i])[0],((short*)&dl_ch0[i])[1]); + printf("re count %d\n",rb*12+i);*/ + } + + dl_ch0_ext+=12; + rxF_ext+=12; + } else { + j=0; + + for (i=0; i<12; i++) { + if ((i&1)!=frame_parms->nushift){ + rxF_ext[j]=rxF[i]; + //printf("extract rb %d, re %d => rxF (%d,%d) rxF ext (%d,%d) addr %p\n",rb,i,*(short *)&rxF[i],*(1+(short*)&rxF[i]),*(short *)&rxF_ext[j],*(1+(short*)&rxF_ext[j]),rxF_ext); + dl_ch0_ext[j]=dl_ch0[i]; + //printf("rb %ld %p: dl_ch (%d,%d)\n",(rxF+i-&rxdataF[aarx][( (symbol*(frame_parms->ofdm_symbol_size)))]),rxF, + //((short*)&dl_ch0[i])[0],((short*)&dl_ch0[i])[1]); + j++; + } + } + + dl_ch0_ext+=6; + rxF_ext+=6; + } + + dl_ch0+=12; + rxF+=12; + k+=12; + } + } + + + return(nb_rb_pdsch/frame_parms->nb_antennas_rx); +} + +//============================================================================================== + +#ifdef USER_MODE + + +void dump_dlsch2(PHY_VARS_UE *ue,uint8_t eNB_id,uint8_t nr_tti_rx,unsigned int *coded_bits_per_codeword,int round, unsigned char harq_pid) +{ + unsigned int nsymb = (ue->frame_parms.Ncp == 0) ? 14 : 12; + char fname[32],vname[32]; + int N_RB_DL=ue->frame_parms.N_RB_DL; + + sprintf(fname,"dlsch%d_rxF_r%d_ext0.m",eNB_id,round); + sprintf(vname,"dl%d_rxF_r%d_ext0",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->rxdataF_ext[0],12*N_RB_DL*nsymb,1,1); + + if (ue->frame_parms.nb_antennas_rx >1) { + sprintf(fname,"dlsch%d_rxF_r%d_ext1.m",eNB_id,round); + sprintf(vname,"dl%d_rxF_r%d_ext1",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->rxdataF_ext[1],12*N_RB_DL*nsymb,1,1); + } + + sprintf(fname,"dlsch%d_ch_r%d_ext00.m",eNB_id,round); + sprintf(vname,"dl%d_ch_r%d_ext00",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_estimates_ext[0],12*N_RB_DL*nsymb,1,1); + + if (ue->transmission_mode[eNB_id]==7){ + sprintf(fname,"dlsch%d_bf_ch_r%d.m",eNB_id,round); + sprintf(vname,"dl%d_bf_ch_r%d",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_bf_ch_estimates[0],512*nsymb,1,1); + //write_output(fname,vname,phy_vars_ue->lte_ue_pdsch_vars[eNB_id]->dl_bf_ch_estimates[0],512,1,1); + + sprintf(fname,"dlsch%d_bf_ch_r%d_ext00.m",eNB_id,round); + sprintf(vname,"dl%d_bf_ch_r%d_ext00",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_bf_ch_estimates_ext[0],12*N_RB_DL*nsymb,1,1); + } + + if (ue->frame_parms.nb_antennas_rx == 2) { + sprintf(fname,"dlsch%d_ch_r%d_ext01.m",eNB_id,round); + sprintf(vname,"dl%d_ch_r%d_ext01",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_estimates_ext[1],12*N_RB_DL*nsymb,1,1); + } + + if (ue->frame_parms.nb_antenna_ports_eNB == 2) { + sprintf(fname,"dlsch%d_ch_r%d_ext10.m",eNB_id,round); + sprintf(vname,"dl%d_ch_r%d_ext10",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_estimates_ext[2],12*N_RB_DL*nsymb,1,1); + + if (ue->frame_parms.nb_antennas_rx == 2) { + sprintf(fname,"dlsch%d_ch_r%d_ext11.m",eNB_id,round); + sprintf(vname,"dl%d_ch_r%d_ext11",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_estimates_ext[3],12*N_RB_DL*nsymb,1,1); + } + } + + sprintf(fname,"dlsch%d_rxF_r%d_uespec0.m",eNB_id,round); + sprintf(vname,"dl%d_rxF_r%d_uespec0",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->rxdataF_uespec_pilots[0],12*N_RB_DL,1,1); + + /* + write_output("dlsch%d_ch_ext01.m","dl01_ch0_ext",pdsch_vars[eNB_id]->dl_ch_estimates_ext[1],12*N_RB_DL*nsymb,1,1); + write_output("dlsch%d_ch_ext10.m","dl10_ch0_ext",pdsch_vars[eNB_id]->dl_ch_estimates_ext[2],12*N_RB_DL*nsymb,1,1); + write_output("dlsch%d_ch_ext11.m","dl11_ch0_ext",pdsch_vars[eNB_id]->dl_ch_estimates_ext[3],12*N_RB_DL*nsymb,1,1); + */ + sprintf(fname,"dlsch%d_r%d_rho.m",eNB_id,round); + sprintf(vname,"dl_rho_r%d_%d",eNB_id,round); + + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_rho_ext[harq_pid][round][0],12*N_RB_DL*nsymb,1,1); + + sprintf(fname,"dlsch%d_r%d_rho2.m",eNB_id,round); + sprintf(vname,"dl_rho2_r%d_%d",eNB_id,round); + + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_rho2_ext[0],12*N_RB_DL*nsymb,1,1); + + sprintf(fname,"dlsch%d_rxF_r%d_comp0.m",eNB_id,round); + sprintf(vname,"dl%d_rxF_r%d_comp0",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->rxdataF_comp0[0],12*N_RB_DL*nsymb,1,1); + if (ue->frame_parms.nb_antenna_ports_eNB == 2) { + sprintf(fname,"dlsch%d_rxF_r%d_comp1.m",eNB_id,round); + sprintf(vname,"dl%d_rxF_r%d_comp1",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->rxdataF_comp1[harq_pid][round][0],12*N_RB_DL*nsymb,1,1); + } + + sprintf(fname,"dlsch%d_rxF_r%d_llr.m",eNB_id,round); + sprintf(vname,"dl%d_r%d_llr",eNB_id,round); + write_output(fname,vname, ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->llr[0],coded_bits_per_codeword[0],1,0); + sprintf(fname,"dlsch%d_r%d_mag1.m",eNB_id,round); + sprintf(vname,"dl%d_r%d_mag1",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_mag0[0],12*N_RB_DL*nsymb,1,1); + sprintf(fname,"dlsch%d_r%d_mag2.m",eNB_id,round); + sprintf(vname,"dl%d_r%d_mag2",eNB_id,round); + write_output(fname,vname,ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dl_ch_magb0[0],12*N_RB_DL*nsymb,1,1); + + // printf("log2_maxh = %d\n",ue->pdsch_vars[eNB_id]->log2_maxh); +} +#endif + +#ifdef DEBUG_DLSCH_DEMOD +/* +void print_bytes(char *s,__m128i *x) +{ + + char *tempb = (char *)x; + + printf("%s : %d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d\n",s, + tempb[0],tempb[1],tempb[2],tempb[3],tempb[4],tempb[5],tempb[6],tempb[7], + tempb[8],tempb[9],tempb[10],tempb[11],tempb[12],tempb[13],tempb[14],tempb[15] + ); + +} + +void print_shorts(char *s,__m128i *x) +{ + + short *tempb = (short *)x; + printf("%s : %d,%d,%d,%d,%d,%d,%d,%d\n",s, + tempb[0],tempb[1],tempb[2],tempb[3],tempb[4],tempb[5],tempb[6],tempb[7]); + +} + +void print_shorts2(char *s,__m64 *x) +{ + + short *tempb = (short *)x; + printf("%s : %d,%d,%d,%d\n",s, + tempb[0],tempb[1],tempb[2],tempb[3]); + +} + +void print_ints(char *s,__m128i *x) +{ + + int *tempb = (int *)x; + printf("%s : %d,%d,%d,%d\n",s, + tempb[0],tempb[1],tempb[2],tempb[3]); + +}*/ +#endif diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_llr_computation.c b/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_llr_computation.c new file mode 100644 index 0000000000..9527bf10e9 --- /dev/null +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_llr_computation.c @@ -0,0 +1,8641 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +/*! \file PHY/NR_UE_TRANSPORT/nr_dlsch_llr_computation.c + * \brief Top-level routines for LLR computation of the PDSCH physical channel + * \author H. WANG + * \date 2018 + * \version 0.1 + * \company Eurecom + * \email: + * \note + * \warning + */ + +#include "PHY/defs_nr_UE.h" +#include "PHY/phy_extern_nr_ue.h" +#include "nr_transport_proto_ue.h" +#include "PHY/TOOLS/tools_defs.h" +#include "PHY/sse_intrin.h" + +//#define DEBUG_LLR_SIC + + +int16_t zeros[8] __attribute__ ((aligned(16))) = {0,0,0,0,0,0,0,0}; +int16_t ones[8] __attribute__ ((aligned(16))) = {0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff}; +#if defined(__x86_64__) || defined(__i386__) +__m128i rho_rpi __attribute__ ((aligned(16))); +__m128i rho_rmi __attribute__ ((aligned(16))); +__m128i rho_rpi_1_1 __attribute__ ((aligned(16))); +__m128i rho_rpi_1_3 __attribute__ ((aligned(16))); +__m128i rho_rpi_1_5 __attribute__ ((aligned(16))); +__m128i rho_rpi_1_7 __attribute__ ((aligned(16))); +__m128i rho_rpi_3_1 __attribute__ ((aligned(16))); +__m128i rho_rpi_3_3 __attribute__ ((aligned(16))); +__m128i rho_rpi_3_5 __attribute__ ((aligned(16))); +__m128i rho_rpi_3_7 __attribute__ ((aligned(16))); +__m128i rho_rpi_5_1 __attribute__ ((aligned(16))); +__m128i rho_rpi_5_3 __attribute__ ((aligned(16))); +__m128i rho_rpi_5_5 __attribute__ ((aligned(16))); +__m128i rho_rpi_5_7 __attribute__ ((aligned(16))); +__m128i rho_rpi_7_1 __attribute__ ((aligned(16))); +__m128i rho_rpi_7_3 __attribute__ ((aligned(16))); +__m128i rho_rpi_7_5 __attribute__ ((aligned(16))); +__m128i rho_rpi_7_7 __attribute__ ((aligned(16))); +__m128i rho_rmi_1_1 __attribute__ ((aligned(16))); +__m128i rho_rmi_1_3 __attribute__ ((aligned(16))); +__m128i rho_rmi_1_5 __attribute__ ((aligned(16))); +__m128i rho_rmi_1_7 __attribute__ ((aligned(16))); +__m128i rho_rmi_3_1 __attribute__ ((aligned(16))); +__m128i rho_rmi_3_3 __attribute__ ((aligned(16))); +__m128i rho_rmi_3_5 __attribute__ ((aligned(16))); +__m128i rho_rmi_3_7 __attribute__ ((aligned(16))); +__m128i rho_rmi_5_1 __attribute__ ((aligned(16))); +__m128i rho_rmi_5_3 __attribute__ ((aligned(16))); +__m128i rho_rmi_5_5 __attribute__ ((aligned(16))); +__m128i rho_rmi_5_7 __attribute__ ((aligned(16))); +__m128i rho_rmi_7_1 __attribute__ ((aligned(16))); +__m128i rho_rmi_7_3 __attribute__ ((aligned(16))); +__m128i rho_rmi_7_5 __attribute__ ((aligned(16))); +__m128i rho_rmi_7_7 __attribute__ ((aligned(16))); + +__m128i psi_r_m7_m7 __attribute__ ((aligned(16))); +__m128i psi_r_m7_m5 __attribute__ ((aligned(16))); +__m128i psi_r_m7_m3 __attribute__ ((aligned(16))); +__m128i psi_r_m7_m1 __attribute__ ((aligned(16))); +__m128i psi_r_m7_p1 __attribute__ ((aligned(16))); +__m128i psi_r_m7_p3 __attribute__ ((aligned(16))); +__m128i psi_r_m7_p5 __attribute__ ((aligned(16))); +__m128i psi_r_m7_p7 __attribute__ ((aligned(16))); +__m128i psi_r_m5_m7 __attribute__ ((aligned(16))); +__m128i psi_r_m5_m5 __attribute__ ((aligned(16))); +__m128i psi_r_m5_m3 __attribute__ ((aligned(16))); +__m128i psi_r_m5_m1 __attribute__ ((aligned(16))); +__m128i psi_r_m5_p1 __attribute__ ((aligned(16))); +__m128i psi_r_m5_p3 __attribute__ ((aligned(16))); +__m128i psi_r_m5_p5 __attribute__ ((aligned(16))); +__m128i psi_r_m5_p7 __attribute__ ((aligned(16))); +__m128i psi_r_m3_m7 __attribute__ ((aligned(16))); +__m128i psi_r_m3_m5 __attribute__ ((aligned(16))); +__m128i psi_r_m3_m3 __attribute__ ((aligned(16))); +__m128i psi_r_m3_m1 __attribute__ ((aligned(16))); +__m128i psi_r_m3_p1 __attribute__ ((aligned(16))); +__m128i psi_r_m3_p3 __attribute__ ((aligned(16))); +__m128i psi_r_m3_p5 __attribute__ ((aligned(16))); +__m128i psi_r_m3_p7 __attribute__ ((aligned(16))); +__m128i psi_r_m1_m7 __attribute__ ((aligned(16))); +__m128i psi_r_m1_m5 __attribute__ ((aligned(16))); +__m128i psi_r_m1_m3 __attribute__ ((aligned(16))); +__m128i psi_r_m1_m1 __attribute__ ((aligned(16))); +__m128i psi_r_m1_p1 __attribute__ ((aligned(16))); +__m128i psi_r_m1_p3 __attribute__ ((aligned(16))); +__m128i psi_r_m1_p5 __attribute__ ((aligned(16))); +__m128i psi_r_m1_p7 __attribute__ ((aligned(16))); +__m128i psi_r_p1_m7 __attribute__ ((aligned(16))); +__m128i psi_r_p1_m5 __attribute__ ((aligned(16))); +__m128i psi_r_p1_m3 __attribute__ ((aligned(16))); +__m128i psi_r_p1_m1 __attribute__ ((aligned(16))); +__m128i psi_r_p1_p1 __attribute__ ((aligned(16))); +__m128i psi_r_p1_p3 __attribute__ ((aligned(16))); +__m128i psi_r_p1_p5 __attribute__ ((aligned(16))); +__m128i psi_r_p1_p7 __attribute__ ((aligned(16))); +__m128i psi_r_p3_m7 __attribute__ ((aligned(16))); +__m128i psi_r_p3_m5 __attribute__ ((aligned(16))); +__m128i psi_r_p3_m3 __attribute__ ((aligned(16))); +__m128i psi_r_p3_m1 __attribute__ ((aligned(16))); +__m128i psi_r_p3_p1 __attribute__ ((aligned(16))); +__m128i psi_r_p3_p3 __attribute__ ((aligned(16))); +__m128i psi_r_p3_p5 __attribute__ ((aligned(16))); +__m128i psi_r_p3_p7 __attribute__ ((aligned(16))); +__m128i psi_r_p5_m7 __attribute__ ((aligned(16))); +__m128i psi_r_p5_m5 __attribute__ ((aligned(16))); +__m128i psi_r_p5_m3 __attribute__ ((aligned(16))); +__m128i psi_r_p5_m1 __attribute__ ((aligned(16))); +__m128i psi_r_p5_p1 __attribute__ ((aligned(16))); +__m128i psi_r_p5_p3 __attribute__ ((aligned(16))); +__m128i psi_r_p5_p5 __attribute__ ((aligned(16))); +__m128i psi_r_p5_p7 __attribute__ ((aligned(16))); +__m128i psi_r_p7_m7 __attribute__ ((aligned(16))); +__m128i psi_r_p7_m5 __attribute__ ((aligned(16))); +__m128i psi_r_p7_m3 __attribute__ ((aligned(16))); +__m128i psi_r_p7_m1 __attribute__ ((aligned(16))); +__m128i psi_r_p7_p1 __attribute__ ((aligned(16))); +__m128i psi_r_p7_p3 __attribute__ ((aligned(16))); +__m128i psi_r_p7_p5 __attribute__ ((aligned(16))); +__m128i psi_r_p7_p7 __attribute__ ((aligned(16))); + +__m128i psi_i_m7_m7 __attribute__ ((aligned(16))); +__m128i psi_i_m7_m5 __attribute__ ((aligned(16))); +__m128i psi_i_m7_m3 __attribute__ ((aligned(16))); +__m128i psi_i_m7_m1 __attribute__ ((aligned(16))); +__m128i psi_i_m7_p1 __attribute__ ((aligned(16))); +__m128i psi_i_m7_p3 __attribute__ ((aligned(16))); +__m128i psi_i_m7_p5 __attribute__ ((aligned(16))); +__m128i psi_i_m7_p7 __attribute__ ((aligned(16))); +__m128i psi_i_m5_m7 __attribute__ ((aligned(16))); +__m128i psi_i_m5_m5 __attribute__ ((aligned(16))); +__m128i psi_i_m5_m3 __attribute__ ((aligned(16))); +__m128i psi_i_m5_m1 __attribute__ ((aligned(16))); +__m128i psi_i_m5_p1 __attribute__ ((aligned(16))); +__m128i psi_i_m5_p3 __attribute__ ((aligned(16))); +__m128i psi_i_m5_p5 __attribute__ ((aligned(16))); +__m128i psi_i_m5_p7 __attribute__ ((aligned(16))); +__m128i psi_i_m3_m7 __attribute__ ((aligned(16))); +__m128i psi_i_m3_m5 __attribute__ ((aligned(16))); +__m128i psi_i_m3_m3 __attribute__ ((aligned(16))); +__m128i psi_i_m3_m1 __attribute__ ((aligned(16))); +__m128i psi_i_m3_p1 __attribute__ ((aligned(16))); +__m128i psi_i_m3_p3 __attribute__ ((aligned(16))); +__m128i psi_i_m3_p5 __attribute__ ((aligned(16))); +__m128i psi_i_m3_p7 __attribute__ ((aligned(16))); +__m128i psi_i_m1_m7 __attribute__ ((aligned(16))); +__m128i psi_i_m1_m5 __attribute__ ((aligned(16))); +__m128i psi_i_m1_m3 __attribute__ ((aligned(16))); +__m128i psi_i_m1_m1 __attribute__ ((aligned(16))); +__m128i psi_i_m1_p1 __attribute__ ((aligned(16))); +__m128i psi_i_m1_p3 __attribute__ ((aligned(16))); +__m128i psi_i_m1_p5 __attribute__ ((aligned(16))); +__m128i psi_i_m1_p7 __attribute__ ((aligned(16))); +__m128i psi_i_p1_m7 __attribute__ ((aligned(16))); +__m128i psi_i_p1_m5 __attribute__ ((aligned(16))); +__m128i psi_i_p1_m3 __attribute__ ((aligned(16))); +__m128i psi_i_p1_m1 __attribute__ ((aligned(16))); +__m128i psi_i_p1_p1 __attribute__ ((aligned(16))); +__m128i psi_i_p1_p3 __attribute__ ((aligned(16))); +__m128i psi_i_p1_p5 __attribute__ ((aligned(16))); +__m128i psi_i_p1_p7 __attribute__ ((aligned(16))); +__m128i psi_i_p3_m7 __attribute__ ((aligned(16))); +__m128i psi_i_p3_m5 __attribute__ ((aligned(16))); +__m128i psi_i_p3_m3 __attribute__ ((aligned(16))); +__m128i psi_i_p3_m1 __attribute__ ((aligned(16))); +__m128i psi_i_p3_p1 __attribute__ ((aligned(16))); +__m128i psi_i_p3_p3 __attribute__ ((aligned(16))); +__m128i psi_i_p3_p5 __attribute__ ((aligned(16))); +__m128i psi_i_p3_p7 __attribute__ ((aligned(16))); +__m128i psi_i_p5_m7 __attribute__ ((aligned(16))); +__m128i psi_i_p5_m5 __attribute__ ((aligned(16))); +__m128i psi_i_p5_m3 __attribute__ ((aligned(16))); +__m128i psi_i_p5_m1 __attribute__ ((aligned(16))); +__m128i psi_i_p5_p1 __attribute__ ((aligned(16))); +__m128i psi_i_p5_p3 __attribute__ ((aligned(16))); +__m128i psi_i_p5_p5 __attribute__ ((aligned(16))); +__m128i psi_i_p5_p7 __attribute__ ((aligned(16))); +__m128i psi_i_p7_m7 __attribute__ ((aligned(16))); +__m128i psi_i_p7_m5 __attribute__ ((aligned(16))); +__m128i psi_i_p7_m3 __attribute__ ((aligned(16))); +__m128i psi_i_p7_m1 __attribute__ ((aligned(16))); +__m128i psi_i_p7_p1 __attribute__ ((aligned(16))); +__m128i psi_i_p7_p3 __attribute__ ((aligned(16))); +__m128i psi_i_p7_p5 __attribute__ ((aligned(16))); +__m128i psi_i_p7_p7 __attribute__ ((aligned(16))); + +__m128i a_r_m7_m7 __attribute__ ((aligned(16))); +__m128i a_r_m7_m5 __attribute__ ((aligned(16))); +__m128i a_r_m7_m3 __attribute__ ((aligned(16))); +__m128i a_r_m7_m1 __attribute__ ((aligned(16))); +__m128i a_r_m7_p1 __attribute__ ((aligned(16))); +__m128i a_r_m7_p3 __attribute__ ((aligned(16))); +__m128i a_r_m7_p5 __attribute__ ((aligned(16))); +__m128i a_r_m7_p7 __attribute__ ((aligned(16))); +__m128i a_r_m5_m7 __attribute__ ((aligned(16))); +__m128i a_r_m5_m5 __attribute__ ((aligned(16))); +__m128i a_r_m5_m3 __attribute__ ((aligned(16))); +__m128i a_r_m5_m1 __attribute__ ((aligned(16))); +__m128i a_r_m5_p1 __attribute__ ((aligned(16))); +__m128i a_r_m5_p3 __attribute__ ((aligned(16))); +__m128i a_r_m5_p5 __attribute__ ((aligned(16))); +__m128i a_r_m5_p7 __attribute__ ((aligned(16))); +__m128i a_r_m3_m7 __attribute__ ((aligned(16))); +__m128i a_r_m3_m5 __attribute__ ((aligned(16))); +__m128i a_r_m3_m3 __attribute__ ((aligned(16))); +__m128i a_r_m3_m1 __attribute__ ((aligned(16))); +__m128i a_r_m3_p1 __attribute__ ((aligned(16))); +__m128i a_r_m3_p3 __attribute__ ((aligned(16))); +__m128i a_r_m3_p5 __attribute__ ((aligned(16))); +__m128i a_r_m3_p7 __attribute__ ((aligned(16))); +__m128i a_r_m1_m7 __attribute__ ((aligned(16))); +__m128i a_r_m1_m5 __attribute__ ((aligned(16))); +__m128i a_r_m1_m3 __attribute__ ((aligned(16))); +__m128i a_r_m1_m1 __attribute__ ((aligned(16))); +__m128i a_r_m1_p1 __attribute__ ((aligned(16))); +__m128i a_r_m1_p3 __attribute__ ((aligned(16))); +__m128i a_r_m1_p5 __attribute__ ((aligned(16))); +__m128i a_r_m1_p7 __attribute__ ((aligned(16))); +__m128i a_r_p1_m7 __attribute__ ((aligned(16))); +__m128i a_r_p1_m5 __attribute__ ((aligned(16))); +__m128i a_r_p1_m3 __attribute__ ((aligned(16))); +__m128i a_r_p1_m1 __attribute__ ((aligned(16))); +__m128i a_r_p1_p1 __attribute__ ((aligned(16))); +__m128i a_r_p1_p3 __attribute__ ((aligned(16))); +__m128i a_r_p1_p5 __attribute__ ((aligned(16))); +__m128i a_r_p1_p7 __attribute__ ((aligned(16))); +__m128i a_r_p3_m7 __attribute__ ((aligned(16))); +__m128i a_r_p3_m5 __attribute__ ((aligned(16))); +__m128i a_r_p3_m3 __attribute__ ((aligned(16))); +__m128i a_r_p3_m1 __attribute__ ((aligned(16))); +__m128i a_r_p3_p1 __attribute__ ((aligned(16))); +__m128i a_r_p3_p3 __attribute__ ((aligned(16))); +__m128i a_r_p3_p5 __attribute__ ((aligned(16))); +__m128i a_r_p3_p7 __attribute__ ((aligned(16))); +__m128i a_r_p5_m7 __attribute__ ((aligned(16))); +__m128i a_r_p5_m5 __attribute__ ((aligned(16))); +__m128i a_r_p5_m3 __attribute__ ((aligned(16))); +__m128i a_r_p5_m1 __attribute__ ((aligned(16))); +__m128i a_r_p5_p1 __attribute__ ((aligned(16))); +__m128i a_r_p5_p3 __attribute__ ((aligned(16))); +__m128i a_r_p5_p5 __attribute__ ((aligned(16))); +__m128i a_r_p5_p7 __attribute__ ((aligned(16))); +__m128i a_r_p7_m7 __attribute__ ((aligned(16))); +__m128i a_r_p7_m5 __attribute__ ((aligned(16))); +__m128i a_r_p7_m3 __attribute__ ((aligned(16))); +__m128i a_r_p7_m1 __attribute__ ((aligned(16))); +__m128i a_r_p7_p1 __attribute__ ((aligned(16))); +__m128i a_r_p7_p3 __attribute__ ((aligned(16))); +__m128i a_r_p7_p5 __attribute__ ((aligned(16))); +__m128i a_r_p7_p7 __attribute__ ((aligned(16))); + +__m128i a_i_m7_m7 __attribute__ ((aligned(16))); +__m128i a_i_m7_m5 __attribute__ ((aligned(16))); +__m128i a_i_m7_m3 __attribute__ ((aligned(16))); +__m128i a_i_m7_m1 __attribute__ ((aligned(16))); +__m128i a_i_m7_p1 __attribute__ ((aligned(16))); +__m128i a_i_m7_p3 __attribute__ ((aligned(16))); +__m128i a_i_m7_p5 __attribute__ ((aligned(16))); +__m128i a_i_m7_p7 __attribute__ ((aligned(16))); +__m128i a_i_m5_m7 __attribute__ ((aligned(16))); +__m128i a_i_m5_m5 __attribute__ ((aligned(16))); +__m128i a_i_m5_m3 __attribute__ ((aligned(16))); +__m128i a_i_m5_m1 __attribute__ ((aligned(16))); +__m128i a_i_m5_p1 __attribute__ ((aligned(16))); +__m128i a_i_m5_p3 __attribute__ ((aligned(16))); +__m128i a_i_m5_p5 __attribute__ ((aligned(16))); +__m128i a_i_m5_p7 __attribute__ ((aligned(16))); +__m128i a_i_m3_m7 __attribute__ ((aligned(16))); +__m128i a_i_m3_m5 __attribute__ ((aligned(16))); +__m128i a_i_m3_m3 __attribute__ ((aligned(16))); +__m128i a_i_m3_m1 __attribute__ ((aligned(16))); +__m128i a_i_m3_p1 __attribute__ ((aligned(16))); +__m128i a_i_m3_p3 __attribute__ ((aligned(16))); +__m128i a_i_m3_p5 __attribute__ ((aligned(16))); +__m128i a_i_m3_p7 __attribute__ ((aligned(16))); +__m128i a_i_m1_m7 __attribute__ ((aligned(16))); +__m128i a_i_m1_m5 __attribute__ ((aligned(16))); +__m128i a_i_m1_m3 __attribute__ ((aligned(16))); +__m128i a_i_m1_m1 __attribute__ ((aligned(16))); +__m128i a_i_m1_p1 __attribute__ ((aligned(16))); +__m128i a_i_m1_p3 __attribute__ ((aligned(16))); +__m128i a_i_m1_p5 __attribute__ ((aligned(16))); +__m128i a_i_m1_p7 __attribute__ ((aligned(16))); +__m128i a_i_p1_m7 __attribute__ ((aligned(16))); +__m128i a_i_p1_m5 __attribute__ ((aligned(16))); +__m128i a_i_p1_m3 __attribute__ ((aligned(16))); +__m128i a_i_p1_m1 __attribute__ ((aligned(16))); +__m128i a_i_p1_p1 __attribute__ ((aligned(16))); +__m128i a_i_p1_p3 __attribute__ ((aligned(16))); +__m128i a_i_p1_p5 __attribute__ ((aligned(16))); +__m128i a_i_p1_p7 __attribute__ ((aligned(16))); +__m128i a_i_p3_m7 __attribute__ ((aligned(16))); +__m128i a_i_p3_m5 __attribute__ ((aligned(16))); +__m128i a_i_p3_m3 __attribute__ ((aligned(16))); +__m128i a_i_p3_m1 __attribute__ ((aligned(16))); +__m128i a_i_p3_p1 __attribute__ ((aligned(16))); +__m128i a_i_p3_p3 __attribute__ ((aligned(16))); +__m128i a_i_p3_p5 __attribute__ ((aligned(16))); +__m128i a_i_p3_p7 __attribute__ ((aligned(16))); +__m128i a_i_p5_m7 __attribute__ ((aligned(16))); +__m128i a_i_p5_m5 __attribute__ ((aligned(16))); +__m128i a_i_p5_m3 __attribute__ ((aligned(16))); +__m128i a_i_p5_m1 __attribute__ ((aligned(16))); +__m128i a_i_p5_p1 __attribute__ ((aligned(16))); +__m128i a_i_p5_p3 __attribute__ ((aligned(16))); +__m128i a_i_p5_p5 __attribute__ ((aligned(16))); +__m128i a_i_p5_p7 __attribute__ ((aligned(16))); +__m128i a_i_p7_m7 __attribute__ ((aligned(16))); +__m128i a_i_p7_m5 __attribute__ ((aligned(16))); +__m128i a_i_p7_m3 __attribute__ ((aligned(16))); +__m128i a_i_p7_m1 __attribute__ ((aligned(16))); +__m128i a_i_p7_p1 __attribute__ ((aligned(16))); +__m128i a_i_p7_p3 __attribute__ ((aligned(16))); +__m128i a_i_p7_p5 __attribute__ ((aligned(16))); +__m128i a_i_p7_p7 __attribute__ ((aligned(16))); + +__m128i psi_a_m7_m7 __attribute__ ((aligned(16))); +__m128i psi_a_m7_m5 __attribute__ ((aligned(16))); +__m128i psi_a_m7_m3 __attribute__ ((aligned(16))); +__m128i psi_a_m7_m1 __attribute__ ((aligned(16))); +__m128i psi_a_m7_p1 __attribute__ ((aligned(16))); +__m128i psi_a_m7_p3 __attribute__ ((aligned(16))); +__m128i psi_a_m7_p5 __attribute__ ((aligned(16))); +__m128i psi_a_m7_p7 __attribute__ ((aligned(16))); +__m128i psi_a_m5_m7 __attribute__ ((aligned(16))); +__m128i psi_a_m5_m5 __attribute__ ((aligned(16))); +__m128i psi_a_m5_m3 __attribute__ ((aligned(16))); +__m128i psi_a_m5_m1 __attribute__ ((aligned(16))); +__m128i psi_a_m5_p1 __attribute__ ((aligned(16))); +__m128i psi_a_m5_p3 __attribute__ ((aligned(16))); +__m128i psi_a_m5_p5 __attribute__ ((aligned(16))); +__m128i psi_a_m5_p7 __attribute__ ((aligned(16))); +__m128i psi_a_m3_m7 __attribute__ ((aligned(16))); +__m128i psi_a_m3_m5 __attribute__ ((aligned(16))); +__m128i psi_a_m3_m3 __attribute__ ((aligned(16))); +__m128i psi_a_m3_m1 __attribute__ ((aligned(16))); +__m128i psi_a_m3_p1 __attribute__ ((aligned(16))); +__m128i psi_a_m3_p3 __attribute__ ((aligned(16))); +__m128i psi_a_m3_p5 __attribute__ ((aligned(16))); +__m128i psi_a_m3_p7 __attribute__ ((aligned(16))); +__m128i psi_a_m1_m7 __attribute__ ((aligned(16))); +__m128i psi_a_m1_m5 __attribute__ ((aligned(16))); +__m128i psi_a_m1_m3 __attribute__ ((aligned(16))); +__m128i psi_a_m1_m1 __attribute__ ((aligned(16))); +__m128i psi_a_m1_p1 __attribute__ ((aligned(16))); +__m128i psi_a_m1_p3 __attribute__ ((aligned(16))); +__m128i psi_a_m1_p5 __attribute__ ((aligned(16))); +__m128i psi_a_m1_p7 __attribute__ ((aligned(16))); +__m128i psi_a_p1_m7 __attribute__ ((aligned(16))); +__m128i psi_a_p1_m5 __attribute__ ((aligned(16))); +__m128i psi_a_p1_m3 __attribute__ ((aligned(16))); +__m128i psi_a_p1_m1 __attribute__ ((aligned(16))); +__m128i psi_a_p1_p1 __attribute__ ((aligned(16))); +__m128i psi_a_p1_p3 __attribute__ ((aligned(16))); +__m128i psi_a_p1_p5 __attribute__ ((aligned(16))); +__m128i psi_a_p1_p7 __attribute__ ((aligned(16))); +__m128i psi_a_p3_m7 __attribute__ ((aligned(16))); +__m128i psi_a_p3_m5 __attribute__ ((aligned(16))); +__m128i psi_a_p3_m3 __attribute__ ((aligned(16))); +__m128i psi_a_p3_m1 __attribute__ ((aligned(16))); +__m128i psi_a_p3_p1 __attribute__ ((aligned(16))); +__m128i psi_a_p3_p3 __attribute__ ((aligned(16))); +__m128i psi_a_p3_p5 __attribute__ ((aligned(16))); +__m128i psi_a_p3_p7 __attribute__ ((aligned(16))); +__m128i psi_a_p5_m7 __attribute__ ((aligned(16))); +__m128i psi_a_p5_m5 __attribute__ ((aligned(16))); +__m128i psi_a_p5_m3 __attribute__ ((aligned(16))); +__m128i psi_a_p5_m1 __attribute__ ((aligned(16))); +__m128i psi_a_p5_p1 __attribute__ ((aligned(16))); +__m128i psi_a_p5_p3 __attribute__ ((aligned(16))); +__m128i psi_a_p5_p5 __attribute__ ((aligned(16))); +__m128i psi_a_p5_p7 __attribute__ ((aligned(16))); +__m128i psi_a_p7_m7 __attribute__ ((aligned(16))); +__m128i psi_a_p7_m5 __attribute__ ((aligned(16))); +__m128i psi_a_p7_m3 __attribute__ ((aligned(16))); +__m128i psi_a_p7_m1 __attribute__ ((aligned(16))); +__m128i psi_a_p7_p1 __attribute__ ((aligned(16))); +__m128i psi_a_p7_p3 __attribute__ ((aligned(16))); +__m128i psi_a_p7_p5 __attribute__ ((aligned(16))); +__m128i psi_a_p7_p7 __attribute__ ((aligned(16))); + +__m128i a_sq_m7_m7 __attribute__ ((aligned(16))); +__m128i a_sq_m7_m5 __attribute__ ((aligned(16))); +__m128i a_sq_m7_m3 __attribute__ ((aligned(16))); +__m128i a_sq_m7_m1 __attribute__ ((aligned(16))); +__m128i a_sq_m7_p1 __attribute__ ((aligned(16))); +__m128i a_sq_m7_p3 __attribute__ ((aligned(16))); +__m128i a_sq_m7_p5 __attribute__ ((aligned(16))); +__m128i a_sq_m7_p7 __attribute__ ((aligned(16))); +__m128i a_sq_m5_m7 __attribute__ ((aligned(16))); +__m128i a_sq_m5_m5 __attribute__ ((aligned(16))); +__m128i a_sq_m5_m3 __attribute__ ((aligned(16))); +__m128i a_sq_m5_m1 __attribute__ ((aligned(16))); +__m128i a_sq_m5_p1 __attribute__ ((aligned(16))); +__m128i a_sq_m5_p3 __attribute__ ((aligned(16))); +__m128i a_sq_m5_p5 __attribute__ ((aligned(16))); +__m128i a_sq_m5_p7 __attribute__ ((aligned(16))); +__m128i a_sq_m3_m7 __attribute__ ((aligned(16))); +__m128i a_sq_m3_m5 __attribute__ ((aligned(16))); +__m128i a_sq_m3_m3 __attribute__ ((aligned(16))); +__m128i a_sq_m3_m1 __attribute__ ((aligned(16))); +__m128i a_sq_m3_p1 __attribute__ ((aligned(16))); +__m128i a_sq_m3_p3 __attribute__ ((aligned(16))); +__m128i a_sq_m3_p5 __attribute__ ((aligned(16))); +__m128i a_sq_m3_p7 __attribute__ ((aligned(16))); +__m128i a_sq_m1_m7 __attribute__ ((aligned(16))); +__m128i a_sq_m1_m5 __attribute__ ((aligned(16))); +__m128i a_sq_m1_m3 __attribute__ ((aligned(16))); +__m128i a_sq_m1_m1 __attribute__ ((aligned(16))); +__m128i a_sq_m1_p1 __attribute__ ((aligned(16))); +__m128i a_sq_m1_p3 __attribute__ ((aligned(16))); +__m128i a_sq_m1_p5 __attribute__ ((aligned(16))); +__m128i a_sq_m1_p7 __attribute__ ((aligned(16))); +__m128i a_sq_p1_m7 __attribute__ ((aligned(16))); +__m128i a_sq_p1_m5 __attribute__ ((aligned(16))); +__m128i a_sq_p1_m3 __attribute__ ((aligned(16))); +__m128i a_sq_p1_m1 __attribute__ ((aligned(16))); +__m128i a_sq_p1_p1 __attribute__ ((aligned(16))); +__m128i a_sq_p1_p3 __attribute__ ((aligned(16))); +__m128i a_sq_p1_p5 __attribute__ ((aligned(16))); +__m128i a_sq_p1_p7 __attribute__ ((aligned(16))); +__m128i a_sq_p3_m7 __attribute__ ((aligned(16))); +__m128i a_sq_p3_m5 __attribute__ ((aligned(16))); +__m128i a_sq_p3_m3 __attribute__ ((aligned(16))); +__m128i a_sq_p3_m1 __attribute__ ((aligned(16))); +__m128i a_sq_p3_p1 __attribute__ ((aligned(16))); +__m128i a_sq_p3_p3 __attribute__ ((aligned(16))); +__m128i a_sq_p3_p5 __attribute__ ((aligned(16))); +__m128i a_sq_p3_p7 __attribute__ ((aligned(16))); +__m128i a_sq_p5_m7 __attribute__ ((aligned(16))); +__m128i a_sq_p5_m5 __attribute__ ((aligned(16))); +__m128i a_sq_p5_m3 __attribute__ ((aligned(16))); +__m128i a_sq_p5_m1 __attribute__ ((aligned(16))); +__m128i a_sq_p5_p1 __attribute__ ((aligned(16))); +__m128i a_sq_p5_p3 __attribute__ ((aligned(16))); +__m128i a_sq_p5_p5 __attribute__ ((aligned(16))); +__m128i a_sq_p5_p7 __attribute__ ((aligned(16))); +__m128i a_sq_p7_m7 __attribute__ ((aligned(16))); +__m128i a_sq_p7_m5 __attribute__ ((aligned(16))); +__m128i a_sq_p7_m3 __attribute__ ((aligned(16))); +__m128i a_sq_p7_m1 __attribute__ ((aligned(16))); +__m128i a_sq_p7_p1 __attribute__ ((aligned(16))); +__m128i a_sq_p7_p3 __attribute__ ((aligned(16))); +__m128i a_sq_p7_p5 __attribute__ ((aligned(16))); +__m128i a_sq_p7_p7 __attribute__ ((aligned(16))); + +__m128i bit_met_m7_m7 __attribute__ ((aligned(16))); +__m128i bit_met_m7_m5 __attribute__ ((aligned(16))); +__m128i bit_met_m7_m3 __attribute__ ((aligned(16))); +__m128i bit_met_m7_m1 __attribute__ ((aligned(16))); +__m128i bit_met_m7_p1 __attribute__ ((aligned(16))); +__m128i bit_met_m7_p3 __attribute__ ((aligned(16))); +__m128i bit_met_m7_p5 __attribute__ ((aligned(16))); +__m128i bit_met_m7_p7 __attribute__ ((aligned(16))); +__m128i bit_met_m5_m7 __attribute__ ((aligned(16))); +__m128i bit_met_m5_m5 __attribute__ ((aligned(16))); +__m128i bit_met_m5_m3 __attribute__ ((aligned(16))); +__m128i bit_met_m5_m1 __attribute__ ((aligned(16))); +__m128i bit_met_m5_p1 __attribute__ ((aligned(16))); +__m128i bit_met_m5_p3 __attribute__ ((aligned(16))); +__m128i bit_met_m5_p5 __attribute__ ((aligned(16))); +__m128i bit_met_m5_p7 __attribute__ ((aligned(16))); +__m128i bit_met_m3_m7 __attribute__ ((aligned(16))); +__m128i bit_met_m3_m5 __attribute__ ((aligned(16))); +__m128i bit_met_m3_m3 __attribute__ ((aligned(16))); +__m128i bit_met_m3_m1 __attribute__ ((aligned(16))); +__m128i bit_met_m3_p1 __attribute__ ((aligned(16))); +__m128i bit_met_m3_p3 __attribute__ ((aligned(16))); +__m128i bit_met_m3_p5 __attribute__ ((aligned(16))); +__m128i bit_met_m3_p7 __attribute__ ((aligned(16))); +__m128i bit_met_m1_m7 __attribute__ ((aligned(16))); +__m128i bit_met_m1_m5 __attribute__ ((aligned(16))); +__m128i bit_met_m1_m3 __attribute__ ((aligned(16))); +__m128i bit_met_m1_m1 __attribute__ ((aligned(16))); +__m128i bit_met_m1_p1 __attribute__ ((aligned(16))); +__m128i bit_met_m1_p3 __attribute__ ((aligned(16))); +__m128i bit_met_m1_p5 __attribute__ ((aligned(16))); +__m128i bit_met_m1_p7 __attribute__ ((aligned(16))); +__m128i bit_met_p1_m7 __attribute__ ((aligned(16))); +__m128i bit_met_p1_m5 __attribute__ ((aligned(16))); +__m128i bit_met_p1_m3 __attribute__ ((aligned(16))); +__m128i bit_met_p1_m1 __attribute__ ((aligned(16))); +__m128i bit_met_p1_p1 __attribute__ ((aligned(16))); +__m128i bit_met_p1_p3 __attribute__ ((aligned(16))); +__m128i bit_met_p1_p5 __attribute__ ((aligned(16))); +__m128i bit_met_p1_p7 __attribute__ ((aligned(16))); +__m128i bit_met_p3_m7 __attribute__ ((aligned(16))); +__m128i bit_met_p3_m5 __attribute__ ((aligned(16))); +__m128i bit_met_p3_m3 __attribute__ ((aligned(16))); +__m128i bit_met_p3_m1 __attribute__ ((aligned(16))); +__m128i bit_met_p3_p1 __attribute__ ((aligned(16))); +__m128i bit_met_p3_p3 __attribute__ ((aligned(16))); +__m128i bit_met_p3_p5 __attribute__ ((aligned(16))); +__m128i bit_met_p3_p7 __attribute__ ((aligned(16))); +__m128i bit_met_p5_m7 __attribute__ ((aligned(16))); +__m128i bit_met_p5_m5 __attribute__ ((aligned(16))); +__m128i bit_met_p5_m3 __attribute__ ((aligned(16))); +__m128i bit_met_p5_m1 __attribute__ ((aligned(16))); +__m128i bit_met_p5_p1 __attribute__ ((aligned(16))); +__m128i bit_met_p5_p3 __attribute__ ((aligned(16))); +__m128i bit_met_p5_p5 __attribute__ ((aligned(16))); +__m128i bit_met_p5_p7 __attribute__ ((aligned(16))); +__m128i bit_met_p7_m7 __attribute__ ((aligned(16))); +__m128i bit_met_p7_m5 __attribute__ ((aligned(16))); +__m128i bit_met_p7_m3 __attribute__ ((aligned(16))); +__m128i bit_met_p7_m1 __attribute__ ((aligned(16))); +__m128i bit_met_p7_p1 __attribute__ ((aligned(16))); +__m128i bit_met_p7_p3 __attribute__ ((aligned(16))); +__m128i bit_met_p7_p5 __attribute__ ((aligned(16))); +__m128i bit_met_p7_p7 __attribute__ ((aligned(16))); + +__m128i y0_p_1_1 __attribute__ ((aligned(16))); +__m128i y0_p_1_3 __attribute__ ((aligned(16))); +__m128i y0_p_1_5 __attribute__ ((aligned(16))); +__m128i y0_p_1_7 __attribute__ ((aligned(16))); +__m128i y0_p_3_1 __attribute__ ((aligned(16))); +__m128i y0_p_3_3 __attribute__ ((aligned(16))); +__m128i y0_p_3_5 __attribute__ ((aligned(16))); +__m128i y0_p_3_7 __attribute__ ((aligned(16))); +__m128i y0_p_5_1 __attribute__ ((aligned(16))); +__m128i y0_p_5_3 __attribute__ ((aligned(16))); +__m128i y0_p_5_5 __attribute__ ((aligned(16))); +__m128i y0_p_5_7 __attribute__ ((aligned(16))); +__m128i y0_p_7_1 __attribute__ ((aligned(16))); +__m128i y0_p_7_3 __attribute__ ((aligned(16))); +__m128i y0_p_7_5 __attribute__ ((aligned(16))); +__m128i y0_p_7_7 __attribute__ ((aligned(16))); +__m128i y0_m_1_1 __attribute__ ((aligned(16))); +__m128i y0_m_1_3 __attribute__ ((aligned(16))); +__m128i y0_m_1_5 __attribute__ ((aligned(16))); +__m128i y0_m_1_7 __attribute__ ((aligned(16))); +__m128i y0_m_3_1 __attribute__ ((aligned(16))); +__m128i y0_m_3_3 __attribute__ ((aligned(16))); +__m128i y0_m_3_5 __attribute__ ((aligned(16))); +__m128i y0_m_3_7 __attribute__ ((aligned(16))); +__m128i y0_m_5_1 __attribute__ ((aligned(16))); +__m128i y0_m_5_3 __attribute__ ((aligned(16))); +__m128i y0_m_5_5 __attribute__ ((aligned(16))); +__m128i y0_m_5_7 __attribute__ ((aligned(16))); +__m128i y0_m_7_1 __attribute__ ((aligned(16))); +__m128i y0_m_7_3 __attribute__ ((aligned(16))); +__m128i y0_m_7_5 __attribute__ ((aligned(16))); +__m128i y0_m_7_7 __attribute__ ((aligned(16))); + +__m128i xmm0 __attribute__ ((aligned(16))); +__m128i xmm1 __attribute__ ((aligned(16))); +__m128i xmm2 __attribute__ ((aligned(16))); +__m128i xmm3 __attribute__ ((aligned(16))); +__m128i xmm4 __attribute__ ((aligned(16))); +__m128i xmm5 __attribute__ ((aligned(16))); +__m128i xmm6 __attribute__ ((aligned(16))); +__m128i xmm7 __attribute__ ((aligned(16))); +__m128i xmm8 __attribute__ ((aligned(16))); + +__m128i y0r __attribute__ ((aligned(16))); +__m128i y0i __attribute__ ((aligned(16))); +__m128i y1r __attribute__ ((aligned(16))); +__m128i y1i __attribute__ ((aligned(16))); +__m128i y2r __attribute__ ((aligned(16))); +__m128i y2i __attribute__ ((aligned(16))); + +__m128i logmax_num_re0 __attribute__ ((aligned(16))); +__m128i logmax_num_im0 __attribute__ ((aligned(16))); +__m128i logmax_den_re0 __attribute__ ((aligned(16))); +__m128i logmax_den_im0 __attribute__ ((aligned(16))); +__m128i logmax_num_re1 __attribute__ ((aligned(16))); +__m128i logmax_num_im1 __attribute__ ((aligned(16))); +__m128i logmax_den_re1 __attribute__ ((aligned(16))); +__m128i logmax_den_im1 __attribute__ ((aligned(16))); + +__m128i tmp_result __attribute__ ((aligned(16))); +__m128i tmp_result2 __attribute__ ((aligned(16))); +__m128i tmp_result3 __attribute__ ((aligned(16))); +__m128i tmp_result4 __attribute__ ((aligned(16))); + + +//============================================================================================== +// Auxiliary Makros + +// calculates psi_a = psi_r*a_r + psi_i*a_i +#define prodsum_psi_a_epi16(psi_r,a_r,psi_i,a_i,psi_a) tmp_result = _mm_mulhi_epi16(psi_r,a_r); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result2 = _mm_mulhi_epi16(psi_i,a_i); tmp_result2 = _mm_slli_epi16(tmp_result2,1); psi_a = _mm_adds_epi16(tmp_result,tmp_result2); + +// calculate interference magnitude +#define interference_abs_epi16(psi,int_ch_mag,int_mag,c1,c2) tmp_result = _mm_cmplt_epi16(psi,int_ch_mag); tmp_result2 = _mm_xor_si128(tmp_result,(*(__m128i*)&ones[0])); tmp_result = _mm_and_si128(tmp_result,c1); tmp_result2 = _mm_and_si128(tmp_result2,c2); int_mag = _mm_or_si128(tmp_result,tmp_result2); + +// calculate interference magnitude +// tmp_result = ones in shorts corr. to interval 2<=x<=4, tmp_result2 interval < 2, tmp_result3 interval 46 +#define interference_abs_64qam_epi16(psi,int_ch_mag,int_two_ch_mag,int_three_ch_mag,a,c1,c3,c5,c7) tmp_result = _mm_cmplt_epi16(psi,int_two_ch_mag); tmp_result3 = _mm_xor_si128(tmp_result,(*(__m128i*)&ones[0])); tmp_result2 = _mm_cmplt_epi16(psi,int_ch_mag); tmp_result = _mm_xor_si128(tmp_result,tmp_result2); tmp_result4 = _mm_cmpgt_epi16(psi,int_three_ch_mag); tmp_result3 = _mm_xor_si128(tmp_result3,tmp_result4); tmp_result = _mm_and_si128(tmp_result,c3); tmp_result2 = _mm_and_si128(tmp_result2,c1); tmp_result3 = _mm_and_si128(tmp_result3,c5); tmp_result4 = _mm_and_si128(tmp_result4,c7); tmp_result = _mm_or_si128(tmp_result,tmp_result2); tmp_result3 = _mm_or_si128(tmp_result3,tmp_result4); a = _mm_or_si128(tmp_result,tmp_result3); + +// calculates a_sq = int_ch_mag*(a_r^2 + a_i^2)*scale_factor +#define square_a_epi16(a_r,a_i,int_ch_mag,scale_factor,a_sq) tmp_result = _mm_mulhi_epi16(a_r,a_r); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result = _mm_mulhi_epi16(tmp_result,scale_factor); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result = _mm_mulhi_epi16(tmp_result,int_ch_mag); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result2 = _mm_mulhi_epi16(a_i,a_i); tmp_result2 = _mm_slli_epi16(tmp_result2,1); tmp_result2 = _mm_mulhi_epi16(tmp_result2,scale_factor); tmp_result2 = _mm_slli_epi16(tmp_result2,1); tmp_result2 = _mm_mulhi_epi16(tmp_result2,int_ch_mag); tmp_result2 = _mm_slli_epi16(tmp_result2,1); a_sq = _mm_adds_epi16(tmp_result,tmp_result2); + +// calculates a_sq = int_ch_mag*(a_r^2 + a_i^2)*scale_factor for 64-QAM +#define square_a_64qam_epi16(a_r,a_i,int_ch_mag,scale_factor,a_sq) tmp_result = _mm_mulhi_epi16(a_r,a_r); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result = _mm_mulhi_epi16(tmp_result,scale_factor); tmp_result = _mm_slli_epi16(tmp_result,3); tmp_result = _mm_mulhi_epi16(tmp_result,int_ch_mag); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result2 = _mm_mulhi_epi16(a_i,a_i); tmp_result2 = _mm_slli_epi16(tmp_result2,1); tmp_result2 = _mm_mulhi_epi16(tmp_result2,scale_factor); tmp_result2 = _mm_slli_epi16(tmp_result2,3); tmp_result2 = _mm_mulhi_epi16(tmp_result2,int_ch_mag); tmp_result2 = _mm_slli_epi16(tmp_result2,1); a_sq = _mm_adds_epi16(tmp_result,tmp_result2); + +#elif defined(__arm__) + +#endif + +//============================================================================================== +// SINGLE-STREAM +//============================================================================================== + +//---------------------------------------------------------------------------------------------- +// QPSK +//---------------------------------------------------------------------------------------------- + +int nr_dlsch_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint8_t beamforming_mode) +{ + + uint32_t *rxF = (uint32_t*)&rxdataF_comp[0][((int32_t)symbol*nb_rb*12)]; + uint32_t *llr32; + int i,len; + + llr32 = (uint32_t*)dlsch_llr; + if (!llr32) { + LOG_E(PHY,"nr_dlsch_qpsk_llr: llr is null, symbol %d, llr32=%p\n",symbol, llr32); + return(-1); + } + + if (symbol ==2) //to update from config + len = nb_rb*6; + else + len = nb_rb*12; + /* + LOG_I(PHY,"dlsch_qpsk_llr: [symb %d / Length %d]: @LLR Buff %x, @LLR Buff(symb) %x \n", + symbol, + len, + dlsch_llr, + llr32); + */ + for (i=0; i>=2; // length in quad words (4 REs) + // printf("len>>=2=%d\n", len); + len+=(len_mod4==0 ? 0 : 1); + // printf("len+=%d\n", len); + for (i=0; iN_RB_DL*12)]; + ch_magb = (__m128i*)&dl_ch_magb[0][(symbol*frame_parms->N_RB_DL*12)]; +#elif defined(__arm__) + ch_mag = (int16x8_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; + ch_magb = (int16x8_t*)&dl_ch_magb[0][(symbol*frame_parms->N_RB_DL*12)]; +#endif + + if (symbol ==2) //to update from config + len = nb_rb*6; + else + len = nb_rb*12; + +// printf("nr_dlsch_64qam_llr: symbol %d,nb_rb %d, len %d,pbch_pss_sss_adjust %d\n",symbol,nb_rb,len,pbch_pss_sss_adjust); + +/* LOG_I(PHY,"nr_dlsch_64qam_llr [symb %d / FirstSym %d / Length %d]: @LLR Buff %x \n", + symbol, + first_symbol_flag, + len, + dlsch_llr, + pllr_symbol);*/ + + llr2 = llr; + llr += (len*6); + + len_mod4 =len&3; + len2=len>>2; // length in quad words (4 REs) + len2+=((len_mod4==0)?0:1); + + for (i=0; iN_RB_DL*12]; + int16_t rho_rho_amp_x0[2*frame_parms->N_RB_DL*12]; + uint16_t amp_tmp; + uint16_t *llr32=(uint16_t*)dlsch_llr; + int i, len, nsymb, len2; + uint8_t symbol, symbol_mod; + int len_acc=0; + uint16_t *sic_data; + uint16_t pbch_pss_sss_adjust; + unsigned char len_mod4=0; + uint16_t *llr2; + __m128i *ch_mag,*ch_magb; + + nsymb = (frame_parms->Ncp==0) ? 14:12; + + for (symbol=num_pdcch_symbols; symbolN_RB_DL*12)]); + int16_t *rho_1=(int16_t*)(&rho_i[0][((int16_t)symbol*frame_parms->N_RB_DL*12)]); + ch_mag = (__m128i*)(&dl_ch_mag[0][((int16_t)symbol*frame_parms->N_RB_DL*12)]); + ch_magb = (__m128i*)(&dl_ch_magb[0][((int16_t)symbol*frame_parms->N_RB_DL*12)]); + sic_data = (uint16_t*)(&sic_buffer[0][((int16_t)len_acc)]); + + symbol_mod = (symbol>=(7-frame_parms->Ncp)) ? symbol-(7-frame_parms->Ncp) : symbol; + + pbch_pss_sss_adjust=adjust_G2(frame_parms,&rb_alloc,6,subframe,symbol); + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + amp_tmp = 0x1fff;//dlsch0->sqrt_rho_b; already taken into account + if (frame_parms->nb_antenna_ports_eNB!=1) + len = nb_rb*8 - (2*pbch_pss_sss_adjust/3); + else + len = nb_rb*10 - (5*pbch_pss_sss_adjust/6); + } else { + amp_tmp = 0x1fff; //dlsch0->sqrt_rho_a; already taken into account + len = nb_rb*12 - pbch_pss_sss_adjust; + } + + if (mod_order_0==6) + amp_tmp=amp_tmp<<1; // to compensate for >> 1 shift in modulation + + len_acc+=len; + + multadd_complex_vector_real_scalar((int16_t *)sic_data, + amp_tmp, + (int16_t *)rho_amp_x0, //this is in Q13 + 1, + len); + + mult_cpx_vector((int16_t *)rho_1, //Q15 + (int16_t *)rho_amp_x0, //Q13 + (int16_t*)rho_rho_amp_x0, + len, + 13); + + sub_cpx_vector16((int16_t *)rxF, + (int16_t *)rho_rho_amp_x0, + //(int16_t *)clean_x1, + (int16_t *)rxF, + len*2); + + llr2 = llr32; + llr32 += (len*6); + + len_mod4 =len&3; + len2=len>>2; // length in quad words (4 REs) + len2+=(len_mod4?0:1); + + + + for (i=0; iN_RB_DL*12)]; + int16_t *rxF_i=(int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho=(int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + AssertFatal(llr16!=NULL,"nr_dlsch_qpsk_qpsk_llr: llr is null, symbol %d\n",symbol); + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + // printf("nr_dlsch_qpsk_qpsk_llr: symbol %d,nb_rb %d, len %d,pbch_pss_sss_adjust %d\n",symbol,nb_rb,len,pbch_pss_sss_adjust); + // printf("qpsk_qpsk: len %d, llr16 %p\n",len,llr16); + qpsk_qpsk((short *)rxF, + (short *)rxF_i, + (short *)llr16, + (short *)rho, + len); + + llr16 += (len<<1); + *llr16p = (short *)llr16; + + return(0); +} + +//__m128i ONE_OVER_SQRT_8 __attribute__((aligned(16))); + +void qpsk_qpsk(short *stream0_in, + short *stream1_in, + short *stream0_out, + short *rho01, + int length + ) +{ + + /* + This function computes the LLRs of stream 0 (s_0) in presence of the interfering stream 1 (s_1) assuming that both symbols are QPSK. It can be used for both MU-MIMO interference-aware receiver or for SU-MIMO receivers. + + Parameters: + stream0_in = Matched filter output y0' = (h0*g0)*y0 + stream1_in = Matched filter output y1' = (h0*g1)*y0 + stream0_out = LLRs + rho01 = Correlation between the two effective channels \rho_{10} = (h1*g1)*(h0*g0) + length = number of resource elements + */ + +#if defined(__x86_64__) || defined(__i386__) + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *stream0_128i_out = (__m128i *)stream0_out; + __m128i ONE_OVER_SQRT_8 = _mm_set1_epi16(23170); //round(2^16/sqrt(8)) +#elif defined(__arm__) + int16x8_t *rho01_128i = (int16x8_t *)rho01; + int16x8_t *stream0_128i_in = (int16x8_t *)stream0_in; + int16x8_t *stream1_128i_in = (int16x8_t *)stream1_in; + int16x8_t *stream0_128i_out = (int16x8_t *)stream0_out; + int16x8_t ONE_OVER_SQRT_8 = vdupq_n_s16(23170); //round(2^16/sqrt(8)) +#endif + + int i; + + + for (i=0; i>2; i+=2) { + // in each iteration, we take 8 complex samples +#if defined(__x86_64__) || defined(__i386__) + xmm0 = rho01_128i[i]; // 4 symbols + xmm1 = rho01_128i[i+1]; + + // put (rho_r + rho_i)/2sqrt2 in rho_rpi + // put (rho_r - rho_i)/2sqrt2 in rho_rmi + + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // divide by sqrt(8), no shift needed ONE_OVER_SQRT_8 = Q1.16 + rho_rpi = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_8); + rho_rmi = _mm_mulhi_epi16(rho_rmi,ONE_OVER_SQRT_8); +#elif defined(__arm__) + + +#endif + // Compute LLR for first bit of stream 0 + + // Compute real and imaginary parts of MF output for stream 0 +#if defined(__x86_64__) || defined(__i386__) + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + y0r_over2 = _mm_srai_epi16(y0r,1); // divide by 2 + y0i_over2 = _mm_srai_epi16(y0i,1); // divide by 2 +#elif defined(__arm__) + + +#endif + // Compute real and imaginary parts of MF output for stream 1 +#if defined(__x86_64__) || defined(__i386__) + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + y1r_over2 = _mm_srai_epi16(y1r,1); // divide by 2 + y1i_over2 = _mm_srai_epi16(y1i,1); // divide by 2 + + // Compute the terms for the LLR of first bit + + xmm0 = _mm_setzero_si128(); // ZERO + + // 1 term for numerator of LLR + xmm3 = _mm_subs_epi16(y1r_over2,rho_rpi); + A = _mm_abs_epi16(xmm3); // A = |y1r/2 - rho/sqrt(8)| + xmm2 = _mm_adds_epi16(A,y0i_over2); // = |y1r/2 - rho/sqrt(8)| + y0i/2 + xmm3 = _mm_subs_epi16(y1i_over2,rho_rmi); + B = _mm_abs_epi16(xmm3); // B = |y1i/2 - rho*/sqrt(8)| + logmax_num_re0 = _mm_adds_epi16(B,xmm2); // = |y1r/2 - rho/sqrt(8)|+|y1i/2 - rho*/sqrt(8)| + y0i/2 + + // 2 term for numerator of LLR + xmm3 = _mm_subs_epi16(y1r_over2,rho_rmi); + C = _mm_abs_epi16(xmm3); // C = |y1r/2 - rho*/4| + xmm2 = _mm_subs_epi16(C,y0i_over2); // = |y1r/2 - rho*/4| - y0i/2 + xmm3 = _mm_adds_epi16(y1i_over2,rho_rpi); + D = _mm_abs_epi16(xmm3); // D = |y1i/2 + rho/4| + xmm2 = _mm_adds_epi16(xmm2,D); // |y1r/2 - rho*/4| + |y1i/2 + rho/4| - y0i/2 + logmax_num_re0 = _mm_max_epi16(logmax_num_re0,xmm2); // max, numerator done + + // 1 term for denominator of LLR + xmm3 = _mm_adds_epi16(y1r_over2,rho_rmi); + E = _mm_abs_epi16(xmm3); // E = |y1r/2 + rho*/4| + xmm2 = _mm_adds_epi16(E,y0i_over2); // = |y1r/2 + rho*/4| + y0i/2 + xmm3 = _mm_subs_epi16(y1i_over2,rho_rpi); + F = _mm_abs_epi16(xmm3); // F = |y1i/2 - rho/4| + logmax_den_re0 = _mm_adds_epi16(F,xmm2); // = |y1r/2 + rho*/4| + |y1i/2 - rho/4| + y0i/2 + + // 2 term for denominator of LLR + xmm3 = _mm_adds_epi16(y1r_over2,rho_rpi); + G = _mm_abs_epi16(xmm3); // G = |y1r/2 + rho/4| + xmm2 = _mm_subs_epi16(G,y0i_over2); // = |y1r/2 + rho/4| - y0i/2 + xmm3 = _mm_adds_epi16(y1i_over2,rho_rmi); + H = _mm_abs_epi16(xmm3); // H = |y1i/2 + rho*/4| + xmm2 = _mm_adds_epi16(xmm2,H); // = |y1r/2 + rho/4| + |y1i/2 + rho*/4| - y0i/2 + logmax_den_re0 = _mm_max_epi16(logmax_den_re0,xmm2); // max, denominator done + + // Compute the terms for the LLR of first bit + + // 1 term for nominator of LLR + xmm2 = _mm_adds_epi16(A,y0r_over2); + logmax_num_im0 = _mm_adds_epi16(B,xmm2); // = |y1r/2 - rho/4| + |y1i/2 - rho*/4| + y0r/2 + + // 2 term for nominator of LLR + xmm2 = _mm_subs_epi16(E,y0r_over2); + xmm2 = _mm_adds_epi16(xmm2,F); // = |y1r/2 + rho*/4| + |y1i/2 - rho/4| - y0r/2 + + logmax_num_im0 = _mm_max_epi16(logmax_num_im0,xmm2); // max, nominator done + + // 1 term for denominator of LLR + xmm2 = _mm_adds_epi16(C,y0r_over2); + logmax_den_im0 = _mm_adds_epi16(D,xmm2); // = |y1r/2 - rho*/4| + |y1i/2 + rho/4| - y0r/2 + + xmm2 = _mm_subs_epi16(G,y0r_over2); + xmm2 = _mm_adds_epi16(xmm2,H); // = |y1r/2 + rho/4| + |y1i/2 + rho*/4| - y0r/2 + + logmax_den_im0 = _mm_max_epi16(logmax_den_im0,xmm2); // max, denominator done + + // LLR of first bit [L1(1), L1(2), L1(3), L1(4)] + y0r = _mm_adds_epi16(y0r,logmax_num_re0); + y0r = _mm_subs_epi16(y0r,logmax_den_re0); + + // LLR of second bit [L2(1), L2(2), L2(3), L2(4)] + y0i = _mm_adds_epi16(y0i,logmax_num_im0); + y0i = _mm_subs_epi16(y0i,logmax_den_im0); + + _mm_storeu_si128(&stream0_128i_out[i],_mm_unpacklo_epi16(y0r,y0i)); // = [L1(1), L2(1), L1(2), L2(2)] + + if (i<((length>>1) - 1)) // false if only 2 REs remain + _mm_storeu_si128(&stream0_128i_out[i+1],_mm_unpackhi_epi16(y0r,y0i)); + +#elif defined(__x86_64__) + +#endif + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif +} + +int nr_dlsch_qpsk_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + int16_t **llr16p) +{ + + int16_t *rxF=(int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i=(int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho=(int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + AssertFatal(llr16!=NULL,"nr_dlsch_qpsk_qpsk_llr: llr is null, symbol %d\n",symbol); + + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + qpsk_qam16((short *)rxF, + (short *)rxF_i, + (short *)ch_mag_i, + (short *)llr16, + (short *)rho, + len); + + llr16 += (len<<1); + *llr16p = (short *)llr16; + + return(0); +} + +/* +#if defined(__x86_64__) || defined(__i386__) +__m128i ONE_OVER_SQRT_2 __attribute__((aligned(16))); +__m128i ONE_OVER_SQRT_10 __attribute__((aligned(16))); +__m128i THREE_OVER_SQRT_10 __attribute__((aligned(16))); +__m128i ONE_OVER_SQRT_10_Q15 __attribute__((aligned(16))); +__m128i SQRT_10_OVER_FOUR __attribute__((aligned(16))); +__m128i ch_mag_int; +#endif +*/ +void qpsk_qam16(int16_t *stream0_in, + int16_t *stream1_in, + int16_t *ch_mag_i, + int16_t *stream0_out, + int16_t *rho01, + int32_t length + ) +{ + /* + This function computes the LLRs of stream 0 (s_0) in presence of the interfering stream 1 (s_1) assuming that both symbols are QPSK. It can be used for both MU-MIMO interference-aware receiver or for SU-MIMO receivers. + + Parameters: + stream0_in = Matched filter output y0' = (h0*g0)*y0 + stream1_in = Matched filter output y1' = (h0*g1)*y0 + stream0_out = LLRs + rho01 = Correlation between the two effective channels \rho_{10} = (h1*g1)*(h0*g0) + length = number of resource elements + */ + +#if defined(__x86_64__) || defined(__i386__) + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *stream0_128i_out = (__m128i *)stream0_out; + __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; + __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) + __m128i ONE_OVER_SQRT_10_Q15 = _mm_set1_epi16(10362); // round(1/sqrt(10)*2^15) + __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) + __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) + __m128i ch_mag_int __attribute__((aligned(16))); +#elif defined(__arm__) + int16x8_t *rho01_128i = (int16x8_t *)rho01; + int16x8_t *stream0_128i_in = (int16x8_t *)stream0_in; + int16x8_t *stream1_128i_in = (int16x8_t *)stream1_in; + int16x8_t *stream0_128i_out = (int16x8_t *)stream0_out; + int16x8_t *ch_mag_128i_i = (int16x8_t *)ch_mag_i; + int16x8_t ONE_OVER_SQRT_2 = vdupq_n_s16(23170); // round(1/sqrt(2)*2^15) + int16x8_t ONE_OVER_SQRT_10_Q15 = vdupq_n_s16(10362); // round(1/sqrt(10)*2^15) + int16x8_t THREE_OVER_SQRT_10 = vdupq_n_s16(31086); // round(3/sqrt(10)*2^15) + int16x8_t SQRT_10_OVER_FOUR = vdupq_n_s16(25905); // round(sqrt(10)/4*2^15) + int16x8_t ch_mag_int __attribute__((aligned(16))); +#endif + +#ifdef DEBUG_LLR + print_shorts2("rho01_128i:\n",rho01_128i); +#endif + + int i; + + + for (i=0; i>2; i+=2) { + // in each iteration, we take 8 complex samples + +#if defined(__x86_64__) || defined(__i386__) + + xmm0 = rho01_128i[i]; // 4 symbols + xmm1 = rho01_128i[i+1]; + + // put (rho_r + rho_i)/2sqrt2 in rho_rpi + // put (rho_r - rho_i)/2sqrt2 in rho_rmi + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // divide by sqrt(2) + rho_rpi = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_2); + rho_rmi = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_2); + rho_rpi = _mm_slli_epi16(rho_rpi,1); + rho_rmi = _mm_slli_epi16(rho_rmi,1); + + // Compute LLR for first bit of stream 0 + + // Compute real and imaginary parts of MF output for stream 0 + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // divide by sqrt(2) + y0r_over2 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_2); + y0i_over2 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_2); + y0r_over2 = _mm_slli_epi16(y0r,1); + y0i_over2 = _mm_slli_epi16(y0i,1); + + y0_p_1_1 = _mm_adds_epi16(y0r_over2, y0i_over2); + y0_m_1_1 = _mm_subs_epi16(y0r_over2, y0i_over2); + + // Compute real and imaginary parts of MF output for stream 1 + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + xmm0 = _mm_setzero_si128(); // ZERO + + // compute psi + xmm3 = _mm_subs_epi16(y1r,rho_rpi); + psi_r_p1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_subs_epi16(y1i,rho_rmi); + psi_i_p1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_subs_epi16(y1r,rho_rmi); + psi_r_p1_m1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1i,rho_rpi); + psi_i_p1_m1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1r,rho_rmi); + psi_r_m1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_subs_epi16(y1i,rho_rpi); + psi_i_m1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1r,rho_rpi); + psi_r_m1_m1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1i,rho_rmi); + psi_i_m1_m1 = _mm_abs_epi16(xmm3); + + // Rearrange interfering channel magnitudes + xmm2 = ch_mag_128i_i[i]; + xmm3 = ch_mag_128i_i[i+1]; + + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + + ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); + + // calculate optimal interference amplitudes + interference_abs_epi16(psi_r_p1_p1 , ch_mag_int, a_r_p1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_p1 , ch_mag_int, a_i_p1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_m1 , ch_mag_int, a_r_p1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_m1 , ch_mag_int, a_i_p1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_p1 , ch_mag_int, a_r_m1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_p1 , ch_mag_int, a_i_m1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_m1 , ch_mag_int, a_r_m1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_m1 , ch_mag_int, a_i_m1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + + // prodsum + prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); + prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); + prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); + prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); + + // squares + square_a_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p1); + square_a_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m1); + square_a_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p1); + square_a_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m1); + + // Computing Metrics + xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); + bit_met_p1_p1 = _mm_adds_epi16(xmm0, y0_p_1_1); + + xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); + bit_met_p1_m1 = _mm_adds_epi16(xmm0, y0_m_1_1); + + xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); + bit_met_m1_p1 = _mm_subs_epi16(xmm0, y0_m_1_1); + + xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); + bit_met_m1_m1 = _mm_subs_epi16(xmm0, y0_p_1_1); + + // MSB + logmax_num_re0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_m1); // bit=0 + logmax_den_re0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_m1); // bit=1 + + y0r = _mm_subs_epi16(logmax_num_re0,logmax_den_re0); + + // LSB + logmax_num_im0 = _mm_max_epi16(bit_met_p1_p1,bit_met_m1_p1); // bit=0 + logmax_den_im0 = _mm_max_epi16(bit_met_p1_m1,bit_met_m1_m1); // bit=1 + + y0i = _mm_subs_epi16(logmax_num_im0,logmax_den_im0); + + stream0_128i_out[i] = _mm_unpacklo_epi16(y0r,y0i); // = [L1(1), L2(1), L1(2), L2(2)] + + if (i<((length>>1) - 1)) // false if only 2 REs remain + stream0_128i_out[i+1] = _mm_unpackhi_epi16(y0r,y0i); + +#elif defined(__arm__) + +#endif + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif +} + +int nr_dlsch_qpsk_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + int16_t **llr16p) +{ + + int16_t *rxF=(int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i=(int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho=(int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + AssertFatal(llr16!=NULL,"nr_dlsch_qpsk_qam64_llr: llr is null, symbol %d\n",symbol); + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + qpsk_qam64((short *)rxF, + (short *)rxF_i, + (short *)ch_mag_i, + (short *)llr16, + (short *)rho, + len); + + llr16 += (len<<1); + *llr16p = (short *)llr16; + + return(0); +} +/* +__m128i ONE_OVER_SQRT_2_42 __attribute__((aligned(16))); +__m128i THREE_OVER_SQRT_2_42 __attribute__((aligned(16))); +__m128i FIVE_OVER_SQRT_2_42 __attribute__((aligned(16))); +__m128i SEVEN_OVER_SQRT_2_42 __attribute__((aligned(16))); + +__m128i ch_mag_int_with_sigma2 __attribute__((aligned(16))); +__m128i two_ch_mag_int_with_sigma2 __attribute__((aligned(16))); +__m128i three_ch_mag_int_with_sigma2 __attribute__((aligned(16))); +__m128i SQRT_42_OVER_FOUR __attribute__((aligned(16))); +*/ +void qpsk_qam64(short *stream0_in, + short *stream1_in, + short *ch_mag_i, + short *stream0_out, + short *rho01, + int length + ) +{ + + /* + This function computes the LLRs of stream 0 (s_0) in presence of the interfering stream 1 (s_1) assuming that both symbols are QPSK. It can be used for both MU-MIMO interference-aware receiver or for SU-MIMO receivers. + + Parameters: + stream0_in = Matched filter output y0' = (h0*g0)*y0 + stream1_in = Matched filter output y1' = (h0*g1)*y0 + stream0_out = LLRs + rho01 = Correlation between the two effective channels \rho_{10} = (h1*g1)*(h0*g0) + length = number of resource elements + */ + +#if defined(__x86_64__) || defined(__i386__) + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *stream0_128i_out = (__m128i *)stream0_out; + __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; + __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) + __m128i ONE_OVER_SQRT_2_42 = _mm_set1_epi16(3575); // round(1/sqrt(2*42)*2^15) + __m128i THREE_OVER_SQRT_2_42 = _mm_set1_epi16(10726); // round(3/sqrt(2*42)*2^15) + __m128i FIVE_OVER_SQRT_2_42 = _mm_set1_epi16(17876); // round(5/sqrt(2*42)*2^15) + __m128i SEVEN_OVER_SQRT_2_42 = _mm_set1_epi16(25027); // round(7/sqrt(2*42)*2^15) + __m128i SQRT_42_OVER_FOUR = _mm_set1_epi16(13272); // round(sqrt(42)/4*2^13), Q3.1 + __m128i ch_mag_int; + __m128i ch_mag_int_with_sigma2; + __m128i two_ch_mag_int_with_sigma2; + __m128i three_ch_mag_int_with_sigma2; +#elif defined(__arm__) + +#endif + +#ifdef DEBUG_LLR + print_shorts2("rho01_128i:\n",rho01_128i); +#endif + + int i; + + + for (i=0; i>2; i+=2) { + // in each iteration, we take 8 complex samples + +#if defined(__x86_64__) || defined(__i386__) + + xmm0 = rho01_128i[i]; // 4 symbols + xmm1 = rho01_128i[i+1]; + + // put (rho_r + rho_i)/sqrt2 in rho_rpi + // put (rho_r - rho_i)/sqrt2 in rho_rmi + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // divide by sqrt(2) + rho_rpi = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_2); + rho_rmi = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_2); + rho_rpi = _mm_slli_epi16(rho_rpi,1); + rho_rmi = _mm_slli_epi16(rho_rmi,1); + + // Compute LLR for first bit of stream 0 + + // Compute real and imaginary parts of MF output for stream 0 + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // divide by sqrt(2) + y0r_over2 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_2); + y0i_over2 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_2); + y0r_over2 = _mm_slli_epi16(y0r,1); + y0i_over2 = _mm_slli_epi16(y0i,1); + + y0_p_1_1 = _mm_adds_epi16(y0r_over2, y0i_over2); + y0_m_1_1 = _mm_subs_epi16(y0r_over2, y0i_over2); + + // Compute real and imaginary parts of MF output for stream 1 + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + xmm0 = _mm_setzero_si128(); // ZERO + + // compute psi + xmm3 = _mm_subs_epi16(y1r,rho_rpi); + psi_r_p1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_subs_epi16(y1i,rho_rmi); + psi_i_p1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_subs_epi16(y1r,rho_rmi); + psi_r_p1_m1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1i,rho_rpi); + psi_i_p1_m1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1r,rho_rmi); + psi_r_m1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_subs_epi16(y1i,rho_rpi); + psi_i_m1_p1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1r,rho_rpi); + psi_r_m1_m1 = _mm_abs_epi16(xmm3); + xmm3 = _mm_adds_epi16(y1i,rho_rmi); + psi_i_m1_m1 = _mm_abs_epi16(xmm3); + + // Rearrange interfering channel magnitudes + xmm2 = ch_mag_128i_i[i]; + xmm3 = ch_mag_128i_i[i+1]; + + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + + ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); + ch_mag_int_with_sigma2 = _mm_srai_epi16(ch_mag_int, 1); // *2 + two_ch_mag_int_with_sigma2 = ch_mag_int; // *4 + three_ch_mag_int_with_sigma2 = _mm_adds_epi16(ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2); // *6 + + interference_abs_64qam_epi16(psi_r_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + + // prodsum + prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); + prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); + prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); + prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); + + // Multiply by sqrt(2) + psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1, ONE_OVER_SQRT_2); + psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1, 2); + psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1, ONE_OVER_SQRT_2); + psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1, 2); + psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1, ONE_OVER_SQRT_2); + psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1, 2); + psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1, ONE_OVER_SQRT_2); + psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1, 2); + + square_a_64qam_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p1); + square_a_64qam_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m1); + square_a_64qam_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p1); + square_a_64qam_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m1); + + // Computing Metrics + xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); + bit_met_p1_p1 = _mm_adds_epi16(xmm0, y0_p_1_1); + + xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); + bit_met_p1_m1 = _mm_adds_epi16(xmm0, y0_m_1_1); + + xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); + bit_met_m1_p1 = _mm_subs_epi16(xmm0, y0_m_1_1); + + xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); + bit_met_m1_m1 = _mm_subs_epi16(xmm0, y0_p_1_1); + + // MSB + logmax_num_re0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_m1); // bit=0 + logmax_den_re0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_m1); // bit=1 + + y0r = _mm_subs_epi16(logmax_num_re0,logmax_den_re0); + + // LSB + logmax_num_im0 = _mm_max_epi16(bit_met_p1_p1,bit_met_m1_p1); // bit=0 + logmax_den_im0 = _mm_max_epi16(bit_met_p1_m1,bit_met_m1_m1); // bit=1 + + y0i = _mm_subs_epi16(logmax_num_im0,logmax_den_im0); + + stream0_128i_out[i] = _mm_unpacklo_epi16(y0r,y0i); // = [L1(1), L2(1), L1(2), L2(2)] + + if (i<((length>>1) - 1)) // false if only 2 REs remain + stream0_128i_out[i+1] = _mm_unpackhi_epi16(y0r,y0i); + +#elif defined(__arm__) + +#endif + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif +} + + +//---------------------------------------------------------------------------------------------- +// 16-QAM +//---------------------------------------------------------------------------------------------- + +/* +__m128i ONE_OVER_TWO_SQRT_10 __attribute__((aligned(16))); +__m128i NINE_OVER_TWO_SQRT_10 __attribute__((aligned(16))); + +__m128i y0r_over_sqrt10 __attribute__ ((aligned(16))); +__m128i y0i_over_sqrt10 __attribute__ ((aligned(16))); +__m128i y0r_three_over_sqrt10 __attribute__ ((aligned(16))); +__m128i y0i_three_over_sqrt10 __attribute__ ((aligned(16))); + +__m128i ch_mag_des __attribute__((aligned(16))); +__m128i ch_mag_over_10 __attribute__ ((aligned(16))); +__m128i ch_mag_over_2 __attribute__ ((aligned(16))); +__m128i ch_mag_9_over_10 __attribute__ ((aligned(16))); +*/ + +void qam16_qpsk(short *stream0_in, + short *stream1_in, + short *ch_mag, + short *stream0_out, + short *rho01, + int length + ) +{ + + /* + Author: Sebastian Wagner + Date: 2012-06-04 + + Input: + stream0_in: MF filter for 1st stream, i.e., y0=h0'*y + stream!_in: MF filter for 2nd stream, i.e., y1=h1'*y + ch_mag: 2*h0/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + ch_mag_i: 2*h1/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + rho01: Channel cross correlation, i.e., h1'*h0 + + Output: + stream0_out: output LLRs for 1st stream + */ + +#if defined(__x86_64__) || defined(__i386__) + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *stream0_128i_out = (__m128i *)stream0_out; + __m128i *ch_mag_128i = (__m128i *)ch_mag; + __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) + __m128i ONE_OVER_SQRT_10 = _mm_set1_epi16(20724); // round(1/sqrt(10)*2^16) + __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) + __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) + __m128i ONE_OVER_TWO_SQRT_10 = _mm_set1_epi16(10362); // round(1/2/sqrt(10)*2^16) + __m128i NINE_OVER_TWO_SQRT_10 = _mm_set1_epi16(23315); // round(9/2/sqrt(10)*2^14) + __m128i y0r_over_sqrt10; + __m128i y0i_over_sqrt10; + __m128i y0r_three_over_sqrt10; + __m128i y0i_three_over_sqrt10; + + __m128i ch_mag_des; + __m128i ch_mag_over_10; + __m128i ch_mag_over_2; + __m128i ch_mag_9_over_10; +#elif defined(__arm__) + +#endif + + int i; + + + for (i=0; i>2; i+=2) { + // In one iteration, we deal with 8 REs + +#if defined(__x86_64__) || defined(__i386__) + // Get rho + xmm0 = rho01_128i[i]; + xmm1 = rho01_128i[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // Compute the different rhos + rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_10); + rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi,ONE_OVER_SQRT_10); + rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi,THREE_OVER_SQRT_10); + rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi,THREE_OVER_SQRT_10); + rho_rpi_3_3 = _mm_slli_epi16(rho_rpi_3_3,1); + rho_rmi_3_3 = _mm_slli_epi16(rho_rmi_3_3,1); + + xmm4 = _mm_mulhi_epi16(xmm2,ONE_OVER_SQRT_10); // Re(rho) + xmm5 = _mm_mulhi_epi16(xmm3,THREE_OVER_SQRT_10); // Im(rho) + xmm5 = _mm_slli_epi16(xmm5,1); + + rho_rpi_1_3 = _mm_adds_epi16(xmm4,xmm5); + rho_rmi_1_3 = _mm_subs_epi16(xmm4,xmm5); + + xmm6 = _mm_mulhi_epi16(xmm2,THREE_OVER_SQRT_10); // Re(rho) + xmm7 = _mm_mulhi_epi16(xmm3,ONE_OVER_SQRT_10); // Im(rho) + xmm6 = _mm_slli_epi16(xmm6,1); + + rho_rpi_3_1 = _mm_adds_epi16(xmm6,xmm7); + rho_rmi_3_1 = _mm_subs_epi16(xmm6,xmm7); + + // Rearrange interfering MF output + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + xmm0 = _mm_setzero_si128(); // ZERO + xmm2 = _mm_subs_epi16(rho_rpi_1_1,y1r); // = [Re(rho)+ Im(rho)]/sqrt(10) - y1r + psi_r_p1_p1 = _mm_abs_epi16(xmm2); // = |[Re(rho)+ Im(rho)]/sqrt(10) - y1r| + + xmm2= _mm_subs_epi16(rho_rmi_1_1,y1r); + psi_r_p1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_1,y1i); + psi_i_p1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_3,y1r); + psi_r_p1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_3,y1r); + psi_r_p1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_1,y1i); + psi_i_p1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_1,y1r); + psi_r_p3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_1,y1r); + psi_r_p3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_3,y1i); + psi_i_p3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_3,y1r); + psi_r_p3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_3,y1r); + psi_r_p3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_3,y1i); + psi_i_p3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_1,y1i); + psi_i_m1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_1,y1i); + psi_i_m1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_3,y1i); + psi_i_m3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_3,y1i); + psi_i_m3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_1,y1i); + psi_i_p1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_1,y1i); + psi_i_p1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_3,y1i); + psi_i_p3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_3,y1i); + psi_i_p3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_1,y1r); + psi_r_m1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_3,y1r); + psi_r_m1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_1,y1r); + psi_r_m3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_3,y1r); + psi_r_m3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_1_1); + psi_r_m1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_1_3); + psi_r_m1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_1_1); + psi_i_m1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_3_1); + psi_i_m1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_3_1); + psi_r_m3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_3_3); + psi_r_m3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_1_3); + psi_i_m3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_3_3); + psi_i_m3_m3 = _mm_abs_epi16(xmm2); + + // Rearrange desired MF output + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // Rearrange desired channel magnitudes + xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) + xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + + ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); // = [|h|^2(1),|h|^2(2),|h|^2(3),|h|^2(4)]*(2/sqrt(10)) + + // Scale MF output of desired signal + y0r_over_sqrt10 = _mm_mulhi_epi16(y0r,ONE_OVER_SQRT_10); + y0i_over_sqrt10 = _mm_mulhi_epi16(y0i,ONE_OVER_SQRT_10); + y0r_three_over_sqrt10 = _mm_mulhi_epi16(y0r,THREE_OVER_SQRT_10); + y0i_three_over_sqrt10 = _mm_mulhi_epi16(y0i,THREE_OVER_SQRT_10); + y0r_three_over_sqrt10 = _mm_slli_epi16(y0r_three_over_sqrt10,1); + y0i_three_over_sqrt10 = _mm_slli_epi16(y0i_three_over_sqrt10,1); + + // Compute necessary combination of required terms + y0_p_1_1 = _mm_adds_epi16(y0r_over_sqrt10,y0i_over_sqrt10); + y0_m_1_1 = _mm_subs_epi16(y0r_over_sqrt10,y0i_over_sqrt10); + + y0_p_1_3 = _mm_adds_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); + y0_m_1_3 = _mm_subs_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); + + y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); + y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); + + y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); + y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); + + // Add psi + psi_a_p1_p1 = _mm_adds_epi16(psi_r_p1_p1 ,psi_i_p1_p1); + psi_a_p1_p3 = _mm_adds_epi16(psi_r_p1_p3 ,psi_i_p1_p3); + psi_a_p3_p1 = _mm_adds_epi16(psi_r_p3_p1 ,psi_i_p3_p1); + psi_a_p3_p3 = _mm_adds_epi16(psi_r_p3_p3 ,psi_i_p3_p3); + psi_a_p1_m1 = _mm_adds_epi16(psi_r_p1_m1 ,psi_i_p1_m1); + psi_a_p1_m3 = _mm_adds_epi16(psi_r_p1_m3 ,psi_i_p1_m3); + psi_a_p3_m1 = _mm_adds_epi16(psi_r_p3_m1 ,psi_i_p3_m1); + psi_a_p3_m3 = _mm_adds_epi16(psi_r_p3_m3 ,psi_i_p3_m3); + psi_a_m1_p1 = _mm_adds_epi16(psi_r_m1_p1 ,psi_i_m1_p1); + psi_a_m1_p3 = _mm_adds_epi16(psi_r_m1_p3 ,psi_i_m1_p3); + psi_a_m3_p1 = _mm_adds_epi16(psi_r_m3_p1 ,psi_i_m3_p1); + psi_a_m3_p3 = _mm_adds_epi16(psi_r_m3_p3 ,psi_i_m3_p3); + psi_a_m1_m1 = _mm_adds_epi16(psi_r_m1_m1 ,psi_i_m1_m1); + psi_a_m1_m3 = _mm_adds_epi16(psi_r_m1_m3 ,psi_i_m1_m3); + psi_a_m3_m1 = _mm_adds_epi16(psi_r_m3_m1 ,psi_i_m3_m1); + psi_a_m3_m3 = _mm_adds_epi16(psi_r_m3_m3 ,psi_i_m3_m3); + + // scale by sqrt(2) + psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1,ONE_OVER_SQRT_2); + psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1,1); + psi_a_p1_p3 = _mm_mulhi_epi16(psi_a_p1_p3,ONE_OVER_SQRT_2); + psi_a_p1_p3 = _mm_slli_epi16(psi_a_p1_p3,1); + psi_a_p3_p1 = _mm_mulhi_epi16(psi_a_p3_p1,ONE_OVER_SQRT_2); + psi_a_p3_p1 = _mm_slli_epi16(psi_a_p3_p1,1); + psi_a_p3_p3 = _mm_mulhi_epi16(psi_a_p3_p3,ONE_OVER_SQRT_2); + psi_a_p3_p3 = _mm_slli_epi16(psi_a_p3_p3,1); + + psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1,ONE_OVER_SQRT_2); + psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1,1); + psi_a_p1_m3 = _mm_mulhi_epi16(psi_a_p1_m3,ONE_OVER_SQRT_2); + psi_a_p1_m3 = _mm_slli_epi16(psi_a_p1_m3,1); + psi_a_p3_m1 = _mm_mulhi_epi16(psi_a_p3_m1,ONE_OVER_SQRT_2); + psi_a_p3_m1 = _mm_slli_epi16(psi_a_p3_m1,1); + psi_a_p3_m3 = _mm_mulhi_epi16(psi_a_p3_m3,ONE_OVER_SQRT_2); + psi_a_p3_m3 = _mm_slli_epi16(psi_a_p3_m3,1); + + psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1,ONE_OVER_SQRT_2); + psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1,1); + psi_a_m1_p3 = _mm_mulhi_epi16(psi_a_m1_p3,ONE_OVER_SQRT_2); + psi_a_m1_p3 = _mm_slli_epi16(psi_a_m1_p3,1); + psi_a_m3_p1 = _mm_mulhi_epi16(psi_a_m3_p1,ONE_OVER_SQRT_2); + psi_a_m3_p1 = _mm_slli_epi16(psi_a_m3_p1,1); + psi_a_m3_p3 = _mm_mulhi_epi16(psi_a_m3_p3,ONE_OVER_SQRT_2); + psi_a_m3_p3 = _mm_slli_epi16(psi_a_m3_p3,1); + + psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1,ONE_OVER_SQRT_2); + psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1,1); + psi_a_m1_m3 = _mm_mulhi_epi16(psi_a_m1_m3,ONE_OVER_SQRT_2); + psi_a_m1_m3 = _mm_slli_epi16(psi_a_m1_m3,1); + psi_a_m3_m1 = _mm_mulhi_epi16(psi_a_m3_m1,ONE_OVER_SQRT_2); + psi_a_m3_m1 = _mm_slli_epi16(psi_a_m3_m1,1); + psi_a_m3_m3 = _mm_mulhi_epi16(psi_a_m3_m3,ONE_OVER_SQRT_2); + psi_a_m3_m3 = _mm_slli_epi16(psi_a_m3_m3,1); + + // Computing different multiples of channel norms + ch_mag_over_10=_mm_mulhi_epi16(ch_mag_des, ONE_OVER_TWO_SQRT_10); + ch_mag_over_2=_mm_mulhi_epi16(ch_mag_des, SQRT_10_OVER_FOUR); + ch_mag_over_2=_mm_slli_epi16(ch_mag_over_2, 1); + ch_mag_9_over_10=_mm_mulhi_epi16(ch_mag_des, NINE_OVER_TWO_SQRT_10); + ch_mag_9_over_10=_mm_slli_epi16(ch_mag_9_over_10, 2); + + // Computing Metrics + xmm1 = _mm_adds_epi16(psi_a_p1_p1, y0_p_1_1); + bit_met_p1_p1= _mm_subs_epi16(xmm1, ch_mag_over_10); + + xmm1 = _mm_adds_epi16(psi_a_p1_p3, y0_p_1_3); + bit_met_p1_p3= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_adds_epi16(psi_a_p1_m1, y0_m_1_1); + bit_met_p1_m1= _mm_subs_epi16(xmm1, ch_mag_over_10); + + xmm1 = _mm_adds_epi16(psi_a_p1_m3, y0_m_1_3); + bit_met_p1_m3= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_adds_epi16(psi_a_p3_p1, y0_p_3_1); + bit_met_p3_p1= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_adds_epi16(psi_a_p3_p3, y0_p_3_3); + bit_met_p3_p3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); + + xmm1 = _mm_adds_epi16(psi_a_p3_m1, y0_m_3_1); + bit_met_p3_m1= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_adds_epi16(psi_a_p3_m3, y0_m_3_3); + bit_met_p3_m3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); + + xmm1 = _mm_subs_epi16(psi_a_m1_p1, y0_m_1_1); + bit_met_m1_p1= _mm_subs_epi16(xmm1, ch_mag_over_10); + + xmm1 = _mm_subs_epi16(psi_a_m1_p3, y0_m_1_3); + bit_met_m1_p3= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_subs_epi16(psi_a_m1_m1, y0_p_1_1); + bit_met_m1_m1= _mm_subs_epi16(xmm1, ch_mag_over_10); + + xmm1 = _mm_subs_epi16(psi_a_m1_m3, y0_p_1_3); + bit_met_m1_m3= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_subs_epi16(psi_a_m3_p1, y0_m_3_1); + bit_met_m3_p1= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_subs_epi16(psi_a_m3_p3, y0_m_3_3); + bit_met_m3_p3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); + + xmm1 = _mm_subs_epi16(psi_a_m3_m1, y0_p_3_1); + bit_met_m3_m1= _mm_subs_epi16(xmm1, ch_mag_over_2); + + xmm1 = _mm_subs_epi16(psi_a_m3_m3, y0_p_3_3); + bit_met_m3_m3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); + + // LLR of the first bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); + xmm2 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_re0= _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); + xmm1 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4,xmm5); + + // LLR of first bit [L1(1), L1(2), L1(3), L1(4), L1(5), L1(6), L1(7), L1(8)] + y0r = _mm_subs_epi16(logmax_den_re0,logmax_num_re0); + + // LLR of the second bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); + xmm3 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_re1 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); + xmm1 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); + xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_re1 = _mm_max_epi16(xmm4,xmm5); + + // LLR of second bit [L2(1), L2(2), L2(3), L2(4)] + y1r = _mm_subs_epi16(logmax_den_re1,logmax_num_re1); + + // LLR of the third bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); + xmm1 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); + xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_im0 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); + xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_im0 = _mm_max_epi16(xmm4,xmm5); + + // LLR of third bit [L3(1), L3(2), L3(3), L3(4)] + y0i = _mm_subs_epi16(logmax_den_im0,logmax_num_im0); + + // LLR of the fourth bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); + xmm1 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); + xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_im1 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); + xmm3 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_im1 = _mm_max_epi16(xmm4,xmm5); + + // LLR of fourth bit [L4(1), L4(2), L4(3), L4(4)] + y1i = _mm_subs_epi16(logmax_den_im1,logmax_num_im1); + + // Pack LLRs in output + // [L1(1), L2(1), L1(2), L2(2), L1(3), L2(3), L1(4), L2(4)] + xmm0 = _mm_unpacklo_epi16(y0r,y1r); + // [L1(5), L2(5), L1(6), L2(6), L1(7), L2(7), L1(8), L2(8)] + xmm1 = _mm_unpackhi_epi16(y0r,y1r); + // [L3(1), L4(1), L3(2), L4(2), L3(3), L4(3), L3(4), L4(4)] + xmm2 = _mm_unpacklo_epi16(y0i,y1i); + // [L3(5), L4(5), L3(6), L4(6), L3(7), L4(7), L3(8), L4(8)] + xmm3 = _mm_unpackhi_epi16(y0i,y1i); + + stream0_128i_out[2*i+0] = _mm_unpacklo_epi32(xmm0,xmm2); // 8LLRs, 2REs + stream0_128i_out[2*i+1] = _mm_unpackhi_epi32(xmm0,xmm2); + stream0_128i_out[2*i+2] = _mm_unpacklo_epi32(xmm1,xmm3); + stream0_128i_out[2*i+3] = _mm_unpackhi_epi32(xmm1,xmm3); + +#elif defined(__arm__) + +#endif + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif + +} + +int nr_dlsch_16qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + int16_t **llr16p) +{ + + int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + // first symbol has different structure due to more pilots + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + AssertFatal(llr16!=NULL,"nr_dlsch_16qam_qpsk_llr: llr is null, symbol %d\n",symbol); + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + // printf("symbol %d: qam16_llr, len %d (llr16 %p)\n",symbol,len,llr16); + + qam16_qpsk((short *)rxF, + (short *)rxF_i, + (short *)ch_mag, + (short *)llr16, + (short *)rho, + len); + + llr16 += (len<<2); + *llr16p = (short *)llr16; + + return(0); +} + +void qam16_qam16(short *stream0_in, + short *stream1_in, + short *ch_mag, + short *ch_mag_i, + short *stream0_out, + short *rho01, + int length + ) +{ + + /* + Author: Sebastian Wagner + Date: 2012-06-04 + + Input: + stream0_in: MF filter for 1st stream, i.e., y0=h0'*y + stream!_in: MF filter for 2nd stream, i.e., y1=h1'*y + ch_mag: 2*h0/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + ch_mag_i: 2*h1/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + rho01: Channel cross correlation, i.e., h1'*h0 + + Output: + stream0_out: output LLRs for 1st stream + */ +#if defined(__x86_64__) || defined(__i386__) + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *stream0_128i_out = (__m128i *)stream0_out; + __m128i *ch_mag_128i = (__m128i *)ch_mag; + __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; + + + + __m128i ONE_OVER_SQRT_10 = _mm_set1_epi16(20724); // round(1/sqrt(10)*2^16) + __m128i ONE_OVER_SQRT_10_Q15 = _mm_set1_epi16(10362); // round(1/sqrt(10)*2^15) + __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) + __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) + __m128i ONE_OVER_TWO_SQRT_10 = _mm_set1_epi16(10362); // round(1/2/sqrt(10)*2^16) + __m128i NINE_OVER_TWO_SQRT_10 = _mm_set1_epi16(23315); // round(9/2/sqrt(10)*2^14) + __m128i ch_mag_des,ch_mag_int; + __m128i y0r_over_sqrt10; + __m128i y0i_over_sqrt10; + __m128i y0r_three_over_sqrt10; + __m128i y0i_three_over_sqrt10; + __m128i ch_mag_over_10; + __m128i ch_mag_over_2; + __m128i ch_mag_9_over_10; +#elif defined(__arm__) + +#endif + + int i; + + for (i=0; i>2; i+=2) { + // In one iteration, we deal with 8 REs + +#if defined(__x86_64__) || defined(__i386__) + // Get rho + xmm0 = rho01_128i[i]; + xmm1 = rho01_128i[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // Compute the different rhos + rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_10); + rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi,ONE_OVER_SQRT_10); + rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi,THREE_OVER_SQRT_10); + rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi,THREE_OVER_SQRT_10); + rho_rpi_3_3 = _mm_slli_epi16(rho_rpi_3_3,1); + rho_rmi_3_3 = _mm_slli_epi16(rho_rmi_3_3,1); + + xmm4 = _mm_mulhi_epi16(xmm2,ONE_OVER_SQRT_10); // Re(rho) + xmm5 = _mm_mulhi_epi16(xmm3,THREE_OVER_SQRT_10); // Im(rho) + xmm5 = _mm_slli_epi16(xmm5,1); + + rho_rpi_1_3 = _mm_adds_epi16(xmm4,xmm5); + rho_rmi_1_3 = _mm_subs_epi16(xmm4,xmm5); + + xmm6 = _mm_mulhi_epi16(xmm2,THREE_OVER_SQRT_10); // Re(rho) + xmm7 = _mm_mulhi_epi16(xmm3,ONE_OVER_SQRT_10); // Im(rho) + xmm6 = _mm_slli_epi16(xmm6,1); + + rho_rpi_3_1 = _mm_adds_epi16(xmm6,xmm7); + rho_rmi_3_1 = _mm_subs_epi16(xmm6,xmm7); + + // Rearrange interfering MF output + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + xmm0 = _mm_setzero_si128(); // ZERO + xmm2 = _mm_subs_epi16(rho_rpi_1_1,y1r); // = [Re(rho)+ Im(rho)]/sqrt(10) - y1r + psi_r_p1_p1 = _mm_abs_epi16(xmm2); // = |[Re(rho)+ Im(rho)]/sqrt(10) - y1r| + + xmm2= _mm_subs_epi16(rho_rmi_1_1,y1r); + psi_r_p1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_1,y1i); + psi_i_p1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_3,y1r); + psi_r_p1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_3,y1r); + psi_r_p1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_1,y1i); + psi_i_p1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_1,y1r); + psi_r_p3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_1,y1r); + psi_r_p3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_3,y1i); + psi_i_p3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_3,y1r); + psi_r_p3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_3,y1r); + psi_r_p3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_3,y1i); + psi_i_p3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_1,y1i); + psi_i_m1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_1,y1i); + psi_i_m1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_3,y1i); + psi_i_m3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_3,y1i); + psi_i_m3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_1,y1i); + psi_i_p1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_1,y1i); + psi_i_p1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_3,y1i); + psi_i_p3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_3,y1i); + psi_i_p3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_1,y1r); + psi_r_m1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_3,y1r); + psi_r_m1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_1,y1r); + psi_r_m3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_3,y1r); + psi_r_m3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_1_1); + psi_r_m1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_1_3); + psi_r_m1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_1_1); + psi_i_m1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_3_1); + psi_i_m1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_3_1); + psi_r_m3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_3_3); + psi_r_m3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_1_3); + psi_i_m3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_3_3); + psi_i_m3_m3 = _mm_abs_epi16(xmm2); + + // Rearrange desired MF output + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // Rearrange desired channel magnitudes + xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) + xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + + ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); // = [|h|^2(1),|h|^2(2),|h|^2(3),|h|^2(4)]*(2/sqrt(10)) + + // Rearrange interfering channel magnitudes + xmm2 = ch_mag_128i_i[i]; + xmm3 = ch_mag_128i_i[i+1]; + + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + + ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); + + // Scale MF output of desired signal + y0r_over_sqrt10 = _mm_mulhi_epi16(y0r,ONE_OVER_SQRT_10); + y0i_over_sqrt10 = _mm_mulhi_epi16(y0i,ONE_OVER_SQRT_10); + y0r_three_over_sqrt10 = _mm_mulhi_epi16(y0r,THREE_OVER_SQRT_10); + y0i_three_over_sqrt10 = _mm_mulhi_epi16(y0i,THREE_OVER_SQRT_10); + y0r_three_over_sqrt10 = _mm_slli_epi16(y0r_three_over_sqrt10,1); + y0i_three_over_sqrt10 = _mm_slli_epi16(y0i_three_over_sqrt10,1); + + // Compute necessary combination of required terms + y0_p_1_1 = _mm_adds_epi16(y0r_over_sqrt10,y0i_over_sqrt10); + y0_m_1_1 = _mm_subs_epi16(y0r_over_sqrt10,y0i_over_sqrt10); + + y0_p_1_3 = _mm_adds_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); + y0_m_1_3 = _mm_subs_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); + + y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); + y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); + + y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); + y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); + + // Compute optimal interfering symbol magnitude + interference_abs_epi16(psi_r_p1_p1 ,ch_mag_int,a_r_p1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_p1 ,ch_mag_int,a_i_p1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_p3 ,ch_mag_int,a_r_p1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_p3 ,ch_mag_int,a_i_p1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_m1 ,ch_mag_int,a_r_p1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_m1 ,ch_mag_int,a_i_p1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_m3 ,ch_mag_int,a_r_p1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_m3 ,ch_mag_int,a_i_p1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_p1 ,ch_mag_int,a_r_p3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_p1 ,ch_mag_int,a_i_p3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_p3 ,ch_mag_int,a_r_p3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_p3 ,ch_mag_int,a_i_p3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_m1 ,ch_mag_int,a_r_p3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_m1 ,ch_mag_int,a_i_p3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_m3 ,ch_mag_int,a_r_p3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_m3 ,ch_mag_int,a_i_p3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_p1 ,ch_mag_int,a_r_m1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_p1 ,ch_mag_int,a_i_m1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_p3 ,ch_mag_int,a_r_m1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_p3 ,ch_mag_int,a_i_m1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_m1 ,ch_mag_int,a_r_m1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_m1 ,ch_mag_int,a_i_m1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_m3 ,ch_mag_int,a_r_m1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_m3 ,ch_mag_int,a_i_m1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_p1 ,ch_mag_int,a_r_m3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_p1 ,ch_mag_int,a_i_m3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_p3 ,ch_mag_int,a_r_m3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_p3 ,ch_mag_int,a_i_m3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_m1 ,ch_mag_int,a_r_m3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_m1 ,ch_mag_int,a_i_m3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_m3 ,ch_mag_int,a_r_m3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_m3 ,ch_mag_int,a_i_m3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + + // Calculation of groups of two terms in the bit metric involving product of psi and interference magnitude + prodsum_psi_a_epi16(psi_r_p1_p1,a_r_p1_p1,psi_i_p1_p1,a_i_p1_p1,psi_a_p1_p1); + prodsum_psi_a_epi16(psi_r_p1_p3,a_r_p1_p3,psi_i_p1_p3,a_i_p1_p3,psi_a_p1_p3); + prodsum_psi_a_epi16(psi_r_p3_p1,a_r_p3_p1,psi_i_p3_p1,a_i_p3_p1,psi_a_p3_p1); + prodsum_psi_a_epi16(psi_r_p3_p3,a_r_p3_p3,psi_i_p3_p3,a_i_p3_p3,psi_a_p3_p3); + prodsum_psi_a_epi16(psi_r_p1_m1,a_r_p1_m1,psi_i_p1_m1,a_i_p1_m1,psi_a_p1_m1); + prodsum_psi_a_epi16(psi_r_p1_m3,a_r_p1_m3,psi_i_p1_m3,a_i_p1_m3,psi_a_p1_m3); + prodsum_psi_a_epi16(psi_r_p3_m1,a_r_p3_m1,psi_i_p3_m1,a_i_p3_m1,psi_a_p3_m1); + prodsum_psi_a_epi16(psi_r_p3_m3,a_r_p3_m3,psi_i_p3_m3,a_i_p3_m3,psi_a_p3_m3); + prodsum_psi_a_epi16(psi_r_m1_p1,a_r_m1_p1,psi_i_m1_p1,a_i_m1_p1,psi_a_m1_p1); + prodsum_psi_a_epi16(psi_r_m1_p3,a_r_m1_p3,psi_i_m1_p3,a_i_m1_p3,psi_a_m1_p3); + prodsum_psi_a_epi16(psi_r_m3_p1,a_r_m3_p1,psi_i_m3_p1,a_i_m3_p1,psi_a_m3_p1); + prodsum_psi_a_epi16(psi_r_m3_p3,a_r_m3_p3,psi_i_m3_p3,a_i_m3_p3,psi_a_m3_p3); + prodsum_psi_a_epi16(psi_r_m1_m1,a_r_m1_m1,psi_i_m1_m1,a_i_m1_m1,psi_a_m1_m1); + prodsum_psi_a_epi16(psi_r_m1_m3,a_r_m1_m3,psi_i_m1_m3,a_i_m1_m3,psi_a_m1_m3); + prodsum_psi_a_epi16(psi_r_m3_m1,a_r_m3_m1,psi_i_m3_m1,a_i_m3_m1,psi_a_m3_m1); + prodsum_psi_a_epi16(psi_r_m3_m3,a_r_m3_m3,psi_i_m3_m3,a_i_m3_m3,psi_a_m3_m3); + + + // squared interference magnitude times int. ch. power + square_a_epi16(a_r_p1_p1,a_i_p1_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_p1); + square_a_epi16(a_r_p1_p3,a_i_p1_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_p3); + square_a_epi16(a_r_p3_p1,a_i_p3_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_p1); + square_a_epi16(a_r_p3_p3,a_i_p3_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_p3); + square_a_epi16(a_r_p1_m1,a_i_p1_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_m1); + square_a_epi16(a_r_p1_m3,a_i_p1_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_m3); + square_a_epi16(a_r_p3_m1,a_i_p3_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_m1); + square_a_epi16(a_r_p3_m3,a_i_p3_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_m3); + square_a_epi16(a_r_m1_p1,a_i_m1_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_p1); + square_a_epi16(a_r_m1_p3,a_i_m1_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_p3); + square_a_epi16(a_r_m3_p1,a_i_m3_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_p1); + square_a_epi16(a_r_m3_p3,a_i_m3_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_p3); + square_a_epi16(a_r_m1_m1,a_i_m1_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_m1); + square_a_epi16(a_r_m1_m3,a_i_m1_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_m3); + square_a_epi16(a_r_m3_m1,a_i_m3_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_m1); + square_a_epi16(a_r_m3_m3,a_i_m3_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_m3); + + // Computing different multiples of channel norms + ch_mag_over_10=_mm_mulhi_epi16(ch_mag_des, ONE_OVER_TWO_SQRT_10); + ch_mag_over_2=_mm_mulhi_epi16(ch_mag_des, SQRT_10_OVER_FOUR); + ch_mag_over_2=_mm_slli_epi16(ch_mag_over_2, 1); + ch_mag_9_over_10=_mm_mulhi_epi16(ch_mag_des, NINE_OVER_TWO_SQRT_10); + ch_mag_9_over_10=_mm_slli_epi16(ch_mag_9_over_10, 2); + + // Computing Metrics + xmm0 = _mm_subs_epi16(psi_a_p1_p1,a_sq_p1_p1); + xmm1 = _mm_adds_epi16(xmm0,y0_p_1_1); + bit_met_p1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_p1_p3,a_sq_p1_p3); + xmm1 = _mm_adds_epi16(xmm0,y0_p_1_3); + bit_met_p1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p1_m1,a_sq_p1_m1); + xmm1 = _mm_adds_epi16(xmm0,y0_m_1_1); + bit_met_p1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_p1_m3,a_sq_p1_m3); + xmm1 = _mm_adds_epi16(xmm0,y0_m_1_3); + bit_met_p1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p3_p1,a_sq_p3_p1); + xmm1 = _mm_adds_epi16(xmm0,y0_p_3_1); + bit_met_p3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p3_p3,a_sq_p3_p3); + xmm1 = _mm_adds_epi16(xmm0,y0_p_3_3); + bit_met_p3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + xmm0 = _mm_subs_epi16(psi_a_p3_m1,a_sq_p3_m1); + xmm1 = _mm_adds_epi16(xmm0,y0_m_3_1); + bit_met_p3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p3_m3,a_sq_p3_m3); + xmm1 = _mm_adds_epi16(xmm0,y0_m_3_3); + bit_met_p3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m1_p1,a_sq_m1_p1); + xmm1 = _mm_subs_epi16(xmm0,y0_m_1_1); + bit_met_m1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m1_p3,a_sq_m1_p3); + xmm1 = _mm_subs_epi16(xmm0,y0_m_1_3); + bit_met_m1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m1_m1,a_sq_m1_m1); + xmm1 = _mm_subs_epi16(xmm0,y0_p_1_1); + bit_met_m1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m1_m3,a_sq_m1_m3); + xmm1 = _mm_subs_epi16(xmm0,y0_p_1_3); + bit_met_m1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m3_p1,a_sq_m3_p1); + xmm1 = _mm_subs_epi16(xmm0,y0_m_3_1); + bit_met_m3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m3_p3,a_sq_m3_p3); + xmm1 = _mm_subs_epi16(xmm0,y0_m_3_3); + bit_met_m3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m3_m1,a_sq_m3_m1); + xmm1 = _mm_subs_epi16(xmm0,y0_p_3_1); + bit_met_m3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m3_m3,a_sq_m3_m3); + xmm1 = _mm_subs_epi16(xmm0,y0_p_3_3); + bit_met_m3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + // LLR of the first bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); + xmm2 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_re0= _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); + xmm1 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4,xmm5); + + // LLR of first bit [L1(1), L1(2), L1(3), L1(4), L1(5), L1(6), L1(7), L1(8)] + y0r = _mm_subs_epi16(logmax_den_re0,logmax_num_re0); + + // LLR of the second bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); + xmm3 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_re1 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); + xmm1 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); + xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_re1 = _mm_max_epi16(xmm4,xmm5); + + // LLR of second bit [L2(1), L2(2), L2(3), L2(4)] + y1r = _mm_subs_epi16(logmax_den_re1,logmax_num_re1); + + // LLR of the third bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); + xmm1 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); + xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_im0 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); + xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_im0 = _mm_max_epi16(xmm4,xmm5); + + // LLR of third bit [L3(1), L3(2), L3(3), L3(4)] + y0i = _mm_subs_epi16(logmax_den_im0,logmax_num_im0); + + // LLR of the fourth bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); + xmm1 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); + xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_im1 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); + xmm3 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_im1 = _mm_max_epi16(xmm4,xmm5); + + // LLR of fourth bit [L4(1), L4(2), L4(3), L4(4)] + y1i = _mm_subs_epi16(logmax_den_im1,logmax_num_im1); + + // Pack LLRs in output + // [L1(1), L2(1), L1(2), L2(2), L1(3), L2(3), L1(4), L2(4)] + xmm0 = _mm_unpacklo_epi16(y0r,y1r); + // [L1(5), L2(5), L1(6), L2(6), L1(7), L2(7), L1(8), L2(8)] + xmm1 = _mm_unpackhi_epi16(y0r,y1r); + // [L3(1), L4(1), L3(2), L4(2), L3(3), L4(3), L3(4), L4(4)] + xmm2 = _mm_unpacklo_epi16(y0i,y1i); + // [L3(5), L4(5), L3(6), L4(6), L3(7), L4(7), L3(8), L4(8)] + xmm3 = _mm_unpackhi_epi16(y0i,y1i); + + stream0_128i_out[2*i+0] = _mm_unpacklo_epi32(xmm0,xmm2); // 8LLRs, 2REs + stream0_128i_out[2*i+1] = _mm_unpackhi_epi32(xmm0,xmm2); + stream0_128i_out[2*i+2] = _mm_unpacklo_epi32(xmm1,xmm3); + stream0_128i_out[2*i+3] = _mm_unpackhi_epi32(xmm1,xmm3); +#elif defined(__arm__) + +#endif + + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif +} + +int nr_dlsch_16qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) + int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + int16_t **llr16p) +{ + + int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + // first symbol has different structure due to more pilots + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + + AssertFatal(llr16!=NULL,"nr_dlsch_16qam_16qam_llr: llr is null, symbol %d\n",symbol); + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + // printf("symbol %d: qam16_llr, len %d (llr16 %p)\n",symbol,len,llr16); + + qam16_qam16((short *)rxF, + (short *)rxF_i, + (short *)ch_mag, + (short *)ch_mag_i, + (short *)llr16, + (short *)rho, + len); + + llr16 += (len<<2); + *llr16p = (short *)llr16; + + return(0); +} + +void qam16_qam64(int16_t *stream0_in, + int16_t *stream1_in, + int16_t *ch_mag, + int16_t *ch_mag_i, + int16_t *stream0_out, + int16_t *rho01, + int32_t length + ) +{ + + /* + Author: Sebastian Wagner + Date: 2012-06-04 + + Input: + stream0_in: MF filter for 1st stream, i.e., y0=h0'*y + stream!_in: MF filter for 2nd stream, i.e., y1=h1'*y + ch_mag: 2*h0/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + ch_mag_i: 2*h1/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + rho01: Channel cross correlation, i.e., h1'*h0 + + Output: + stream0_out: output LLRs for 1st stream + */ + +#if defined(__x86_64__) || defined(__i386__) + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *stream0_128i_out = (__m128i *)stream0_out; + __m128i *ch_mag_128i = (__m128i *)ch_mag; + __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; + + + __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) + __m128i ONE_OVER_SQRT_10 = _mm_set1_epi16(20724); // round(1/sqrt(10)*2^16) + __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) + __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) + __m128i ONE_OVER_TWO_SQRT_10 = _mm_set1_epi16(10362); // round(1/2/sqrt(10)*2^16) + __m128i NINE_OVER_TWO_SQRT_10 = _mm_set1_epi16(23315); // round(9/2/sqrt(10)*2^14) + __m128i ONE_OVER_SQRT_2_42 = _mm_set1_epi16(3575); // round(1/sqrt(2*42)*2^15) + __m128i THREE_OVER_SQRT_2_42 = _mm_set1_epi16(10726); // round(3/sqrt(2*42)*2^15) + __m128i FIVE_OVER_SQRT_2_42 = _mm_set1_epi16(17876); // round(5/sqrt(2*42)*2^15) + __m128i SEVEN_OVER_SQRT_2_42 = _mm_set1_epi16(25027); // round(7/sqrt(2*42)*2^15) + __m128i SQRT_42_OVER_FOUR = _mm_set1_epi16(13272); // round(sqrt(42)/4*2^13), Q3. + __m128i ch_mag_des,ch_mag_int; + __m128i y0r_over_sqrt10; + __m128i y0i_over_sqrt10; + __m128i y0r_three_over_sqrt10; + __m128i y0i_three_over_sqrt10; + __m128i ch_mag_over_10; + __m128i ch_mag_over_2; + __m128i ch_mag_9_over_10; + __m128i ch_mag_int_with_sigma2; + __m128i two_ch_mag_int_with_sigma2; + __m128i three_ch_mag_int_with_sigma2; + +#elif defined(__arm__) + +#endif + int i; + + for (i=0; i>2; i+=2) { + // In one iteration, we deal with 8 REs + +#if defined(__x86_64__) || defined(__i386__) + // Get rho + xmm0 = rho01_128i[i]; + xmm1 = rho01_128i[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // Compute the different rhos + rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_10); + rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi,ONE_OVER_SQRT_10); + rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi,THREE_OVER_SQRT_10); + rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi,THREE_OVER_SQRT_10); + rho_rpi_3_3 = _mm_slli_epi16(rho_rpi_3_3,1); + rho_rmi_3_3 = _mm_slli_epi16(rho_rmi_3_3,1); + + xmm4 = _mm_mulhi_epi16(xmm2,ONE_OVER_SQRT_10); // Re(rho) + xmm5 = _mm_mulhi_epi16(xmm3,THREE_OVER_SQRT_10); // Im(rho) + xmm5 = _mm_slli_epi16(xmm5,1); + + rho_rpi_1_3 = _mm_adds_epi16(xmm4,xmm5); + rho_rmi_1_3 = _mm_subs_epi16(xmm4,xmm5); + + xmm6 = _mm_mulhi_epi16(xmm2,THREE_OVER_SQRT_10); // Re(rho) + xmm7 = _mm_mulhi_epi16(xmm3,ONE_OVER_SQRT_10); // Im(rho) + xmm6 = _mm_slli_epi16(xmm6,1); + + rho_rpi_3_1 = _mm_adds_epi16(xmm6,xmm7); + rho_rmi_3_1 = _mm_subs_epi16(xmm6,xmm7); + + // Rearrange interfering MF output + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + xmm0 = _mm_setzero_si128(); // ZERO + xmm2 = _mm_subs_epi16(rho_rpi_1_1,y1r); // = [Re(rho)+ Im(rho)]/sqrt(10) - y1r + psi_r_p1_p1 = _mm_abs_epi16(xmm2); // = |[Re(rho)+ Im(rho)]/sqrt(10) - y1r| + + xmm2= _mm_subs_epi16(rho_rmi_1_1,y1r); + psi_r_p1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_1,y1i); + psi_i_p1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_3,y1r); + psi_r_p1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_3,y1r); + psi_r_p1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_1,y1i); + psi_i_p1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_1,y1r); + psi_r_p3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_1,y1r); + psi_r_p3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_1_3,y1i); + psi_i_p3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_3,y1r); + psi_r_p3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_3,y1r); + psi_r_p3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rmi_3_3,y1i); + psi_i_p3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_1,y1i); + psi_i_m1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_1,y1i); + psi_i_m1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_1_3,y1i); + psi_i_m3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_subs_epi16(rho_rpi_3_3,y1i); + psi_i_m3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_1,y1i); + psi_i_p1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_1,y1i); + psi_i_p1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_3,y1i); + psi_i_p3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_3,y1i); + psi_i_p3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_1,y1r); + psi_r_m1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_1_3,y1r); + psi_r_m1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_1,y1r); + psi_r_m3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(rho_rpi_3_3,y1r); + psi_r_m3_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_1_1); + psi_r_m1_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_1_3); + psi_r_m1_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_1_1); + psi_i_m1_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_3_1); + psi_i_m1_m3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_3_1); + psi_r_m3_p1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1r,rho_rmi_3_3); + psi_r_m3_p3 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_1_3); + psi_i_m3_m1 = _mm_abs_epi16(xmm2); + xmm2= _mm_adds_epi16(y1i,rho_rmi_3_3); + psi_i_m3_m3 = _mm_abs_epi16(xmm2); + + // Rearrange desired MF output + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // Rearrange desired channel magnitudes + xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) + xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + + ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); // = [|h|^2(1),|h|^2(2),|h|^2(3),|h|^2(4)]*(2/sqrt(10)) + + // Rearrange interfering channel magnitudes + xmm2 = ch_mag_128i_i[i]; + xmm3 = ch_mag_128i_i[i+1]; + + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + + ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); + + // Scale MF output of desired signal + y0r_over_sqrt10 = _mm_mulhi_epi16(y0r,ONE_OVER_SQRT_10); + y0i_over_sqrt10 = _mm_mulhi_epi16(y0i,ONE_OVER_SQRT_10); + y0r_three_over_sqrt10 = _mm_mulhi_epi16(y0r,THREE_OVER_SQRT_10); + y0i_three_over_sqrt10 = _mm_mulhi_epi16(y0i,THREE_OVER_SQRT_10); + y0r_three_over_sqrt10 = _mm_slli_epi16(y0r_three_over_sqrt10,1); + y0i_three_over_sqrt10 = _mm_slli_epi16(y0i_three_over_sqrt10,1); + + // Compute necessary combination of required terms + y0_p_1_1 = _mm_adds_epi16(y0r_over_sqrt10,y0i_over_sqrt10); + y0_m_1_1 = _mm_subs_epi16(y0r_over_sqrt10,y0i_over_sqrt10); + + y0_p_1_3 = _mm_adds_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); + y0_m_1_3 = _mm_subs_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); + + y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); + y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); + + y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); + y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); + + // Compute optimal interfering symbol magnitude + ch_mag_int_with_sigma2 = _mm_srai_epi16(ch_mag_int, 1); // *2 + two_ch_mag_int_with_sigma2 = ch_mag_int; // *4 + three_ch_mag_int_with_sigma2 = _mm_adds_epi16(ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2); // *6 + + interference_abs_64qam_epi16(psi_r_p1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + + // Calculation of groups of two terms in the bit metric involving product of psi and interference magnitude + prodsum_psi_a_epi16(psi_r_p1_p1,a_r_p1_p1,psi_i_p1_p1,a_i_p1_p1,psi_a_p1_p1); + prodsum_psi_a_epi16(psi_r_p1_p3,a_r_p1_p3,psi_i_p1_p3,a_i_p1_p3,psi_a_p1_p3); + prodsum_psi_a_epi16(psi_r_p3_p1,a_r_p3_p1,psi_i_p3_p1,a_i_p3_p1,psi_a_p3_p1); + prodsum_psi_a_epi16(psi_r_p3_p3,a_r_p3_p3,psi_i_p3_p3,a_i_p3_p3,psi_a_p3_p3); + prodsum_psi_a_epi16(psi_r_p1_m1,a_r_p1_m1,psi_i_p1_m1,a_i_p1_m1,psi_a_p1_m1); + prodsum_psi_a_epi16(psi_r_p1_m3,a_r_p1_m3,psi_i_p1_m3,a_i_p1_m3,psi_a_p1_m3); + prodsum_psi_a_epi16(psi_r_p3_m1,a_r_p3_m1,psi_i_p3_m1,a_i_p3_m1,psi_a_p3_m1); + prodsum_psi_a_epi16(psi_r_p3_m3,a_r_p3_m3,psi_i_p3_m3,a_i_p3_m3,psi_a_p3_m3); + prodsum_psi_a_epi16(psi_r_m1_p1,a_r_m1_p1,psi_i_m1_p1,a_i_m1_p1,psi_a_m1_p1); + prodsum_psi_a_epi16(psi_r_m1_p3,a_r_m1_p3,psi_i_m1_p3,a_i_m1_p3,psi_a_m1_p3); + prodsum_psi_a_epi16(psi_r_m3_p1,a_r_m3_p1,psi_i_m3_p1,a_i_m3_p1,psi_a_m3_p1); + prodsum_psi_a_epi16(psi_r_m3_p3,a_r_m3_p3,psi_i_m3_p3,a_i_m3_p3,psi_a_m3_p3); + prodsum_psi_a_epi16(psi_r_m1_m1,a_r_m1_m1,psi_i_m1_m1,a_i_m1_m1,psi_a_m1_m1); + prodsum_psi_a_epi16(psi_r_m1_m3,a_r_m1_m3,psi_i_m1_m3,a_i_m1_m3,psi_a_m1_m3); + prodsum_psi_a_epi16(psi_r_m3_m1,a_r_m3_m1,psi_i_m3_m1,a_i_m3_m1,psi_a_m3_m1); + prodsum_psi_a_epi16(psi_r_m3_m3,a_r_m3_m3,psi_i_m3_m3,a_i_m3_m3,psi_a_m3_m3); + + // Multiply by sqrt(2) + psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1, ONE_OVER_SQRT_2); + psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1, 2); + psi_a_p1_p3 = _mm_mulhi_epi16(psi_a_p1_p3, ONE_OVER_SQRT_2); + psi_a_p1_p3 = _mm_slli_epi16(psi_a_p1_p3, 2); + psi_a_p3_p1 = _mm_mulhi_epi16(psi_a_p3_p1, ONE_OVER_SQRT_2); + psi_a_p3_p1 = _mm_slli_epi16(psi_a_p3_p1, 2); + psi_a_p3_p3 = _mm_mulhi_epi16(psi_a_p3_p3, ONE_OVER_SQRT_2); + psi_a_p3_p3 = _mm_slli_epi16(psi_a_p3_p3, 2); + psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1, ONE_OVER_SQRT_2); + psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1, 2); + psi_a_p1_m3 = _mm_mulhi_epi16(psi_a_p1_m3, ONE_OVER_SQRT_2); + psi_a_p1_m3 = _mm_slli_epi16(psi_a_p1_m3, 2); + psi_a_p3_m1 = _mm_mulhi_epi16(psi_a_p3_m1, ONE_OVER_SQRT_2); + psi_a_p3_m1 = _mm_slli_epi16(psi_a_p3_m1, 2); + psi_a_p3_m3 = _mm_mulhi_epi16(psi_a_p3_m3, ONE_OVER_SQRT_2); + psi_a_p3_m3 = _mm_slli_epi16(psi_a_p3_m3, 2); + psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1, ONE_OVER_SQRT_2); + psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1, 2); + psi_a_m1_p3 = _mm_mulhi_epi16(psi_a_m1_p3, ONE_OVER_SQRT_2); + psi_a_m1_p3 = _mm_slli_epi16(psi_a_m1_p3, 2); + psi_a_m3_p1 = _mm_mulhi_epi16(psi_a_m3_p1, ONE_OVER_SQRT_2); + psi_a_m3_p1 = _mm_slli_epi16(psi_a_m3_p1, 2); + psi_a_m3_p3 = _mm_mulhi_epi16(psi_a_m3_p3, ONE_OVER_SQRT_2); + psi_a_m3_p3 = _mm_slli_epi16(psi_a_m3_p3, 2); + psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1, ONE_OVER_SQRT_2); + psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1, 2); + psi_a_m1_m3 = _mm_mulhi_epi16(psi_a_m1_m3, ONE_OVER_SQRT_2); + psi_a_m1_m3 = _mm_slli_epi16(psi_a_m1_m3, 2); + psi_a_m3_m1 = _mm_mulhi_epi16(psi_a_m3_m1, ONE_OVER_SQRT_2); + psi_a_m3_m1 = _mm_slli_epi16(psi_a_m3_m1, 2); + psi_a_m3_m3 = _mm_mulhi_epi16(psi_a_m3_m3, ONE_OVER_SQRT_2); + psi_a_m3_m3 = _mm_slli_epi16(psi_a_m3_m3, 2); + + // squared interference magnitude times int. ch. power + square_a_64qam_epi16(a_r_p1_p1,a_i_p1_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_p1); + square_a_64qam_epi16(a_r_p1_p3,a_i_p1_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_p3); + square_a_64qam_epi16(a_r_p3_p1,a_i_p3_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_p1); + square_a_64qam_epi16(a_r_p3_p3,a_i_p3_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_p3); + square_a_64qam_epi16(a_r_p1_m1,a_i_p1_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_m1); + square_a_64qam_epi16(a_r_p1_m3,a_i_p1_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_m3); + square_a_64qam_epi16(a_r_p3_m1,a_i_p3_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_m1); + square_a_64qam_epi16(a_r_p3_m3,a_i_p3_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_m3); + square_a_64qam_epi16(a_r_m1_p1,a_i_m1_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_p1); + square_a_64qam_epi16(a_r_m1_p3,a_i_m1_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_p3); + square_a_64qam_epi16(a_r_m3_p1,a_i_m3_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_p1); + square_a_64qam_epi16(a_r_m3_p3,a_i_m3_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_p3); + square_a_64qam_epi16(a_r_m1_m1,a_i_m1_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_m1); + square_a_64qam_epi16(a_r_m1_m3,a_i_m1_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_m3); + square_a_64qam_epi16(a_r_m3_m1,a_i_m3_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_m1); + square_a_64qam_epi16(a_r_m3_m3,a_i_m3_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_m3); + + // Computing different multiples of channel norms + ch_mag_over_10=_mm_mulhi_epi16(ch_mag_des, ONE_OVER_TWO_SQRT_10); + ch_mag_over_2=_mm_mulhi_epi16(ch_mag_des, SQRT_10_OVER_FOUR); + ch_mag_over_2=_mm_slli_epi16(ch_mag_over_2, 1); + ch_mag_9_over_10=_mm_mulhi_epi16(ch_mag_des, NINE_OVER_TWO_SQRT_10); + ch_mag_9_over_10=_mm_slli_epi16(ch_mag_9_over_10, 2); + + // Computing Metrics + xmm0 = _mm_subs_epi16(psi_a_p1_p1,a_sq_p1_p1); + xmm1 = _mm_adds_epi16(xmm0,y0_p_1_1); + bit_met_p1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_p1_p3,a_sq_p1_p3); + xmm1 = _mm_adds_epi16(xmm0,y0_p_1_3); + bit_met_p1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p1_m1,a_sq_p1_m1); + xmm1 = _mm_adds_epi16(xmm0,y0_m_1_1); + bit_met_p1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_p1_m3,a_sq_p1_m3); + xmm1 = _mm_adds_epi16(xmm0,y0_m_1_3); + bit_met_p1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p3_p1,a_sq_p3_p1); + xmm1 = _mm_adds_epi16(xmm0,y0_p_3_1); + bit_met_p3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p3_p3,a_sq_p3_p3); + xmm1 = _mm_adds_epi16(xmm0,y0_p_3_3); + bit_met_p3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + xmm0 = _mm_subs_epi16(psi_a_p3_m1,a_sq_p3_m1); + xmm1 = _mm_adds_epi16(xmm0,y0_m_3_1); + bit_met_p3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_p3_m3,a_sq_p3_m3); + xmm1 = _mm_adds_epi16(xmm0,y0_m_3_3); + bit_met_p3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m1_p1,a_sq_m1_p1); + xmm1 = _mm_subs_epi16(xmm0,y0_m_1_1); + bit_met_m1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m1_p3,a_sq_m1_p3); + xmm1 = _mm_subs_epi16(xmm0,y0_m_1_3); + bit_met_m1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m1_m1,a_sq_m1_m1); + xmm1 = _mm_subs_epi16(xmm0,y0_p_1_1); + bit_met_m1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m1_m3,a_sq_m1_m3); + xmm1 = _mm_subs_epi16(xmm0,y0_p_1_3); + bit_met_m1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m3_p1,a_sq_m3_p1); + xmm1 = _mm_subs_epi16(xmm0,y0_m_3_1); + bit_met_m3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m3_p3,a_sq_m3_p3); + xmm1 = _mm_subs_epi16(xmm0,y0_m_3_3); + bit_met_m3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + xmm0 = _mm_subs_epi16(psi_a_m3_m1,a_sq_m3_m1); + xmm1 = _mm_subs_epi16(xmm0,y0_p_3_1); + bit_met_m3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); + + xmm0 = _mm_subs_epi16(psi_a_m3_m3,a_sq_m3_m3); + xmm1 = _mm_subs_epi16(xmm0,y0_p_3_3); + bit_met_m3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); + + // LLR of the first bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); + xmm2 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_re0= _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); + xmm1 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4,xmm5); + + // LLR of first bit [L1(1), L1(2), L1(3), L1(4), L1(5), L1(6), L1(7), L1(8)] + y0r = _mm_subs_epi16(logmax_den_re0,logmax_num_re0); + + // LLR of the second bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); + xmm3 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_re1 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); + xmm1 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); + xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_re1 = _mm_max_epi16(xmm4,xmm5); + + // LLR of second bit [L2(1), L2(2), L2(3), L2(4)] + y1r = _mm_subs_epi16(logmax_den_re1,logmax_num_re1); + + // LLR of the third bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); + xmm1 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); + xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_im0 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); + xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_im0 = _mm_max_epi16(xmm4,xmm5); + + // LLR of third bit [L3(1), L3(2), L3(3), L3(4)] + y0i = _mm_subs_epi16(logmax_den_im0,logmax_num_im0); + + // LLR of the fourth bit + // Bit = 1 + xmm0 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); + xmm1 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); + xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_num_im1 = _mm_max_epi16(xmm4,xmm5); + + // Bit = 0 + xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); + xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); + xmm3 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); + xmm4 = _mm_max_epi16(xmm0,xmm1); + xmm5 = _mm_max_epi16(xmm2,xmm3); + logmax_den_im1 = _mm_max_epi16(xmm4,xmm5); + + // LLR of fourth bit [L4(1), L4(2), L4(3), L4(4)] + y1i = _mm_subs_epi16(logmax_den_im1,logmax_num_im1); + + // Pack LLRs in output + // [L1(1), L2(1), L1(2), L2(2), L1(3), L2(3), L1(4), L2(4)] + xmm0 = _mm_unpacklo_epi16(y0r,y1r); + // [L1(5), L2(5), L1(6), L2(6), L1(7), L2(7), L1(8), L2(8)] + xmm1 = _mm_unpackhi_epi16(y0r,y1r); + // [L3(1), L4(1), L3(2), L4(2), L3(3), L4(3), L3(4), L4(4)] + xmm2 = _mm_unpacklo_epi16(y0i,y1i); + // [L3(5), L4(5), L3(6), L4(6), L3(7), L4(7), L3(8), L4(8)] + xmm3 = _mm_unpackhi_epi16(y0i,y1i); + + stream0_128i_out[2*i+0] = _mm_unpacklo_epi32(xmm0,xmm2); // 8LLRs, 2REs + stream0_128i_out[2*i+1] = _mm_unpackhi_epi32(xmm0,xmm2); + stream0_128i_out[2*i+2] = _mm_unpacklo_epi32(xmm1,xmm3); + stream0_128i_out[2*i+3] = _mm_unpackhi_epi32(xmm1,xmm3); +#elif defined(__arm__) + +#endif + + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif +} + +int nr_dlsch_16qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) + int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + int16_t **llr16p) +{ + + int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + // first symbol has different structure due to more pilots + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + + AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); + + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + // printf("symbol %d: qam16_llr, len %d (llr16 %p)\n",symbol,len,llr16); + + qam16_qam64((short *)rxF, + (short *)rxF_i, + (short *)ch_mag, + (short *)ch_mag_i, + (short *)llr16, + (short *)rho, + len); + + llr16 += (len<<2); + *llr16p = (short *)llr16; + + return(0); +} + +//---------------------------------------------------------------------------------------------- +// 64-QAM +//---------------------------------------------------------------------------------------------- + +/* +__m128i ONE_OVER_SQRT_42 __attribute__((aligned(16))); +__m128i THREE_OVER_SQRT_42 __attribute__((aligned(16))); +__m128i FIVE_OVER_SQRT_42 __attribute__((aligned(16))); +__m128i SEVEN_OVER_SQRT_42 __attribute__((aligned(16))); + +__m128i FORTYNINE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i TWENTYNINE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i TWENTYFIVE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i SEVENTEEN_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i NINE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i THIRTEEN_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i FIVE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); +__m128i ONE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); + +__m128i y0r_one_over_sqrt_21 __attribute__((aligned(16))); +__m128i y0r_three_over_sqrt_21 __attribute__((aligned(16))); +__m128i y0r_five_over_sqrt_21 __attribute__((aligned(16))); +__m128i y0r_seven_over_sqrt_21 __attribute__((aligned(16))); +__m128i y0i_one_over_sqrt_21 __attribute__((aligned(16))); +__m128i y0i_three_over_sqrt_21 __attribute__((aligned(16))); +__m128i y0i_five_over_sqrt_21 __attribute__((aligned(16))); +__m128i y0i_seven_over_sqrt_21 __attribute__((aligned(16))); + +__m128i ch_mag_98_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_74_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_58_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_50_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_34_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_18_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_26_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_10_over_42_with_sigma2 __attribute__((aligned(16))); +__m128i ch_mag_2_over_42_with_sigma2 __attribute__((aligned(16))); + +*/ + +void qam64_qpsk(int16_t *stream0_in, + int16_t *stream1_in, + int16_t *ch_mag, + int16_t *stream0_out, + int16_t *rho01, + int32_t length + ) +{ + + /* + Author: S. Wagner + Date: 31-07-12 + + Input: + stream0_in: MF filter for 1st stream, i.e., y0=h0'*y + stream1_in: MF filter for 2nd stream, i.e., y1=h1'*y + ch_mag: 4*h0/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + ch_mag_i: 4*h1/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + rho01: Channel cross correlation, i.e., h1'*h0 + + Output: + stream0_out: output LLRs for 1st stream + */ + +#if defined(__x86_64__) || defined(__i386__) + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *ch_mag_128i = (__m128i *)ch_mag; + + + __m128i ONE_OVER_SQRT_42 = _mm_set1_epi16(10112); // round(1/sqrt(42)*2^16) + __m128i THREE_OVER_SQRT_42 = _mm_set1_epi16(30337); // round(3/sqrt(42)*2^16) + __m128i FIVE_OVER_SQRT_42 = _mm_set1_epi16(25281); // round(5/sqrt(42)*2^15) + __m128i SEVEN_OVER_SQRT_42 = _mm_set1_epi16(17697); // round(5/sqrt(42)*2^15) + __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) + __m128i FORTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(30969); // round(49/(4*sqrt(42))*2^14), Q2.14 + __m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(23385); // round(37/(4*sqrt(42))*2^14), Q2.14 + __m128i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(31601); // round(25/(4*sqrt(42))*2^15) + __m128i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(18329); // round(29/(4*sqrt(42))*2^15), Q2.14 + __m128i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(21489); // round(17/(4*sqrt(42))*2^15) + __m128i NINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(11376); // round(9/(4*sqrt(42))*2^15) + __m128i THIRTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(16433); // round(13/(4*sqrt(42))*2^15) + __m128i FIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(6320); // round(5/(4*sqrt(42))*2^15) + __m128i ONE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(1264); // round(1/(4*sqrt(42))*2^15) + + + __m128i ch_mag_des; + __m128i ch_mag_98_over_42_with_sigma2; + __m128i ch_mag_74_over_42_with_sigma2; + __m128i ch_mag_58_over_42_with_sigma2; + __m128i ch_mag_50_over_42_with_sigma2; + __m128i ch_mag_34_over_42_with_sigma2; + __m128i ch_mag_18_over_42_with_sigma2; + __m128i ch_mag_26_over_42_with_sigma2; + __m128i ch_mag_10_over_42_with_sigma2; + __m128i ch_mag_2_over_42_with_sigma2; + __m128i y0r_one_over_sqrt_21; + __m128i y0r_three_over_sqrt_21; + __m128i y0r_five_over_sqrt_21; + __m128i y0r_seven_over_sqrt_21; + __m128i y0i_one_over_sqrt_21; + __m128i y0i_three_over_sqrt_21; + __m128i y0i_five_over_sqrt_21; + __m128i y0i_seven_over_sqrt_21; +#elif defined(__arm__) + +#endif + + int i,j; + + for (i=0; i>2; i+=2) { + +#if defined(__x86_64) || defined(__i386__) + // Get rho + xmm0 = rho01_128i[i]; + xmm1 = rho01_128i[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // Compute the different rhos + rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42); + rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42); + rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42); + rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42); + rho_rpi_5_5 = _mm_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42); + rho_rmi_5_5 = _mm_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42); + rho_rpi_7_7 = _mm_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42); + rho_rmi_7_7 = _mm_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42); + + rho_rpi_5_5 = _mm_slli_epi16(rho_rpi_5_5, 1); + rho_rmi_5_5 = _mm_slli_epi16(rho_rmi_5_5, 1); + rho_rpi_7_7 = _mm_slli_epi16(rho_rpi_7_7, 2); + rho_rmi_7_7 = _mm_slli_epi16(rho_rmi_7_7, 2); + + xmm4 = _mm_mulhi_epi16(xmm2, ONE_OVER_SQRT_42); + xmm5 = _mm_mulhi_epi16(xmm3, ONE_OVER_SQRT_42); + xmm6 = _mm_mulhi_epi16(xmm3, THREE_OVER_SQRT_42); + xmm7 = _mm_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42); + xmm8 = _mm_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42); + xmm7 = _mm_slli_epi16(xmm7, 1); + xmm8 = _mm_slli_epi16(xmm8, 2); + + rho_rpi_1_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_1_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_1_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_1_5 = _mm_subs_epi16(xmm4, xmm7); + rho_rpi_1_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_1_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, THREE_OVER_SQRT_42); + rho_rpi_3_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_3_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_3_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_3_5 = _mm_subs_epi16(xmm4, xmm7); + rho_rpi_3_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_3_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42); + xmm4 = _mm_slli_epi16(xmm4, 1); + rho_rpi_5_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_5_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_5_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_5_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_5_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_5_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42); + xmm4 = _mm_slli_epi16(xmm4, 2); + rho_rpi_7_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_7_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_7_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_7_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_7_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_7_5 = _mm_subs_epi16(xmm4, xmm7); + + // Rearrange interfering MF output + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + // Psi_r calculation from rho_rpi or rho_rmi + xmm0 = _mm_setzero_si128(); // ZERO for abs_pi16 + xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1r); + psi_r_p7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1r); + psi_r_p7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1r); + psi_r_p7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1r); + psi_r_p7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1r); + psi_r_p7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1r); + psi_r_p7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1r); + psi_r_p7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1r); + psi_r_p7_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1r); + psi_r_p5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1r); + psi_r_p5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1r); + psi_r_p5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1r); + psi_r_p5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1r); + psi_r_p5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1r); + psi_r_p5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1r); + psi_r_p5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1r); + psi_r_p5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1r); + psi_r_p3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1r); + psi_r_p3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1r); + psi_r_p3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1r); + psi_r_p3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1r); + psi_r_p3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1r); + psi_r_p3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1r); + psi_r_p3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1r); + psi_r_p3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1r); + psi_r_p1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1r); + psi_r_p1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1r); + psi_r_p1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1r); + psi_r_p1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1r); + psi_r_p1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1r); + psi_r_p1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1r); + psi_r_p1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1r); + psi_r_p1_m7 = _mm_abs_epi16(xmm2); + + xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1r); + psi_r_m1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1r); + psi_r_m1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1r); + psi_r_m1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1r); + psi_r_m1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1r); + psi_r_m1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1r); + psi_r_m1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1r); + psi_r_m1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1r); + psi_r_m1_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1r); + psi_r_m3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1r); + psi_r_m3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1r); + psi_r_m3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1r); + psi_r_m3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1r); + psi_r_m3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1r); + psi_r_m3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1r); + psi_r_m3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1r); + psi_r_m3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1r); + psi_r_m5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1r); + psi_r_m5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1r); + psi_r_m5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1r); + psi_r_m5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1r); + psi_r_m5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1r); + psi_r_m5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1r); + psi_r_m5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1r); + psi_r_m5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1r); + psi_r_m7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1r); + psi_r_m7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1r); + psi_r_m7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1r); + psi_r_m7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1r); + psi_r_m7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1r); + psi_r_m7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1r); + psi_r_m7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1r); + psi_r_m7_m7 = _mm_abs_epi16(xmm2); + + // Psi_i calculation from rho_rpi or rho_rmi + xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1i); + psi_i_p7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1i); + psi_i_p7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1i); + psi_i_p7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1i); + psi_i_p7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1i); + psi_i_p7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1i); + psi_i_p7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1i); + psi_i_p7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1i); + psi_i_p7_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1i); + psi_i_p5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1i); + psi_i_p5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1i); + psi_i_p5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1i); + psi_i_p5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1i); + psi_i_p5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1i); + psi_i_p5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1i); + psi_i_p5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1i); + psi_i_p5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1i); + psi_i_p3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1i); + psi_i_p3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1i); + psi_i_p3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1i); + psi_i_p3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1i); + psi_i_p3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1i); + psi_i_p3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1i); + psi_i_p3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1i); + psi_i_p3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1i); + psi_i_p1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1i); + psi_i_p1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1i); + psi_i_p1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1i); + psi_i_p1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1i); + psi_i_p1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1i); + psi_i_p1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1i); + psi_i_p1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1i); + psi_i_p1_m7 = _mm_abs_epi16(xmm2); + + xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1i); + psi_i_m1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1i); + psi_i_m1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1i); + psi_i_m1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1i); + psi_i_m1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1i); + psi_i_m1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1i); + psi_i_m1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1i); + psi_i_m1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1i); + psi_i_m1_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1i); + psi_i_m3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1i); + psi_i_m3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1i); + psi_i_m3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1i); + psi_i_m3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1i); + psi_i_m3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1i); + psi_i_m3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1i); + psi_i_m3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1i); + psi_i_m3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1i); + psi_i_m5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1i); + psi_i_m5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1i); + psi_i_m5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1i); + psi_i_m5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1i); + psi_i_m5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1i); + psi_i_m5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1i); + psi_i_m5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1i); + psi_i_m5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1i); + psi_i_m7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1i); + psi_i_m7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1i); + psi_i_m7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1i); + psi_i_m7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1i); + psi_i_m7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1i); + psi_i_m7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1i); + psi_i_m7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1i); + psi_i_m7_m7 = _mm_abs_epi16(xmm2); + + + // Rearrange desired MF output + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // Rearrange desired channel magnitudes + xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) + xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); + + y0r_one_over_sqrt_21 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_42); + y0r_three_over_sqrt_21 = _mm_mulhi_epi16(y0r, THREE_OVER_SQRT_42); + y0r_five_over_sqrt_21 = _mm_mulhi_epi16(y0r, FIVE_OVER_SQRT_42); + y0r_five_over_sqrt_21 = _mm_slli_epi16(y0r_five_over_sqrt_21, 1); + y0r_seven_over_sqrt_21 = _mm_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42); + y0r_seven_over_sqrt_21 = _mm_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14 + + y0i_one_over_sqrt_21 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_42); + y0i_three_over_sqrt_21 = _mm_mulhi_epi16(y0i, THREE_OVER_SQRT_42); + y0i_five_over_sqrt_21 = _mm_mulhi_epi16(y0i, FIVE_OVER_SQRT_42); + y0i_five_over_sqrt_21 = _mm_slli_epi16(y0i_five_over_sqrt_21, 1); + y0i_seven_over_sqrt_21 = _mm_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42); + y0i_seven_over_sqrt_21 = _mm_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14 + + y0_p_7_1 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_7_3 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_7_5 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_7_7 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_5_1 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_5_3 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_5_5 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_5_7 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_3_5 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_3_7 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_1_1 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_1_3 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_1_5 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_1_7 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); + + y0_m_1_1 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_1_3 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_1_5 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_1_7 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_3_5 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_3_7 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_5_1 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_5_3 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_5_5 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_5_7 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_7_1 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_7_3 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_7_5 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_7_7 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); + + // divide by sqrt(2) + psi_r_p7_p7 = _mm_mulhi_epi16(psi_r_p7_p7, ONE_OVER_SQRT_2); + psi_r_p7_p7 = _mm_slli_epi16(psi_r_p7_p7, 1); + psi_r_p7_p5 = _mm_mulhi_epi16(psi_r_p7_p5, ONE_OVER_SQRT_2); + psi_r_p7_p5 = _mm_slli_epi16(psi_r_p7_p5, 1); + psi_r_p7_p3 = _mm_mulhi_epi16(psi_r_p7_p3, ONE_OVER_SQRT_2); + psi_r_p7_p3 = _mm_slli_epi16(psi_r_p7_p3, 1); + psi_r_p7_p1 = _mm_mulhi_epi16(psi_r_p7_p1, ONE_OVER_SQRT_2); + psi_r_p7_p1 = _mm_slli_epi16(psi_r_p7_p1, 1); + psi_r_p7_m1 = _mm_mulhi_epi16(psi_r_p7_m1, ONE_OVER_SQRT_2); + psi_r_p7_m1 = _mm_slli_epi16(psi_r_p7_m1, 1); + psi_r_p7_m3 = _mm_mulhi_epi16(psi_r_p7_m3, ONE_OVER_SQRT_2); + psi_r_p7_m3 = _mm_slli_epi16(psi_r_p7_m3, 1); + psi_r_p7_m5 = _mm_mulhi_epi16(psi_r_p7_m5, ONE_OVER_SQRT_2); + psi_r_p7_m5 = _mm_slli_epi16(psi_r_p7_m5, 1); + psi_r_p7_m7 = _mm_mulhi_epi16(psi_r_p7_m7, ONE_OVER_SQRT_2); + psi_r_p7_m7 = _mm_slli_epi16(psi_r_p7_m7, 1); + psi_r_p5_p7 = _mm_mulhi_epi16(psi_r_p5_p7, ONE_OVER_SQRT_2); + psi_r_p5_p7 = _mm_slli_epi16(psi_r_p5_p7, 1); + psi_r_p5_p5 = _mm_mulhi_epi16(psi_r_p5_p5, ONE_OVER_SQRT_2); + psi_r_p5_p5 = _mm_slli_epi16(psi_r_p5_p5, 1); + psi_r_p5_p3 = _mm_mulhi_epi16(psi_r_p5_p3, ONE_OVER_SQRT_2); + psi_r_p5_p3 = _mm_slli_epi16(psi_r_p5_p3, 1); + psi_r_p5_p1 = _mm_mulhi_epi16(psi_r_p5_p1, ONE_OVER_SQRT_2); + psi_r_p5_p1 = _mm_slli_epi16(psi_r_p5_p1, 1); + psi_r_p5_m1 = _mm_mulhi_epi16(psi_r_p5_m1, ONE_OVER_SQRT_2); + psi_r_p5_m1 = _mm_slli_epi16(psi_r_p5_m1, 1); + psi_r_p5_m3 = _mm_mulhi_epi16(psi_r_p5_m3, ONE_OVER_SQRT_2); + psi_r_p5_m3 = _mm_slli_epi16(psi_r_p5_m3, 1); + psi_r_p5_m5 = _mm_mulhi_epi16(psi_r_p5_m5, ONE_OVER_SQRT_2); + psi_r_p5_m5 = _mm_slli_epi16(psi_r_p5_m5, 1); + psi_r_p5_m7 = _mm_mulhi_epi16(psi_r_p5_m7, ONE_OVER_SQRT_2); + psi_r_p5_m7 = _mm_slli_epi16(psi_r_p5_m7, 1); + psi_r_p3_p7 = _mm_mulhi_epi16(psi_r_p3_p7, ONE_OVER_SQRT_2); + psi_r_p3_p7 = _mm_slli_epi16(psi_r_p3_p7, 1); + psi_r_p3_p5 = _mm_mulhi_epi16(psi_r_p3_p5, ONE_OVER_SQRT_2); + psi_r_p3_p5 = _mm_slli_epi16(psi_r_p3_p5, 1); + psi_r_p3_p3 = _mm_mulhi_epi16(psi_r_p3_p3, ONE_OVER_SQRT_2); + psi_r_p3_p3 = _mm_slli_epi16(psi_r_p3_p3, 1); + psi_r_p3_p1 = _mm_mulhi_epi16(psi_r_p3_p1, ONE_OVER_SQRT_2); + psi_r_p3_p1 = _mm_slli_epi16(psi_r_p3_p1, 1); + psi_r_p3_m1 = _mm_mulhi_epi16(psi_r_p3_m1, ONE_OVER_SQRT_2); + psi_r_p3_m1 = _mm_slli_epi16(psi_r_p3_m1, 1); + psi_r_p3_m3 = _mm_mulhi_epi16(psi_r_p3_m3, ONE_OVER_SQRT_2); + psi_r_p3_m3 = _mm_slli_epi16(psi_r_p3_m3, 1); + psi_r_p3_m5 = _mm_mulhi_epi16(psi_r_p3_m5, ONE_OVER_SQRT_2); + psi_r_p3_m5 = _mm_slli_epi16(psi_r_p3_m5, 1); + psi_r_p3_m7 = _mm_mulhi_epi16(psi_r_p3_m7, ONE_OVER_SQRT_2); + psi_r_p3_m7 = _mm_slli_epi16(psi_r_p3_m7, 1); + psi_r_p1_p7 = _mm_mulhi_epi16(psi_r_p1_p7, ONE_OVER_SQRT_2); + psi_r_p1_p7 = _mm_slli_epi16(psi_r_p1_p7, 1); + psi_r_p1_p5 = _mm_mulhi_epi16(psi_r_p1_p5, ONE_OVER_SQRT_2); + psi_r_p1_p5 = _mm_slli_epi16(psi_r_p1_p5, 1); + psi_r_p1_p3 = _mm_mulhi_epi16(psi_r_p1_p3, ONE_OVER_SQRT_2); + psi_r_p1_p3 = _mm_slli_epi16(psi_r_p1_p3, 1); + psi_r_p1_p1 = _mm_mulhi_epi16(psi_r_p1_p1, ONE_OVER_SQRT_2); + psi_r_p1_p1 = _mm_slli_epi16(psi_r_p1_p1, 1); + psi_r_p1_m1 = _mm_mulhi_epi16(psi_r_p1_m1, ONE_OVER_SQRT_2); + psi_r_p1_m1 = _mm_slli_epi16(psi_r_p1_m1, 1); + psi_r_p1_m3 = _mm_mulhi_epi16(psi_r_p1_m3, ONE_OVER_SQRT_2); + psi_r_p1_m3 = _mm_slli_epi16(psi_r_p1_m3, 1); + psi_r_p1_m5 = _mm_mulhi_epi16(psi_r_p1_m5, ONE_OVER_SQRT_2); + psi_r_p1_m5 = _mm_slli_epi16(psi_r_p1_m5, 1); + psi_r_p1_m7 = _mm_mulhi_epi16(psi_r_p1_m7, ONE_OVER_SQRT_2); + psi_r_p1_m7 = _mm_slli_epi16(psi_r_p1_m7, 1); + psi_r_m1_p7 = _mm_mulhi_epi16(psi_r_m1_p7, ONE_OVER_SQRT_2); + psi_r_m1_p7 = _mm_slli_epi16(psi_r_m1_p7, 1); + psi_r_m1_p5 = _mm_mulhi_epi16(psi_r_m1_p5, ONE_OVER_SQRT_2); + psi_r_m1_p5 = _mm_slli_epi16(psi_r_m1_p5, 1); + psi_r_m1_p3 = _mm_mulhi_epi16(psi_r_m1_p3, ONE_OVER_SQRT_2); + psi_r_m1_p3 = _mm_slli_epi16(psi_r_m1_p3, 1); + psi_r_m1_p1 = _mm_mulhi_epi16(psi_r_m1_p1, ONE_OVER_SQRT_2); + psi_r_m1_p1 = _mm_slli_epi16(psi_r_m1_p1, 1); + psi_r_m1_m1 = _mm_mulhi_epi16(psi_r_m1_m1, ONE_OVER_SQRT_2); + psi_r_m1_m1 = _mm_slli_epi16(psi_r_m1_m1, 1); + psi_r_m1_m3 = _mm_mulhi_epi16(psi_r_m1_m3, ONE_OVER_SQRT_2); + psi_r_m1_m3 = _mm_slli_epi16(psi_r_m1_m3, 1); + psi_r_m1_m5 = _mm_mulhi_epi16(psi_r_m1_m5, ONE_OVER_SQRT_2); + psi_r_m1_m5 = _mm_slli_epi16(psi_r_m1_m5, 1); + psi_r_m1_m7 = _mm_mulhi_epi16(psi_r_m1_m7, ONE_OVER_SQRT_2); + psi_r_m1_m7 = _mm_slli_epi16(psi_r_m1_m7, 1); + psi_r_m3_p7 = _mm_mulhi_epi16(psi_r_m3_p7, ONE_OVER_SQRT_2); + psi_r_m3_p7 = _mm_slli_epi16(psi_r_m3_p7, 1); + psi_r_m3_p5 = _mm_mulhi_epi16(psi_r_m3_p5, ONE_OVER_SQRT_2); + psi_r_m3_p5 = _mm_slli_epi16(psi_r_m3_p5, 1); + psi_r_m3_p3 = _mm_mulhi_epi16(psi_r_m3_p3, ONE_OVER_SQRT_2); + psi_r_m3_p3 = _mm_slli_epi16(psi_r_m3_p3, 1); + psi_r_m3_p1 = _mm_mulhi_epi16(psi_r_m3_p1, ONE_OVER_SQRT_2); + psi_r_m3_p1 = _mm_slli_epi16(psi_r_m3_p1, 1); + psi_r_m3_m1 = _mm_mulhi_epi16(psi_r_m3_m1, ONE_OVER_SQRT_2); + psi_r_m3_m1 = _mm_slli_epi16(psi_r_m3_m1, 1); + psi_r_m3_m3 = _mm_mulhi_epi16(psi_r_m3_m3, ONE_OVER_SQRT_2); + psi_r_m3_m3 = _mm_slli_epi16(psi_r_m3_m3, 1); + psi_r_m3_m5 = _mm_mulhi_epi16(psi_r_m3_m5, ONE_OVER_SQRT_2); + psi_r_m3_m5 = _mm_slli_epi16(psi_r_m3_m5, 1); + psi_r_m3_m7 = _mm_mulhi_epi16(psi_r_m3_m7, ONE_OVER_SQRT_2); + psi_r_m3_m7 = _mm_slli_epi16(psi_r_m3_m7, 1); + psi_r_m5_p7 = _mm_mulhi_epi16(psi_r_m5_p7, ONE_OVER_SQRT_2); + psi_r_m5_p7 = _mm_slli_epi16(psi_r_m5_p7, 1); + psi_r_m5_p5 = _mm_mulhi_epi16(psi_r_m5_p5, ONE_OVER_SQRT_2); + psi_r_m5_p5 = _mm_slli_epi16(psi_r_m5_p5, 1); + psi_r_m5_p3 = _mm_mulhi_epi16(psi_r_m5_p3, ONE_OVER_SQRT_2); + psi_r_m5_p3 = _mm_slli_epi16(psi_r_m5_p3, 1); + psi_r_m5_p1 = _mm_mulhi_epi16(psi_r_m5_p1, ONE_OVER_SQRT_2); + psi_r_m5_p1 = _mm_slli_epi16(psi_r_m5_p1, 1); + psi_r_m5_m1 = _mm_mulhi_epi16(psi_r_m5_m1, ONE_OVER_SQRT_2); + psi_r_m5_m1 = _mm_slli_epi16(psi_r_m5_m1, 1); + psi_r_m5_m3 = _mm_mulhi_epi16(psi_r_m5_m3, ONE_OVER_SQRT_2); + psi_r_m5_m3 = _mm_slli_epi16(psi_r_m5_m3, 1); + psi_r_m5_m5 = _mm_mulhi_epi16(psi_r_m5_m5, ONE_OVER_SQRT_2); + psi_r_m5_m5 = _mm_slli_epi16(psi_r_m5_m5, 1); + psi_r_m5_m7 = _mm_mulhi_epi16(psi_r_m5_m7, ONE_OVER_SQRT_2); + psi_r_m5_m7 = _mm_slli_epi16(psi_r_m5_m7, 1); + psi_r_m7_p7 = _mm_mulhi_epi16(psi_r_m7_p7, ONE_OVER_SQRT_2); + psi_r_m7_p7 = _mm_slli_epi16(psi_r_m7_p7, 1); + psi_r_m7_p5 = _mm_mulhi_epi16(psi_r_m7_p5, ONE_OVER_SQRT_2); + psi_r_m7_p5 = _mm_slli_epi16(psi_r_m7_p5, 1); + psi_r_m7_p3 = _mm_mulhi_epi16(psi_r_m7_p3, ONE_OVER_SQRT_2); + psi_r_m7_p3 = _mm_slli_epi16(psi_r_m7_p3, 1); + psi_r_m7_p1 = _mm_mulhi_epi16(psi_r_m7_p1, ONE_OVER_SQRT_2); + psi_r_m7_p1 = _mm_slli_epi16(psi_r_m7_p1, 1); + psi_r_m7_m1 = _mm_mulhi_epi16(psi_r_m7_m1, ONE_OVER_SQRT_2); + psi_r_m7_m1 = _mm_slli_epi16(psi_r_m7_m1, 1); + psi_r_m7_m3 = _mm_mulhi_epi16(psi_r_m7_m3, ONE_OVER_SQRT_2); + psi_r_m7_m3 = _mm_slli_epi16(psi_r_m7_m3, 1); + psi_r_m7_m5 = _mm_mulhi_epi16(psi_r_m7_m5, ONE_OVER_SQRT_2); + psi_r_m7_m5 = _mm_slli_epi16(psi_r_m7_m5, 1); + psi_r_m7_m7 = _mm_mulhi_epi16(psi_r_m7_m7, ONE_OVER_SQRT_2); + psi_r_m7_m7 = _mm_slli_epi16(psi_r_m7_m7, 1); + + psi_i_p7_p7 = _mm_mulhi_epi16(psi_i_p7_p7, ONE_OVER_SQRT_2); + psi_i_p7_p7 = _mm_slli_epi16(psi_i_p7_p7, 1); + psi_i_p7_p5 = _mm_mulhi_epi16(psi_i_p7_p5, ONE_OVER_SQRT_2); + psi_i_p7_p5 = _mm_slli_epi16(psi_i_p7_p5, 1); + psi_i_p7_p3 = _mm_mulhi_epi16(psi_i_p7_p3, ONE_OVER_SQRT_2); + psi_i_p7_p3 = _mm_slli_epi16(psi_i_p7_p3, 1); + psi_i_p7_p1 = _mm_mulhi_epi16(psi_i_p7_p1, ONE_OVER_SQRT_2); + psi_i_p7_p1 = _mm_slli_epi16(psi_i_p7_p1, 1); + psi_i_p7_m1 = _mm_mulhi_epi16(psi_i_p7_m1, ONE_OVER_SQRT_2); + psi_i_p7_m1 = _mm_slli_epi16(psi_i_p7_m1, 1); + psi_i_p7_m3 = _mm_mulhi_epi16(psi_i_p7_m3, ONE_OVER_SQRT_2); + psi_i_p7_m3 = _mm_slli_epi16(psi_i_p7_m3, 1); + psi_i_p7_m5 = _mm_mulhi_epi16(psi_i_p7_m5, ONE_OVER_SQRT_2); + psi_i_p7_m5 = _mm_slli_epi16(psi_i_p7_m5, 1); + psi_i_p7_m7 = _mm_mulhi_epi16(psi_i_p7_m7, ONE_OVER_SQRT_2); + psi_i_p7_m7 = _mm_slli_epi16(psi_i_p7_m7, 1); + psi_i_p5_p7 = _mm_mulhi_epi16(psi_i_p5_p7, ONE_OVER_SQRT_2); + psi_i_p5_p7 = _mm_slli_epi16(psi_i_p5_p7, 1); + psi_i_p5_p5 = _mm_mulhi_epi16(psi_i_p5_p5, ONE_OVER_SQRT_2); + psi_i_p5_p5 = _mm_slli_epi16(psi_i_p5_p5, 1); + psi_i_p5_p3 = _mm_mulhi_epi16(psi_i_p5_p3, ONE_OVER_SQRT_2); + psi_i_p5_p3 = _mm_slli_epi16(psi_i_p5_p3, 1); + psi_i_p5_p1 = _mm_mulhi_epi16(psi_i_p5_p1, ONE_OVER_SQRT_2); + psi_i_p5_p1 = _mm_slli_epi16(psi_i_p5_p1, 1); + psi_i_p5_m1 = _mm_mulhi_epi16(psi_i_p5_m1, ONE_OVER_SQRT_2); + psi_i_p5_m1 = _mm_slli_epi16(psi_i_p5_m1, 1); + psi_i_p5_m3 = _mm_mulhi_epi16(psi_i_p5_m3, ONE_OVER_SQRT_2); + psi_i_p5_m3 = _mm_slli_epi16(psi_i_p5_m3, 1); + psi_i_p5_m5 = _mm_mulhi_epi16(psi_i_p5_m5, ONE_OVER_SQRT_2); + psi_i_p5_m5 = _mm_slli_epi16(psi_i_p5_m5, 1); + psi_i_p5_m7 = _mm_mulhi_epi16(psi_i_p5_m7, ONE_OVER_SQRT_2); + psi_i_p5_m7 = _mm_slli_epi16(psi_i_p5_m7, 1); + psi_i_p3_p7 = _mm_mulhi_epi16(psi_i_p3_p7, ONE_OVER_SQRT_2); + psi_i_p3_p7 = _mm_slli_epi16(psi_i_p3_p7, 1); + psi_i_p3_p5 = _mm_mulhi_epi16(psi_i_p3_p5, ONE_OVER_SQRT_2); + psi_i_p3_p5 = _mm_slli_epi16(psi_i_p3_p5, 1); + psi_i_p3_p3 = _mm_mulhi_epi16(psi_i_p3_p3, ONE_OVER_SQRT_2); + psi_i_p3_p3 = _mm_slli_epi16(psi_i_p3_p3, 1); + psi_i_p3_p1 = _mm_mulhi_epi16(psi_i_p3_p1, ONE_OVER_SQRT_2); + psi_i_p3_p1 = _mm_slli_epi16(psi_i_p3_p1, 1); + psi_i_p3_m1 = _mm_mulhi_epi16(psi_i_p3_m1, ONE_OVER_SQRT_2); + psi_i_p3_m1 = _mm_slli_epi16(psi_i_p3_m1, 1); + psi_i_p3_m3 = _mm_mulhi_epi16(psi_i_p3_m3, ONE_OVER_SQRT_2); + psi_i_p3_m3 = _mm_slli_epi16(psi_i_p3_m3, 1); + psi_i_p3_m5 = _mm_mulhi_epi16(psi_i_p3_m5, ONE_OVER_SQRT_2); + psi_i_p3_m5 = _mm_slli_epi16(psi_i_p3_m5, 1); + psi_i_p3_m7 = _mm_mulhi_epi16(psi_i_p3_m7, ONE_OVER_SQRT_2); + psi_i_p3_m7 = _mm_slli_epi16(psi_i_p3_m7, 1); + psi_i_p1_p7 = _mm_mulhi_epi16(psi_i_p1_p7, ONE_OVER_SQRT_2); + psi_i_p1_p7 = _mm_slli_epi16(psi_i_p1_p7, 1); + psi_i_p1_p5 = _mm_mulhi_epi16(psi_i_p1_p5, ONE_OVER_SQRT_2); + psi_i_p1_p5 = _mm_slli_epi16(psi_i_p1_p5, 1); + psi_i_p1_p3 = _mm_mulhi_epi16(psi_i_p1_p3, ONE_OVER_SQRT_2); + psi_i_p1_p3 = _mm_slli_epi16(psi_i_p1_p3, 1); + psi_i_p1_p1 = _mm_mulhi_epi16(psi_i_p1_p1, ONE_OVER_SQRT_2); + psi_i_p1_p1 = _mm_slli_epi16(psi_i_p1_p1, 1); + psi_i_p1_m1 = _mm_mulhi_epi16(psi_i_p1_m1, ONE_OVER_SQRT_2); + psi_i_p1_m1 = _mm_slli_epi16(psi_i_p1_m1, 1); + psi_i_p1_m3 = _mm_mulhi_epi16(psi_i_p1_m3, ONE_OVER_SQRT_2); + psi_i_p1_m3 = _mm_slli_epi16(psi_i_p1_m3, 1); + psi_i_p1_m5 = _mm_mulhi_epi16(psi_i_p1_m5, ONE_OVER_SQRT_2); + psi_i_p1_m5 = _mm_slli_epi16(psi_i_p1_m5, 1); + psi_i_p1_m7 = _mm_mulhi_epi16(psi_i_p1_m7, ONE_OVER_SQRT_2); + psi_i_p1_m7 = _mm_slli_epi16(psi_i_p1_m7, 1); + psi_i_m1_p7 = _mm_mulhi_epi16(psi_i_m1_p7, ONE_OVER_SQRT_2); + psi_i_m1_p7 = _mm_slli_epi16(psi_i_m1_p7, 1); + psi_i_m1_p5 = _mm_mulhi_epi16(psi_i_m1_p5, ONE_OVER_SQRT_2); + psi_i_m1_p5 = _mm_slli_epi16(psi_i_m1_p5, 1); + psi_i_m1_p3 = _mm_mulhi_epi16(psi_i_m1_p3, ONE_OVER_SQRT_2); + psi_i_m1_p3 = _mm_slli_epi16(psi_i_m1_p3, 1); + psi_i_m1_p1 = _mm_mulhi_epi16(psi_i_m1_p1, ONE_OVER_SQRT_2); + psi_i_m1_p1 = _mm_slli_epi16(psi_i_m1_p1, 1); + psi_i_m1_m1 = _mm_mulhi_epi16(psi_i_m1_m1, ONE_OVER_SQRT_2); + psi_i_m1_m1 = _mm_slli_epi16(psi_i_m1_m1, 1); + psi_i_m1_m3 = _mm_mulhi_epi16(psi_i_m1_m3, ONE_OVER_SQRT_2); + psi_i_m1_m3 = _mm_slli_epi16(psi_i_m1_m3, 1); + psi_i_m1_m5 = _mm_mulhi_epi16(psi_i_m1_m5, ONE_OVER_SQRT_2); + psi_i_m1_m5 = _mm_slli_epi16(psi_i_m1_m5, 1); + psi_i_m1_m7 = _mm_mulhi_epi16(psi_i_m1_m7, ONE_OVER_SQRT_2); + psi_i_m1_m7 = _mm_slli_epi16(psi_i_m1_m7, 1); + psi_i_m3_p7 = _mm_mulhi_epi16(psi_i_m3_p7, ONE_OVER_SQRT_2); + psi_i_m3_p7 = _mm_slli_epi16(psi_i_m3_p7, 1); + psi_i_m3_p5 = _mm_mulhi_epi16(psi_i_m3_p5, ONE_OVER_SQRT_2); + psi_i_m3_p5 = _mm_slli_epi16(psi_i_m3_p5, 1); + psi_i_m3_p3 = _mm_mulhi_epi16(psi_i_m3_p3, ONE_OVER_SQRT_2); + psi_i_m3_p3 = _mm_slli_epi16(psi_i_m3_p3, 1); + psi_i_m3_p1 = _mm_mulhi_epi16(psi_i_m3_p1, ONE_OVER_SQRT_2); + psi_i_m3_p1 = _mm_slli_epi16(psi_i_m3_p1, 1); + psi_i_m3_m1 = _mm_mulhi_epi16(psi_i_m3_m1, ONE_OVER_SQRT_2); + psi_i_m3_m1 = _mm_slli_epi16(psi_i_m3_m1, 1); + psi_i_m3_m3 = _mm_mulhi_epi16(psi_i_m3_m3, ONE_OVER_SQRT_2); + psi_i_m3_m3 = _mm_slli_epi16(psi_i_m3_m3, 1); + psi_i_m3_m5 = _mm_mulhi_epi16(psi_i_m3_m5, ONE_OVER_SQRT_2); + psi_i_m3_m5 = _mm_slli_epi16(psi_i_m3_m5, 1); + psi_i_m3_m7 = _mm_mulhi_epi16(psi_i_m3_m7, ONE_OVER_SQRT_2); + psi_i_m3_m7 = _mm_slli_epi16(psi_i_m3_m7, 1); + psi_i_m5_p7 = _mm_mulhi_epi16(psi_i_m5_p7, ONE_OVER_SQRT_2); + psi_i_m5_p7 = _mm_slli_epi16(psi_i_m5_p7, 1); + psi_i_m5_p5 = _mm_mulhi_epi16(psi_i_m5_p5, ONE_OVER_SQRT_2); + psi_i_m5_p5 = _mm_slli_epi16(psi_i_m5_p5, 1); + psi_i_m5_p3 = _mm_mulhi_epi16(psi_i_m5_p3, ONE_OVER_SQRT_2); + psi_i_m5_p3 = _mm_slli_epi16(psi_i_m5_p3, 1); + psi_i_m5_p1 = _mm_mulhi_epi16(psi_i_m5_p1, ONE_OVER_SQRT_2); + psi_i_m5_p1 = _mm_slli_epi16(psi_i_m5_p1, 1); + psi_i_m5_m1 = _mm_mulhi_epi16(psi_i_m5_m1, ONE_OVER_SQRT_2); + psi_i_m5_m1 = _mm_slli_epi16(psi_i_m5_m1, 1); + psi_i_m5_m3 = _mm_mulhi_epi16(psi_i_m5_m3, ONE_OVER_SQRT_2); + psi_i_m5_m3 = _mm_slli_epi16(psi_i_m5_m3, 1); + psi_i_m5_m5 = _mm_mulhi_epi16(psi_i_m5_m5, ONE_OVER_SQRT_2); + psi_i_m5_m5 = _mm_slli_epi16(psi_i_m5_m5, 1); + psi_i_m5_m7 = _mm_mulhi_epi16(psi_i_m5_m7, ONE_OVER_SQRT_2); + psi_i_m5_m7 = _mm_slli_epi16(psi_i_m5_m7, 1); + psi_i_m7_p7 = _mm_mulhi_epi16(psi_i_m7_p7, ONE_OVER_SQRT_2); + psi_i_m7_p7 = _mm_slli_epi16(psi_i_m7_p7, 1); + psi_i_m7_p5 = _mm_mulhi_epi16(psi_i_m7_p5, ONE_OVER_SQRT_2); + psi_i_m7_p5 = _mm_slli_epi16(psi_i_m7_p5, 1); + psi_i_m7_p3 = _mm_mulhi_epi16(psi_i_m7_p3, ONE_OVER_SQRT_2); + psi_i_m7_p3 = _mm_slli_epi16(psi_i_m7_p3, 1); + psi_i_m7_p1 = _mm_mulhi_epi16(psi_i_m7_p1, ONE_OVER_SQRT_2); + psi_i_m7_p1 = _mm_slli_epi16(psi_i_m7_p1, 1); + psi_i_m7_m1 = _mm_mulhi_epi16(psi_i_m7_m1, ONE_OVER_SQRT_2); + psi_i_m7_m1 = _mm_slli_epi16(psi_i_m7_m1, 1); + psi_i_m7_m3 = _mm_mulhi_epi16(psi_i_m7_m3, ONE_OVER_SQRT_2); + psi_i_m7_m3 = _mm_slli_epi16(psi_i_m7_m3, 1); + psi_i_m7_m5 = _mm_mulhi_epi16(psi_i_m7_m5, ONE_OVER_SQRT_2); + psi_i_m7_m5 = _mm_slli_epi16(psi_i_m7_m5, 1); + psi_i_m7_m7 = _mm_mulhi_epi16(psi_i_m7_m7, ONE_OVER_SQRT_2); + psi_i_m7_m7 = _mm_slli_epi16(psi_i_m7_m7, 1); + + psi_a_p7_p7 = _mm_adds_epi16(psi_r_p7_p7, psi_i_p7_p7); + psi_a_p7_p5 = _mm_adds_epi16(psi_r_p7_p5, psi_i_p7_p5); + psi_a_p7_p3 = _mm_adds_epi16(psi_r_p7_p3, psi_i_p7_p3); + psi_a_p7_p1 = _mm_adds_epi16(psi_r_p7_p1, psi_i_p7_p1); + psi_a_p7_m1 = _mm_adds_epi16(psi_r_p7_m1, psi_i_p7_m1); + psi_a_p7_m3 = _mm_adds_epi16(psi_r_p7_m3, psi_i_p7_m3); + psi_a_p7_m5 = _mm_adds_epi16(psi_r_p7_m5, psi_i_p7_m5); + psi_a_p7_m7 = _mm_adds_epi16(psi_r_p7_m7, psi_i_p7_m7); + psi_a_p5_p7 = _mm_adds_epi16(psi_r_p5_p7, psi_i_p5_p7); + psi_a_p5_p5 = _mm_adds_epi16(psi_r_p5_p5, psi_i_p5_p5); + psi_a_p5_p3 = _mm_adds_epi16(psi_r_p5_p3, psi_i_p5_p3); + psi_a_p5_p1 = _mm_adds_epi16(psi_r_p5_p1, psi_i_p5_p1); + psi_a_p5_m1 = _mm_adds_epi16(psi_r_p5_m1, psi_i_p5_m1); + psi_a_p5_m3 = _mm_adds_epi16(psi_r_p5_m3, psi_i_p5_m3); + psi_a_p5_m5 = _mm_adds_epi16(psi_r_p5_m5, psi_i_p5_m5); + psi_a_p5_m7 = _mm_adds_epi16(psi_r_p5_m7, psi_i_p5_m7); + psi_a_p3_p7 = _mm_adds_epi16(psi_r_p3_p7, psi_i_p3_p7); + psi_a_p3_p5 = _mm_adds_epi16(psi_r_p3_p5, psi_i_p3_p5); + psi_a_p3_p3 = _mm_adds_epi16(psi_r_p3_p3, psi_i_p3_p3); + psi_a_p3_p1 = _mm_adds_epi16(psi_r_p3_p1, psi_i_p3_p1); + psi_a_p3_m1 = _mm_adds_epi16(psi_r_p3_m1, psi_i_p3_m1); + psi_a_p3_m3 = _mm_adds_epi16(psi_r_p3_m3, psi_i_p3_m3); + psi_a_p3_m5 = _mm_adds_epi16(psi_r_p3_m5, psi_i_p3_m5); + psi_a_p3_m7 = _mm_adds_epi16(psi_r_p3_m7, psi_i_p3_m7); + psi_a_p1_p7 = _mm_adds_epi16(psi_r_p1_p7, psi_i_p1_p7); + psi_a_p1_p5 = _mm_adds_epi16(psi_r_p1_p5, psi_i_p1_p5); + psi_a_p1_p3 = _mm_adds_epi16(psi_r_p1_p3, psi_i_p1_p3); + psi_a_p1_p1 = _mm_adds_epi16(psi_r_p1_p1, psi_i_p1_p1); + psi_a_p1_m1 = _mm_adds_epi16(psi_r_p1_m1, psi_i_p1_m1); + psi_a_p1_m3 = _mm_adds_epi16(psi_r_p1_m3, psi_i_p1_m3); + psi_a_p1_m5 = _mm_adds_epi16(psi_r_p1_m5, psi_i_p1_m5); + psi_a_p1_m7 = _mm_adds_epi16(psi_r_p1_m7, psi_i_p1_m7); + psi_a_m1_p7 = _mm_adds_epi16(psi_r_m1_p7, psi_i_m1_p7); + psi_a_m1_p5 = _mm_adds_epi16(psi_r_m1_p5, psi_i_m1_p5); + psi_a_m1_p3 = _mm_adds_epi16(psi_r_m1_p3, psi_i_m1_p3); + psi_a_m1_p1 = _mm_adds_epi16(psi_r_m1_p1, psi_i_m1_p1); + psi_a_m1_m1 = _mm_adds_epi16(psi_r_m1_m1, psi_i_m1_m1); + psi_a_m1_m3 = _mm_adds_epi16(psi_r_m1_m3, psi_i_m1_m3); + psi_a_m1_m5 = _mm_adds_epi16(psi_r_m1_m5, psi_i_m1_m5); + psi_a_m1_m7 = _mm_adds_epi16(psi_r_m1_m7, psi_i_m1_m7); + psi_a_m3_p7 = _mm_adds_epi16(psi_r_m3_p7, psi_i_m3_p7); + psi_a_m3_p5 = _mm_adds_epi16(psi_r_m3_p5, psi_i_m3_p5); + psi_a_m3_p3 = _mm_adds_epi16(psi_r_m3_p3, psi_i_m3_p3); + psi_a_m3_p1 = _mm_adds_epi16(psi_r_m3_p1, psi_i_m3_p1); + psi_a_m3_m1 = _mm_adds_epi16(psi_r_m3_m1, psi_i_m3_m1); + psi_a_m3_m3 = _mm_adds_epi16(psi_r_m3_m3, psi_i_m3_m3); + psi_a_m3_m5 = _mm_adds_epi16(psi_r_m3_m5, psi_i_m3_m5); + psi_a_m3_m7 = _mm_adds_epi16(psi_r_m3_m7, psi_i_m3_m7); + psi_a_m5_p7 = _mm_adds_epi16(psi_r_m5_p7, psi_i_m5_p7); + psi_a_m5_p5 = _mm_adds_epi16(psi_r_m5_p5, psi_i_m5_p5); + psi_a_m5_p3 = _mm_adds_epi16(psi_r_m5_p3, psi_i_m5_p3); + psi_a_m5_p1 = _mm_adds_epi16(psi_r_m5_p1, psi_i_m5_p1); + psi_a_m5_m1 = _mm_adds_epi16(psi_r_m5_m1, psi_i_m5_m1); + psi_a_m5_m3 = _mm_adds_epi16(psi_r_m5_m3, psi_i_m5_m3); + psi_a_m5_m5 = _mm_adds_epi16(psi_r_m5_m5, psi_i_m5_m5); + psi_a_m5_m7 = _mm_adds_epi16(psi_r_m5_m7, psi_i_m5_m7); + psi_a_m7_p7 = _mm_adds_epi16(psi_r_m7_p7, psi_i_m7_p7); + psi_a_m7_p5 = _mm_adds_epi16(psi_r_m7_p5, psi_i_m7_p5); + psi_a_m7_p3 = _mm_adds_epi16(psi_r_m7_p3, psi_i_m7_p3); + psi_a_m7_p1 = _mm_adds_epi16(psi_r_m7_p1, psi_i_m7_p1); + psi_a_m7_m1 = _mm_adds_epi16(psi_r_m7_m1, psi_i_m7_m1); + psi_a_m7_m3 = _mm_adds_epi16(psi_r_m7_m3, psi_i_m7_m3); + psi_a_m7_m5 = _mm_adds_epi16(psi_r_m7_m5, psi_i_m7_m5); + psi_a_m7_m7 = _mm_adds_epi16(psi_r_m7_m7, psi_i_m7_m7); + + // Computing different multiples of ||h0||^2 + // x=1, y=1 + ch_mag_2_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42); + ch_mag_2_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_2_over_42_with_sigma2,1); + // x=1, y=3 + ch_mag_10_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42); + ch_mag_10_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_10_over_42_with_sigma2,1); + // x=1, x=5 + ch_mag_26_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42); + ch_mag_26_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_26_over_42_with_sigma2,1); + // x=1, y=7 + ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); + ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); + // x=3, y=3 + ch_mag_18_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42); + ch_mag_18_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_18_over_42_with_sigma2,1); + // x=3, y=5 + ch_mag_34_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42); + ch_mag_34_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_34_over_42_with_sigma2,1); + // x=3, y=7 + ch_mag_58_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42); + ch_mag_58_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_58_over_42_with_sigma2,2); + // x=5, y=5 + ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); + ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); + // x=5, y=7 + ch_mag_74_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42); + ch_mag_74_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_74_over_42_with_sigma2,2); + // x=7, y=7 + ch_mag_98_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42); + ch_mag_98_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_98_over_42_with_sigma2,2); + + // Computing Metrics + xmm1 = _mm_adds_epi16(psi_a_p7_p7, y0_p_7_7); + bit_met_p7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p7_p5, y0_p_7_5); + bit_met_p7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p7_p3, y0_p_7_3); + bit_met_p7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p7_p1, y0_p_7_1); + bit_met_p7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p7_m1, y0_m_7_1); + bit_met_p7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p7_m3, y0_m_7_3); + bit_met_p7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p7_m5, y0_m_7_5); + bit_met_p7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p7_m7, y0_m_7_7); + bit_met_p7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_p7, y0_p_5_7); + bit_met_p5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_p5, y0_p_5_5); + bit_met_p5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_p3, y0_p_5_3); + bit_met_p5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_p1, y0_p_5_1); + bit_met_p5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_m1, y0_m_5_1); + bit_met_p5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_m3, y0_m_5_3); + bit_met_p5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_m5, y0_m_5_5); + bit_met_p5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p5_m7, y0_m_5_7); + bit_met_p5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_p7, y0_p_3_7); + bit_met_p3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_p5, y0_p_3_5); + bit_met_p3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_p3, y0_p_3_3); + bit_met_p3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_p1, y0_p_3_1); + bit_met_p3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_m1, y0_m_3_1); + bit_met_p3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_m3, y0_m_3_3); + bit_met_p3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_m5, y0_m_3_5); + bit_met_p3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p3_m7, y0_m_3_7); + bit_met_p3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_p7, y0_p_1_7); + bit_met_p1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_p5, y0_p_1_5); + bit_met_p1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_p3, y0_p_1_3); + bit_met_p1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_p1, y0_p_1_1); + bit_met_p1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_m1, y0_m_1_1); + bit_met_p1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_m3, y0_m_1_3); + bit_met_p1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_m5, y0_m_1_5); + bit_met_p1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_adds_epi16(psi_a_p1_m7, y0_m_1_7); + bit_met_p1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + + xmm1 = _mm_subs_epi16(psi_a_m1_p7, y0_m_1_7); + bit_met_m1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m1_p5, y0_m_1_5); + bit_met_m1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m1_p3, y0_m_1_3); + bit_met_m1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m1_p1, y0_m_1_1); + bit_met_m1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m1_m1, y0_p_1_1); + bit_met_m1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m1_m3, y0_p_1_3); + bit_met_m1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m1_m5, y0_p_1_5); + bit_met_m1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m1_m7, y0_p_1_7); + bit_met_m1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_p7, y0_m_3_7); + bit_met_m3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_p5, y0_m_3_5); + bit_met_m3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_p3, y0_m_3_3); + bit_met_m3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_p1, y0_m_3_1); + bit_met_m3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_m1, y0_p_3_1); + bit_met_m3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_m3, y0_p_3_3); + bit_met_m3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_m5, y0_p_3_5); + bit_met_m3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m3_m7, y0_p_3_7); + bit_met_m3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_p7, y0_m_5_7); + bit_met_m5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_p5, y0_m_5_5); + bit_met_m5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_p3, y0_m_5_3); + bit_met_m5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_p1, y0_m_5_1); + bit_met_m5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_m1, y0_p_5_1); + bit_met_m5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_m3, y0_p_5_3); + bit_met_m5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_m5, y0_p_5_5); + bit_met_m5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m5_m7, y0_p_5_7); + bit_met_m5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_p7, y0_m_7_7); + bit_met_m7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_p5, y0_m_7_5); + bit_met_m7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_p3, y0_m_7_3); + bit_met_m7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_p1, y0_m_7_1); + bit_met_m7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_m1, y0_p_7_1); + bit_met_m7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_m3, y0_p_7_3); + bit_met_m7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_m5, y0_p_7_5); + bit_met_m7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm1 = _mm_subs_epi16(psi_a_m7_m7, y0_p_7_7); + bit_met_m7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + + // Detection for 1st bit (LTE mapping) + // bit = 1 + xmm0 = _mm_max_epi16(bit_met_m7_p7, bit_met_m7_p5); + xmm1 = _mm_max_epi16(bit_met_m7_p3, bit_met_m7_p1); + xmm2 = _mm_max_epi16(bit_met_m7_m1, bit_met_m7_m3); + xmm3 = _mm_max_epi16(bit_met_m7_m5, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m5_p7, bit_met_m5_p5); + xmm1 = _mm_max_epi16(bit_met_m5_p3, bit_met_m5_p1); + xmm2 = _mm_max_epi16(bit_met_m5_m1, bit_met_m5_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m5_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_m3_p7, bit_met_m3_p5); + xmm1 = _mm_max_epi16(bit_met_m3_p3, bit_met_m3_p1); + xmm2 = _mm_max_epi16(bit_met_m3_m1, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m3_m5, bit_met_m3_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_p7, bit_met_m1_p5); + xmm1 = _mm_max_epi16(bit_met_m1_p3, bit_met_m1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m1_m3); + xmm3 = _mm_max_epi16(bit_met_m1_m5, bit_met_m1_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + // bit = 0 + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p7_p5); + xmm1 = _mm_max_epi16(bit_met_p7_p3, bit_met_p7_p1); + xmm2 = _mm_max_epi16(bit_met_p7_m1, bit_met_p7_m3); + xmm3 = _mm_max_epi16(bit_met_p7_m5, bit_met_p7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_p7, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p5_p3, bit_met_p5_p1); + xmm2 = _mm_max_epi16(bit_met_p5_m1, bit_met_p5_m3); + xmm3 = _mm_max_epi16(bit_met_p5_m5, bit_met_p5_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_p7, bit_met_p3_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p3_p1); + xmm2 = _mm_max_epi16(bit_met_p3_m1, bit_met_p3_m3); + xmm3 = _mm_max_epi16(bit_met_p3_m5, bit_met_p3_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_p7, bit_met_p1_p5); + xmm1 = _mm_max_epi16(bit_met_p1_p3, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_p1_m1, bit_met_p1_m3); + xmm3 = _mm_max_epi16(bit_met_p1_m5, bit_met_p1_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y0r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 2nd bit (LTE mapping) + // bit = 1 + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + // bit = 0 + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y1r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 3rd bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); + xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); + xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); + xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); + xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); + xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); + xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); + xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); + xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); + xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); + xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); + xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); + xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); + xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); + xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y2r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 4th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y0i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + + // Detection for 5th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); + xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); + xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); + xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); + xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); + xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); + xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); + xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); + xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); + xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); + xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); + xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); + xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); + xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); + xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y1i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 6th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y2i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + + // map to output stream, difficult to do in SIMD since we have 6 16bit LLRs + // RE 1 + j = 24*i; + stream0_out[j + 0] = ((short *)&y0r)[0]; + stream0_out[j + 1] = ((short *)&y1r)[0]; + stream0_out[j + 2] = ((short *)&y2r)[0]; + stream0_out[j + 3] = ((short *)&y0i)[0]; + stream0_out[j + 4] = ((short *)&y1i)[0]; + stream0_out[j + 5] = ((short *)&y2i)[0]; + // RE 2 + stream0_out[j + 6] = ((short *)&y0r)[1]; + stream0_out[j + 7] = ((short *)&y1r)[1]; + stream0_out[j + 8] = ((short *)&y2r)[1]; + stream0_out[j + 9] = ((short *)&y0i)[1]; + stream0_out[j + 10] = ((short *)&y1i)[1]; + stream0_out[j + 11] = ((short *)&y2i)[1]; + // RE 3 + stream0_out[j + 12] = ((short *)&y0r)[2]; + stream0_out[j + 13] = ((short *)&y1r)[2]; + stream0_out[j + 14] = ((short *)&y2r)[2]; + stream0_out[j + 15] = ((short *)&y0i)[2]; + stream0_out[j + 16] = ((short *)&y1i)[2]; + stream0_out[j + 17] = ((short *)&y2i)[2]; + // RE 4 + stream0_out[j + 18] = ((short *)&y0r)[3]; + stream0_out[j + 19] = ((short *)&y1r)[3]; + stream0_out[j + 20] = ((short *)&y2r)[3]; + stream0_out[j + 21] = ((short *)&y0i)[3]; + stream0_out[j + 22] = ((short *)&y1i)[3]; + stream0_out[j + 23] = ((short *)&y2i)[3]; + // RE 5 + stream0_out[j + 24] = ((short *)&y0r)[4]; + stream0_out[j + 25] = ((short *)&y1r)[4]; + stream0_out[j + 26] = ((short *)&y2r)[4]; + stream0_out[j + 27] = ((short *)&y0i)[4]; + stream0_out[j + 28] = ((short *)&y1i)[4]; + stream0_out[j + 29] = ((short *)&y2i)[4]; + // RE 6 + stream0_out[j + 30] = ((short *)&y0r)[5]; + stream0_out[j + 31] = ((short *)&y1r)[5]; + stream0_out[j + 32] = ((short *)&y2r)[5]; + stream0_out[j + 33] = ((short *)&y0i)[5]; + stream0_out[j + 34] = ((short *)&y1i)[5]; + stream0_out[j + 35] = ((short *)&y2i)[5]; + // RE 7 + stream0_out[j + 36] = ((short *)&y0r)[6]; + stream0_out[j + 37] = ((short *)&y1r)[6]; + stream0_out[j + 38] = ((short *)&y2r)[6]; + stream0_out[j + 39] = ((short *)&y0i)[6]; + stream0_out[j + 40] = ((short *)&y1i)[6]; + stream0_out[j + 41] = ((short *)&y2i)[6]; + // RE 8 + stream0_out[j + 42] = ((short *)&y0r)[7]; + stream0_out[j + 43] = ((short *)&y1r)[7]; + stream0_out[j + 44] = ((short *)&y2r)[7]; + stream0_out[j + 45] = ((short *)&y0i)[7]; + stream0_out[j + 46] = ((short *)&y1i)[7]; + stream0_out[j + 47] = ((short *)&y2i)[7]; +#elif defined(__arm__) + +#endif + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif +} + + +int nr_dlsch_64qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag, + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + int16_t **llr16p) +{ + + int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + //first symbol has different structure due to more pilots + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + qam64_qpsk((short *)rxF, + (short *)rxF_i, + (short *)ch_mag, + (short *)llr16, + (short *)rho, + len); + + llr16 += (6*len); + *llr16p = (short *)llr16; + return(0); +} + + + +void qam64_qam16(short *stream0_in, + short *stream1_in, + short *ch_mag, + short *ch_mag_i, + short *stream0_out, + short *rho01, + int length + ) +{ + + /* + Author: S. Wagner + Date: 31-07-12 + + Input: + stream0_in: MF filter for 1st stream, i.e., y0=h0'*y + stream1_in: MF filter for 2nd stream, i.e., y1=h1'*y + ch_mag: 4*h0/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + ch_mag_i: 4*h1/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + rho01: Channel cross correlation, i.e., h1'*h0 + + Output: + stream0_out: output LLRs for 1st stream + */ + +#if defined(__x86_64__) || defined(__i386__) + + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *ch_mag_128i = (__m128i *)ch_mag; + __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; + + __m128i ONE_OVER_SQRT_42 = _mm_set1_epi16(10112); // round(1/sqrt(42)*2^16) + __m128i THREE_OVER_SQRT_42 = _mm_set1_epi16(30337); // round(3/sqrt(42)*2^16) + __m128i FIVE_OVER_SQRT_42 = _mm_set1_epi16(25281); // round(5/sqrt(42)*2^15) + __m128i SEVEN_OVER_SQRT_42 = _mm_set1_epi16(17697); // round(5/sqrt(42)*2^15) + __m128i FORTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(30969); // round(49/(4*sqrt(42))*2^14), Q2.14 + __m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(23385); // round(37/(4*sqrt(42))*2^14), Q2.14 + __m128i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(31601); // round(25/(4*sqrt(42))*2^15) + __m128i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(18329); // round(29/(4*sqrt(42))*2^15), Q2.14 + __m128i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(21489); // round(17/(4*sqrt(42))*2^15) + __m128i NINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(11376); // round(9/(4*sqrt(42))*2^15) + __m128i THIRTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(16433); // round(13/(4*sqrt(42))*2^15) + __m128i FIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(6320); // round(5/(4*sqrt(42))*2^15) + __m128i ONE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(1264); // round(1/(4*sqrt(42))*2^15) + __m128i ONE_OVER_SQRT_10_Q15 = _mm_set1_epi16(10362); // round(1/sqrt(10)*2^15) + __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) + __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) + + + __m128i ch_mag_int; + __m128i ch_mag_des; + __m128i ch_mag_98_over_42_with_sigma2; + __m128i ch_mag_74_over_42_with_sigma2; + __m128i ch_mag_58_over_42_with_sigma2; + __m128i ch_mag_50_over_42_with_sigma2; + __m128i ch_mag_34_over_42_with_sigma2; + __m128i ch_mag_18_over_42_with_sigma2; + __m128i ch_mag_26_over_42_with_sigma2; + __m128i ch_mag_10_over_42_with_sigma2; + __m128i ch_mag_2_over_42_with_sigma2; + __m128i y0r_one_over_sqrt_21; + __m128i y0r_three_over_sqrt_21; + __m128i y0r_five_over_sqrt_21; + __m128i y0r_seven_over_sqrt_21; + __m128i y0i_one_over_sqrt_21; + __m128i y0i_three_over_sqrt_21; + __m128i y0i_five_over_sqrt_21; + __m128i y0i_seven_over_sqrt_21; + +#elif defined(__arm__) + +#endif + int i,j; + + + + for (i=0; i>2; i+=2) { + +#if defined(__x86_64__) || defined(__i386__) + // Get rho + xmm0 = rho01_128i[i]; + xmm1 = rho01_128i[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // Compute the different rhos + rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42); + rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42); + rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42); + rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42); + rho_rpi_5_5 = _mm_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42); + rho_rmi_5_5 = _mm_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42); + rho_rpi_7_7 = _mm_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42); + rho_rmi_7_7 = _mm_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42); + + rho_rpi_5_5 = _mm_slli_epi16(rho_rpi_5_5, 1); + rho_rmi_5_5 = _mm_slli_epi16(rho_rmi_5_5, 1); + rho_rpi_7_7 = _mm_slli_epi16(rho_rpi_7_7, 2); + rho_rmi_7_7 = _mm_slli_epi16(rho_rmi_7_7, 2); + + xmm4 = _mm_mulhi_epi16(xmm2, ONE_OVER_SQRT_42); + xmm5 = _mm_mulhi_epi16(xmm3, ONE_OVER_SQRT_42); + xmm6 = _mm_mulhi_epi16(xmm3, THREE_OVER_SQRT_42); + xmm7 = _mm_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42); + xmm8 = _mm_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42); + xmm7 = _mm_slli_epi16(xmm7, 1); + xmm8 = _mm_slli_epi16(xmm8, 2); + + rho_rpi_1_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_1_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_1_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_1_5 = _mm_subs_epi16(xmm4, xmm7); + rho_rpi_1_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_1_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, THREE_OVER_SQRT_42); + rho_rpi_3_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_3_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_3_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_3_5 = _mm_subs_epi16(xmm4, xmm7); + rho_rpi_3_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_3_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42); + xmm4 = _mm_slli_epi16(xmm4, 1); + rho_rpi_5_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_5_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_5_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_5_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_5_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_5_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42); + xmm4 = _mm_slli_epi16(xmm4, 2); + rho_rpi_7_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_7_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_7_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_7_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_7_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_7_5 = _mm_subs_epi16(xmm4, xmm7); + + // Rearrange interfering MF output + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + // Psi_r calculation from rho_rpi or rho_rmi + xmm0 = _mm_setzero_si128(); // ZERO for abs_pi16 + xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1r); + psi_r_p7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1r); + psi_r_p7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1r); + psi_r_p7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1r); + psi_r_p7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1r); + psi_r_p7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1r); + psi_r_p7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1r); + psi_r_p7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1r); + psi_r_p7_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1r); + psi_r_p5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1r); + psi_r_p5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1r); + psi_r_p5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1r); + psi_r_p5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1r); + psi_r_p5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1r); + psi_r_p5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1r); + psi_r_p5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1r); + psi_r_p5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1r); + psi_r_p3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1r); + psi_r_p3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1r); + psi_r_p3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1r); + psi_r_p3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1r); + psi_r_p3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1r); + psi_r_p3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1r); + psi_r_p3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1r); + psi_r_p3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1r); + psi_r_p1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1r); + psi_r_p1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1r); + psi_r_p1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1r); + psi_r_p1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1r); + psi_r_p1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1r); + psi_r_p1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1r); + psi_r_p1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1r); + psi_r_p1_m7 = _mm_abs_epi16(xmm2); + + xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1r); + psi_r_m1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1r); + psi_r_m1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1r); + psi_r_m1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1r); + psi_r_m1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1r); + psi_r_m1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1r); + psi_r_m1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1r); + psi_r_m1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1r); + psi_r_m1_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1r); + psi_r_m3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1r); + psi_r_m3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1r); + psi_r_m3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1r); + psi_r_m3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1r); + psi_r_m3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1r); + psi_r_m3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1r); + psi_r_m3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1r); + psi_r_m3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1r); + psi_r_m5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1r); + psi_r_m5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1r); + psi_r_m5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1r); + psi_r_m5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1r); + psi_r_m5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1r); + psi_r_m5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1r); + psi_r_m5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1r); + psi_r_m5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1r); + psi_r_m7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1r); + psi_r_m7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1r); + psi_r_m7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1r); + psi_r_m7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1r); + psi_r_m7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1r); + psi_r_m7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1r); + psi_r_m7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1r); + psi_r_m7_m7 = _mm_abs_epi16(xmm2); + + // Psi_i calculation from rho_rpi or rho_rmi + xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1i); + psi_i_p7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1i); + psi_i_p7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1i); + psi_i_p7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1i); + psi_i_p7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1i); + psi_i_p7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1i); + psi_i_p7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1i); + psi_i_p7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1i); + psi_i_p7_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1i); + psi_i_p5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1i); + psi_i_p5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1i); + psi_i_p5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1i); + psi_i_p5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1i); + psi_i_p5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1i); + psi_i_p5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1i); + psi_i_p5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1i); + psi_i_p5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1i); + psi_i_p3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1i); + psi_i_p3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1i); + psi_i_p3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1i); + psi_i_p3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1i); + psi_i_p3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1i); + psi_i_p3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1i); + psi_i_p3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1i); + psi_i_p3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1i); + psi_i_p1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1i); + psi_i_p1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1i); + psi_i_p1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1i); + psi_i_p1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1i); + psi_i_p1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1i); + psi_i_p1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1i); + psi_i_p1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1i); + psi_i_p1_m7 = _mm_abs_epi16(xmm2); + + xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1i); + psi_i_m1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1i); + psi_i_m1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1i); + psi_i_m1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1i); + psi_i_m1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1i); + psi_i_m1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1i); + psi_i_m1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1i); + psi_i_m1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1i); + psi_i_m1_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1i); + psi_i_m3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1i); + psi_i_m3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1i); + psi_i_m3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1i); + psi_i_m3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1i); + psi_i_m3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1i); + psi_i_m3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1i); + psi_i_m3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1i); + psi_i_m3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1i); + psi_i_m5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1i); + psi_i_m5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1i); + psi_i_m5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1i); + psi_i_m5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1i); + psi_i_m5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1i); + psi_i_m5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1i); + psi_i_m5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1i); + psi_i_m5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1i); + psi_i_m7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1i); + psi_i_m7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1i); + psi_i_m7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1i); + psi_i_m7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1i); + psi_i_m7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1i); + psi_i_m7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1i); + psi_i_m7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1i); + psi_i_m7_m7 = _mm_abs_epi16(xmm2); + + + // Rearrange desired MF output + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // Rearrange desired channel magnitudes + xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) + xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); + + // Rearrange interfering channel magnitudes + xmm2 = ch_mag_128i_i[i]; + xmm3 = ch_mag_128i_i[i+1]; + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); + + y0r_one_over_sqrt_21 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_42); + y0r_three_over_sqrt_21 = _mm_mulhi_epi16(y0r, THREE_OVER_SQRT_42); + y0r_five_over_sqrt_21 = _mm_mulhi_epi16(y0r, FIVE_OVER_SQRT_42); + y0r_five_over_sqrt_21 = _mm_slli_epi16(y0r_five_over_sqrt_21, 1); + y0r_seven_over_sqrt_21 = _mm_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42); + y0r_seven_over_sqrt_21 = _mm_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14 + + y0i_one_over_sqrt_21 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_42); + y0i_three_over_sqrt_21 = _mm_mulhi_epi16(y0i, THREE_OVER_SQRT_42); + y0i_five_over_sqrt_21 = _mm_mulhi_epi16(y0i, FIVE_OVER_SQRT_42); + y0i_five_over_sqrt_21 = _mm_slli_epi16(y0i_five_over_sqrt_21, 1); + y0i_seven_over_sqrt_21 = _mm_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42); + y0i_seven_over_sqrt_21 = _mm_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14 + + y0_p_7_1 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_7_3 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_7_5 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_7_7 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_5_1 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_5_3 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_5_5 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_5_7 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_3_5 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_3_7 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_1_1 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_1_3 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_1_5 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_1_7 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); + + y0_m_1_1 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_1_3 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_1_5 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_1_7 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_3_5 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_3_7 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_5_1 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_5_3 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_5_5 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_5_7 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_7_1 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_7_3 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_7_5 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_7_7 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); + + interference_abs_epi16(psi_r_p7_p7, ch_mag_int, a_r_p7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p7_p5, ch_mag_int, a_r_p7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p7_p3, ch_mag_int, a_r_p7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p7_p1, ch_mag_int, a_r_p7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p7_m1, ch_mag_int, a_r_p7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p7_m3, ch_mag_int, a_r_p7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p7_m5, ch_mag_int, a_r_p7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p7_m7, ch_mag_int, a_r_p7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_p7, ch_mag_int, a_r_p5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_p5, ch_mag_int, a_r_p5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_p3, ch_mag_int, a_r_p5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_p1, ch_mag_int, a_r_p5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_m1, ch_mag_int, a_r_p5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_m3, ch_mag_int, a_r_p5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_m5, ch_mag_int, a_r_p5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p5_m7, ch_mag_int, a_r_p5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_p7, ch_mag_int, a_r_p3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_p5, ch_mag_int, a_r_p3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_p3, ch_mag_int, a_r_p3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_p1, ch_mag_int, a_r_p3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_m1, ch_mag_int, a_r_p3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_m3, ch_mag_int, a_r_p3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_m5, ch_mag_int, a_r_p3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p3_m7, ch_mag_int, a_r_p3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_p7, ch_mag_int, a_r_p1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_p5, ch_mag_int, a_r_p1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_p3, ch_mag_int, a_r_p1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_p1, ch_mag_int, a_r_p1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_m1, ch_mag_int, a_r_p1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_m3, ch_mag_int, a_r_p1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_m5, ch_mag_int, a_r_p1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_p1_m7, ch_mag_int, a_r_p1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_p7, ch_mag_int, a_r_m1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_p5, ch_mag_int, a_r_m1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_p3, ch_mag_int, a_r_m1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_p1, ch_mag_int, a_r_m1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_m1, ch_mag_int, a_r_m1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_m3, ch_mag_int, a_r_m1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_m5, ch_mag_int, a_r_m1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m1_m7, ch_mag_int, a_r_m1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_p7, ch_mag_int, a_r_m3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_p5, ch_mag_int, a_r_m3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_p3, ch_mag_int, a_r_m3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_p1, ch_mag_int, a_r_m3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_m1, ch_mag_int, a_r_m3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_m3, ch_mag_int, a_r_m3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_m5, ch_mag_int, a_r_m3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m3_m7, ch_mag_int, a_r_m3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_p7, ch_mag_int, a_r_m5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_p5, ch_mag_int, a_r_m5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_p3, ch_mag_int, a_r_m5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_p1, ch_mag_int, a_r_m5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_m1, ch_mag_int, a_r_m5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_m3, ch_mag_int, a_r_m5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_m5, ch_mag_int, a_r_m5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m5_m7, ch_mag_int, a_r_m5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_p7, ch_mag_int, a_r_m7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_p5, ch_mag_int, a_r_m7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_p3, ch_mag_int, a_r_m7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_p1, ch_mag_int, a_r_m7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_m1, ch_mag_int, a_r_m7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_m3, ch_mag_int, a_r_m7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_m5, ch_mag_int, a_r_m7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_r_m7_m7, ch_mag_int, a_r_m7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + + interference_abs_epi16(psi_i_p7_p7, ch_mag_int, a_i_p7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p7_p5, ch_mag_int, a_i_p7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p7_p3, ch_mag_int, a_i_p7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p7_p1, ch_mag_int, a_i_p7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p7_m1, ch_mag_int, a_i_p7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p7_m3, ch_mag_int, a_i_p7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p7_m5, ch_mag_int, a_i_p7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p7_m7, ch_mag_int, a_i_p7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_p7, ch_mag_int, a_i_p5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_p5, ch_mag_int, a_i_p5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_p3, ch_mag_int, a_i_p5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_p1, ch_mag_int, a_i_p5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_m1, ch_mag_int, a_i_p5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_m3, ch_mag_int, a_i_p5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_m5, ch_mag_int, a_i_p5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p5_m7, ch_mag_int, a_i_p5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_p7, ch_mag_int, a_i_p3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_p5, ch_mag_int, a_i_p3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_p3, ch_mag_int, a_i_p3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_p1, ch_mag_int, a_i_p3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_m1, ch_mag_int, a_i_p3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_m3, ch_mag_int, a_i_p3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_m5, ch_mag_int, a_i_p3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p3_m7, ch_mag_int, a_i_p3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_p7, ch_mag_int, a_i_p1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_p5, ch_mag_int, a_i_p1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_p3, ch_mag_int, a_i_p1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_p1, ch_mag_int, a_i_p1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_m1, ch_mag_int, a_i_p1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_m3, ch_mag_int, a_i_p1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_m5, ch_mag_int, a_i_p1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_p1_m7, ch_mag_int, a_i_p1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_p7, ch_mag_int, a_i_m1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_p5, ch_mag_int, a_i_m1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_p3, ch_mag_int, a_i_m1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_p1, ch_mag_int, a_i_m1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_m1, ch_mag_int, a_i_m1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_m3, ch_mag_int, a_i_m1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_m5, ch_mag_int, a_i_m1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m1_m7, ch_mag_int, a_i_m1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_p7, ch_mag_int, a_i_m3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_p5, ch_mag_int, a_i_m3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_p3, ch_mag_int, a_i_m3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_p1, ch_mag_int, a_i_m3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_m1, ch_mag_int, a_i_m3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_m3, ch_mag_int, a_i_m3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_m5, ch_mag_int, a_i_m3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m3_m7, ch_mag_int, a_i_m3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_p7, ch_mag_int, a_i_m5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_p5, ch_mag_int, a_i_m5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_p3, ch_mag_int, a_i_m5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_p1, ch_mag_int, a_i_m5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_m1, ch_mag_int, a_i_m5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_m3, ch_mag_int, a_i_m5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_m5, ch_mag_int, a_i_m5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m5_m7, ch_mag_int, a_i_m5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_p7, ch_mag_int, a_i_m7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_p5, ch_mag_int, a_i_m7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_p3, ch_mag_int, a_i_m7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_p1, ch_mag_int, a_i_m7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_m1, ch_mag_int, a_i_m7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_m3, ch_mag_int, a_i_m7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_m5, ch_mag_int, a_i_m7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + interference_abs_epi16(psi_i_m7_m7, ch_mag_int, a_i_m7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); + + // Calculation of a group of two terms in the bit metric involving product of psi and interference + prodsum_psi_a_epi16(psi_r_p7_p7, a_r_p7_p7, psi_i_p7_p7, a_i_p7_p7, psi_a_p7_p7); + prodsum_psi_a_epi16(psi_r_p7_p5, a_r_p7_p5, psi_i_p7_p5, a_i_p7_p5, psi_a_p7_p5); + prodsum_psi_a_epi16(psi_r_p7_p3, a_r_p7_p3, psi_i_p7_p3, a_i_p7_p3, psi_a_p7_p3); + prodsum_psi_a_epi16(psi_r_p7_p1, a_r_p7_p1, psi_i_p7_p1, a_i_p7_p1, psi_a_p7_p1); + prodsum_psi_a_epi16(psi_r_p7_m1, a_r_p7_m1, psi_i_p7_m1, a_i_p7_m1, psi_a_p7_m1); + prodsum_psi_a_epi16(psi_r_p7_m3, a_r_p7_m3, psi_i_p7_m3, a_i_p7_m3, psi_a_p7_m3); + prodsum_psi_a_epi16(psi_r_p7_m5, a_r_p7_m5, psi_i_p7_m5, a_i_p7_m5, psi_a_p7_m5); + prodsum_psi_a_epi16(psi_r_p7_m7, a_r_p7_m7, psi_i_p7_m7, a_i_p7_m7, psi_a_p7_m7); + prodsum_psi_a_epi16(psi_r_p5_p7, a_r_p5_p7, psi_i_p5_p7, a_i_p5_p7, psi_a_p5_p7); + prodsum_psi_a_epi16(psi_r_p5_p5, a_r_p5_p5, psi_i_p5_p5, a_i_p5_p5, psi_a_p5_p5); + prodsum_psi_a_epi16(psi_r_p5_p3, a_r_p5_p3, psi_i_p5_p3, a_i_p5_p3, psi_a_p5_p3); + prodsum_psi_a_epi16(psi_r_p5_p1, a_r_p5_p1, psi_i_p5_p1, a_i_p5_p1, psi_a_p5_p1); + prodsum_psi_a_epi16(psi_r_p5_m1, a_r_p5_m1, psi_i_p5_m1, a_i_p5_m1, psi_a_p5_m1); + prodsum_psi_a_epi16(psi_r_p5_m3, a_r_p5_m3, psi_i_p5_m3, a_i_p5_m3, psi_a_p5_m3); + prodsum_psi_a_epi16(psi_r_p5_m5, a_r_p5_m5, psi_i_p5_m5, a_i_p5_m5, psi_a_p5_m5); + prodsum_psi_a_epi16(psi_r_p5_m7, a_r_p5_m7, psi_i_p5_m7, a_i_p5_m7, psi_a_p5_m7); + prodsum_psi_a_epi16(psi_r_p3_p7, a_r_p3_p7, psi_i_p3_p7, a_i_p3_p7, psi_a_p3_p7); + prodsum_psi_a_epi16(psi_r_p3_p5, a_r_p3_p5, psi_i_p3_p5, a_i_p3_p5, psi_a_p3_p5); + prodsum_psi_a_epi16(psi_r_p3_p3, a_r_p3_p3, psi_i_p3_p3, a_i_p3_p3, psi_a_p3_p3); + prodsum_psi_a_epi16(psi_r_p3_p1, a_r_p3_p1, psi_i_p3_p1, a_i_p3_p1, psi_a_p3_p1); + prodsum_psi_a_epi16(psi_r_p3_m1, a_r_p3_m1, psi_i_p3_m1, a_i_p3_m1, psi_a_p3_m1); + prodsum_psi_a_epi16(psi_r_p3_m3, a_r_p3_m3, psi_i_p3_m3, a_i_p3_m3, psi_a_p3_m3); + prodsum_psi_a_epi16(psi_r_p3_m5, a_r_p3_m5, psi_i_p3_m5, a_i_p3_m5, psi_a_p3_m5); + prodsum_psi_a_epi16(psi_r_p3_m7, a_r_p3_m7, psi_i_p3_m7, a_i_p3_m7, psi_a_p3_m7); + prodsum_psi_a_epi16(psi_r_p1_p7, a_r_p1_p7, psi_i_p1_p7, a_i_p1_p7, psi_a_p1_p7); + prodsum_psi_a_epi16(psi_r_p1_p5, a_r_p1_p5, psi_i_p1_p5, a_i_p1_p5, psi_a_p1_p5); + prodsum_psi_a_epi16(psi_r_p1_p3, a_r_p1_p3, psi_i_p1_p3, a_i_p1_p3, psi_a_p1_p3); + prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); + prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); + prodsum_psi_a_epi16(psi_r_p1_m3, a_r_p1_m3, psi_i_p1_m3, a_i_p1_m3, psi_a_p1_m3); + prodsum_psi_a_epi16(psi_r_p1_m5, a_r_p1_m5, psi_i_p1_m5, a_i_p1_m5, psi_a_p1_m5); + prodsum_psi_a_epi16(psi_r_p1_m7, a_r_p1_m7, psi_i_p1_m7, a_i_p1_m7, psi_a_p1_m7); + prodsum_psi_a_epi16(psi_r_m1_p7, a_r_m1_p7, psi_i_m1_p7, a_i_m1_p7, psi_a_m1_p7); + prodsum_psi_a_epi16(psi_r_m1_p5, a_r_m1_p5, psi_i_m1_p5, a_i_m1_p5, psi_a_m1_p5); + prodsum_psi_a_epi16(psi_r_m1_p3, a_r_m1_p3, psi_i_m1_p3, a_i_m1_p3, psi_a_m1_p3); + prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); + prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); + prodsum_psi_a_epi16(psi_r_m1_m3, a_r_m1_m3, psi_i_m1_m3, a_i_m1_m3, psi_a_m1_m3); + prodsum_psi_a_epi16(psi_r_m1_m5, a_r_m1_m5, psi_i_m1_m5, a_i_m1_m5, psi_a_m1_m5); + prodsum_psi_a_epi16(psi_r_m1_m7, a_r_m1_m7, psi_i_m1_m7, a_i_m1_m7, psi_a_m1_m7); + prodsum_psi_a_epi16(psi_r_m3_p7, a_r_m3_p7, psi_i_m3_p7, a_i_m3_p7, psi_a_m3_p7); + prodsum_psi_a_epi16(psi_r_m3_p5, a_r_m3_p5, psi_i_m3_p5, a_i_m3_p5, psi_a_m3_p5); + prodsum_psi_a_epi16(psi_r_m3_p3, a_r_m3_p3, psi_i_m3_p3, a_i_m3_p3, psi_a_m3_p3); + prodsum_psi_a_epi16(psi_r_m3_p1, a_r_m3_p1, psi_i_m3_p1, a_i_m3_p1, psi_a_m3_p1); + prodsum_psi_a_epi16(psi_r_m3_m1, a_r_m3_m1, psi_i_m3_m1, a_i_m3_m1, psi_a_m3_m1); + prodsum_psi_a_epi16(psi_r_m3_m3, a_r_m3_m3, psi_i_m3_m3, a_i_m3_m3, psi_a_m3_m3); + prodsum_psi_a_epi16(psi_r_m3_m5, a_r_m3_m5, psi_i_m3_m5, a_i_m3_m5, psi_a_m3_m5); + prodsum_psi_a_epi16(psi_r_m3_m7, a_r_m3_m7, psi_i_m3_m7, a_i_m3_m7, psi_a_m3_m7); + prodsum_psi_a_epi16(psi_r_m5_p7, a_r_m5_p7, psi_i_m5_p7, a_i_m5_p7, psi_a_m5_p7); + prodsum_psi_a_epi16(psi_r_m5_p5, a_r_m5_p5, psi_i_m5_p5, a_i_m5_p5, psi_a_m5_p5); + prodsum_psi_a_epi16(psi_r_m5_p3, a_r_m5_p3, psi_i_m5_p3, a_i_m5_p3, psi_a_m5_p3); + prodsum_psi_a_epi16(psi_r_m5_p1, a_r_m5_p1, psi_i_m5_p1, a_i_m5_p1, psi_a_m5_p1); + prodsum_psi_a_epi16(psi_r_m5_m1, a_r_m5_m1, psi_i_m5_m1, a_i_m5_m1, psi_a_m5_m1); + prodsum_psi_a_epi16(psi_r_m5_m3, a_r_m5_m3, psi_i_m5_m3, a_i_m5_m3, psi_a_m5_m3); + prodsum_psi_a_epi16(psi_r_m5_m5, a_r_m5_m5, psi_i_m5_m5, a_i_m5_m5, psi_a_m5_m5); + prodsum_psi_a_epi16(psi_r_m5_m7, a_r_m5_m7, psi_i_m5_m7, a_i_m5_m7, psi_a_m5_m7); + prodsum_psi_a_epi16(psi_r_m7_p7, a_r_m7_p7, psi_i_m7_p7, a_i_m7_p7, psi_a_m7_p7); + prodsum_psi_a_epi16(psi_r_m7_p5, a_r_m7_p5, psi_i_m7_p5, a_i_m7_p5, psi_a_m7_p5); + prodsum_psi_a_epi16(psi_r_m7_p3, a_r_m7_p3, psi_i_m7_p3, a_i_m7_p3, psi_a_m7_p3); + prodsum_psi_a_epi16(psi_r_m7_p1, a_r_m7_p1, psi_i_m7_p1, a_i_m7_p1, psi_a_m7_p1); + prodsum_psi_a_epi16(psi_r_m7_m1, a_r_m7_m1, psi_i_m7_m1, a_i_m7_m1, psi_a_m7_m1); + prodsum_psi_a_epi16(psi_r_m7_m3, a_r_m7_m3, psi_i_m7_m3, a_i_m7_m3, psi_a_m7_m3); + prodsum_psi_a_epi16(psi_r_m7_m5, a_r_m7_m5, psi_i_m7_m5, a_i_m7_m5, psi_a_m7_m5); + prodsum_psi_a_epi16(psi_r_m7_m7, a_r_m7_m7, psi_i_m7_m7, a_i_m7_m7, psi_a_m7_m7); + + // Calculation of a group of two terms in the bit metric involving squares of interference + square_a_epi16(a_r_p7_p7, a_i_p7_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p7); + square_a_epi16(a_r_p7_p5, a_i_p7_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p5); + square_a_epi16(a_r_p7_p3, a_i_p7_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p3); + square_a_epi16(a_r_p7_p1, a_i_p7_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p1); + square_a_epi16(a_r_p7_m1, a_i_p7_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m1); + square_a_epi16(a_r_p7_m3, a_i_p7_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m3); + square_a_epi16(a_r_p7_m5, a_i_p7_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m5); + square_a_epi16(a_r_p7_m7, a_i_p7_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m7); + square_a_epi16(a_r_p5_p7, a_i_p5_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p7); + square_a_epi16(a_r_p5_p5, a_i_p5_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p5); + square_a_epi16(a_r_p5_p3, a_i_p5_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p3); + square_a_epi16(a_r_p5_p1, a_i_p5_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p1); + square_a_epi16(a_r_p5_m1, a_i_p5_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m1); + square_a_epi16(a_r_p5_m3, a_i_p5_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m3); + square_a_epi16(a_r_p5_m5, a_i_p5_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m5); + square_a_epi16(a_r_p5_m7, a_i_p5_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m7); + square_a_epi16(a_r_p3_p7, a_i_p3_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p7); + square_a_epi16(a_r_p3_p5, a_i_p3_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p5); + square_a_epi16(a_r_p3_p3, a_i_p3_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p3); + square_a_epi16(a_r_p3_p1, a_i_p3_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p1); + square_a_epi16(a_r_p3_m1, a_i_p3_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m1); + square_a_epi16(a_r_p3_m3, a_i_p3_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m3); + square_a_epi16(a_r_p3_m5, a_i_p3_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m5); + square_a_epi16(a_r_p3_m7, a_i_p3_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m7); + square_a_epi16(a_r_p1_p7, a_i_p1_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p7); + square_a_epi16(a_r_p1_p5, a_i_p1_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p5); + square_a_epi16(a_r_p1_p3, a_i_p1_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p3); + square_a_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p1); + square_a_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m1); + square_a_epi16(a_r_p1_m3, a_i_p1_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m3); + square_a_epi16(a_r_p1_m5, a_i_p1_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m5); + square_a_epi16(a_r_p1_m7, a_i_p1_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m7); + square_a_epi16(a_r_m1_p7, a_i_m1_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p7); + square_a_epi16(a_r_m1_p5, a_i_m1_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p5); + square_a_epi16(a_r_m1_p3, a_i_m1_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p3); + square_a_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p1); + square_a_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m1); + square_a_epi16(a_r_m1_m3, a_i_m1_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m3); + square_a_epi16(a_r_m1_m5, a_i_m1_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m5); + square_a_epi16(a_r_m1_m7, a_i_m1_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m7); + square_a_epi16(a_r_m3_p7, a_i_m3_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p7); + square_a_epi16(a_r_m3_p5, a_i_m3_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p5); + square_a_epi16(a_r_m3_p3, a_i_m3_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p3); + square_a_epi16(a_r_m3_p1, a_i_m3_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p1); + square_a_epi16(a_r_m3_m1, a_i_m3_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m1); + square_a_epi16(a_r_m3_m3, a_i_m3_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m3); + square_a_epi16(a_r_m3_m5, a_i_m3_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m5); + square_a_epi16(a_r_m3_m7, a_i_m3_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m7); + square_a_epi16(a_r_m5_p7, a_i_m5_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p7); + square_a_epi16(a_r_m5_p5, a_i_m5_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p5); + square_a_epi16(a_r_m5_p3, a_i_m5_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p3); + square_a_epi16(a_r_m5_p1, a_i_m5_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p1); + square_a_epi16(a_r_m5_m1, a_i_m5_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m1); + square_a_epi16(a_r_m5_m3, a_i_m5_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m3); + square_a_epi16(a_r_m5_m5, a_i_m5_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m5); + square_a_epi16(a_r_m5_m7, a_i_m5_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m7); + square_a_epi16(a_r_m7_p7, a_i_m7_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p7); + square_a_epi16(a_r_m7_p5, a_i_m7_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p5); + square_a_epi16(a_r_m7_p3, a_i_m7_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p3); + square_a_epi16(a_r_m7_p1, a_i_m7_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p1); + square_a_epi16(a_r_m7_m1, a_i_m7_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m1); + square_a_epi16(a_r_m7_m3, a_i_m7_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m3); + square_a_epi16(a_r_m7_m5, a_i_m7_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m5); + square_a_epi16(a_r_m7_m7, a_i_m7_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m7); + + // Computing different multiples of ||h0||^2 + // x=1, y=1 + ch_mag_2_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42); + ch_mag_2_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_2_over_42_with_sigma2,1); + // x=1, y=3 + ch_mag_10_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42); + ch_mag_10_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_10_over_42_with_sigma2,1); + // x=1, x=5 + ch_mag_26_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42); + ch_mag_26_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_26_over_42_with_sigma2,1); + // x=1, y=7 + ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); + ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); + // x=3, y=3 + ch_mag_18_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42); + ch_mag_18_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_18_over_42_with_sigma2,1); + // x=3, y=5 + ch_mag_34_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42); + ch_mag_34_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_34_over_42_with_sigma2,1); + // x=3, y=7 + ch_mag_58_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42); + ch_mag_58_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_58_over_42_with_sigma2,2); + // x=5, y=5 + ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); + ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); + // x=5, y=7 + ch_mag_74_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42); + ch_mag_74_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_74_over_42_with_sigma2,2); + // x=7, y=7 + ch_mag_98_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42); + ch_mag_98_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_98_over_42_with_sigma2,2); + + // Computing Metrics + xmm0 = _mm_subs_epi16(psi_a_p7_p7, a_sq_p7_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_7); + bit_met_p7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_p5, a_sq_p7_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_5); + bit_met_p7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_p3, a_sq_p7_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_3); + bit_met_p7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_p1, a_sq_p7_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_1); + bit_met_p7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m1, a_sq_p7_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_1); + bit_met_p7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m3, a_sq_p7_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_3); + bit_met_p7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m5, a_sq_p7_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_5); + bit_met_p7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m7, a_sq_p7_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_7); + bit_met_p7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p7, a_sq_p5_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_7); + bit_met_p5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p5, a_sq_p5_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_5); + bit_met_p5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p3, a_sq_p5_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_3); + bit_met_p5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p1, a_sq_p5_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_1); + bit_met_p5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m1, a_sq_p5_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_1); + bit_met_p5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m3, a_sq_p5_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_3); + bit_met_p5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m5, a_sq_p5_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_5); + bit_met_p5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m7, a_sq_p5_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_7); + bit_met_p5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p7, a_sq_p3_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_7); + bit_met_p3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p5, a_sq_p3_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_5); + bit_met_p3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p3, a_sq_p3_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_3); + bit_met_p3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p1, a_sq_p3_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_1); + bit_met_p3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m1, a_sq_p3_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_1); + bit_met_p3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m3, a_sq_p3_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_3); + bit_met_p3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m5, a_sq_p3_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_5); + bit_met_p3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m7, a_sq_p3_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_7); + bit_met_p3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p7, a_sq_p1_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_7); + bit_met_p1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p5, a_sq_p1_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_5); + bit_met_p1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p3, a_sq_p1_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_3); + bit_met_p1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_1); + bit_met_p1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_1); + bit_met_p1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m3, a_sq_p1_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_3); + bit_met_p1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m5, a_sq_p1_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_5); + bit_met_p1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m7, a_sq_p1_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_7); + bit_met_p1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + + xmm0 = _mm_subs_epi16(psi_a_m1_p7, a_sq_m1_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_7); + bit_met_m1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_p5, a_sq_m1_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_5); + bit_met_m1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_p3, a_sq_m1_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_3); + bit_met_m1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_1); + bit_met_m1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_1); + bit_met_m1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m3, a_sq_m1_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_3); + bit_met_m1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m5, a_sq_m1_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_5); + bit_met_m1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m7, a_sq_m1_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_7); + bit_met_m1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p7, a_sq_m3_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_7); + bit_met_m3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p5, a_sq_m3_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_5); + bit_met_m3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p3, a_sq_m3_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_3); + bit_met_m3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p1, a_sq_m3_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_1); + bit_met_m3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m1, a_sq_m3_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_1); + bit_met_m3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m3, a_sq_m3_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_3); + bit_met_m3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m5, a_sq_m3_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_5); + bit_met_m3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m7, a_sq_m3_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_7); + bit_met_m3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p7, a_sq_m5_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_7); + bit_met_m5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p5, a_sq_m5_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_5); + bit_met_m5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p3, a_sq_m5_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_3); + bit_met_m5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p1, a_sq_m5_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_1); + bit_met_m5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m1, a_sq_m5_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_1); + bit_met_m5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m3, a_sq_m5_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_3); + bit_met_m5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m5, a_sq_m5_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_5); + bit_met_m5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m7, a_sq_m5_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_7); + bit_met_m5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p7, a_sq_m7_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_7); + bit_met_m7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p5, a_sq_m7_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_5); + bit_met_m7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p3, a_sq_m7_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_3); + bit_met_m7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p1, a_sq_m7_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_1); + bit_met_m7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m1, a_sq_m7_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_1); + bit_met_m7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m3, a_sq_m7_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_3); + bit_met_m7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m5, a_sq_m7_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_5); + bit_met_m7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m7, a_sq_m7_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_7); + bit_met_m7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + + // Detection for 1st bit (LTE mapping) + // bit = 1 + xmm0 = _mm_max_epi16(bit_met_m7_p7, bit_met_m7_p5); + xmm1 = _mm_max_epi16(bit_met_m7_p3, bit_met_m7_p1); + xmm2 = _mm_max_epi16(bit_met_m7_m1, bit_met_m7_m3); + xmm3 = _mm_max_epi16(bit_met_m7_m5, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m5_p7, bit_met_m5_p5); + xmm1 = _mm_max_epi16(bit_met_m5_p3, bit_met_m5_p1); + xmm2 = _mm_max_epi16(bit_met_m5_m1, bit_met_m5_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m5_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_m3_p7, bit_met_m3_p5); + xmm1 = _mm_max_epi16(bit_met_m3_p3, bit_met_m3_p1); + xmm2 = _mm_max_epi16(bit_met_m3_m1, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m3_m5, bit_met_m3_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_p7, bit_met_m1_p5); + xmm1 = _mm_max_epi16(bit_met_m1_p3, bit_met_m1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m1_m3); + xmm3 = _mm_max_epi16(bit_met_m1_m5, bit_met_m1_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + // bit = 0 + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p7_p5); + xmm1 = _mm_max_epi16(bit_met_p7_p3, bit_met_p7_p1); + xmm2 = _mm_max_epi16(bit_met_p7_m1, bit_met_p7_m3); + xmm3 = _mm_max_epi16(bit_met_p7_m5, bit_met_p7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_p7, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p5_p3, bit_met_p5_p1); + xmm2 = _mm_max_epi16(bit_met_p5_m1, bit_met_p5_m3); + xmm3 = _mm_max_epi16(bit_met_p5_m5, bit_met_p5_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_p7, bit_met_p3_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p3_p1); + xmm2 = _mm_max_epi16(bit_met_p3_m1, bit_met_p3_m3); + xmm3 = _mm_max_epi16(bit_met_p3_m5, bit_met_p3_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_p7, bit_met_p1_p5); + xmm1 = _mm_max_epi16(bit_met_p1_p3, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_p1_m1, bit_met_p1_m3); + xmm3 = _mm_max_epi16(bit_met_p1_m5, bit_met_p1_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y0r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 2nd bit (LTE mapping) + // bit = 1 + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + // bit = 0 + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y1r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 3rd bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); + xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); + xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); + xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); + xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); + xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); + xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); + xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); + xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); + xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); + xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); + xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); + xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); + xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); + xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y2r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 4th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y0i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + + // Detection for 5th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); + xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); + xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); + xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); + xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); + xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); + xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); + xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); + xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); + xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); + xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); + xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); + xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); + xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); + xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y1i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 6th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y2i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + + // map to output stream, difficult to do in SIMD since we have 6 16bit LLRs + // RE 1 + j = 24*i; + stream0_out[j + 0] = ((short *)&y0r)[0]; + stream0_out[j + 1] = ((short *)&y1r)[0]; + stream0_out[j + 2] = ((short *)&y2r)[0]; + stream0_out[j + 3] = ((short *)&y0i)[0]; + stream0_out[j + 4] = ((short *)&y1i)[0]; + stream0_out[j + 5] = ((short *)&y2i)[0]; + // RE 2 + stream0_out[j + 6] = ((short *)&y0r)[1]; + stream0_out[j + 7] = ((short *)&y1r)[1]; + stream0_out[j + 8] = ((short *)&y2r)[1]; + stream0_out[j + 9] = ((short *)&y0i)[1]; + stream0_out[j + 10] = ((short *)&y1i)[1]; + stream0_out[j + 11] = ((short *)&y2i)[1]; + // RE 3 + stream0_out[j + 12] = ((short *)&y0r)[2]; + stream0_out[j + 13] = ((short *)&y1r)[2]; + stream0_out[j + 14] = ((short *)&y2r)[2]; + stream0_out[j + 15] = ((short *)&y0i)[2]; + stream0_out[j + 16] = ((short *)&y1i)[2]; + stream0_out[j + 17] = ((short *)&y2i)[2]; + // RE 4 + stream0_out[j + 18] = ((short *)&y0r)[3]; + stream0_out[j + 19] = ((short *)&y1r)[3]; + stream0_out[j + 20] = ((short *)&y2r)[3]; + stream0_out[j + 21] = ((short *)&y0i)[3]; + stream0_out[j + 22] = ((short *)&y1i)[3]; + stream0_out[j + 23] = ((short *)&y2i)[3]; + // RE 5 + stream0_out[j + 24] = ((short *)&y0r)[4]; + stream0_out[j + 25] = ((short *)&y1r)[4]; + stream0_out[j + 26] = ((short *)&y2r)[4]; + stream0_out[j + 27] = ((short *)&y0i)[4]; + stream0_out[j + 28] = ((short *)&y1i)[4]; + stream0_out[j + 29] = ((short *)&y2i)[4]; + // RE 6 + stream0_out[j + 30] = ((short *)&y0r)[5]; + stream0_out[j + 31] = ((short *)&y1r)[5]; + stream0_out[j + 32] = ((short *)&y2r)[5]; + stream0_out[j + 33] = ((short *)&y0i)[5]; + stream0_out[j + 34] = ((short *)&y1i)[5]; + stream0_out[j + 35] = ((short *)&y2i)[5]; + // RE 7 + stream0_out[j + 36] = ((short *)&y0r)[6]; + stream0_out[j + 37] = ((short *)&y1r)[6]; + stream0_out[j + 38] = ((short *)&y2r)[6]; + stream0_out[j + 39] = ((short *)&y0i)[6]; + stream0_out[j + 40] = ((short *)&y1i)[6]; + stream0_out[j + 41] = ((short *)&y2i)[6]; + // RE 8 + stream0_out[j + 42] = ((short *)&y0r)[7]; + stream0_out[j + 43] = ((short *)&y1r)[7]; + stream0_out[j + 44] = ((short *)&y2r)[7]; + stream0_out[j + 45] = ((short *)&y0i)[7]; + stream0_out[j + 46] = ((short *)&y1i)[7]; + stream0_out[j + 47] = ((short *)&y2i)[7]; + +#elif defined(__arm__) + +#endif + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif + +} + + +int nr_dlsch_64qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag, + int32_t **dl_ch_mag_i, + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + int16_t **llr16p) +{ + + int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + //first symbol has different structure due to more pilots + if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + } + + AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + qam64_qam16((short *)rxF, + (short *)rxF_i, + (short *)ch_mag, + (short *)ch_mag_i, + (short *)llr16, + (short *)rho, + len); + + llr16 += (6*len); + *llr16p = (short *)llr16; + return(0); +} + +void qam64_qam64(short *stream0_in, + short *stream1_in, + short *ch_mag, + short *ch_mag_i, + short *stream0_out, + short *rho01, + int length + ) +{ + + /* + Author: S. Wagner + Date: 31-07-12 + + Input: + stream0_in: MF filter for 1st stream, i.e., y0=h0'*y + stream1_in: MF filter for 2nd stream, i.e., y1=h1'*y + ch_mag: 4*h0/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + ch_mag_i: 4*h1/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc + rho01: Channel cross correlation, i.e., h1'*h0 + + Output: + stream0_out: output LLRs for 1st stream + */ + +#if defined(__x86_64__) || defined(__i386__) + + __m128i *rho01_128i = (__m128i *)rho01; + __m128i *stream0_128i_in = (__m128i *)stream0_in; + __m128i *stream1_128i_in = (__m128i *)stream1_in; + __m128i *ch_mag_128i = (__m128i *)ch_mag; + __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; + + __m128i ONE_OVER_SQRT_42 = _mm_set1_epi16(10112); // round(1/sqrt(42)*2^16) + __m128i THREE_OVER_SQRT_42 = _mm_set1_epi16(30337); // round(3/sqrt(42)*2^16) + __m128i FIVE_OVER_SQRT_42 = _mm_set1_epi16(25281); // round(5/sqrt(42)*2^15) + __m128i SEVEN_OVER_SQRT_42 = _mm_set1_epi16(17697); // round(7/sqrt(42)*2^14) Q2.14 + __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) + __m128i ONE_OVER_SQRT_2_42 = _mm_set1_epi16(3575); // round(1/sqrt(2*42)*2^15) + __m128i THREE_OVER_SQRT_2_42 = _mm_set1_epi16(10726); // round(3/sqrt(2*42)*2^15) + __m128i FIVE_OVER_SQRT_2_42 = _mm_set1_epi16(17876); // round(5/sqrt(2*42)*2^15) + __m128i SEVEN_OVER_SQRT_2_42 = _mm_set1_epi16(25027); // round(7/sqrt(2*42)*2^15) + __m128i FORTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(30969); // round(49/(4*sqrt(42))*2^14), Q2.14 + __m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(23385); // round(37/(4*sqrt(42))*2^14), Q2.14 + __m128i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(31601); // round(25/(4*sqrt(42))*2^15) + __m128i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(18329); // round(29/(4*sqrt(42))*2^15), Q2.14 + __m128i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(21489); // round(17/(4*sqrt(42))*2^15) + __m128i NINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(11376); // round(9/(4*sqrt(42))*2^15) + __m128i THIRTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(16433); // round(13/(4*sqrt(42))*2^15) + __m128i FIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(6320); // round(5/(4*sqrt(42))*2^15) + __m128i ONE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(1264); // round(1/(4*sqrt(42))*2^15) + __m128i SQRT_42_OVER_FOUR = _mm_set1_epi16(13272); // round(sqrt(42)/4*2^13), Q3.12 + + __m128i ch_mag_des; + __m128i ch_mag_int; + __m128i ch_mag_98_over_42_with_sigma2; + __m128i ch_mag_74_over_42_with_sigma2; + __m128i ch_mag_58_over_42_with_sigma2; + __m128i ch_mag_50_over_42_with_sigma2; + __m128i ch_mag_34_over_42_with_sigma2; + __m128i ch_mag_18_over_42_with_sigma2; + __m128i ch_mag_26_over_42_with_sigma2; + __m128i ch_mag_10_over_42_with_sigma2; + __m128i ch_mag_2_over_42_with_sigma2; + __m128i y0r_one_over_sqrt_21; + __m128i y0r_three_over_sqrt_21; + __m128i y0r_five_over_sqrt_21; + __m128i y0r_seven_over_sqrt_21; + __m128i y0i_one_over_sqrt_21; + __m128i y0i_three_over_sqrt_21; + __m128i y0i_five_over_sqrt_21; + __m128i y0i_seven_over_sqrt_21; + __m128i ch_mag_int_with_sigma2; + __m128i two_ch_mag_int_with_sigma2; + __m128i three_ch_mag_int_with_sigma2; +#elif defined(__arm__) + +#endif + + int i,j; + + + for (i=0; i>2; i+=2) { + +#if defined(__x86_64__) || defined(__i386__) + + // Get rho + xmm0 = rho01_128i[i]; + xmm1 = rho01_128i[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) + xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) + rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) + rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + + // Compute the different rhos + rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42); + rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42); + rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42); + rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42); + rho_rpi_5_5 = _mm_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42); + rho_rmi_5_5 = _mm_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42); + rho_rpi_7_7 = _mm_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42); + rho_rmi_7_7 = _mm_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42); + + rho_rpi_5_5 = _mm_slli_epi16(rho_rpi_5_5, 1); + rho_rmi_5_5 = _mm_slli_epi16(rho_rmi_5_5, 1); + rho_rpi_7_7 = _mm_slli_epi16(rho_rpi_7_7, 2); + rho_rmi_7_7 = _mm_slli_epi16(rho_rmi_7_7, 2); + + xmm4 = _mm_mulhi_epi16(xmm2, ONE_OVER_SQRT_42); + xmm5 = _mm_mulhi_epi16(xmm3, ONE_OVER_SQRT_42); + xmm6 = _mm_mulhi_epi16(xmm3, THREE_OVER_SQRT_42); + xmm7 = _mm_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42); + xmm8 = _mm_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42); + xmm7 = _mm_slli_epi16(xmm7, 1); + xmm8 = _mm_slli_epi16(xmm8, 2); + + rho_rpi_1_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_1_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_1_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_1_5 = _mm_subs_epi16(xmm4, xmm7); + rho_rpi_1_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_1_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, THREE_OVER_SQRT_42); + rho_rpi_3_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_3_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_3_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_3_5 = _mm_subs_epi16(xmm4, xmm7); + rho_rpi_3_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_3_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42); + xmm4 = _mm_slli_epi16(xmm4, 1); + rho_rpi_5_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_5_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_5_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_5_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_5_7 = _mm_adds_epi16(xmm4, xmm8); + rho_rmi_5_7 = _mm_subs_epi16(xmm4, xmm8); + + xmm4 = _mm_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42); + xmm4 = _mm_slli_epi16(xmm4, 2); + rho_rpi_7_1 = _mm_adds_epi16(xmm4, xmm5); + rho_rmi_7_1 = _mm_subs_epi16(xmm4, xmm5); + rho_rpi_7_3 = _mm_adds_epi16(xmm4, xmm6); + rho_rmi_7_3 = _mm_subs_epi16(xmm4, xmm6); + rho_rpi_7_5 = _mm_adds_epi16(xmm4, xmm7); + rho_rmi_7_5 = _mm_subs_epi16(xmm4, xmm7); + + // Rearrange interfering MF output + xmm0 = stream1_128i_in[i]; + xmm1 = stream1_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + + // Psi_r calculation from rho_rpi or rho_rmi + xmm0 = _mm_setzero_si128(); // ZERO for abs_pi16 + xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1r); + psi_r_p7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1r); + psi_r_p7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1r); + psi_r_p7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1r); + psi_r_p7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1r); + psi_r_p7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1r); + psi_r_p7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1r); + psi_r_p7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1r); + psi_r_p7_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1r); + psi_r_p5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1r); + psi_r_p5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1r); + psi_r_p5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1r); + psi_r_p5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1r); + psi_r_p5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1r); + psi_r_p5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1r); + psi_r_p5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1r); + psi_r_p5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1r); + psi_r_p3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1r); + psi_r_p3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1r); + psi_r_p3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1r); + psi_r_p3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1r); + psi_r_p3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1r); + psi_r_p3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1r); + psi_r_p3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1r); + psi_r_p3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1r); + psi_r_p1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1r); + psi_r_p1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1r); + psi_r_p1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1r); + psi_r_p1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1r); + psi_r_p1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1r); + psi_r_p1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1r); + psi_r_p1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1r); + psi_r_p1_m7 = _mm_abs_epi16(xmm2); + + xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1r); + psi_r_m1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1r); + psi_r_m1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1r); + psi_r_m1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1r); + psi_r_m1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1r); + psi_r_m1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1r); + psi_r_m1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1r); + psi_r_m1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1r); + psi_r_m1_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1r); + psi_r_m3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1r); + psi_r_m3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1r); + psi_r_m3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1r); + psi_r_m3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1r); + psi_r_m3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1r); + psi_r_m3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1r); + psi_r_m3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1r); + psi_r_m3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1r); + psi_r_m5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1r); + psi_r_m5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1r); + psi_r_m5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1r); + psi_r_m5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1r); + psi_r_m5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1r); + psi_r_m5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1r); + psi_r_m5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1r); + psi_r_m5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1r); + psi_r_m7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1r); + psi_r_m7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1r); + psi_r_m7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1r); + psi_r_m7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1r); + psi_r_m7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1r); + psi_r_m7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1r); + psi_r_m7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1r); + psi_r_m7_m7 = _mm_abs_epi16(xmm2); + + // Psi_i calculation from rho_rpi or rho_rmi + xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1i); + psi_i_p7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1i); + psi_i_p7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1i); + psi_i_p7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1i); + psi_i_p7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1i); + psi_i_p7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1i); + psi_i_p7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1i); + psi_i_p7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1i); + psi_i_p7_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1i); + psi_i_p5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1i); + psi_i_p5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1i); + psi_i_p5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1i); + psi_i_p5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1i); + psi_i_p5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1i); + psi_i_p5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1i); + psi_i_p5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1i); + psi_i_p5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1i); + psi_i_p3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1i); + psi_i_p3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1i); + psi_i_p3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1i); + psi_i_p3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1i); + psi_i_p3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1i); + psi_i_p3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1i); + psi_i_p3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1i); + psi_i_p3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1i); + psi_i_p1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1i); + psi_i_p1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1i); + psi_i_p1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1i); + psi_i_p1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1i); + psi_i_p1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1i); + psi_i_p1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1i); + psi_i_p1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1i); + psi_i_p1_m7 = _mm_abs_epi16(xmm2); + + xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1i); + psi_i_m1_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1i); + psi_i_m1_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1i); + psi_i_m1_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1i); + psi_i_m1_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1i); + psi_i_m1_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1i); + psi_i_m1_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1i); + psi_i_m1_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1i); + psi_i_m1_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1i); + psi_i_m3_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1i); + psi_i_m3_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1i); + psi_i_m3_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1i); + psi_i_m3_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1i); + psi_i_m3_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1i); + psi_i_m3_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1i); + psi_i_m3_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1i); + psi_i_m3_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1i); + psi_i_m5_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1i); + psi_i_m5_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1i); + psi_i_m5_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1i); + psi_i_m5_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1i); + psi_i_m5_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1i); + psi_i_m5_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1i); + psi_i_m5_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1i); + psi_i_m5_m7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1i); + psi_i_m7_p7 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1i); + psi_i_m7_p5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1i); + psi_i_m7_p3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1i); + psi_i_m7_p1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1i); + psi_i_m7_m1 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1i); + psi_i_m7_m3 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1i); + psi_i_m7_m5 = _mm_abs_epi16(xmm2); + xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1i); + psi_i_m7_m7 = _mm_abs_epi16(xmm2); + + + // Rearrange desired MF output + xmm0 = stream0_128i_in[i]; + xmm1 = stream0_128i_in[i+1]; + xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); + //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] + //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] + y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + y0i = _mm_unpackhi_epi64(xmm0,xmm1); + + // Rearrange desired channel magnitudes + xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) + xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); + + // Rearrange interfering channel magnitudes + xmm2 = ch_mag_128i_i[i]; + xmm3 = ch_mag_128i_i[i+1]; + xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); + ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); + + y0r_one_over_sqrt_21 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_42); + y0r_three_over_sqrt_21 = _mm_mulhi_epi16(y0r, THREE_OVER_SQRT_42); + y0r_five_over_sqrt_21 = _mm_mulhi_epi16(y0r, FIVE_OVER_SQRT_42); + y0r_five_over_sqrt_21 = _mm_slli_epi16(y0r_five_over_sqrt_21, 1); + y0r_seven_over_sqrt_21 = _mm_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42); + y0r_seven_over_sqrt_21 = _mm_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14 + + y0i_one_over_sqrt_21 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_42); + y0i_three_over_sqrt_21 = _mm_mulhi_epi16(y0i, THREE_OVER_SQRT_42); + y0i_five_over_sqrt_21 = _mm_mulhi_epi16(y0i, FIVE_OVER_SQRT_42); + y0i_five_over_sqrt_21 = _mm_slli_epi16(y0i_five_over_sqrt_21, 1); + y0i_seven_over_sqrt_21 = _mm_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42); + y0i_seven_over_sqrt_21 = _mm_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14 + + y0_p_7_1 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_7_3 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_7_5 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_7_7 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_5_1 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_5_3 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_5_5 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_5_7 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_3_5 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_3_7 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_p_1_1 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); + y0_p_1_3 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); + y0_p_1_5 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); + y0_p_1_7 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); + + y0_m_1_1 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_1_3 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_1_5 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_1_7 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_3_5 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_3_7 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_5_1 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_5_3 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_5_5 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_5_7 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); + y0_m_7_1 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); + y0_m_7_3 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); + y0_m_7_5 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); + y0_m_7_7 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); + + // Detection of interference term + ch_mag_int_with_sigma2 = _mm_srai_epi16(ch_mag_int, 1); // *2 + two_ch_mag_int_with_sigma2 = ch_mag_int; // *4 + three_ch_mag_int_with_sigma2 = _mm_adds_epi16(ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2); // *6 + + interference_abs_64qam_epi16(psi_r_p7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_p1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_r_m7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + + interference_abs_64qam_epi16(psi_i_p7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_p1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + interference_abs_64qam_epi16(psi_i_m7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, + SEVEN_OVER_SQRT_2_42); + + // Calculation of a group of two terms in the bit metric involving product of psi and interference + prodsum_psi_a_epi16(psi_r_p7_p7, a_r_p7_p7, psi_i_p7_p7, a_i_p7_p7, psi_a_p7_p7); + prodsum_psi_a_epi16(psi_r_p7_p5, a_r_p7_p5, psi_i_p7_p5, a_i_p7_p5, psi_a_p7_p5); + prodsum_psi_a_epi16(psi_r_p7_p3, a_r_p7_p3, psi_i_p7_p3, a_i_p7_p3, psi_a_p7_p3); + prodsum_psi_a_epi16(psi_r_p7_p1, a_r_p7_p1, psi_i_p7_p1, a_i_p7_p1, psi_a_p7_p1); + prodsum_psi_a_epi16(psi_r_p7_m1, a_r_p7_m1, psi_i_p7_m1, a_i_p7_m1, psi_a_p7_m1); + prodsum_psi_a_epi16(psi_r_p7_m3, a_r_p7_m3, psi_i_p7_m3, a_i_p7_m3, psi_a_p7_m3); + prodsum_psi_a_epi16(psi_r_p7_m5, a_r_p7_m5, psi_i_p7_m5, a_i_p7_m5, psi_a_p7_m5); + prodsum_psi_a_epi16(psi_r_p7_m7, a_r_p7_m7, psi_i_p7_m7, a_i_p7_m7, psi_a_p7_m7); + prodsum_psi_a_epi16(psi_r_p5_p7, a_r_p5_p7, psi_i_p5_p7, a_i_p5_p7, psi_a_p5_p7); + prodsum_psi_a_epi16(psi_r_p5_p5, a_r_p5_p5, psi_i_p5_p5, a_i_p5_p5, psi_a_p5_p5); + prodsum_psi_a_epi16(psi_r_p5_p3, a_r_p5_p3, psi_i_p5_p3, a_i_p5_p3, psi_a_p5_p3); + prodsum_psi_a_epi16(psi_r_p5_p1, a_r_p5_p1, psi_i_p5_p1, a_i_p5_p1, psi_a_p5_p1); + prodsum_psi_a_epi16(psi_r_p5_m1, a_r_p5_m1, psi_i_p5_m1, a_i_p5_m1, psi_a_p5_m1); + prodsum_psi_a_epi16(psi_r_p5_m3, a_r_p5_m3, psi_i_p5_m3, a_i_p5_m3, psi_a_p5_m3); + prodsum_psi_a_epi16(psi_r_p5_m5, a_r_p5_m5, psi_i_p5_m5, a_i_p5_m5, psi_a_p5_m5); + prodsum_psi_a_epi16(psi_r_p5_m7, a_r_p5_m7, psi_i_p5_m7, a_i_p5_m7, psi_a_p5_m7); + prodsum_psi_a_epi16(psi_r_p3_p7, a_r_p3_p7, psi_i_p3_p7, a_i_p3_p7, psi_a_p3_p7); + prodsum_psi_a_epi16(psi_r_p3_p5, a_r_p3_p5, psi_i_p3_p5, a_i_p3_p5, psi_a_p3_p5); + prodsum_psi_a_epi16(psi_r_p3_p3, a_r_p3_p3, psi_i_p3_p3, a_i_p3_p3, psi_a_p3_p3); + prodsum_psi_a_epi16(psi_r_p3_p1, a_r_p3_p1, psi_i_p3_p1, a_i_p3_p1, psi_a_p3_p1); + prodsum_psi_a_epi16(psi_r_p3_m1, a_r_p3_m1, psi_i_p3_m1, a_i_p3_m1, psi_a_p3_m1); + prodsum_psi_a_epi16(psi_r_p3_m3, a_r_p3_m3, psi_i_p3_m3, a_i_p3_m3, psi_a_p3_m3); + prodsum_psi_a_epi16(psi_r_p3_m5, a_r_p3_m5, psi_i_p3_m5, a_i_p3_m5, psi_a_p3_m5); + prodsum_psi_a_epi16(psi_r_p3_m7, a_r_p3_m7, psi_i_p3_m7, a_i_p3_m7, psi_a_p3_m7); + prodsum_psi_a_epi16(psi_r_p1_p7, a_r_p1_p7, psi_i_p1_p7, a_i_p1_p7, psi_a_p1_p7); + prodsum_psi_a_epi16(psi_r_p1_p5, a_r_p1_p5, psi_i_p1_p5, a_i_p1_p5, psi_a_p1_p5); + prodsum_psi_a_epi16(psi_r_p1_p3, a_r_p1_p3, psi_i_p1_p3, a_i_p1_p3, psi_a_p1_p3); + prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); + prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); + prodsum_psi_a_epi16(psi_r_p1_m3, a_r_p1_m3, psi_i_p1_m3, a_i_p1_m3, psi_a_p1_m3); + prodsum_psi_a_epi16(psi_r_p1_m5, a_r_p1_m5, psi_i_p1_m5, a_i_p1_m5, psi_a_p1_m5); + prodsum_psi_a_epi16(psi_r_p1_m7, a_r_p1_m7, psi_i_p1_m7, a_i_p1_m7, psi_a_p1_m7); + prodsum_psi_a_epi16(psi_r_m1_p7, a_r_m1_p7, psi_i_m1_p7, a_i_m1_p7, psi_a_m1_p7); + prodsum_psi_a_epi16(psi_r_m1_p5, a_r_m1_p5, psi_i_m1_p5, a_i_m1_p5, psi_a_m1_p5); + prodsum_psi_a_epi16(psi_r_m1_p3, a_r_m1_p3, psi_i_m1_p3, a_i_m1_p3, psi_a_m1_p3); + prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); + prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); + prodsum_psi_a_epi16(psi_r_m1_m3, a_r_m1_m3, psi_i_m1_m3, a_i_m1_m3, psi_a_m1_m3); + prodsum_psi_a_epi16(psi_r_m1_m5, a_r_m1_m5, psi_i_m1_m5, a_i_m1_m5, psi_a_m1_m5); + prodsum_psi_a_epi16(psi_r_m1_m7, a_r_m1_m7, psi_i_m1_m7, a_i_m1_m7, psi_a_m1_m7); + prodsum_psi_a_epi16(psi_r_m3_p7, a_r_m3_p7, psi_i_m3_p7, a_i_m3_p7, psi_a_m3_p7); + prodsum_psi_a_epi16(psi_r_m3_p5, a_r_m3_p5, psi_i_m3_p5, a_i_m3_p5, psi_a_m3_p5); + prodsum_psi_a_epi16(psi_r_m3_p3, a_r_m3_p3, psi_i_m3_p3, a_i_m3_p3, psi_a_m3_p3); + prodsum_psi_a_epi16(psi_r_m3_p1, a_r_m3_p1, psi_i_m3_p1, a_i_m3_p1, psi_a_m3_p1); + prodsum_psi_a_epi16(psi_r_m3_m1, a_r_m3_m1, psi_i_m3_m1, a_i_m3_m1, psi_a_m3_m1); + prodsum_psi_a_epi16(psi_r_m3_m3, a_r_m3_m3, psi_i_m3_m3, a_i_m3_m3, psi_a_m3_m3); + prodsum_psi_a_epi16(psi_r_m3_m5, a_r_m3_m5, psi_i_m3_m5, a_i_m3_m5, psi_a_m3_m5); + prodsum_psi_a_epi16(psi_r_m3_m7, a_r_m3_m7, psi_i_m3_m7, a_i_m3_m7, psi_a_m3_m7); + prodsum_psi_a_epi16(psi_r_m5_p7, a_r_m5_p7, psi_i_m5_p7, a_i_m5_p7, psi_a_m5_p7); + prodsum_psi_a_epi16(psi_r_m5_p5, a_r_m5_p5, psi_i_m5_p5, a_i_m5_p5, psi_a_m5_p5); + prodsum_psi_a_epi16(psi_r_m5_p3, a_r_m5_p3, psi_i_m5_p3, a_i_m5_p3, psi_a_m5_p3); + prodsum_psi_a_epi16(psi_r_m5_p1, a_r_m5_p1, psi_i_m5_p1, a_i_m5_p1, psi_a_m5_p1); + prodsum_psi_a_epi16(psi_r_m5_m1, a_r_m5_m1, psi_i_m5_m1, a_i_m5_m1, psi_a_m5_m1); + prodsum_psi_a_epi16(psi_r_m5_m3, a_r_m5_m3, psi_i_m5_m3, a_i_m5_m3, psi_a_m5_m3); + prodsum_psi_a_epi16(psi_r_m5_m5, a_r_m5_m5, psi_i_m5_m5, a_i_m5_m5, psi_a_m5_m5); + prodsum_psi_a_epi16(psi_r_m5_m7, a_r_m5_m7, psi_i_m5_m7, a_i_m5_m7, psi_a_m5_m7); + prodsum_psi_a_epi16(psi_r_m7_p7, a_r_m7_p7, psi_i_m7_p7, a_i_m7_p7, psi_a_m7_p7); + prodsum_psi_a_epi16(psi_r_m7_p5, a_r_m7_p5, psi_i_m7_p5, a_i_m7_p5, psi_a_m7_p5); + prodsum_psi_a_epi16(psi_r_m7_p3, a_r_m7_p3, psi_i_m7_p3, a_i_m7_p3, psi_a_m7_p3); + prodsum_psi_a_epi16(psi_r_m7_p1, a_r_m7_p1, psi_i_m7_p1, a_i_m7_p1, psi_a_m7_p1); + prodsum_psi_a_epi16(psi_r_m7_m1, a_r_m7_m1, psi_i_m7_m1, a_i_m7_m1, psi_a_m7_m1); + prodsum_psi_a_epi16(psi_r_m7_m3, a_r_m7_m3, psi_i_m7_m3, a_i_m7_m3, psi_a_m7_m3); + prodsum_psi_a_epi16(psi_r_m7_m5, a_r_m7_m5, psi_i_m7_m5, a_i_m7_m5, psi_a_m7_m5); + prodsum_psi_a_epi16(psi_r_m7_m7, a_r_m7_m7, psi_i_m7_m7, a_i_m7_m7, psi_a_m7_m7); + + // Multiply by sqrt(2) + psi_a_p7_p7 = _mm_mulhi_epi16(psi_a_p7_p7, ONE_OVER_SQRT_2); + psi_a_p7_p7 = _mm_slli_epi16(psi_a_p7_p7, 2); + psi_a_p7_p5 = _mm_mulhi_epi16(psi_a_p7_p5, ONE_OVER_SQRT_2); + psi_a_p7_p5 = _mm_slli_epi16(psi_a_p7_p5, 2); + psi_a_p7_p3 = _mm_mulhi_epi16(psi_a_p7_p3, ONE_OVER_SQRT_2); + psi_a_p7_p3 = _mm_slli_epi16(psi_a_p7_p3, 2); + psi_a_p7_p1 = _mm_mulhi_epi16(psi_a_p7_p1, ONE_OVER_SQRT_2); + psi_a_p7_p1 = _mm_slli_epi16(psi_a_p7_p1, 2); + psi_a_p7_m1 = _mm_mulhi_epi16(psi_a_p7_m1, ONE_OVER_SQRT_2); + psi_a_p7_m1 = _mm_slli_epi16(psi_a_p7_m1, 2); + psi_a_p7_m3 = _mm_mulhi_epi16(psi_a_p7_m3, ONE_OVER_SQRT_2); + psi_a_p7_m3 = _mm_slli_epi16(psi_a_p7_m3, 2); + psi_a_p7_m5 = _mm_mulhi_epi16(psi_a_p7_m5, ONE_OVER_SQRT_2); + psi_a_p7_m5 = _mm_slli_epi16(psi_a_p7_m5, 2); + psi_a_p7_m7 = _mm_mulhi_epi16(psi_a_p7_m7, ONE_OVER_SQRT_2); + psi_a_p7_m7 = _mm_slli_epi16(psi_a_p7_m7, 2); + psi_a_p5_p7 = _mm_mulhi_epi16(psi_a_p5_p7, ONE_OVER_SQRT_2); + psi_a_p5_p7 = _mm_slli_epi16(psi_a_p5_p7, 2); + psi_a_p5_p5 = _mm_mulhi_epi16(psi_a_p5_p5, ONE_OVER_SQRT_2); + psi_a_p5_p5 = _mm_slli_epi16(psi_a_p5_p5, 2); + psi_a_p5_p3 = _mm_mulhi_epi16(psi_a_p5_p3, ONE_OVER_SQRT_2); + psi_a_p5_p3 = _mm_slli_epi16(psi_a_p5_p3, 2); + psi_a_p5_p1 = _mm_mulhi_epi16(psi_a_p5_p1, ONE_OVER_SQRT_2); + psi_a_p5_p1 = _mm_slli_epi16(psi_a_p5_p1, 2); + psi_a_p5_m1 = _mm_mulhi_epi16(psi_a_p5_m1, ONE_OVER_SQRT_2); + psi_a_p5_m1 = _mm_slli_epi16(psi_a_p5_m1, 2); + psi_a_p5_m3 = _mm_mulhi_epi16(psi_a_p5_m3, ONE_OVER_SQRT_2); + psi_a_p5_m3 = _mm_slli_epi16(psi_a_p5_m3, 2); + psi_a_p5_m5 = _mm_mulhi_epi16(psi_a_p5_m5, ONE_OVER_SQRT_2); + psi_a_p5_m5 = _mm_slli_epi16(psi_a_p5_m5, 2); + psi_a_p5_m7 = _mm_mulhi_epi16(psi_a_p5_m7, ONE_OVER_SQRT_2); + psi_a_p5_m7 = _mm_slli_epi16(psi_a_p5_m7, 2); + psi_a_p3_p7 = _mm_mulhi_epi16(psi_a_p3_p7, ONE_OVER_SQRT_2); + psi_a_p3_p7 = _mm_slli_epi16(psi_a_p3_p7, 2); + psi_a_p3_p5 = _mm_mulhi_epi16(psi_a_p3_p5, ONE_OVER_SQRT_2); + psi_a_p3_p5 = _mm_slli_epi16(psi_a_p3_p5, 2); + psi_a_p3_p3 = _mm_mulhi_epi16(psi_a_p3_p3, ONE_OVER_SQRT_2); + psi_a_p3_p3 = _mm_slli_epi16(psi_a_p3_p3, 2); + psi_a_p3_p1 = _mm_mulhi_epi16(psi_a_p3_p1, ONE_OVER_SQRT_2); + psi_a_p3_p1 = _mm_slli_epi16(psi_a_p3_p1, 2); + psi_a_p3_m1 = _mm_mulhi_epi16(psi_a_p3_m1, ONE_OVER_SQRT_2); + psi_a_p3_m1 = _mm_slli_epi16(psi_a_p3_m1, 2); + psi_a_p3_m3 = _mm_mulhi_epi16(psi_a_p3_m3, ONE_OVER_SQRT_2); + psi_a_p3_m3 = _mm_slli_epi16(psi_a_p3_m3, 2); + psi_a_p3_m5 = _mm_mulhi_epi16(psi_a_p3_m5, ONE_OVER_SQRT_2); + psi_a_p3_m5 = _mm_slli_epi16(psi_a_p3_m5, 2); + psi_a_p3_m7 = _mm_mulhi_epi16(psi_a_p3_m7, ONE_OVER_SQRT_2); + psi_a_p3_m7 = _mm_slli_epi16(psi_a_p3_m7, 2); + psi_a_p1_p7 = _mm_mulhi_epi16(psi_a_p1_p7, ONE_OVER_SQRT_2); + psi_a_p1_p7 = _mm_slli_epi16(psi_a_p1_p7, 2); + psi_a_p1_p5 = _mm_mulhi_epi16(psi_a_p1_p5, ONE_OVER_SQRT_2); + psi_a_p1_p5 = _mm_slli_epi16(psi_a_p1_p5, 2); + psi_a_p1_p3 = _mm_mulhi_epi16(psi_a_p1_p3, ONE_OVER_SQRT_2); + psi_a_p1_p3 = _mm_slli_epi16(psi_a_p1_p3, 2); + psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1, ONE_OVER_SQRT_2); + psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1, 2); + psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1, ONE_OVER_SQRT_2); + psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1, 2); + psi_a_p1_m3 = _mm_mulhi_epi16(psi_a_p1_m3, ONE_OVER_SQRT_2); + psi_a_p1_m3 = _mm_slli_epi16(psi_a_p1_m3, 2); + psi_a_p1_m5 = _mm_mulhi_epi16(psi_a_p1_m5, ONE_OVER_SQRT_2); + psi_a_p1_m5 = _mm_slli_epi16(psi_a_p1_m5, 2); + psi_a_p1_m7 = _mm_mulhi_epi16(psi_a_p1_m7, ONE_OVER_SQRT_2); + psi_a_p1_m7 = _mm_slli_epi16(psi_a_p1_m7, 2); + psi_a_m1_p7 = _mm_mulhi_epi16(psi_a_m1_p7, ONE_OVER_SQRT_2); + psi_a_m1_p7 = _mm_slli_epi16(psi_a_m1_p7, 2); + psi_a_m1_p5 = _mm_mulhi_epi16(psi_a_m1_p5, ONE_OVER_SQRT_2); + psi_a_m1_p5 = _mm_slli_epi16(psi_a_m1_p5, 2); + psi_a_m1_p3 = _mm_mulhi_epi16(psi_a_m1_p3, ONE_OVER_SQRT_2); + psi_a_m1_p3 = _mm_slli_epi16(psi_a_m1_p3, 2); + psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1, ONE_OVER_SQRT_2); + psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1, 2); + psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1, ONE_OVER_SQRT_2); + psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1, 2); + psi_a_m1_m3 = _mm_mulhi_epi16(psi_a_m1_m3, ONE_OVER_SQRT_2); + psi_a_m1_m3 = _mm_slli_epi16(psi_a_m1_m3, 2); + psi_a_m1_m5 = _mm_mulhi_epi16(psi_a_m1_m5, ONE_OVER_SQRT_2); + psi_a_m1_m5 = _mm_slli_epi16(psi_a_m1_m5, 2); + psi_a_m1_m7 = _mm_mulhi_epi16(psi_a_m1_m7, ONE_OVER_SQRT_2); + psi_a_m1_m7 = _mm_slli_epi16(psi_a_m1_m7, 2); + psi_a_m3_p7 = _mm_mulhi_epi16(psi_a_m3_p7, ONE_OVER_SQRT_2); + psi_a_m3_p7 = _mm_slli_epi16(psi_a_m3_p7, 2); + psi_a_m3_p5 = _mm_mulhi_epi16(psi_a_m3_p5, ONE_OVER_SQRT_2); + psi_a_m3_p5 = _mm_slli_epi16(psi_a_m3_p5, 2); + psi_a_m3_p3 = _mm_mulhi_epi16(psi_a_m3_p3, ONE_OVER_SQRT_2); + psi_a_m3_p3 = _mm_slli_epi16(psi_a_m3_p3, 2); + psi_a_m3_p1 = _mm_mulhi_epi16(psi_a_m3_p1, ONE_OVER_SQRT_2); + psi_a_m3_p1 = _mm_slli_epi16(psi_a_m3_p1, 2); + psi_a_m3_m1 = _mm_mulhi_epi16(psi_a_m3_m1, ONE_OVER_SQRT_2); + psi_a_m3_m1 = _mm_slli_epi16(psi_a_m3_m1, 2); + psi_a_m3_m3 = _mm_mulhi_epi16(psi_a_m3_m3, ONE_OVER_SQRT_2); + psi_a_m3_m3 = _mm_slli_epi16(psi_a_m3_m3, 2); + psi_a_m3_m5 = _mm_mulhi_epi16(psi_a_m3_m5, ONE_OVER_SQRT_2); + psi_a_m3_m5 = _mm_slli_epi16(psi_a_m3_m5, 2); + psi_a_m3_m7 = _mm_mulhi_epi16(psi_a_m3_m7, ONE_OVER_SQRT_2); + psi_a_m3_m7 = _mm_slli_epi16(psi_a_m3_m7, 2); + psi_a_m5_p7 = _mm_mulhi_epi16(psi_a_m5_p7, ONE_OVER_SQRT_2); + psi_a_m5_p7 = _mm_slli_epi16(psi_a_m5_p7, 2); + psi_a_m5_p5 = _mm_mulhi_epi16(psi_a_m5_p5, ONE_OVER_SQRT_2); + psi_a_m5_p5 = _mm_slli_epi16(psi_a_m5_p5, 2); + psi_a_m5_p3 = _mm_mulhi_epi16(psi_a_m5_p3, ONE_OVER_SQRT_2); + psi_a_m5_p3 = _mm_slli_epi16(psi_a_m5_p3, 2); + psi_a_m5_p1 = _mm_mulhi_epi16(psi_a_m5_p1, ONE_OVER_SQRT_2); + psi_a_m5_p1 = _mm_slli_epi16(psi_a_m5_p1, 2); + psi_a_m5_m1 = _mm_mulhi_epi16(psi_a_m5_m1, ONE_OVER_SQRT_2); + psi_a_m5_m1 = _mm_slli_epi16(psi_a_m5_m1, 2); + psi_a_m5_m3 = _mm_mulhi_epi16(psi_a_m5_m3, ONE_OVER_SQRT_2); + psi_a_m5_m3 = _mm_slli_epi16(psi_a_m5_m3, 2); + psi_a_m5_m5 = _mm_mulhi_epi16(psi_a_m5_m5, ONE_OVER_SQRT_2); + psi_a_m5_m5 = _mm_slli_epi16(psi_a_m5_m5, 2); + psi_a_m5_m7 = _mm_mulhi_epi16(psi_a_m5_m7, ONE_OVER_SQRT_2); + psi_a_m5_m7 = _mm_slli_epi16(psi_a_m5_m7, 2); + psi_a_m7_p7 = _mm_mulhi_epi16(psi_a_m7_p7, ONE_OVER_SQRT_2); + psi_a_m7_p7 = _mm_slli_epi16(psi_a_m7_p7, 2); + psi_a_m7_p5 = _mm_mulhi_epi16(psi_a_m7_p5, ONE_OVER_SQRT_2); + psi_a_m7_p5 = _mm_slli_epi16(psi_a_m7_p5, 2); + psi_a_m7_p3 = _mm_mulhi_epi16(psi_a_m7_p3, ONE_OVER_SQRT_2); + psi_a_m7_p3 = _mm_slli_epi16(psi_a_m7_p3, 2); + psi_a_m7_p1 = _mm_mulhi_epi16(psi_a_m7_p1, ONE_OVER_SQRT_2); + psi_a_m7_p1 = _mm_slli_epi16(psi_a_m7_p1, 2); + psi_a_m7_m1 = _mm_mulhi_epi16(psi_a_m7_m1, ONE_OVER_SQRT_2); + psi_a_m7_m1 = _mm_slli_epi16(psi_a_m7_m1, 2); + psi_a_m7_m3 = _mm_mulhi_epi16(psi_a_m7_m3, ONE_OVER_SQRT_2); + psi_a_m7_m3 = _mm_slli_epi16(psi_a_m7_m3, 2); + psi_a_m7_m5 = _mm_mulhi_epi16(psi_a_m7_m5, ONE_OVER_SQRT_2); + psi_a_m7_m5 = _mm_slli_epi16(psi_a_m7_m5, 2); + psi_a_m7_m7 = _mm_mulhi_epi16(psi_a_m7_m7, ONE_OVER_SQRT_2); + psi_a_m7_m7 = _mm_slli_epi16(psi_a_m7_m7, 2); + + // Calculation of a group of two terms in the bit metric involving squares of interference + square_a_64qam_epi16(a_r_p7_p7, a_i_p7_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p7); + square_a_64qam_epi16(a_r_p7_p5, a_i_p7_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p5); + square_a_64qam_epi16(a_r_p7_p3, a_i_p7_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p3); + square_a_64qam_epi16(a_r_p7_p1, a_i_p7_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p1); + square_a_64qam_epi16(a_r_p7_m1, a_i_p7_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m1); + square_a_64qam_epi16(a_r_p7_m3, a_i_p7_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m3); + square_a_64qam_epi16(a_r_p7_m5, a_i_p7_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m5); + square_a_64qam_epi16(a_r_p7_m7, a_i_p7_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m7); + square_a_64qam_epi16(a_r_p5_p7, a_i_p5_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p7); + square_a_64qam_epi16(a_r_p5_p5, a_i_p5_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p5); + square_a_64qam_epi16(a_r_p5_p3, a_i_p5_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p3); + square_a_64qam_epi16(a_r_p5_p1, a_i_p5_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p1); + square_a_64qam_epi16(a_r_p5_m1, a_i_p5_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m1); + square_a_64qam_epi16(a_r_p5_m3, a_i_p5_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m3); + square_a_64qam_epi16(a_r_p5_m5, a_i_p5_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m5); + square_a_64qam_epi16(a_r_p5_m7, a_i_p5_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m7); + square_a_64qam_epi16(a_r_p3_p7, a_i_p3_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p7); + square_a_64qam_epi16(a_r_p3_p5, a_i_p3_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p5); + square_a_64qam_epi16(a_r_p3_p3, a_i_p3_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p3); + square_a_64qam_epi16(a_r_p3_p1, a_i_p3_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p1); + square_a_64qam_epi16(a_r_p3_m1, a_i_p3_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m1); + square_a_64qam_epi16(a_r_p3_m3, a_i_p3_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m3); + square_a_64qam_epi16(a_r_p3_m5, a_i_p3_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m5); + square_a_64qam_epi16(a_r_p3_m7, a_i_p3_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m7); + square_a_64qam_epi16(a_r_p1_p7, a_i_p1_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p7); + square_a_64qam_epi16(a_r_p1_p5, a_i_p1_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p5); + square_a_64qam_epi16(a_r_p1_p3, a_i_p1_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p3); + square_a_64qam_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p1); + square_a_64qam_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m1); + square_a_64qam_epi16(a_r_p1_m3, a_i_p1_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m3); + square_a_64qam_epi16(a_r_p1_m5, a_i_p1_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m5); + square_a_64qam_epi16(a_r_p1_m7, a_i_p1_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m7); + square_a_64qam_epi16(a_r_m1_p7, a_i_m1_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p7); + square_a_64qam_epi16(a_r_m1_p5, a_i_m1_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p5); + square_a_64qam_epi16(a_r_m1_p3, a_i_m1_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p3); + square_a_64qam_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p1); + square_a_64qam_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m1); + square_a_64qam_epi16(a_r_m1_m3, a_i_m1_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m3); + square_a_64qam_epi16(a_r_m1_m5, a_i_m1_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m5); + square_a_64qam_epi16(a_r_m1_m7, a_i_m1_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m7); + square_a_64qam_epi16(a_r_m3_p7, a_i_m3_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p7); + square_a_64qam_epi16(a_r_m3_p5, a_i_m3_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p5); + square_a_64qam_epi16(a_r_m3_p3, a_i_m3_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p3); + square_a_64qam_epi16(a_r_m3_p1, a_i_m3_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p1); + square_a_64qam_epi16(a_r_m3_m1, a_i_m3_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m1); + square_a_64qam_epi16(a_r_m3_m3, a_i_m3_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m3); + square_a_64qam_epi16(a_r_m3_m5, a_i_m3_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m5); + square_a_64qam_epi16(a_r_m3_m7, a_i_m3_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m7); + square_a_64qam_epi16(a_r_m5_p7, a_i_m5_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p7); + square_a_64qam_epi16(a_r_m5_p5, a_i_m5_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p5); + square_a_64qam_epi16(a_r_m5_p3, a_i_m5_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p3); + square_a_64qam_epi16(a_r_m5_p1, a_i_m5_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p1); + square_a_64qam_epi16(a_r_m5_m1, a_i_m5_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m1); + square_a_64qam_epi16(a_r_m5_m3, a_i_m5_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m3); + square_a_64qam_epi16(a_r_m5_m5, a_i_m5_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m5); + square_a_64qam_epi16(a_r_m5_m7, a_i_m5_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m7); + square_a_64qam_epi16(a_r_m7_p7, a_i_m7_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p7); + square_a_64qam_epi16(a_r_m7_p5, a_i_m7_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p5); + square_a_64qam_epi16(a_r_m7_p3, a_i_m7_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p3); + square_a_64qam_epi16(a_r_m7_p1, a_i_m7_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p1); + square_a_64qam_epi16(a_r_m7_m1, a_i_m7_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m1); + square_a_64qam_epi16(a_r_m7_m3, a_i_m7_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m3); + square_a_64qam_epi16(a_r_m7_m5, a_i_m7_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m5); + square_a_64qam_epi16(a_r_m7_m7, a_i_m7_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m7); + + // Computing different multiples of ||h0||^2 + // x=1, y=1 + ch_mag_2_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42); + ch_mag_2_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_2_over_42_with_sigma2,1); + // x=1, y=3 + ch_mag_10_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42); + ch_mag_10_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_10_over_42_with_sigma2,1); + // x=1, x=5 + ch_mag_26_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42); + ch_mag_26_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_26_over_42_with_sigma2,1); + // x=1, y=7 + ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); + ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); + // x=3, y=3 + ch_mag_18_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42); + ch_mag_18_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_18_over_42_with_sigma2,1); + // x=3, y=5 + ch_mag_34_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42); + ch_mag_34_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_34_over_42_with_sigma2,1); + // x=3, y=7 + ch_mag_58_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42); + ch_mag_58_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_58_over_42_with_sigma2,2); + // x=5, y=5 + ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); + ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); + // x=5, y=7 + ch_mag_74_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42); + ch_mag_74_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_74_over_42_with_sigma2,2); + // x=7, y=7 + ch_mag_98_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42); + ch_mag_98_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_98_over_42_with_sigma2,2); + + // Computing Metrics + xmm0 = _mm_subs_epi16(psi_a_p7_p7, a_sq_p7_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_7); + bit_met_p7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_p5, a_sq_p7_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_5); + bit_met_p7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_p3, a_sq_p7_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_3); + bit_met_p7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_p1, a_sq_p7_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_7_1); + bit_met_p7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m1, a_sq_p7_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_1); + bit_met_p7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m3, a_sq_p7_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_3); + bit_met_p7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m5, a_sq_p7_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_5); + bit_met_p7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p7_m7, a_sq_p7_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_7_7); + bit_met_p7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p7, a_sq_p5_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_7); + bit_met_p5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p5, a_sq_p5_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_5); + bit_met_p5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p3, a_sq_p5_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_3); + bit_met_p5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_p1, a_sq_p5_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_5_1); + bit_met_p5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m1, a_sq_p5_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_1); + bit_met_p5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m3, a_sq_p5_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_3); + bit_met_p5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m5, a_sq_p5_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_5); + bit_met_p5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p5_m7, a_sq_p5_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_5_7); + bit_met_p5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p7, a_sq_p3_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_7); + bit_met_p3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p5, a_sq_p3_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_5); + bit_met_p3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p3, a_sq_p3_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_3); + bit_met_p3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_p1, a_sq_p3_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_3_1); + bit_met_p3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m1, a_sq_p3_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_1); + bit_met_p3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m3, a_sq_p3_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_3); + bit_met_p3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m5, a_sq_p3_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_5); + bit_met_p3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p3_m7, a_sq_p3_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_3_7); + bit_met_p3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p7, a_sq_p1_p7); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_7); + bit_met_p1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p5, a_sq_p1_p5); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_5); + bit_met_p1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p3, a_sq_p1_p3); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_3); + bit_met_p1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); + xmm1 = _mm_adds_epi16(xmm0, y0_p_1_1); + bit_met_p1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_1); + bit_met_p1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m3, a_sq_p1_m3); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_3); + bit_met_p1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m5, a_sq_p1_m5); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_5); + bit_met_p1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_p1_m7, a_sq_p1_m7); + xmm1 = _mm_adds_epi16(xmm0, y0_m_1_7); + bit_met_p1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + + xmm0 = _mm_subs_epi16(psi_a_m1_p7, a_sq_m1_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_7); + bit_met_m1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_p5, a_sq_m1_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_5); + bit_met_m1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_p3, a_sq_m1_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_3); + bit_met_m1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_1_1); + bit_met_m1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_1); + bit_met_m1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m3, a_sq_m1_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_3); + bit_met_m1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m5, a_sq_m1_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_5); + bit_met_m1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m1_m7, a_sq_m1_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_1_7); + bit_met_m1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p7, a_sq_m3_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_7); + bit_met_m3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p5, a_sq_m3_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_5); + bit_met_m3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p3, a_sq_m3_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_3); + bit_met_m3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_p1, a_sq_m3_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_3_1); + bit_met_m3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m1, a_sq_m3_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_1); + bit_met_m3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m3, a_sq_m3_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_3); + bit_met_m3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m5, a_sq_m3_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_5); + bit_met_m3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m3_m7, a_sq_m3_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_3_7); + bit_met_m3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p7, a_sq_m5_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_7); + bit_met_m5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p5, a_sq_m5_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_5); + bit_met_m5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p3, a_sq_m5_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_3); + bit_met_m5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_p1, a_sq_m5_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_5_1); + bit_met_m5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m1, a_sq_m5_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_1); + bit_met_m5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m3, a_sq_m5_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_3); + bit_met_m5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m5, a_sq_m5_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_5); + bit_met_m5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m5_m7, a_sq_m5_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_5_7); + bit_met_m5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p7, a_sq_m7_p7); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_7); + bit_met_m7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p5, a_sq_m7_p5); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_5); + bit_met_m7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p3, a_sq_m7_p3); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_3); + bit_met_m7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_p1, a_sq_m7_p1); + xmm1 = _mm_subs_epi16(xmm0, y0_m_7_1); + bit_met_m7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m1, a_sq_m7_m1); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_1); + bit_met_m7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m3, a_sq_m7_m3); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_3); + bit_met_m7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m5, a_sq_m7_m5); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_5); + bit_met_m7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); + xmm0 = _mm_subs_epi16(psi_a_m7_m7, a_sq_m7_m7); + xmm1 = _mm_subs_epi16(xmm0, y0_p_7_7); + bit_met_m7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); + + // Detection for 1st bit (LTE mapping) + // bit = 1 + xmm0 = _mm_max_epi16(bit_met_m7_p7, bit_met_m7_p5); + xmm1 = _mm_max_epi16(bit_met_m7_p3, bit_met_m7_p1); + xmm2 = _mm_max_epi16(bit_met_m7_m1, bit_met_m7_m3); + xmm3 = _mm_max_epi16(bit_met_m7_m5, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m5_p7, bit_met_m5_p5); + xmm1 = _mm_max_epi16(bit_met_m5_p3, bit_met_m5_p1); + xmm2 = _mm_max_epi16(bit_met_m5_m1, bit_met_m5_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m5_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_m3_p7, bit_met_m3_p5); + xmm1 = _mm_max_epi16(bit_met_m3_p3, bit_met_m3_p1); + xmm2 = _mm_max_epi16(bit_met_m3_m1, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m3_m5, bit_met_m3_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_p7, bit_met_m1_p5); + xmm1 = _mm_max_epi16(bit_met_m1_p3, bit_met_m1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m1_m3); + xmm3 = _mm_max_epi16(bit_met_m1_m5, bit_met_m1_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + // bit = 0 + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p7_p5); + xmm1 = _mm_max_epi16(bit_met_p7_p3, bit_met_p7_p1); + xmm2 = _mm_max_epi16(bit_met_p7_m1, bit_met_p7_m3); + xmm3 = _mm_max_epi16(bit_met_p7_m5, bit_met_p7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_p7, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p5_p3, bit_met_p5_p1); + xmm2 = _mm_max_epi16(bit_met_p5_m1, bit_met_p5_m3); + xmm3 = _mm_max_epi16(bit_met_p5_m5, bit_met_p5_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_p7, bit_met_p3_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p3_p1); + xmm2 = _mm_max_epi16(bit_met_p3_m1, bit_met_p3_m3); + xmm3 = _mm_max_epi16(bit_met_p3_m5, bit_met_p3_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_p7, bit_met_p1_p5); + xmm1 = _mm_max_epi16(bit_met_p1_p3, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_p1_m1, bit_met_p1_m3); + xmm3 = _mm_max_epi16(bit_met_p1_m5, bit_met_p1_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y0r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 2nd bit (LTE mapping) + // bit = 1 + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + // bit = 0 + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y1r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 3rd bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); + xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); + xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); + xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); + xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); + xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); + xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); + xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); + xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); + xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); + xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); + xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); + xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); + xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); + xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y2r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 4th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y0i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + + // Detection for 5th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); + xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); + xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); + xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); + xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); + xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); + xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); + xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); + xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); + xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); + xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); + xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); + xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); + xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); + xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); + xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); + xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); + xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); + xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y1i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + // Detection for 6th bit (LTE mapping) + xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); + xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); + xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); + xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); + xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); + xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); + xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); + xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); + xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); + xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); + xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); + xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); + xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); + logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); + + xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); + xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); + xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); + xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); + xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); + xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); + xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); + xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); + xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); + xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); + xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); + xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); + xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); + xmm4 = _mm_max_epi16(xmm0, xmm1); + xmm5 = _mm_max_epi16(xmm2, xmm3); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); + logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); + + y2i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); + + + // map to output stream, difficult to do in SIMD since we have 6 16bit LLRs + // RE 1 + j = 24*i; + stream0_out[j + 0] = ((short *)&y0r)[0]; + stream0_out[j + 1] = ((short *)&y1r)[0]; + stream0_out[j + 2] = ((short *)&y2r)[0]; + stream0_out[j + 3] = ((short *)&y0i)[0]; + stream0_out[j + 4] = ((short *)&y1i)[0]; + stream0_out[j + 5] = ((short *)&y2i)[0]; + // RE 2 + stream0_out[j + 6] = ((short *)&y0r)[1]; + stream0_out[j + 7] = ((short *)&y1r)[1]; + stream0_out[j + 8] = ((short *)&y2r)[1]; + stream0_out[j + 9] = ((short *)&y0i)[1]; + stream0_out[j + 10] = ((short *)&y1i)[1]; + stream0_out[j + 11] = ((short *)&y2i)[1]; + // RE 3 + stream0_out[j + 12] = ((short *)&y0r)[2]; + stream0_out[j + 13] = ((short *)&y1r)[2]; + stream0_out[j + 14] = ((short *)&y2r)[2]; + stream0_out[j + 15] = ((short *)&y0i)[2]; + stream0_out[j + 16] = ((short *)&y1i)[2]; + stream0_out[j + 17] = ((short *)&y2i)[2]; + // RE 4 + stream0_out[j + 18] = ((short *)&y0r)[3]; + stream0_out[j + 19] = ((short *)&y1r)[3]; + stream0_out[j + 20] = ((short *)&y2r)[3]; + stream0_out[j + 21] = ((short *)&y0i)[3]; + stream0_out[j + 22] = ((short *)&y1i)[3]; + stream0_out[j + 23] = ((short *)&y2i)[3]; + // RE 5 + stream0_out[j + 24] = ((short *)&y0r)[4]; + stream0_out[j + 25] = ((short *)&y1r)[4]; + stream0_out[j + 26] = ((short *)&y2r)[4]; + stream0_out[j + 27] = ((short *)&y0i)[4]; + stream0_out[j + 28] = ((short *)&y1i)[4]; + stream0_out[j + 29] = ((short *)&y2i)[4]; + // RE 6 + stream0_out[j + 30] = ((short *)&y0r)[5]; + stream0_out[j + 31] = ((short *)&y1r)[5]; + stream0_out[j + 32] = ((short *)&y2r)[5]; + stream0_out[j + 33] = ((short *)&y0i)[5]; + stream0_out[j + 34] = ((short *)&y1i)[5]; + stream0_out[j + 35] = ((short *)&y2i)[5]; + // RE 7 + stream0_out[j + 36] = ((short *)&y0r)[6]; + stream0_out[j + 37] = ((short *)&y1r)[6]; + stream0_out[j + 38] = ((short *)&y2r)[6]; + stream0_out[j + 39] = ((short *)&y0i)[6]; + stream0_out[j + 40] = ((short *)&y1i)[6]; + stream0_out[j + 41] = ((short *)&y2i)[6]; + // RE 8 + stream0_out[j + 42] = ((short *)&y0r)[7]; + stream0_out[j + 43] = ((short *)&y1r)[7]; + stream0_out[j + 44] = ((short *)&y2r)[7]; + stream0_out[j + 45] = ((short *)&y0i)[7]; + stream0_out[j + 46] = ((short *)&y1i)[7]; + stream0_out[j + 47] = ((short *)&y2i)[7]; + +#elif defined(__arm__) + +#endif + + } + +#if defined(__x86_64__) || defined(__i386__) + _mm_empty(); + _m_empty(); +#endif +} + + +int nr_dlsch_64qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int32_t **rxdataF_comp_i, + int32_t **dl_ch_mag, + int32_t **dl_ch_mag_i, + int32_t **rho_i, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint16_t pbch_pss_sss_adjust, + //int16_t **llr16p, + uint32_t llr_offset) +{ + + int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; + int16_t *llr16; + int8_t *pllr_symbol; // pointer where llrs should filled for this ofdm symbol + int len; + uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; + + //first symbol has different structure due to more pilots + /*if (first_symbol_flag == 1) { + llr16 = (int16_t*)dlsch_llr; + } else { + llr16 = (int16_t*)(*llr16p); + }*/ + + llr16 = (int16_t*)dlsch_llr; + + AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); + + + if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { + // if symbol has pilots + if (frame_parms->nb_antenna_ports_eNB!=1) + // in 2 antenna ports we have 8 REs per symbol per RB + len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); + else + // for 1 antenna port we have 10 REs per symbol per RB + len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); + } else { + // symbol has no pilots + len = (nb_rb*12) - pbch_pss_sss_adjust; + } + + pllr_symbol = (int8_t*)dlsch_llr; + pllr_symbol += llr_offset; + //printf("nr_dlsch_64qam_64qam_llr: symbol %d,nb_rb %d, len %d,pbch_pss_sss_adjust %d\n",symbol,nb_rb,len,pbch_pss_sss_adjust); + /*LOG_I(PHY,"nr_dlsch_64qam_64qam_llr [symb %d / FirstSym %d / Length %d / LLR Offset %d]: @LLR Buff %x, @LLR Buff(symb) %x, , @Compute LLR Buff(symb) %x \n", + symbol, + first_symbol_flag, + len, + llr_offset, + (int16_t*)dlsch_llr, + llr16, + pllr_symbol);*/ + +#ifdef __AVX2__ + + // Round length up to multiple of 16 words + uint32_t len256i = ((len+16)>>4)*16; + int32_t *rxF_256i = (int32_t*) malloc16_clear(len256i*4); + int32_t *rxF_i_256i = (int32_t*) malloc16_clear(len256i*4); + int32_t *ch_mag_256i = (int32_t*) malloc16_clear(len256i*4); + int32_t *ch_mag_i_256i = (int32_t*) malloc16_clear(len256i*4); + int32_t *rho_256i = (int32_t*) malloc16_clear(len256i*4); + + memcpy(rxF_256i, rxF, len*4); + memcpy(rxF_i_256i, rxF_i, len*4); + memcpy(ch_mag_256i, ch_mag, len*4); + memcpy(ch_mag_i_256i, ch_mag_i, len*4); + memcpy(rho_256i, rho, len*4); + +#if 0 + qam64_qam16_avx2((short *)rxF_256i, + (short *)rxF_i_256i, + (short *)ch_mag_256i, + (short *)ch_mag_i_256i, + (short *)llr16, + (short *) rho_256i, + len); +#else + qam64_qam64_avx2((int32_t *)rxF_256i, + (int32_t *)rxF_i_256i, + (int32_t *)ch_mag_256i, + (int32_t *)ch_mag_i_256i, + (int16_t *)llr16, + (int32_t *) rho_256i, + len); +#endif + + free16(rxF_256i, sizeof(rxF_256i)); + free16(rxF_i_256i, sizeof(rxF_i_256i)); + free16(ch_mag_256i, sizeof(ch_mag_256i)); + free16(ch_mag_i_256i, sizeof(ch_mag_i_256i)); + free16(rho_256i, sizeof(rho_256i)); + +#else + qam64_qam64((short *)rxF, + (short *)rxF_i, + (short *)ch_mag, + (short *)ch_mag_i, + (short *)llr16, + (short *)rho, + len); +#endif + + llr16 += (6*len); + //*llr16p = (short *)llr16; + + return(0); +} diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_initial_sync.c b/openair1/PHY/NR_UE_TRANSPORT/nr_initial_sync.c index 80c71bd665..e5f079ebc3 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/nr_initial_sync.c +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_initial_sync.c @@ -39,6 +39,7 @@ //#include "SCHED/extern.h" #include "common_lib.h" +#include #include "PHY/NR_REFSIG/pss_nr.h" #include "PHY/NR_REFSIG/sss_nr.h" @@ -56,6 +57,7 @@ int nr_pbch_detection(PHY_VARS_NR_UE *ue, runmode_t mode) NR_DL_FRAME_PARMS *frame_parms=&ue->frame_parms; int ret =-1; + #ifdef DEBUG_INITIAL_SYNCH LOG_I(PHY,"[UE%d] Initial sync: starting PBCH detection (rx_offset %d)\n",ue->Mod_id, ue->rx_offset); @@ -65,6 +67,7 @@ int nr_pbch_detection(PHY_VARS_NR_UE *ue, runmode_t mode) int nb_prefix_samples0 = frame_parms->nb_prefix_samples0; frame_parms->nb_prefix_samples0 = frame_parms->nb_prefix_samples; + //symbol 1 nr_slot_fep(ue, 1, @@ -140,9 +143,10 @@ char prefix_string[2][9] = {"NORMAL","EXTENDED"}; int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode) { - int32_t sync_pos, sync_pos2, sync_pos_slot; // k_ssb, N_ssb_crb, + int32_t sync_pos, sync_pos_slot; // k_ssb, N_ssb_crb, sync_pos2, int32_t metric_tdd_ncp=0; uint8_t phase_tdd_ncp; + double im, re; NR_DL_FRAME_PARMS *fp = &ue->frame_parms; int ret=-1; @@ -175,7 +179,6 @@ int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode) * sync_pos SS/PBCH block */ - cnt++; if (1){ // (cnt>100) cnt =0; @@ -183,19 +186,41 @@ int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode) /* process pss search on received buffer */ sync_pos = pss_synchro_nr(ue, NO_RATE_CHANGE); - - if (sync_pos >= fp->nb_prefix_samples) - ue->ssb_offset = sync_pos - fp->nb_prefix_samples; - else - ue->ssb_offset = sync_pos + (fp->samples_per_subframe * 10) - fp->nb_prefix_samples; - + sync_pos_slot = (fp->samples_per_subframe/fp->slots_per_subframe) - 10*(fp->ofdm_symbol_size + fp->nb_prefix_samples); + + if (sync_pos >= fp->nb_prefix_samples){ + ue->ssb_offset = sync_pos - fp->nb_prefix_samples;} + else{ + ue->ssb_offset = sync_pos + (fp->samples_per_subframe * 10) - fp->nb_prefix_samples;} + + ue->rx_offset = ue->ssb_offset - sync_pos_slot; + //write_output("rxdata1.m","rxd1",ue->common_vars.rxdata[0],10*fp->samples_per_subframe,1,1); #ifdef DEBUG_INITIAL_SYNCH LOG_I(PHY,"[UE%d] Initial sync : Estimated PSS position %d, Nid2 %d\n", ue->Mod_id, sync_pos,ue->common_vars.eNb_id); - LOG_I(PHY,"sync_pos %d ssb_offset %d\n",sync_pos,ue->ssb_offset); + LOG_I(PHY,"sync_pos %d ssb_offset %d sync_pos_slot %d \n",sync_pos,ue->ssb_offset,sync_pos_slot); #endif + // digital compensation of FFO for SSB symbols + if (ue->UE_fo_compensation){ + double s_time = 1/(1.0e3*fp->samples_per_subframe); // sampling time + double off_angle = -2*M_PI*s_time*(ue->common_vars.freq_offset); // offset rotation angle compensation per sample + + int start = ue->ssb_offset; // start for offset correction is at ssb_offset (pss time position) + int end = start + 4*(fp->ofdm_symbol_size + fp->nb_prefix_samples); // loop over samples in 4 symbols (ssb size), including prefix + + for(int n=start; nnb_antennas_rx; ar++) { + re = ((double)(((short *)ue->common_vars.rxdata[ar]))[2*n]); + im = ((double)(((short *)ue->common_vars.rxdata[ar]))[2*n+1]); + ((short *)ue->common_vars.rxdata[ar])[2*n] = (short)(round(re*cos(n*off_angle) - im*sin(n*off_angle))); + ((short *)ue->common_vars.rxdata[ar])[2*n+1] = (short)(round(re*sin(n*off_angle) + im*cos(n*off_angle))); + } + } + } + + /* check that SSS/PBCH block is continuous inside the received buffer */ if (sync_pos < (NR_NUMBER_OF_SUBFRAMES_PER_FRAME*fp->samples_per_subframe - (NB_SYMBOLS_PBCH * fp->ofdm_symbol_size))) { @@ -242,7 +267,8 @@ int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode) ret = -1; } - /* Consider this is a false detection if the offset is > 1000 Hz */ + /* Consider this is a false detection if the offset is > 1000 Hz + Not to be used now that offest estimation is in place if( (abs(ue->common_vars.freq_offset) > 150) && (ret == 0) ) { ret=-1; @@ -251,7 +277,7 @@ int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode) #else LOG_E(HW, "Ignore MIB with high freq offset [%d Hz] estimation \n",ue->common_vars.freq_offset); #endif - } + }*/ if (ret==0) { // PBCH found so indicate sync to higher layers and configure frame parameters @@ -327,13 +353,13 @@ int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode) printf("[UE %d] Frame %d Measured Carrier Frequency %.0f Hz (offset %d Hz)\n", ue->Mod_id, ue->proc.proc_rxtx[0].frame_rx, - openair0_cfg[0].rx_freq[0]-ue->common_vars.freq_offset, + openair0_cfg[0].rx_freq[0]+ue->common_vars.freq_offset, ue->common_vars.freq_offset); # else LOG_I(PHY, "[UE %d] Frame %d Measured Carrier Frequency %.0f Hz (offset %d Hz)\n", ue->Mod_id, ue->proc.proc_rxtx[0].frame_rx, - openair0_cfg[0].rx_freq[0]-ue->common_vars.freq_offset, + openair0_cfg[0].rx_freq[0]+ue->common_vars.freq_offset, ue->common_vars.freq_offset); # endif #endif diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_pbch.c b/openair1/PHY/NR_UE_TRANSPORT/nr_pbch.c index 1e8e12307a..a0b150eda1 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/nr_pbch.c +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_pbch.c @@ -51,8 +51,8 @@ uint16_t nr_pbch_extract(int **rxdataF, int **rxdataF_ext, int **dl_ch_estimates_ext, uint32_t symbol, + uint32_t s_offset, uint32_t high_speed_flag, - int is_synchronized, NR_DL_FRAME_PARMS *frame_parms) { @@ -65,18 +65,15 @@ uint16_t nr_pbch_extract(int **rxdataF, unsigned int rx_offset = frame_parms->first_carrier_offset + frame_parms->ssb_start_subcarrier; //and if (rx_offset>= frame_parms->ofdm_symbol_size) rx_offset-=frame_parms->ofdm_symbol_size; - int s_offset=0; AssertFatal(symbol>=1 && symbol<5, "symbol %d illegal for PBCH extraction\n", symbol); - if (is_synchronized==1) s_offset=4; - for (aarx=0; aarxnb_antennas_rx; aarx++) { rxF = &rxdataF[aarx][(symbol+s_offset)*frame_parms->ofdm_symbol_size]; - rxF_ext = &rxdataF_ext[aarx][(symbol+s_offset)*(20*12)]; + rxF_ext = &rxdataF_ext[aarx][symbol*20*12]; #ifdef DEBUG_PBCH printf("extract_rbs (nushift %d): rx_offset=%d, symbol %d\n",frame_parms->nushift, (rx_offset + ((symbol+s_offset)*(frame_parms->ofdm_symbol_size))),symbol); @@ -143,7 +140,7 @@ uint16_t nr_pbch_extract(int **rxdataF, //printf("dl_ch0 addr %p\n",dl_ch0); - dl_ch0_ext = &dl_ch_estimates_ext[aarx][(symbol+s_offset)*(20*12)]; + dl_ch0_ext = &dl_ch_estimates_ext[aarx][symbol*20*12]; for (rb=0; rb<20; rb++) { j=0; @@ -268,8 +265,7 @@ void nr_pbch_channel_compensation(int **rxdataF_ext, int **dl_ch_estimates_ext, int **rxdataF_comp, NR_DL_FRAME_PARMS *frame_parms, - uint8_t symbol, - int is_synchronized, + uint32_t symbol, uint8_t output_shift) { @@ -283,20 +279,18 @@ void nr_pbch_channel_compensation(int **rxdataF_ext, uint16_t nb_re=180; uint8_t aarx; + #if defined(__x86_64__) || defined(__i386__) __m128i *dl_ch128,*rxdataF128,*rxdataF_comp128; #elif defined(__arm__) #endif - AssertFatal((symbol > 0 && symbol < 4 && is_synchronized == 0) || - (symbol > 4 && symbol < 8 && is_synchronized == 1), - "symbol %d is illegal for PBCH DM-RS (is_synchronized %d)\n", - symbol,is_synchronized); + AssertFatal((symbol > 0 && symbol < 4), + "symbol %d is illegal for PBCH DM-RS\n", + symbol); - - - if (symbol == 2 || symbol == 6) nb_re = 72; + if (symbol == 2) nb_re = 72; // printf("comp: symbol %d : nb_re %d\n",symbol,nb_re); @@ -515,7 +509,7 @@ unsigned char sign(int8_t x) { */ uint8_t pbch_deinterleaving_pattern[32] = {28,0,31,30,7,29,25,27,5,8,24,9,10,11,12,13,1,4,3,14,15,16,17,2,26,18,19,20,21,22,6,23}; - + int nr_rx_pbch( PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, NR_UE_PBCH *nr_ue_pbch_vars, @@ -529,10 +523,10 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, int max_h=0; - int symbol,i; + int symbol; //uint8_t pbch_a[64]; uint8_t *pbch_a = malloc(sizeof(uint8_t) * 32); - uint32_t pbch_a_prime; + //uint32_t pbch_a_prime; int16_t *pbch_e_rx; uint8_t *decoded_output = nr_ue_pbch_vars->decoded_output; uint8_t nushift; @@ -540,12 +534,12 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, uint8_t Lmax=8; //to update uint8_t ssb_index=0; //uint16_t crc; - unsigned short idx_demod =0; - int8_t decoderState=0; - uint8_t decoderListSize = 8, pathMetricAppr = 0; + //unsigned short idx_demod =0; + uint32_t decoderState=0; + //uint8_t decoderListSize = 8, pathMetricAppr = 0; - time_stats_t polar_decoder_init,polar_rate_matching,decoding,bit_extraction,deinterleaving; - time_stats_t path_metric,sorting,update_LLR; + //time_stats_t polar_decoder_init,polar_rate_matching,decoding,bit_extraction,deinterleaving; + //time_stats_t path_metric,sorting,update_LLR; memset(&pbch_a[0], 0, sizeof(uint8_t) * NR_POLAR_PBCH_PAYLOAD_BITS); //printf("nr_pbch_ue nid_cell %d\n",frame_parms->Nid_cell); @@ -558,32 +552,35 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, // clear LLR buffer memset(nr_ue_pbch_vars->llr,0,NR_POLAR_PBCH_E); - int first_symbol=1; - if (ue->is_synchronized > 0) first_symbol+=4; + int symbol_offset=1; + if (ue->is_synchronized > 0) + symbol_offset=4; + else + symbol_offset=0; #ifdef DEBUG_PBCH //printf("address dataf %p",nr_ue_common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[subframe_rx]].rxdataF); write_output("rxdataF0_pbch.m","rxF0pbch", - &nr_ue_common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[subframe_rx]].rxdataF[0][first_symbol*frame_parms->ofdm_symbol_size],frame_parms->ofdm_symbol_size*3,1,1); + &nr_ue_common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[subframe_rx]].rxdataF[0][(symbol_offset+1)*frame_parms->ofdm_symbol_size],frame_parms->ofdm_symbol_size*3,1,1); #endif - for (symbol=first_symbol; symbol<(first_symbol+3); symbol++) { + // symbol refers to symbol within SSB. symbol_offset is the offset of the SSB wrt start of slot + for (symbol=1; symbol<4; symbol++) { - nr_pbch_extract(nr_ue_common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[subframe_rx]].rxdataF, nr_ue_common_vars->common_vars_rx_data_per_thread[ue->current_thread_id[subframe_rx]].dl_ch_estimates[eNB_id], nr_ue_pbch_vars->rxdataF_ext, nr_ue_pbch_vars->dl_ch_estimates_ext, - symbol-first_symbol+1, + symbol, + symbol_offset, high_speed_flag, - ue->is_synchronized, frame_parms); #ifdef DEBUG_PBCH LOG_I(PHY,"[PHY] PBCH Symbol %d ofdm size %d\n",symbol, frame_parms->ofdm_symbol_size ); LOG_I(PHY,"[PHY] PBCH starting channel_level\n"); #endif - if (symbol == 1 || symbol == 5) { + if (symbol == 1) { max_h = nr_pbch_channel_level(nr_ue_pbch_vars->dl_ch_estimates_ext, frame_parms, symbol); @@ -599,7 +596,6 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, nr_ue_pbch_vars->rxdataF_comp, frame_parms, symbol, - ue->is_synchronized, nr_ue_pbch_vars->log2_maxh); // log2_maxh+I0_shift /*if (frame_parms->nb_antennas_rx > 1) @@ -615,7 +611,7 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, return(-1); } */ - if (symbol==(first_symbol+1)) { + if (symbol==1) { nr_pbch_quantize(pbch_e_rx, (short*)&(nr_ue_pbch_vars->rxdataF_comp[0][symbol*240]), 144); @@ -633,7 +629,7 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, } #ifdef DEBUG_PBCH - write_output("rxdataF_comp.m","rxFcomp",&nr_ue_pbch_vars->rxdataF_comp[0][240*first_symbol],240*3,1,1); + write_output("rxdataF_comp.m","rxFcomp",&nr_ue_pbch_vars->rxdataF_comp[0][240],240*3,1,1); #endif @@ -644,7 +640,7 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, #ifdef DEBUG_PBCH //pbch_e_rx = &nr_ue_pbch_vars->llr[0]; - short *p = (short *)&(nr_ue_pbch_vars->rxdataF_comp[0][first_symbol*20*12]); + short *p = (short *)&(nr_ue_pbch_vars->rxdataF_comp[0][20*12]); for (int cnt = 0; cnt < 864 ; cnt++) printf("pbch rx llr %d\n",*(pbch_e_rx+cnt)); @@ -661,24 +657,34 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, //polar decoding de-rate matching + nr_polar_init(&nr_ue_pbch_vars->nrPolar_params, + NR_POLAR_PBCH_MESSAGE_TYPE, + NR_POLAR_PBCH_PAYLOAD_BITS, + NR_POLAR_PBCH_AGGREGATION_LEVEL); - AssertFatal(ue->nrPolar_params != NULL,"ue->nrPolar_params is null\n"); + AssertFatal(nr_ue_pbch_vars->nrPolar_params != NULL,"nr_ue_pbch_vars->nrPolar_params is null\n"); - t_nrPolar_params *currentPtr = nr_polar_params(ue->nrPolar_params, NR_POLAR_PBCH_MESSAGE_TYPE, NR_POLAR_PBCH_PAYLOAD_BITS, NR_POLAR_PBCH_AGGREGATION_LEVEL); + t_nrPolar_params *currentPtr = nr_polar_params(nr_ue_pbch_vars->nrPolar_params, NR_POLAR_PBCH_MESSAGE_TYPE, NR_POLAR_PBCH_PAYLOAD_BITS, NR_POLAR_PBCH_AGGREGATION_LEVEL); - decoderState = polar_decoder_int16(pbch_e_rx,(uint8_t*)&nr_ue_pbch_vars->pbch_a_prime,currentPtr); - - - if(decoderState == -1) - return(decoderState); + decoderState = polar_decoder_int16(pbch_e_rx,(uint64_t*)&nr_ue_pbch_vars->pbch_a_prime,currentPtr); + + if(decoderState) return(decoderState); - //printf("polar decoder output 0x%08x\n",nr_ue_pbch_vars->pbch_a_prime); + // printf("polar decoder output 0x%08x\n",nr_ue_pbch_vars->pbch_a_prime); + + // Decoder reversal + uint32_t a_reversed=0; + for (int i=0; ipbch_a_prime>>i)&1)<<(31-i); + + nr_ue_pbch_vars->pbch_a_prime = a_reversed; //payload un-scrambling memset(&nr_ue_pbch_vars->pbch_a_interleaved, 0, sizeof(uint32_t) ); M = (Lmax == 64)? (NR_POLAR_PBCH_PAYLOAD_BITS - 6) : (NR_POLAR_PBCH_PAYLOAD_BITS - 3); nushift = ((nr_ue_pbch_vars->pbch_a_prime>>24)&1) ^ (((nr_ue_pbch_vars->pbch_a_prime>>6)&1)<<1); nr_pbch_unscrambling(nr_ue_pbch_vars,frame_parms->Nid_cell,nushift,M,NR_POLAR_PBCH_PAYLOAD_BITS,1,unscrambling_mask); + //printf("nushift %d sfn 3rd %d 2nd %d", nushift,((nr_ue_pbch_vars->pbch_a_prime>>6)&1), ((nr_ue_pbch_vars->pbch_a_prime>>24)&1) ); //payload deinterleaving //uint32_t in=0; @@ -710,14 +716,13 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue, ue->dl_indication.rx_ind = &ue->rx_ind; // hang on rx_ind instance //ue->rx_ind.sfn_slot = 0; //should be set by higher-1-layer, i.e. clean_and_set_if_instance() - ue->rx_ind.number_pdus = ue->rx_ind.number_pdus + 1; - ue->rx_ind.rx_indication_body = (fapi_nr_rx_indication_body_t *)malloc(sizeof(fapi_nr_rx_indication_body_t)); - ue->rx_ind.rx_indication_body->pdu_type = FAPI_NR_RX_PDU_TYPE_MIB; - ue->rx_ind.rx_indication_body->mib_pdu.pdu = &decoded_output[0]; - ue->rx_ind.rx_indication_body->mib_pdu.additional_bits = xtra_byte; - ue->rx_ind.rx_indication_body->mib_pdu.ssb_index = ssb_index; // confirm with TCL - ue->rx_ind.rx_indication_body->mib_pdu.ssb_length = Lmax; // confirm with TCL - ue->rx_ind.rx_indication_body->mib_pdu.cell_id = frame_parms->Nid_cell; // confirm with TCL + ue->rx_ind.rx_indication_body[0].pdu_type = FAPI_NR_RX_PDU_TYPE_MIB; + ue->rx_ind.rx_indication_body[0].mib_pdu.pdu = &decoded_output[0]; + ue->rx_ind.rx_indication_body[0].mib_pdu.additional_bits = xtra_byte; + ue->rx_ind.rx_indication_body[0].mib_pdu.ssb_index = ssb_index; // confirm with TCL + ue->rx_ind.rx_indication_body[0].mib_pdu.ssb_length = Lmax; // confirm with TCL + ue->rx_ind.rx_indication_body[0].mib_pdu.cell_id = frame_parms->Nid_cell; // confirm with TCL + ue->rx_ind.number_pdus = 1; if (ue->if_inst && ue->if_inst->dl_indication) ue->if_inst->dl_indication(&ue->dl_indication); diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_prach.c b/openair1/PHY/NR_UE_TRANSPORT/nr_prach.c index 0a662304dd..9c7146ca03 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/nr_prach.c +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_prach.c @@ -40,6 +40,7 @@ //#include "PHY/defs_nr_common.h" #include "PHY/defs_nr_UE.h" #include "PHY/NR_UE_TRANSPORT/nr_prach.h" +#include "PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h" //#include "PHY/extern.h" //#include "LAYER2/MAC/extern.h" //#include "PHY/NR_UE_TRANSPORT/pucch_nr.h" @@ -574,7 +575,7 @@ int32_t generate_nr_prach( PHY_VARS_NR_UE *ue, uint8_t eNB_id, uint8_t subframe, uint8_t restricted_set = ue->frame_parms.prach_config_common.prach_ConfigInfo.highSpeedFlag; //uint8_t n_ra_prboffset = ue->frame_parms.prach_config_common.prach_ConfigInfo.prach_FreqOffset; uint8_t preamble_index = ue->prach_resources[eNB_id]->ra_PreambleIndex; - uint8_t tdd_mapindex = ue->prach_resources[eNB_id]->ra_TDD_map_index; + //uint8_t tdd_mapindex = ue->prach_resources[eNB_id]->ra_TDD_map_index; int16_t *prachF = ue->prach_vars[eNB_id]->prachF; static int16_t prach_tmp[45600*2] __attribute__((aligned(32))); int16_t *prach = prach_tmp; @@ -582,7 +583,7 @@ int32_t generate_nr_prach( PHY_VARS_NR_UE *ue, uint8_t eNB_id, uint8_t subframe, int16_t amp = ue->prach_vars[eNB_id]->amp; int16_t Ncp; uint8_t n_ra_prb; - uint16_t NCS; + uint16_t NCS=0; uint16_t *prach_root_sequence_map; uint16_t preamble_offset,preamble_shift; uint16_t preamble_index0,n_shift_ra,n_shift_ra_bar; @@ -599,7 +600,7 @@ int32_t generate_nr_prach( PHY_VARS_NR_UE *ue, uint8_t eNB_id, uint8_t subframe, int32_t Xu_re,Xu_im; uint16_t offset,offset2; int prach_start; - int i, prach_len; + int i, prach_len=0; uint16_t first_nonzero_root_idx=0; #if defined(EXMIMO) || defined(OAI_USRP) diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h b/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h index 79248b08c6..3c272b9c17 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h @@ -32,6 +32,7 @@ #ifndef __NR_TRANSPORT_PROTO_UE__H__ #define __NR_TRANSPORT_PROTO_UE__H__ #include "PHY/defs_nr_UE.h" +#include "SCHED_NR_UE/defs.h" //#include "PHY/LTE_TRANSPORT/transport_common_proto.h" #include #include "nfapi_interface.h" @@ -46,7 +47,7 @@ \brief This function frees memory allocated for a particular DLSCH at UE @param dlsch Pointer to DLSCH to be removed */ -void free_ue_dlsch(NR_UE_DLSCH_t *dlsch); +void free_nr_ue_dlsch(NR_UE_DLSCH_t *dlsch); /** \fn new_ue_dlsch(uint8_t Kmimo,uint8_t Mdlharq,uint32_t Nsoft,uint8_t abstraction_flag) \brief This function allocates structures for a particular DLSCH at UE @@ -57,7 +58,7 @@ void free_ue_dlsch(NR_UE_DLSCH_t *dlsch); @params N_RB_DL total number of resource blocks (determine the operating BW) @param abstraction_flag Flag to indicate abstracted interface */ -NR_UE_DLSCH_t *new_ue_dlsch(uint8_t Kmimo,uint8_t Mdlharq,uint32_t Nsoft,uint8_t max_turbo_iterations,uint8_t N_RB_DL, uint8_t abstraction_flag); +NR_UE_DLSCH_t *new_nr_ue_dlsch(uint8_t Kmimo,uint8_t Mdlharq,uint32_t Nsoft,uint8_t max_turbo_iterations,uint8_t N_RB_DL, uint8_t abstraction_flag); void free_ue_ulsch(NR_UE_ULSCH_t *ulsch); @@ -106,7 +107,7 @@ void qpsk_qpsk(int16_t *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr128p pointer to pointer to symbol in dlsch_llr*/ -int32_t dlsch_qpsk_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, +int32_t nr_dlsch_qpsk_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, int32_t **rxdataF_comp, int32_t **rxdataF_comp_i, int32_t **rho_i, @@ -142,7 +143,7 @@ void qpsk_qam16(int16_t *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr128p pointer to pointer to symbol in dlsch_llr*/ -int32_t dlsch_qpsk_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, +int32_t nr_dlsch_qpsk_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, int32_t **rxdataF_comp, int32_t **rxdataF_comp_i, int **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) @@ -154,6 +155,7 @@ int32_t dlsch_qpsk_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, uint16_t pbch_pss_sss_adj, int16_t **llr128p); + /** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream QPSK/64QAM reception. @param stream0_in Input from channel compensated (MR combined) stream 0 @param stream1_in Input from channel compensated (MR combined) stream 1 @@ -179,7 +181,7 @@ void qpsk_qam64(int16_t *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr128p pointer to pointer to symbol in dlsch_llr*/ -int32_t dlsch_qpsk_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, +int32_t nr_dlsch_qpsk_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, int32_t **rxdataF_comp, int32_t **rxdataF_comp_i, int **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) @@ -217,7 +219,7 @@ void qam16_qpsk(short *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int dlsch_16qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, +int nr_dlsch_16qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int **rxdataF_comp_i, int **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) @@ -258,7 +260,7 @@ void qam16_qam16(short *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int dlsch_16qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, +int nr_dlsch_16qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int **rxdataF_comp_i, int **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) @@ -300,7 +302,7 @@ void qam16_qam64(short *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int dlsch_16qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, +int nr_dlsch_16qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int **rxdataF_comp_i, int **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) @@ -339,7 +341,7 @@ void qam64_qpsk(short *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int dlsch_64qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, +int nr_dlsch_64qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int **rxdataF_comp_i, int **dl_ch_mag, @@ -396,7 +398,7 @@ void qam64_qam16_avx2(short *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int dlsch_64qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, +int nr_dlsch_64qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int **rxdataF_comp_i, int **dl_ch_mag, @@ -454,7 +456,7 @@ void qam64_qam64_avx2(int32_t *stream0_in, @param nb_rb number of RBs for this allocation @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int dlsch_64qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, +int nr_dlsch_64qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int **rxdataF_comp_i, int **dl_ch_mag, @@ -480,15 +482,13 @@ int dlsch_64qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, @param llr128p pointer to pointer to symbol in dlsch_llr @param beamforming_mode beamforming mode */ -int32_t dlsch_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adj, - //int16_t **llr128p, - uint8_t beamforming_mode); +int32_t nr_dlsch_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, + int32_t **rxdataF_comp, + int16_t *dlsch_llr, + uint8_t symbol, + uint8_t first_symbol_flag, + uint16_t nb_rb, + uint8_t beamforming_mode); /** \brief This function generates log-likelihood ratios (decoder input) for single-stream 16QAM received waveforms @@ -504,7 +504,7 @@ int32_t dlsch_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, @param beamforming_mode beamforming mode */ -int32_t dlsch_qpsk_llr_SIC(NR_DL_FRAME_PARMS *frame_parms, +int32_t nr_dlsch_qpsk_llr_SIC(NR_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int32_t **sic_buffer, int **rho_i, @@ -515,15 +515,14 @@ int32_t dlsch_qpsk_llr_SIC(NR_DL_FRAME_PARMS *frame_parms, uint16_t mod_order_0, uint32_t rb_alloc); -void dlsch_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, +void nr_dlsch_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, int32_t **rxdataF_comp, int16_t *dlsch_llr, int32_t **dl_ch_mag, uint8_t symbol, uint8_t first_symbol_flag, uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr128p, + int16_t **llr32p, uint8_t beamforming_mode); /** \brief This function generates log-likelihood ratios (decoder input) for single-stream 16QAM received waveforms @@ -538,7 +537,7 @@ void dlsch_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, @param pbch_pss_sss_adjust PBCH/PSS/SSS RE adjustment (in REs) @param beamforming_mode beamforming mode */ -void dlsch_16qam_llr_SIC (NR_DL_FRAME_PARMS *frame_parms, +void nr_dlsch_16qam_llr_SIC (NR_DL_FRAME_PARMS *frame_parms, int32_t **rxdataF_comp, int32_t **sic_buffer, //Q15 int32_t **rho_i, @@ -563,8 +562,7 @@ void dlsch_64qam_llr_SIC(NR_DL_FRAME_PARMS *frame_parms, uint16_t mod_order_0, uint32_t rb_alloc); - -void dlsch_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, +void nr_dlsch_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, int32_t **rxdataF_comp, int16_t *dlsch_llr, int32_t **dl_ch_mag, @@ -572,8 +570,6 @@ void dlsch_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, uint8_t symbol, uint8_t first_symbol_flag, uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - //int16_t **llr_save, uint32_t llr_offset, uint8_t beamforming_mode); @@ -708,7 +704,7 @@ void dlsch_detection_mrc_TM34(NR_DL_FRAME_PARMS *frame_parms, @param high_speed_flag @param frame_parms Pointer to frame descriptor */ -uint16_t dlsch_extract_rbs_single(int32_t **rxdataF, +/*uint16_t nr_dlsch_extract_rbs_single(int32_t **rxdataF, int32_t **dl_ch_estimates, int32_t **rxdataF_ext, int32_t **dl_ch_estimates_ext, @@ -718,7 +714,21 @@ uint16_t dlsch_extract_rbs_single(int32_t **rxdataF, uint8_t symbol, uint8_t subframe, uint32_t high_speed_flag, - NR_DL_FRAME_PARMS *frame_parms); + NR_DL_FRAME_PARMS *frame_parms);*/ + +unsigned short nr_dlsch_extract_rbs_single(int **rxdataF, + int **dl_ch_estimates, + int **rxdataF_ext, + int **dl_ch_estimates_ext, + unsigned short pmi, + unsigned char *pmi_ext, + unsigned int *rb_alloc, + unsigned char symbol, + unsigned short start_rb, + unsigned short nb_pdsch_rb, + unsigned char nr_tti_rx, + uint32_t high_speed_flag, + NR_DL_FRAME_PARMS *frame_parms); /** \fn dlsch_extract_rbs_dual(int32_t **rxdataF, int32_t **dl_ch_estimates, @@ -802,7 +812,7 @@ uint16_t dlsch_extract_rbs_TM7(int32_t **rxdataF, @param output_shift Rescaling for compensated output (should be energy-normalizing) @param phy_measurements Pointer to UE PHY measurements */ -void dlsch_channel_compensation(int32_t **rxdataF_ext, +void nr_dlsch_channel_compensation(int32_t **rxdataF_ext, int32_t **dl_ch_estimates_ext, int32_t **dl_ch_mag, int32_t **dl_ch_magb, @@ -816,6 +826,12 @@ void dlsch_channel_compensation(int32_t **rxdataF_ext, uint8_t output_shift, PHY_NR_MEASUREMENTS *phy_measurements); +void nr_dlsch_deinterleaving(uint8_t symbol, + uint16_t L, + uint16_t *llr, + uint16_t *llr_deint, + uint16_t nb_rb_pdsch); + void dlsch_dual_stream_correlation(NR_DL_FRAME_PARMS *frame_parms, unsigned char symbol, unsigned short nb_rb, @@ -923,7 +939,7 @@ void dlsch_channel_compensation_TM34(NR_DL_FRAME_PARMS *frame_parms, @param pilots_flag Flag to indicate pilots in symbol @param nb_rb Number of allocated RBs */ -void dlsch_channel_level(int32_t **dl_ch_estimates_ext, +void nr_dlsch_channel_level(int32_t **dl_ch_estimates_ext, NR_DL_FRAME_PARMS *frame_parms, int32_t *avg, uint8_t pilots_flag, @@ -953,7 +969,7 @@ void dlsch_channel_level_TM7(int32_t **dl_bf_ch_estimates_ext, uint8_t pilots_flag, uint16_t nb_rb); -void dlsch_scale_channel(int32_t **dl_ch_estimates_ext, +void nr_dlsch_scale_channel(int32_t **dl_ch_estimates_ext, NR_DL_FRAME_PARMS *frame_parms, NR_UE_DLSCH_t **dlsch_ue, uint8_t symbol_mod, @@ -977,17 +993,42 @@ void dlsch_scale_channel(int32_t **dl_ch_estimates_ext, @param llr8_flag If 1, indicate that the 8-bit turbo decoder should be used @returns 0 on success, 1 on unsuccessful decoding */ -uint32_t dlsch_decoding(PHY_VARS_NR_UE *phy_vars_ue, - int16_t *dlsch_llr, - NR_DL_FRAME_PARMS *lte_frame_parms, - NR_UE_DLSCH_t *dlsch, - NR_DL_UE_HARQ_t *harq_process, - uint32_t frame, - uint8_t subframe, - uint8_t harq_pid, - uint8_t is_crnti, - uint8_t llr8_flag); +uint32_t nr_dlsch_decoding(PHY_VARS_NR_UE *phy_vars_ue, + short *dlsch_llr, + NR_DL_FRAME_PARMS *frame_parms, + NR_UE_DLSCH_t *dlsch, + NR_DL_UE_HARQ_t *harq_process, + uint32_t frame, + uint16_t nb_symb_sch, + uint8_t nr_tti_rx, + uint8_t harq_pid, + uint8_t is_crnti, + uint8_t llr8_flag); + +uint32_t nr_dlsch_decoding_mthread(PHY_VARS_NR_UE *phy_vars_ue, + UE_nr_rxtx_proc_t *proc, + int eNB_id, + short *dlsch_llr, + NR_DL_FRAME_PARMS *frame_parms, + NR_UE_DLSCH_t *dlsch, + NR_DL_UE_HARQ_t *harq_process, + uint32_t frame, + uint16_t nb_symb_sch, + uint8_t nr_tti_rx, + uint8_t harq_pid, + uint8_t is_crnti, + uint8_t llr8_flag); + +void *nr_dlsch_decoding_2thread0(void *arg); + +void *nr_dlsch_decoding_2thread1(void *arg); + +void nr_dlsch_unscrambling(int16_t* llr, + uint32_t size, + uint8_t q, + uint32_t Nid, + uint32_t n_RNTI); uint32_t dlsch_decoding_emul(PHY_VARS_NR_UE *phy_vars_ue, uint8_t subframe, @@ -1011,7 +1052,7 @@ uint32_t dlsch_decoding_emul(PHY_VARS_NR_UE *phy_vars_ue, @param rx_type. rx_type=RX_IC_single_stream will enable interference cancellation of a second stream when decoding the first stream. In case of TM1, 2, 5, and this can cancel interference from a neighbouring cell given by eNB_id_i. In case of TM5, eNB_id_i should be set to n_connected_eNB to perform multi-user interference cancellation. In case of TM3, eNB_id_i should be set to eNB_id to perform co-channel interference cancellation; this option should be used together with an interference cancellation step [...]. In case of TM3, if rx_type=RX_IC_dual_stream, both streams will be decoded by applying the IC single stream receiver twice. @param i_mod Modulation order of the interfering stream */ -int32_t rx_pdsch(PHY_VARS_NR_UE *phy_vars_ue, +int32_t nr_rx_pdsch(PHY_VARS_NR_UE *phy_vars_ue, PDSCH_t type, uint8_t eNB_id, uint8_t eNB_id_i, @@ -1023,12 +1064,16 @@ int32_t rx_pdsch(PHY_VARS_NR_UE *phy_vars_ue, uint8_t i_mod, uint8_t harq_pid); -int32_t rx_pdcch(PHY_VARS_NR_UE *ue, - uint32_t frame, - uint8_t subframe, - uint8_t eNB_id, - MIMO_mode_t mimo_mode, - uint32_t high_speed_flag); +int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, + uint32_t frame, + uint8_t nr_tti_rx, + uint8_t eNB_id, + MIMO_mode_t mimo_mode, + uint32_t high_speed_flag, + uint8_t is_secondary_ue, + int nb_coreset_active, + uint16_t symbol_mon, + NR_SEARCHSPACE_TYPE_t searchSpaceType); /*! \brief Extract PSS and SSS resource elements @param phy_vars_ue Pointer to UE variables @@ -1386,37 +1431,6 @@ uint32_t ulsch_encoding(uint8_t *a, uint8_t control_only_flag, uint8_t Nbundled); - - - -/* \brief This routine demodulates the PHICH and updates PUSCH/ULSCH parameters. - @param phy_vars_ue Pointer to UE variables - @param proc Pointer to RXN_TXNp4 proc - @param subframe Subframe of received PDCCH/PHICH - @param eNB_id Index of eNB -*/ - -void rx_phich(PHY_VARS_NR_UE *phy_vars_ue, - UE_nr_rxtx_proc_t *proc, - uint8_t subframe, - uint8_t eNB_id); - - -/** \brief This routine provides the relationship between a PHICH TXOp and its corresponding PUSCH subframe (Table 8.3.-1 from 36.213). - @param frame_parms Pointer to DL frame configuration parameters - @param subframe Subframe of received/transmitted PHICH - @returns subframe of PUSCH transmission -*/ -uint8_t phich_subframe2_pusch_subframe(NR_DL_FRAME_PARMS *frame_parms,uint8_t subframe); - -/** \brief This routine provides the relationship between a PHICH TXOp and its corresponding PUSCH frame (Table 8.3.-1 from 36.213). - @param frame_parms Pointer to DL frame configuration parameters - @param frame Frame of received/transmitted PHICH - @param subframe Subframe of received/transmitted PHICH - @returns frame of PUSCH transmission -*/ -int phich_frame2_pusch_frame(NR_DL_FRAME_PARMS *frame_parms, int frame, int subframe); - void print_CQI(void *o,UCI_format_t uci_format,uint8_t eNB_id,int N_RB_DL); void fill_CQI(NR_UE_ULSCH_t *ulsch,PHY_NR_MEASUREMENTS *meas,uint8_t eNB_id, uint8_t harq_pid,int N_RB_DL, rnti_t rnti, uint8_t trans_mode,double sinr_eff); @@ -1638,10 +1652,71 @@ uint8_t get_prach_prb_offset(NR_DL_FRAME_PARMS *frame_parms, uint8_t tdd_mapindex, uint16_t Nf); void nr_pdcch_unscrambling(uint16_t crnti, NR_DL_FRAME_PARMS *frame_parms, uint8_t nr_tti_rx, - uint16_t *z, uint32_t length, uint16_t pdcch_DMRS_scrambling_id, int do_common); + int16_t *z, uint32_t length, uint16_t pdcch_DMRS_scrambling_id, int do_common); uint32_t lte_gold_generic(uint32_t *x1, uint32_t *x2, uint8_t reset); +uint8_t nr_dci_decoding_procedure(int s, + int p, + PHY_VARS_NR_UE *ue, + NR_DCI_ALLOC_t *dci_alloc, + NR_SEARCHSPACE_TYPE_t searchSpacetype, + int16_t eNB_id, + uint8_t nr_tti_rx, + uint8_t dci_fields_sizes_cnt[MAX_NR_DCI_DECODED_SLOT][NBR_NR_DCI_FIELDS][NBR_NR_FORMATS], + uint16_t n_RB_ULBWP, + uint16_t n_RB_DLBWP, + crc_scrambled_t *crc_scrambled, + format_found_t *format_found, + uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES]); + +int nr_generate_ue_ul_dlsch_params_from_dci(PHY_VARS_NR_UE *ue, + uint8_t eNB_id, + int frame, + uint8_t nr_tti_rx, + uint32_t dci_pdu[4], + uint16_t rnti, + uint8_t dci_length, + NR_DCI_format_t dci_format, + NR_UE_PDCCH *pdcch_vars, + NR_UE_PDSCH *pdsch_vars, + NR_UE_DLSCH_t **dlsch, + NR_UE_ULSCH_t *ulsch, + NR_DL_FRAME_PARMS *frame_parms, + PDSCH_CONFIG_DEDICATED *pdsch_config_dedicated, + uint8_t beamforming_mode, + uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS][NBR_NR_FORMATS], + uint16_t n_RB_ULBWP, + uint16_t n_RB_DLBWP, + uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES], + NR_DCI_INFO_EXTRACTED_t *nr_dci_info_extracted); + +int nr_rx_pdsch(PHY_VARS_NR_UE *ue, + PDSCH_t type, + unsigned char eNB_id, + unsigned char eNB_id_i, //if this == ue->n_connected_eNB, we assume MU interference + uint32_t frame, + uint8_t nr_tti_rx, + unsigned char symbol, + unsigned char first_symbol_flag, + RX_type_t rx_type, + unsigned char i_mod, + unsigned char harq_pid); + +uint32_t nr_get_G(uint16_t nb_rb, uint16_t nb_symb_sch,uint8_t nb_re_dmrs,uint16_t length_dmrs, uint8_t Qm, uint8_t Nl) ; + +uint32_t nr_dlsch_decoding(PHY_VARS_NR_UE *phy_vars_ue, + short *dlsch_llr, + NR_DL_FRAME_PARMS *frame_parms, + NR_UE_DLSCH_t *dlsch, + NR_DL_UE_HARQ_t *harq_process, + uint32_t frame, + uint16_t nb_symb_sch, + uint8_t nr_tti_rx, + uint8_t harq_pid, + uint8_t is_crnti, + uint8_t llr8_flag); + /**@}*/ #endif diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_transport_ue.h b/openair1/PHY/NR_UE_TRANSPORT/nr_transport_ue.h index 28667df2ca..b41b23dd32 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/nr_transport_ue.h +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_transport_ue.h @@ -34,13 +34,10 @@ #include #include "PHY/impl_defs_top.h" - +#include "PHY/CODING/nrLDPC_decoder/nrLDPC_types.h" //#include "PHY/defs_nr_UE.h" -//#include "../LTE_TRANSPORT/dci.h" -//#include "../LTE_TRANSPORT/mdci.h" -//#include "../LTE_TRANSPORT/uci_common.h" -//#include "../LTE_TRANSPORT/transport_common.h" +#include "../NR_TRANSPORT/nr_transport_common_proto.h" #ifndef STANDALONE_COMPILE #include "UTIL/LISTS/list.h" #endif @@ -117,39 +114,31 @@ typedef struct { uint8_t *b; /// Pointers to transport block segments uint8_t *c[MAX_NUM_ULSCH_SEGMENTS]; - /// RTC values for each segment (for definition see 36-212 V8.6 2009-03, p.15) - uint32_t RTC[MAX_NUM_ULSCH_SEGMENTS]; /// Index of current HARQ round for this ULSCH uint8_t round; /// MCS format of this ULSCH uint8_t mcs; /// Redundancy-version of the current sub-frame uint8_t rvidx; - /// Turbo-code outputs (36-212 V8.6 2009-03, p.12 - uint8_t d[MAX_NUM_ULSCH_SEGMENTS][(96+3+(3*6144))]; - /// Sub-block interleaver outputs (36-212 V8.6 2009-03, p.16-17) - uint8_t w[MAX_NUM_ULSCH_SEGMENTS][3*6144]; - /// Number of code segments (for definition see 36-212 V8.6 2009-03, p.9) + /// LDPC-code outputs + uint8_t d[MAX_NUM_ULSCH_SEGMENTS][3*8448]; + /// Interleaver outputs + uint8_t w[MAX_NUM_ULSCH_SEGMENTS][3*8448]; + /// Number of code segments uint32_t C; - /// Number of "small" code segments (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Cminus; - /// Number of "large" code segments (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Cplus; - /// Number of bits in "small" code segments (<6144) (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Kminus; - /// Number of bits in "large" code segments (<6144) (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Kplus; + /// Number of bits in code segments + uint32_t K; /// Total number of bits across all segments uint32_t sumKr; - /// Number of "Filler" bits (for definition see 36-212 V8.6 2009-03, p.10) + /// Number of "Filler" bits uint32_t F; - /// Msc_initial, Initial number of subcarriers for ULSCH (36-212, v8.6 2009-03, p.26-27) + /// Msc_initial, Initial number of subcarriers for ULSCH uint16_t Msc_initial; - /// Nsymb_initial, Initial number of symbols for ULSCH (36-212, v8.6 2009-03, p.26-27) + /// Nsymb_initial, Initial number of symbols for ULSCH uint8_t Nsymb_initial; - /// n_DMRS for cyclic shift of DMRS (36.213 Table 9.1.2-2) + /// n_DMRS for cyclic shift of DMRS uint8_t n_DMRS; - /// n_DMRS2 for cyclic shift of DMRS (36.211 Table 5.5.1.1.-1) + /// n_DMRS2 for cyclic shift of DMRS uint8_t n_DMRS2; /// Flag to indicate that this is a control only ULSCH (i.e. no MAC SDU) uint8_t control_only; @@ -261,9 +250,7 @@ typedef struct { /// Pointer to the payload uint8_t *b; /// Pointers to transport block segments - uint8_t *c[MAX_NUM_DLSCH_SEGMENTS]; - /// RTC values for each segment (for definition see 36-212 V8.6 2009-03, p.15) - uint32_t RTC[MAX_NUM_DLSCH_SEGMENTS]; + uint8_t *c[MAX_NUM_NR_DLSCH_SEGMENTS]; /// Index of current HARQ round for this DLSCH uint8_t round; /// MCS format for this DLSCH @@ -275,24 +262,22 @@ typedef struct { /// MIMO mode for this DLSCH MIMO_mode_t mimo_mode; /// soft bits for each received segment ("w"-sequence)(for definition see 36-212 V8.6 2009-03, p.15) - int16_t w[MAX_NUM_DLSCH_SEGMENTS][3*(6144+64)]; + int16_t w[MAX_NUM_NR_DLSCH_SEGMENTS][3*8448]; /// for abstraction soft bits for each received segment ("w"-sequence)(for definition see 36-212 V8.6 2009-03, p.15) - double w_abs[MAX_NUM_DLSCH_SEGMENTS][3*(6144+64)]; + double w_abs[MAX_NUM_NR_DLSCH_SEGMENTS][3*8448]; /// soft bits for each received segment ("d"-sequence)(for definition see 36-212 V8.6 2009-03, p.15) - int16_t *d[MAX_NUM_DLSCH_SEGMENTS]; - /// Number of code segments (for definition see 36-212 V8.6 2009-03, p.9) + int16_t *d[MAX_NUM_NR_DLSCH_SEGMENTS]; + /// LDPC processing buffers + t_nrLDPC_procBuf* p_nrLDPC_procBuf[MAX_NUM_DLSCH_SEGMENTS]; + /// Number of code segments uint32_t C; - /// Number of "small" code segments (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Cminus; - /// Number of "large" code segments (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Cplus; - /// Number of bits in "small" code segments (<6144) (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Kminus; - /// Number of bits in "large" code segments (<6144) (for definition see 36-212 V8.6 2009-03, p.10) - uint32_t Kplus; - /// Number of "Filler" bits (for definition see 36-212 V8.6 2009-03, p.10) + /// Number of bits in code segments + uint32_t K; + /// Number of "Filler" bits uint32_t F; - /// Number of MIMO layers (streams) (for definition see 36-212 V8.6 2009-03, p.17) + /// LDPC lifting factor + uint32_t Z; + /// Number of MIMO layers (streams) uint8_t Nl; /// current delta_pucch int8_t delta_PUCCH; @@ -353,29 +338,24 @@ typedef struct { uint32_t cqi_alloc2; /// saved subband PMI allocation from last PUSCH/PUCCH report uint16_t pmi_alloc; -//#if defined(UPGRADE_RAT_NR) -#if 1 /// Pointers to up to HARQ processes - NR_DL_UE_HARQ_t harq_processes[NR_MAX_DLSCH_HARQ_PROCESSES]; + NR_DL_UE_HARQ_t *harq_processes[NR_MAX_DLSCH_HARQ_PROCESSES]; // DL number of harq processes uint8_t number_harq_processes_for_pdsch; /* higher layer parameter for reception of two transport blocks TS 38.213 9.1.3.1 Type-2 HARQ-ACK codebook dtermination */ uint8_t Number_MCS_HARQ_DL_DCI; /* spatial bundling of PUCCH */ uint8_t HARQ_ACK_spatial_bundling_PUCCH; -#endif /// Maximum number of HARQ processes(for definition see 36-212 V8.6 2009-03, p.17 uint8_t Mdlharq; /// MIMO transmission mode indicator for this sub-frame (for definition see 36-212 V8.6 2009-03, p.17) uint8_t Kmimo; /// Nsoft parameter related to UE Category uint32_t Nsoft; - /// Maximum number of Turbo iterations - uint8_t max_turbo_iterations; + /// Maximum number of LDPC iterations + uint8_t max_ldpc_iterations; /// number of iterations used in last turbo decoding - uint8_t last_iteration_cnt; - - + uint8_t last_iteration_cnt; } NR_UE_DLSCH_t; typedef enum {format0_0, diff --git a/openair1/PHY/NR_UE_TRANSPORT/pss_nr.c b/openair1/PHY/NR_UE_TRANSPORT/pss_nr.c index 682d0233d8..9b09a846bb 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/pss_nr.c +++ b/openair1/PHY/NR_UE_TRANSPORT/pss_nr.c @@ -33,6 +33,7 @@ #include #include #include +#include #include "PHY/defs_nr_UE.h" @@ -383,8 +384,8 @@ void init_context_pss_nr(NR_DL_FRAME_PARMS *frame_parms_ue) bzero( primary_synchro_nr[i], sizePss); } else { - msg("Fatal memory allocation problem \n"); - assert(0); + LOG_E(PHY,"Fatal memory allocation problem \n"); + assert(0); } p = malloc(LENGTH_PSS_NR*2); if (p != NULL) { @@ -397,7 +398,7 @@ void init_context_pss_nr(NR_DL_FRAME_PARMS *frame_parms_ue) bzero( primary_synchro_time_nr[i], size); } else { - msg("Fatal memory allocation problem \n"); + LOG_E(PHY,"Fatal memory allocation problem \n"); assert(0); } @@ -408,7 +409,7 @@ void init_context_pss_nr(NR_DL_FRAME_PARMS *frame_parms_ue) bzero( pss_corr_ue[i], size); } else { - msg("Fatal memory allocation problem \n"); + LOG_E(PHY,"Fatal memory allocation problem \n"); assert(0); } @@ -437,7 +438,7 @@ void free_context_pss_nr(void) primary_synchro_time_nr[i] = NULL; } else { - msg("Fatal memory deallocation problem \n"); + LOG_E(PHY,"Fatal memory deallocation problem \n"); assert(0); } @@ -446,7 +447,7 @@ void free_context_pss_nr(void) primary_synchro_nr[i] = NULL; } else { - msg("Fatal memory deallocation problem \n"); + LOG_E(PHY,"Fatal memory deallocation problem \n"); assert(0); } @@ -455,7 +456,7 @@ void free_context_pss_nr(void) pss_corr_ue[i] = NULL; } else { - msg("Fatal memory deallocation problem \n"); + LOG_E(PHY,"Fatal memory deallocation problem \n"); assert(0); } } @@ -480,13 +481,13 @@ void init_context_synchro_nr(NR_DL_FRAME_PARMS *frame_parms_ue) /* initialise global buffers for synchronisation */ synchroF_tmp = malloc16(SYNCF_TMP_SIZE); if (synchroF_tmp == NULL) { - msg("Fatal memory allocation problem \n"); + LOG_E(PHY,"Fatal memory allocation problem \n"); assert(0); } synchro_tmp = malloc16(SYNC_TMP_SIZE); if (synchro_tmp == NULL) { - msg("Fatal memory allocation problem \n"); + LOG_E(PHY,"Fatal memory allocation problem \n"); assert(0); } @@ -518,7 +519,7 @@ void free_context_synchro_nr(void) synchroF_tmp = NULL; } else { - msg("Fatal memory deallocation problem \n"); + LOG_E(PHY,"Fatal memory deallocation problem \n"); assert(0); } @@ -527,7 +528,7 @@ void free_context_synchro_nr(void) synchro_tmp = NULL; } else { - msg("Fatal memory deallocation problem \n"); + LOG_E(PHY,"Fatal memory deallocation problem \n"); assert(0); } @@ -662,6 +663,7 @@ int pss_synchro_nr(PHY_VARS_NR_UE *PHY_vars_UE, int rate_change) NR_DL_FRAME_PARMS *frame_parms = &(PHY_vars_UE->frame_parms); int synchro_position; int **rxdata = NULL; + int fo_flag = PHY_vars_UE->UE_fo_compensation; // flag to enable freq offset estimation and compensation #ifdef DBG_PSS_NR @@ -705,7 +707,10 @@ int pss_synchro_nr(PHY_VARS_NR_UE *PHY_vars_UE, int rate_change) synchro_position = pss_search_time_nr(rxdata, frame_parms, - (int *)&PHY_vars_UE->common_vars.eNb_id); + fo_flag, + (int *)&PHY_vars_UE->common_vars.eNb_id, + (int *)&PHY_vars_UE->common_vars.freq_offset); + #if TEST_SYNCHRO_TIMING_PSS @@ -751,6 +756,15 @@ static inline int64_t abs64(int64_t x) return (((int64_t)((int32_t*)&x)[0])*((int64_t)((int32_t*)&x)[0]) + ((int64_t)((int32_t*)&x)[1])*((int64_t)((int32_t*)&x)[1])); } +static inline double angle64(int64_t x) +{ + + double re=((int32_t*)&x)[0]; + double im=((int32_t*)&x)[1]; + return (atan2(im,re)); +} + + /******************************************************************* * * NAME : pss_search_time_nr @@ -805,12 +819,15 @@ static inline int64_t abs64(int64_t x) int pss_search_time_nr(int **rxdata, ///rx data in time domain NR_DL_FRAME_PARMS *frame_parms, - int *eNB_id) + int fo_flag, + int *eNB_id, + int *f_off) { unsigned int n, ar, peak_position, pss_source; int64_t peak_value; int64_t result; int64_t avg[NUMBER_PSS_SEQUENCE]; + double ffo_est=0; unsigned int length = (NR_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe); /* 1 frame for now, it should be 2 TODO_NR */ @@ -831,13 +848,13 @@ int pss_search_time_nr(int **rxdata, ///rx data in time domain maxval = max(maxval,-primary_synchro_time_nr[1][i]); maxval = max(maxval,primary_synchro_time_nr[2][i]); maxval = max(maxval,-primary_synchro_time_nr[2][i]); - } int shift = log2_approx(maxval);//*(frame_parms->ofdm_symbol_size+frame_parms->nb_prefix_samples)*2); /* Search pss in the received buffer each 4 samples which ensures a memory alignment on 128 bits (32 bits x 4 ) */ /* This is required by SIMD (single instruction Multiple Data) Extensions of Intel processors. */ /* Correlation computation is based on a a dot product which is realized thank to SIMS extensions */ + for (int pss_index = 0; pss_index < NUMBER_PSS_SEQUENCE; pss_index++) { avg[pss_index]=0; memset(pss_corr_ue[pss_index],0,length*sizeof(int64_t)); @@ -858,15 +875,14 @@ int pss_search_time_nr(int **rxdata, ///rx data in time domain frame_parms->ofdm_symbol_size, shift); pss_corr_ue[pss_index][n] += abs64(result); - //((short*)pss_corr_ue[pss_index])[2*n] += ((short*) &result)[0]; /* real part */ //((short*)pss_corr_ue[pss_index])[2*n+1] += ((short*) &result)[1]; /* imaginary part */ - //((short*)&synchro_out)[0] += ((short*) &result)[0]; /* real part */ - //((short*)&synchro_out)[1] += ((short*) &result)[1]; /* imaginary part */ + //((short*)&synchro_out)[0] += ((int*) &result)[0]; /* real part */ + //((short*)&synchro_out)[1] += ((int*) &result)[1]; /* imaginary part */ } } - + /* calculate the absolute value of sync_corr[n] */ avg[pss_index]+=pss_corr_ue[pss_index][n]; if (pss_corr_ue[pss_index][n] > peak_value) { @@ -881,11 +897,45 @@ int pss_search_time_nr(int **rxdata, ///rx data in time domain } } + if (fo_flag){ + // fractional frequency offser computation according to Cross-correlation Synchronization Algorithm Using PSS + // Shoujun Huang, Yongtao Su, Ying He and Shan Tang, "Joint time and frequency offset estimation in LTE downlink," 7th International Conference on Communications and Networking in China, 2012. + + int64_t result1,result2; + // Computing cross-correlation at peak on half the symbol size for first half of data + result1 = dot_product64((short*)primary_synchro_time_nr[pss_source], + (short*) &(rxdata[0][peak_position]), + frame_parms->ofdm_symbol_size>>1, + shift); + // Computing cross-correlation at peak on half the symbol size for data shifted by half symbol size + // as it is real and complex it is necessary to shift by a value equal to symbol size to obtain such shift + result2 = dot_product64((short*)primary_synchro_time_nr[pss_source]+(frame_parms->ofdm_symbol_size), + (short*) &(rxdata[0][peak_position])+(frame_parms->ofdm_symbol_size), + frame_parms->ofdm_symbol_size>>1, + shift); + + int64_t re1,re2,im1,im2; + re1=((int*) &result1)[0]; + re2=((int*) &result2)[0]; + im1=((int*) &result1)[1]; + im2=((int*) &result2)[1]; + + // estimation of fractional frequency offset: angle[(result1)'*(result2)]/pi + ffo_est=atan2(re1*im2-re2*im1,re1*re2+im1*im2)/M_PI; + +#ifdef DBG_PSS_NR + printf("ffo %lf\n",ffo_est); +#endif + } + + // computing absolute value of frequency offset + *f_off = ffo_est*frame_parms->subcarrier_spacing; + for (int pss_index = 0; pss_index < NUMBER_PSS_SEQUENCE; pss_index++) avg[pss_index]/=(length/4); *eNB_id = pss_source; - LOG_I(PHY,"[UE] nr_synchro_time: Sync source = %d, Peak found at pos %d, val = %llu (%d dB) avg %d dB\n", pss_source, peak_position, (unsigned long long)peak_value, dB_fixed64(peak_value),dB_fixed64(avg[pss_source])); + LOG_I(PHY,"[UE] nr_synchro_time: Sync source = %d, Peak found at pos %d, val = %llu (%d dB) avg %d dB, ffo %lf\n", pss_source, peak_position, (unsigned long long)peak_value, dB_fixed64(peak_value),dB_fixed64(avg[pss_source]),ffo_est); if (peak_value < 5*avg[pss_source]) return(-1); diff --git a/openair1/PHY/NR_UE_TRANSPORT/pucch_nr.c b/openair1/PHY/NR_UE_TRANSPORT/pucch_nr.c index 944b6f12c0..de594ee3bf 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/pucch_nr.c +++ b/openair1/PHY/NR_UE_TRANSPORT/pucch_nr.c @@ -36,6 +36,7 @@ //#include "PHY/extern.h" //#include "LAYER2/MAC/extern.h" #include "PHY/NR_UE_TRANSPORT/pucch_nr.h" +#include "PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h" #include "common/utils/LOG/log.h" #include "common/utils/LOG/vcd_signal_dumper.h" @@ -152,10 +153,10 @@ double nr_cyclic_shift_hopping(PHY_VARS_NR_UE *ue, */ // alpha_init initialized to 2*PI/12=0.5235987756 double alpha = 0.5235987756; - uint32_t c_init = ue->pucch_config_common_nr->hoppingId; // uint16_t causes error in generating gold sequence later + uint32_t c_init = ue->pucch_config_common_nr->hoppingId; // we initialize c_init again to calculate n_cs #ifdef DEBUG_NR_PUCCH_TX - // initialization to be removed + // initialization to be remo.ved c_init=10; printf("\t\t [nr_cyclic_shift_hopping] initialization c_init=%d -> variable initialization TO BE REMOVED\n",c_init); #endif @@ -215,9 +216,9 @@ void nr_generate_pucch0(PHY_VARS_NR_UE *ue, // alpha is cyclic shift double alpha; // lnormal is the OFDM symbol number in the PUCCH transmission where l=0 corresponds to the first OFDM symbol of the PUCCH transmission - uint8_t lnormal; + //uint8_t lnormal; // lprime is the index of the OFDM symbol in the slot that corresponds to the first OFDM symbol of the PUCCH transmission in the slot given by [5, TS 38.213] - uint8_t lprime; + //uint8_t lprime; // mcs is provided by TC 38.213 subclauses 9.2.3, 9.2.4, 9.2.5 FIXME! //uint8_t mcs; @@ -232,7 +233,7 @@ void nr_generate_pucch0(PHY_VARS_NR_UE *ue, * x(l*12+n) = r_u_v_alpha_delta(n) */ // the value of u,v (delta always 0 for PUCCH) has to be calculated according to TS 38.211 Subclause 6.3.2.2.1 - uint8_t u=0,v=0,delta=0; + uint8_t u=0,v=0;//,delta=0; // if frequency hopping is disabled by the higher-layer parameter PUCCH-frequency-hopping // n_hop = 0 // if frequency hopping is enabled by the higher-layer parameter PUCCH-frequency-hopping @@ -270,8 +271,8 @@ void nr_generate_pucch0(PHY_VARS_NR_UE *ue, /* * Implementing TS 38.211 Subclause 6.3.2.3.2 Mapping to physical resources FIXME! */ - int32_t *txptr; - uint32_t re_offset; + //int32_t *txptr; + uint32_t re_offset=0; for (int l=0; lN_RB_DL>>1)) && ((frame_parms->N_RB_DL & 1) == 0)) { // if number RBs in bandwidth is even and current PRB is lower band re_offset = ((l+startingSymbolIndex)*frame_parms->ofdm_symbol_size) + (12*startingPRB) + frame_parms->first_carrier_offset; @@ -288,7 +289,7 @@ void nr_generate_pucch0(PHY_VARS_NR_UE *ue, if ((startingPRB == (frame_parms->N_RB_DL>>1)) && ((frame_parms->N_RB_DL & 1) == 1)) { // if number RBs in bandwidth is odd and current PRB contains DC re_offset = ((l+startingSymbolIndex)*frame_parms->ofdm_symbol_size) + (12*startingPRB) + frame_parms->first_carrier_offset; } - txptr = &txdataF[0][re_offset]; + //txptr = &txdataF[0][re_offset]; for (int n=0; n<12; n++){ if ((n==6) && (startingPRB == (frame_parms->N_RB_DL>>1)) && ((frame_parms->N_RB_DL & 1) == 1)) { // if number RBs in bandwidth is odd and current PRB contains DC, we need to recalculate the offset when n=6 (for second half PRB) @@ -333,7 +334,7 @@ void nr_generate_pucch1(PHY_VARS_NR_UE *ue, * */ // complex-valued symbol d_re, d_im containing complex-valued symbol d(0): - int16_t d_re, d_im; + int16_t d_re=0, d_im=0; if (nr_bit == 1) { // using BPSK if M_bit=1 according to TC 38.211 Subclause 5.1.2 d_re = (payload&1)==0 ? (int16_t)(((int32_t)amp*ONE_OVER_SQRT2)>>15) : -(int16_t)(((int32_t)amp*ONE_OVER_SQRT2)>>15); d_im = (payload&1)==0 ? (int16_t)(((int32_t)amp*ONE_OVER_SQRT2)>>15) : -(int16_t)(((int32_t)amp*ONE_OVER_SQRT2)>>15); @@ -366,7 +367,7 @@ void nr_generate_pucch1(PHY_VARS_NR_UE *ue, // alpha is cyclic shift double alpha; // lnormal is the OFDM symbol number in the PUCCH transmission where l=0 corresponds to the first OFDM symbol of the PUCCH transmission - uint8_t lnormal = 0 ; + //uint8_t lnormal = 0 ; // lprime is the index of the OFDM symbol in the slot that corresponds to the first OFDM symbol of the PUCCH transmission in the slot given by [5, TS 38.213] uint8_t lprime = startingSymbolIndex; // mcs = 0 except for PUCCH format 0 @@ -384,7 +385,7 @@ void nr_generate_pucch1(PHY_VARS_NR_UE *ue, * the complex-valued symbol d_0 shall be multiplied with a sequence r_u_v_alpha_delta(n): y(n) = d_0 * r_u_v_alpha_delta(n) */ // the value of u,v (delta always 0 for PUCCH) has to be calculated according to TS 38.211 Subclause 6.3.2.2.1 - uint8_t u=0,v=0,delta=0; + uint8_t u=0,v=0;//,delta=0; // if frequency hopping is disabled, intraSlotFrequencyHopping is not provided // n_hop = 0 // if frequency hopping is enabled, intraSlotFrequencyHopping is provided @@ -405,8 +406,8 @@ void nr_generate_pucch1(PHY_VARS_NR_UE *ue, /* * Implementing TS 38.211 Subclause 6.3.2.4.2 Mapping to physical resources */ - int32_t *txptr; - uint32_t re_offset; + //int32_t *txptr; + uint32_t re_offset=0; int i=0; #define MAX_SIZE_Z 168 // this value has to be calculated from mprime*12*table_6_3_2_4_1_1_N_SF_mprime_PUCCH_1_noHop[pucch_symbol_length]+m*12+n int16_t z_re[MAX_SIZE_Z],z_im[MAX_SIZE_Z]; @@ -586,7 +587,7 @@ void nr_generate_pucch1(PHY_VARS_NR_UE *ue, } - txptr = &txdataF[0][re_offset]; + //txptr = &txdataF[0][re_offset]; for (int n=0; n<12; n++){ if ((n==6) && (startingPRB == (frame_parms->N_RB_DL>>1)) && ((frame_parms->N_RB_DL & 1) == 1)) { // if number RBs in bandwidth is odd and current PRB contains DC, we need to recalculate the offset when n=6 (for second half PRB) @@ -938,7 +939,7 @@ void nr_uci_encoding(uint64_t payload, // L is the CRC size uint8_t L; // E is the rate matching output sequence length as given in TS 38.212 subclause 6.3.1.4.1 - uint16_t E,E_init; + uint16_t E=0,E_init; if (fmt == pucch_format2_nr) E = 16*nrofSymbols*nrofPRB; if (fmt == pucch_format3_nr){ E_init = (is_pi_over_2_bpsk_enabled == 0) ? 24:12; @@ -1095,11 +1096,11 @@ void nr_generate_pucch2(PHY_VARS_NR_UE *ue, /* * Implementing TS 38.211 Subclause 6.3.2.5.3 Mapping to physical resources */ - int32_t *txptr; - uint32_t re_offset; + //int32_t *txptr; + uint32_t re_offset=0; uint32_t x1, x2, s=0; int i=0; - int m; + int m=0; for (int l=0; lN_RB_DL>>1)) && ((frame_parms->N_RB_DL & 1) == 1)) { // if number RBs in bandwidth is odd and current PRB contains DC re_offset = ((l+startingSymbolIndex)*frame_parms->ofdm_symbol_size) + (12*(rb+startingPRB)) + frame_parms->first_carrier_offset; } - txptr = &txdataF[0][re_offset]; + //txptr = &txdataF[0][re_offset]; int k=0; int kk=0; for (int n=0; n<12; n++){ @@ -1325,7 +1326,7 @@ void nr_generate_pucch3_4(PHY_VARS_NR_UE *ue, {0, 0, 0,-1,-1,-1, 0, 0, 0, 1, 1, 1}, {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}, {0, 0, 0, 1, 1, 1, 0, 0, 0,-1,-1,-1}}; - uint8_t occ_Length = occ_length_format4; // higher layer parameter occ-Length + //uint8_t occ_Length = occ_length_format4; // higher layer parameter occ-Length uint8_t occ_Index = occ_index_format4; // higher layer parameter occ-Index //occ_Index = 1; //only for testing purposes; to be removed FIXME!!! if (fmt == pucch_format3_nr){ // no block-wise spreading for format 3 @@ -1416,7 +1417,7 @@ void nr_generate_pucch3_4(PHY_VARS_NR_UE *ue, * Implementing TS 38.211 Subclauses 6.3.2.5.3 and 6.3.2.6.5 Mapping to physical resources */ // the value of u,v (delta always 0 for PUCCH) has to be calculated according to TS 38.211 Subclause 6.3.2.2.1 - uint8_t u=0,v=0,delta=0; + uint8_t u=0,v=0;//,delta=0; // if frequency hopping is disabled, intraSlotFrequencyHopping is not provided // n_hop = 0 // if frequency hopping is enabled, intraSlotFrequencyHopping is provided @@ -1424,9 +1425,9 @@ void nr_generate_pucch3_4(PHY_VARS_NR_UE *ue, // n_hop = 1 for second hop uint8_t n_hop = 0; // lnormal is the OFDM symbol number in the PUCCH transmission where l=0 corresponds to the first OFDM symbol of the PUCCH transmission - uint8_t lnormal = 0 ; + //uint8_t lnormal = 0 ; // lprime is the index of the OFDM symbol in the slot that corresponds to the first OFDM symbol of the PUCCH transmission in the slot given by [5, TS 38.213] - uint8_t lprime = startingSymbolIndex; + //uint8_t lprime = startingSymbolIndex; // m0 is the cyclic shift index calculated depending on the Orthogonal sequence index n, according to table 6.4.1.3.3.1-1 from TS 38.211 subclause 6.4.1.3.3.1 uint8_t m0; uint8_t mcs=0; @@ -1443,13 +1444,13 @@ void nr_generate_pucch3_4(PHY_VARS_NR_UE *ue, uint8_t N_ZC = 12*nrofPRB; int16_t *r_u_v_base_re = malloc(sizeof(int16_t)*12*nrofPRB); int16_t *r_u_v_base_im = malloc(sizeof(int16_t)*12*nrofPRB); - int16_t *r_u_v_alpha_delta_re = malloc(sizeof(int16_t)*12*nrofPRB); - int16_t *r_u_v_alpha_delta_im = malloc(sizeof(int16_t)*12*nrofPRB); + //int16_t *r_u_v_alpha_delta_re = malloc(sizeof(int16_t)*12*nrofPRB); + //int16_t *r_u_v_alpha_delta_im = malloc(sizeof(int16_t)*12*nrofPRB); // Next we proceed to mapping to physical resources according to TS 38.211, subclause 6.3.2.6.5 dor PUCCH formats 3 and 4 and subclause 6.4.1.3.3.2 for DM-RS - int32_t *txptr; - uint32_t re_offset; - uint32_t x1, x2, s=0; + //int32_t *txptr; + uint32_t re_offset=0; + //uint32_t x1, x2, s=0; // intraSlotFrequencyHopping // uint8_t intraSlotFrequencyHopping = 0; uint8_t table_6_4_1_3_3_2_1_dmrs_positions[11][14] ={ @@ -1552,7 +1553,7 @@ void nr_generate_pucch3_4(PHY_VARS_NR_UE *ue, #ifdef DEBUG_NR_PUCCH_TX printf("re_offset=%d,(rb+startingPRB)=%d\n",re_offset,(rb+startingPRB)); #endif - txptr = &txdataF[0][re_offset]; + //txptr = &txdataF[0][re_offset]; for (int n=0; n<12; n++){ if ((n==6) && ((rb+startingPRB) == (frame_parms->N_RB_DL>>1)) && ((frame_parms->N_RB_DL & 1) == 1)) { // if number RBs in bandwidth is odd and current PRB contains DC, we need to recalculate the offset when n=6 (for second half PRB) diff --git a/openair1/PHY/TOOLS/cdot_prod.c b/openair1/PHY/TOOLS/cdot_prod.c index 63b74afe6d..843a9605f1 100644 --- a/openair1/PHY/TOOLS/cdot_prod.c +++ b/openair1/PHY/TOOLS/cdot_prod.c @@ -169,7 +169,7 @@ int64_t dot_product64(int16_t *x, #if defined(__x86_64__) || defined(__i386__) __m128i *x128,*y128,mmtmp1,mmtmp2,mmtmp3,mmcumul,mmcumul_re,mmcumul_im; __m128i minus_i = _mm_set_epi16(-1,1,-1,1,-1,1,-1,1); - int32_t result; + int64_t result; x128 = (__m128i*) x; y128 = (__m128i*) y; @@ -180,14 +180,13 @@ int64_t dot_product64(int16_t *x, for (n=0; n<(N>>2); n++) { // printf("n=%d, x128=%p, y128=%p\n",n,x128,y128); - // print_shorts("x",&x128[0]); - // print_shorts("y",&y128[0]); + // print_shorts("x",&x128[0]); + // print_shorts("y",&y128[0]); // this computes Re(z) = Re(x)*Re(y) + Im(x)*Im(y) mmtmp1 = _mm_madd_epi16(x128[0],y128[0]); // print_ints("retmp",&mmtmp1); // mmtmp1 contains real part of 4 consecutive outputs (32-bit) - // shift and accumulate results mmtmp1 = _mm_srai_epi32(mmtmp1,output_shift); mmcumul_re = _mm_add_epi32(mmcumul_re,mmtmp1); @@ -205,7 +204,6 @@ int64_t dot_product64(int16_t *x, mmtmp3 = _mm_madd_epi16(x128[0],mmtmp2); //print_ints("imtmp",&mmtmp3); // mmtmp3 contains imag part of 4 consecutive outputs (32-bit) - // shift and accumulate results mmtmp3 = _mm_srai_epi32(mmtmp3,output_shift); mmcumul_im = _mm_add_epi32(mmcumul_im,mmtmp3); @@ -218,13 +216,10 @@ int64_t dot_product64(int16_t *x, // this gives Re Re Im Im mmcumul = _mm_hadd_epi32(mmcumul_re,mmcumul_im); //print_ints("cumul1",&mmcumul); - // this gives Re Im Re Im mmcumul = _mm_hadd_epi32(mmcumul,mmcumul); - //print_ints("cumul2",&mmcumul); - //mmcumul = _mm_srai_epi32(mmcumul,output_shift); // extract the lower half result = _mm_extract_epi64(mmcumul,0); diff --git a/openair1/PHY/TOOLS/nr_phy_scope.c b/openair1/PHY/TOOLS/nr_phy_scope.c index 4da49702d8..3dd16ff0da 100644 --- a/openair1/PHY/TOOLS/nr_phy_scope.c +++ b/openair1/PHY/TOOLS/nr_phy_scope.c @@ -419,7 +419,7 @@ FD_lte_phy_scope_ue *create_lte_phy_scope_ue( void ) fl_set_object_lcolor( fdui->pbch_llr, FL_WHITE ); // Label color fl_set_xyplot_symbolsize( fdui->pbch_llr,2); fl_set_xyplot_xgrid( fdui->pbch_llr,FL_GRID_MAJOR); - fl_set_xyplot_xbounds( fdui->pbch_llr,0,1920); + //fl_set_xyplot_xbounds( fdui->pbch_llr,0,1920); // I/Q PBCH comp fdui->pbch_comp = fl_add_xyplot( FL_POINTS_XYPLOT, 540, 260, 240, 100, "PBCH I/Q of MF Output" ); @@ -498,24 +498,33 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, float **rxsig_t_dB; float *time; float *corr; - /* int16_t **chest_t; int16_t **chest_f; int16_t *pdsch_llr; int16_t *pdsch_comp; - int16_t *pdsch_mag; + //int16_t *pdsch_mag; int8_t *pdcch_llr; int16_t *pdcch_comp; - int8_t *pbch_llr; + int16_t *pbch_llr; int16_t *pbch_comp; - float Re,Im,ymax=1; - int num_pdcch_symbols=3; - float *llr, *bit, *chest_f_abs, llr_pbch[1920], bit_pbch[1920], *llr_pdcch, *bit_pdcch; + float llr_pbch[1920], bit_pbch[1920]; + float *llr, *bit; + float *llr_pdcch, *bit_pdcch; float *I, *Q; + int num_pdcch_symbols=2; + int num_re = 4500; + int Qm = 2; + int coded_bits_per_codeword = num_re*Qm; + int symbol, first_symbol,nb_re; + int nb_rb_pdsch =50; + float ymax=1; float **chest_t_abs; - float freq[nsymb_ce*nb_antennas_rx*nb_antennas_tx]; + float Re,Im; + float *chest_f_abs; + float *freq; + static int overlay = 0; + /* int frame = phy_vars_ue->proc.proc_rxtx[0].frame_rx; - int coded_bits_per_codeword = 0; int mcs = 0; unsigned char harq_pid = 0; */ @@ -554,21 +563,24 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, } else { coded_bits_per_codeword = 0; //frame_parms->N_RB_DL*12*get_Qm(mcs)*(frame_parms->symbols_per_tti); } + */ + I = (float*) calloc(frame_parms->ofdm_symbol_size*frame_parms->symbols_per_slot*2,sizeof(float)); + Q = (float*) calloc(frame_parms->ofdm_symbol_size*frame_parms->symbols_per_slot*2,sizeof(float)); - I = (float*) calloc(frame_parms->ofdm_symbol_size*frame_parms->symbols_per_tti*2,sizeof(float)); - Q = (float*) calloc(frame_parms->ofdm_symbol_size*frame_parms->symbols_per_tti*2,sizeof(float)); chest_t_abs = (float**) malloc(nb_antennas_rx*sizeof(float*)); for (arx=0; arxofdm_symbol_size,sizeof(float)); } - chest_f_abs = (float*) calloc(nsymb_ce*nb_antennas_rx*nb_antennas_tx,sizeof(float)); + chest_f_abs = (float*) calloc(frame_parms->ofdm_symbol_size,sizeof(float)); + freq = (float*) calloc(frame_parms->ofdm_symbol_size,sizeof(float)); + llr = (float*) calloc(coded_bits_per_codeword,sizeof(float)); // init to zero bit = malloc(coded_bits_per_codeword*sizeof(float)); + llr_pdcch = (float*) calloc(12*frame_parms->N_RB_DL*num_pdcch_symbols*2,sizeof(float)); // init to zero bit_pdcch = (float*) calloc(12*frame_parms->N_RB_DL*num_pdcch_symbols*2,sizeof(float)); - */ rxsig_t = (int16_t**) phy_vars_ue->common_vars.rxdata; rxsig_t_dB = calloc(nb_antennas_rx,sizeof(float*)); @@ -577,19 +589,19 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, } time = calloc(samples_per_frame,sizeof(float)); corr = calloc(samples_per_frame,sizeof(float)); - - /* + chest_t = (int16_t**) phy_vars_ue->common_vars.common_vars_rx_data_per_thread[phy_vars_ue->current_thread_id[subframe]].dl_ch_estimates_time[eNB_id]; chest_f = (int16_t**) phy_vars_ue->common_vars.common_vars_rx_data_per_thread[phy_vars_ue->current_thread_id[subframe]].dl_ch_estimates[eNB_id]; - pbch_llr = (int8_t*) phy_vars_ue->pbch_vars[eNB_id]->llr; + + pbch_llr = (int16_t*) phy_vars_ue->pbch_vars[eNB_id]->llr; pbch_comp = (int16_t*) phy_vars_ue->pbch_vars[eNB_id]->rxdataF_comp[0]; + pdcch_llr = (int8_t*) phy_vars_ue->pdcch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->llr; pdcch_comp = (int16_t*) phy_vars_ue->pdcch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->rxdataF_comp[0]; - pdsch_llr = (int16_t*) phy_vars_ue->pdsch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->llr[0]; // stream 0 + pdsch_llr = (int16_t*) phy_vars_ue->pdsch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->llr[0]; // stream 0 // pdsch_llr = (int16_t*) phy_vars_ue->lte_ue_pdsch_vars_SI[eNB_id]->llr[0]; // stream 0 pdsch_comp = (int16_t*) phy_vars_ue->pdsch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->rxdataF_comp0[0]; - pdsch_mag = (int16_t*) phy_vars_ue->pdsch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->dl_ch_mag0[0]; - */ + //pdsch_mag = (int16_t*) phy_vars_ue->pdsch_vars[phy_vars_ue->current_thread_id[subframe]][eNB_id]->dl_ch_mag0[0]; // Received signal in time domain of receive antenna 0 if (rxsig_t != NULL) { @@ -615,29 +627,38 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, */ } - for (ind=0;ind<3;ind++) { - if (pss_corr_ue[ind]) { - for (i=0; iis_synchronized==0) { + for (ind=0;ind<3;ind++) { + if (pss_corr_ue[ind]) { + for (i=0; ichest_t,time,corr,samples_per_frame,"","",""); + else + fl_add_xyplot_overlay(form->chest_t,ind,time,corr,samples_per_frame,rx_antenna_colors[ind]); + + overlay = 1; } - - if (ind==0) - fl_set_xyplot_data(form->chest_t,time,corr,samples_per_frame,"","",""); - else - fl_add_xyplot_overlay(form->chest_t,ind,time,corr,samples_per_frame,rx_antenna_colors[ind]); } + } + else { + + if (overlay) { //there was a previous overlay + fl_clear_xyplot(form->chest_t); + overlay = 0; } - - /* - // Channel Impulse Response (still repeated format) + // Channel Impulse Response if (chest_t != NULL) { ymax = 0; if (chest_t[0] !=NULL) { for (i=0; i<(frame_parms->ofdm_symbol_size>>3); i++) { - chest_t_abs[0][i] = (float) (chest_t[0][4*i]*chest_t[0][4*i]+chest_t[0][4*i+1]*chest_t[0][4*i+1]); + chest_t_abs[0][i] = (float) (chest_t[0][2*i]*chest_t[0][2*i]+chest_t[0][2*i+1]*chest_t[0][2*i+1]); + time[i] = (float) i; if (chest_t_abs[0][i] > ymax) ymax = chest_t_abs[0][i]; @@ -645,7 +666,7 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, fl_set_xyplot_data(form->chest_t,time,chest_t_abs[0],(frame_parms->ofdm_symbol_size>>3),"","",""); } - + /* for (arx=1; arxofdm_symbol_size>>3); i++) { @@ -659,11 +680,12 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, fl_set_xyplot_overlay_type(form->chest_t,arx,FL_DASHED_XYPLOT); } } - + */ // Avoid flickering effect // fl_get_xyplot_ybounds(form->chest_t,&ymin,&ymax); // Does not always work... fl_set_xyplot_ybounds(form->chest_t,0,(double) ymax); } + } // Channel Frequency Response (includes 5 complex sample for filter) if (chest_f != NULL) { @@ -672,7 +694,7 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, for (atx=0; atxofdm_symbol_size; k++) { freq[ind] = (float)ind; Re = (float)(chest_f[(atx<<1)+arx][(2*k)]); Im = (float)(chest_f[(atx<<1)+arx][(2*k)+1]); @@ -685,12 +707,13 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, } // tx antenna 0 - fl_set_xyplot_xbounds(form->chest_f,0,nb_antennas_rx*nb_antennas_tx*nsymb_ce); + //fl_set_xyplot_xbounds(form->chest_f,0,nb_antennas_rx*nb_antennas_tx*nsymb_ce); //fl_set_xyplot_xtics(form->chest_f,nb_antennas_rx*nb_antennas_tx*frame_parms->symbols_per_tti,2); // fl_set_xyplot_xtics(form->chest_f,nb_antennas_rx*nb_antennas_tx*2,2); - fl_set_xyplot_xgrid(form->chest_f,FL_GRID_MAJOR); - fl_set_xyplot_data(form->chest_f,freq,chest_f_abs,nsymb_ce,"","",""); + //fl_set_xyplot_xgrid(form->chest_f,FL_GRID_MAJOR); + fl_set_xyplot_data(form->chest_f,freq,chest_f_abs,frame_parms->ofdm_symbol_size,"","",""); + /* for (arx=1; arxchest_f,1,&freq[arx*nsymb_ce],&chest_f_abs[arx*nsymb_ce],nsymb_ce,rx_antenna_colors[arx]); } @@ -709,60 +732,53 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, fl_add_xyplot_overlay(form->chest_f,atx,&freq[atx*nsymb_ce],&chest_f_abs[atx*nsymb_ce],nsymb_ce,rx_antenna_colors[arx]); } } + */ } - + // PBCH LLRs if (pbch_llr != NULL) { - for (i=0; i<1920; i++) { + for (i=0; i<864; i++) { llr_pbch[i] = (float) pbch_llr[i]; bit_pbch[i] = (float) i; } - fl_set_xyplot_data(form->pbch_llr,bit_pbch,llr_pbch,1920,"","",""); + fl_set_xyplot_data(form->pbch_llr,bit_pbch,llr_pbch,864,"","",""); } + first_symbol=1; + // PBCH I/Q of MF Output if (pbch_comp!=NULL) { - for (i=0; i<72*2; i++) { - I[i] = pbch_comp[2*i]; - Q[i] = pbch_comp[2*i+1]; + for (symbol=first_symbol; symbol<(first_symbol+3); symbol++) { + if (symbol == 2 || symbol == 6) + nb_re = 72; + else + nb_re = 180; + for (i=0; ipbch_comp,I,Q,72*2,"","",""); + fl_set_xyplot_data(form->pbch_comp,I,Q,432,"","",""); } // PDCCH LLRs if (pdcch_llr != NULL) { - for (i=0; i<12*frame_parms->N_RB_DL*2*num_pdcch_symbols; i++) { - llr_pdcch[i] = (float) pdcch_llr[i]; + for (i=0; i<100; i++) { //12*frame_parms->N_RB_DL*2*num_pdcch_symbols + llr_pdcch[i] = (float) pdcch_llr[2*24*9 +i]; bit_pdcch[i] = (float) i; } - fl_set_xyplot_xbounds(form->pdcch_llr,0,12*frame_parms->N_RB_DL*2*3); - if (frame_parms->N_RB_DL != 100) - { - fl_set_xyplot_data(form->pdcch_llr,bit_pdcch,llr_pdcch,12*frame_parms->N_RB_DL*2*num_pdcch_symbols,"","",""); - } - else - { - LOG_D(PHY,"UE PDCCH LLR plot is bugged in 20 MHz BW, to be fixed !!!\n"); - } + fl_set_xyplot_data(form->pdcch_llr,bit_pdcch,llr_pdcch,12*frame_parms->N_RB_DL*num_pdcch_symbols,"","",""); } // PDCCH I/Q of MF Output if (pdcch_comp!=NULL) { - for (i=0; i<12*frame_parms->N_RB_DL*num_pdcch_symbols; i++) { - I[i] = pdcch_comp[2*i]; - Q[i] = pdcch_comp[2*i+1]; - } - if (frame_parms->N_RB_DL != 100) - { - fl_set_xyplot_data(form->pdcch_comp,I,Q,12*frame_parms->N_RB_DL*num_pdcch_symbols,"","",""); - } - else - { - LOG_D(PHY,"UE PDCCH COMP plot is bugged in 20 MHz BW, to be fixed !!!\n"); + for (i=0; i<100; i++) { + I[i] = pdcch_comp[2*50*12+2*i]; + Q[i] = pdcch_comp[2*50*12+2*i+1]; } + fl_set_xyplot_data(form->pdcch_comp,I,Q,12*frame_parms->N_RB_DL*num_pdcch_symbols,"","",""); } // PDSCH LLRs @@ -772,32 +788,20 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, bit[i] = (float) i; } - fl_set_xyplot_xbounds(form->pdsch_llr,0,coded_bits_per_codeword); - if (frame_parms->N_RB_DL != 100) - { - fl_set_xyplot_data(form->pdsch_llr,bit,llr,coded_bits_per_codeword,"","",""); - } - else - { - LOG_D(PHY,"UE PDSCH LLR plot is bugged in 20 MHz BW, to be fixed !!!\n"); - } + //fl_set_xyplot_xbounds(form->pdsch_llr,0,coded_bits_per_codeword); + fl_set_xyplot_data(form->pdsch_llr,bit,llr,coded_bits_per_codeword,"","",""); } // PDSCH I/Q of MF Output if (pdsch_comp!=NULL) { - ind=0; - - for (k=0; ksymbols_per_tti; k++) { - for (i=0; i<12*frame_parms->N_RB_DL/2; i++) { - int j = (2*frame_parms->N_RB_DL*12*k)+4*i; - I[ind] = (pdsch_mag[j ]!=0? 1.0/pdsch_mag[j ]: 0.0) * pdsch_comp[j ]*1.0; - Q[ind] = (pdsch_mag[j+1]!=0? 1.0/pdsch_mag[j+1]: 0.0) * pdsch_comp[j+1]*1.0; - ind++; - } + for (i=0; ipdsch_comp,I,Q,ind,"","",""); + + fl_set_xyplot_data(form->pdsch_comp,I,Q,nb_rb_pdsch*12,"","",""); } + /* // PDSCH Throughput memmove( tput_time_ue[UE_id], &tput_time_ue[UE_id][1], (TPUT_WINDOW_LENGTH-1)*sizeof(float) ); @@ -824,19 +828,18 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form, } free(rxsig_t_dB); - /* free(I); free(Q); - free(chest_f_abs); free(llr); free(bit); free(bit_pdcch); free(llr_pdcch); + /* + free(chest_f_abs); for (arx=0; arx= MAKE_VERSION(14, 0, 0)) @@ -420,7 +420,7 @@ typedef struct RU_t_s{ /// function pointer to NB entry routine void (*eNB_top)(struct PHY_VARS_eNB_s *eNB, int frame_rx, int subframe_rx, char *string, struct RU_t_s *ru); - void (*gNB_top)(struct PHY_VARS_gNB_s *gNB, int frame_rx, int subframe_rx, char *string, struct RU_t_s *ru); + void (*gNB_top)(struct PHY_VARS_gNB_s *gNB, int frame_rx, int slot_rx, char *string, struct RU_t_s *ru); /// Timing statistics time_stats_t ofdm_demod_stats; diff --git a/openair1/PHY/defs_common.h b/openair1/PHY/defs_common.h index bd172be885..f479c30b8c 100644 --- a/openair1/PHY/defs_common.h +++ b/openair1/PHY/defs_common.h @@ -67,7 +67,10 @@ #include +#include "TOOLS/tools_defs.h" + #include "targets/COMMON/openairinterface5g_limits.h" +#include "common/utils/LOG/log.h" #include "types.h" #include "nfapi_interface.h" @@ -102,6 +105,7 @@ #ifdef OCP_FRAMEWORK #include "enums.h" #else + typedef enum {TDD=1,FDD=0} lte_frame_type_t; typedef enum {EXTENDED=1,NORMAL=0} lte_prefix_type_t; diff --git a/openair1/PHY/defs_eNB.h b/openair1/PHY/defs_eNB.h index ab88fd133c..d23586c634 100644 --- a/openair1/PHY/defs_eNB.h +++ b/openair1/PHY/defs_eNB.h @@ -710,4 +710,5 @@ typedef struct PHY_VARS_eNB_s { int32_t pusch_stats_bsr[NUMBER_OF_UE_MAX][10240]; int32_t pusch_stats_BO[NUMBER_OF_UE_MAX][10240]; } PHY_VARS_eNB; + #endif // __PHY_DEFS_eNB_H__ diff --git a/openair1/PHY/defs_gNB.h b/openair1/PHY/defs_gNB.h index 08945a4695..24a9fd45c8 100644 --- a/openair1/PHY/defs_gNB.h +++ b/openair1/PHY/defs_gNB.h @@ -37,6 +37,7 @@ #include "defs_nr_common.h" #include "CODING/nrPolar_tools/nr_polar_pbch_defs.h" #include "openair2/NR_PHY_INTERFACE/NR_IF_Module.h" +#include "PHY/NR_TRANSPORT/nr_transport_common_proto.h" #define MAX_NUM_RU_PER_gNB MAX_NUM_RU_PER_eNB @@ -61,19 +62,104 @@ typedef struct { uint16_t size; /// Aggregation level uint8_t L; + /// HARQ PID + uint8_t harq_pid; /// PDCCH parameters nfapi_nr_dl_config_pdcch_parameters_rel15_t pdcch_params; /// CCE list nr_cce_t cce_list[NR_MAX_PDCCH_AGG_LEVEL]; /// DCI pdu - uint32_t dci_pdu[4]; + uint64_t dci_pdu[2]; } NR_gNB_DCI_ALLOC_t; typedef struct { uint8_t num_dci; + uint8_t num_pdsch_rnti; NR_gNB_DCI_ALLOC_t dci_alloc[256]; } NR_gNB_PDCCH; + +typedef struct { + /// Nfapi DLSCH PDU + nfapi_nr_dl_config_dlsch_pdu dlsch_pdu; + /// pointer to pdu from MAC interface (this is "a" in 36.212) + uint8_t *pdu; + /// The payload + CRC size in bits, "B" from 36-212 + uint32_t B; + /// Pointer to the payload + uint8_t *b; + /// Pointers to transport block segments + uint8_t *c[MAX_NUM_NR_DLSCH_SEGMENTS]; + /// Frame where current HARQ round was sent + uint32_t frame; + /// Subframe where current HARQ round was sent + uint32_t subframe; + /// Index of current HARQ round for this DLSCH + uint8_t round; + /// MIMO mode for this DLSCH + MIMO_mode_t mimo_mode; + /// Concatenated sequences + uint8_t e[MAX_NUM_NR_CHANNEL_BITS] __attribute__((aligned(32))); + /// LDPC-code outputs + uint8_t *d[MAX_NUM_NR_DLSCH_SEGMENTS]; + /// Interleaver outputs + uint8_t f[MAX_NUM_NR_CHANNEL_BITS] __attribute__((aligned(32))); + /// Number of code segments + uint32_t C; + /// Number of bits in "small" code segments + uint32_t K; + /// Number of "Filler" bits + uint32_t F; +} NR_DL_gNB_HARQ_t; + + +typedef struct { + + /// Pointers to 16 HARQ processes for the DLSCH + NR_DL_gNB_HARQ_t *harq_processes[NR_MAX_NB_HARQ_PROCESSES]; + /// TX buffers for UE-spec transmission (antenna ports 5 or 7..14, prior to precoding) + int32_t *txdataF[NR_MAX_NB_LAYERS]; + /// Modulated symbols buffer + int32_t *mod_symbs[NR_MAX_NB_CODEWORDS]; + /// beamforming weights for UE-spec transmission (antenna ports 5 or 7..14), for each codeword, maximum 4 layers? + int32_t **ue_spec_bf_weights[NR_MAX_NB_LAYERS]; + /// dl channel estimates (estimated from ul channel estimates) + int32_t **calib_dl_ch_estimates; + /// Allocated RNTI (0 means DLSCH_t is not currently used) + uint16_t rnti; + /// Active flag for baseband transmitter processing + uint8_t active; + /// HARQ process mask, indicates which processes are currently active + uint16_t harq_mask; + /// Indicator of TX activation per subframe. Used during PUCCH detection for ACK/NAK. + uint8_t subframe_tx[10]; + /// First CCE of last PDSCH scheduling per subframe. Again used during PUCCH detection for ACK/NAK. + uint8_t nCCE[10]; + /// Process ID's per subframe. Used to associate received ACKs on PUSCH/PUCCH to DLSCH harq process ids + uint8_t harq_ids[10]; + /// Window size (in outgoing transport blocks) for fine-grain rate adaptation + uint8_t ra_window_size; + /// First-round error threshold for fine-grain rate adaptation + uint8_t error_threshold; + /// Number of soft channel bits + uint32_t G; + /// Codebook index for this dlsch (0,1,2,3) + uint8_t codebook_index; + /// Maximum number of HARQ processes + uint8_t Mdlharq; + /// Maximum number of HARQ rounds + uint8_t Mlimit; + /// MIMO transmission mode indicator for this sub-frame + uint8_t Kmimo; + /// Nsoft parameter related to UE Category + uint32_t Nsoft; + /// amplitude of PDSCH (compared to RS) in symbols without pilots + int16_t sqrt_rho_a; + /// amplitude of PDSCH (compared to RS) in symbols containing pilots + int16_t sqrt_rho_b; +} NR_gNB_DLSCH_t; + + typedef struct { /// \brief Pointers (dynamic) to the received data in the time domain. /// - first index: rx antenna [0..nb_antennas_rx[ @@ -92,16 +178,16 @@ typedef struct { } NR_gNB_COMMON; -/// Context data structure for RX/TX portion of subframe processing +/// Context data structure for RX/TX portion of slot processing typedef struct { /// Component Carrier index uint8_t CC_id; /// timestamp transmitted to HW openair0_timestamp timestamp_tx; - /// subframe to act upon for transmission - int subframe_tx; - /// subframe to act upon for reception - int subframe_rx; + /// slot to act upon for transmission + int slot_tx; + /// slot to act upon for reception + int slot_rx; /// frame to act upon for transmission int frame_tx; /// frame to act upon for reception @@ -129,7 +215,7 @@ typedef struct { } gNB_L1_rxtx_proc_t; -/// Context data structure for eNB subframe processing +/// Context data structure for eNB slot processing typedef struct gNB_L1_proc_t_s { /// Component Carrier index uint8_t CC_id; @@ -139,10 +225,10 @@ typedef struct gNB_L1_proc_t_s { openair0_timestamp timestamp_rx; /// timestamp to send to "slave rru" openair0_timestamp timestamp_tx; - /// subframe to act upon for reception - int subframe_rx; - /// subframe to act upon for PRACH - int subframe_prach; + /// slot to act upon for reception + int slot_rx; + /// slot to act upon for PRACH + int slot_prach; /// frame to act upon for reception int frame_rx; /// frame to act upon for transmission @@ -306,7 +392,6 @@ typedef struct PHY_VARS_gNB_s { NR_gNB_PDCCH pdcch_vars; NR_gNB_PBCH pbch; t_nrPolar_paramsPtr nrPolar_params; - nfapi_nr_dl_config_pdcch_parameters_rel15_t pdcch_type0_params; LTE_eNB_PHICH phich_vars[2]; NR_gNB_COMMON common_vars; @@ -314,11 +399,11 @@ typedef struct PHY_VARS_gNB_s { LTE_eNB_SRS srs_vars[NUMBER_OF_UE_MAX]; LTE_eNB_PUSCH *pusch_vars[NUMBER_OF_UE_MAX]; LTE_eNB_PRACH prach_vars; - LTE_eNB_DLSCH_t *dlsch[NUMBER_OF_UE_MAX][2]; // Nusers times two spatial streams + NR_gNB_DLSCH_t *dlsch[NUMBER_OF_UE_MAX][2]; // Nusers times two spatial streams LTE_eNB_ULSCH_t *ulsch[NUMBER_OF_UE_MAX+1]; // Nusers + number of RA - LTE_eNB_DLSCH_t *dlsch_SI,*dlsch_ra,*dlsch_p; - LTE_eNB_DLSCH_t *dlsch_MCH; - LTE_eNB_DLSCH_t *dlsch_PCH; + NR_gNB_DLSCH_t *dlsch_SI,*dlsch_ra,*dlsch_p; + NR_gNB_DLSCH_t *dlsch_MCH; + NR_gNB_DLSCH_t *dlsch_PCH; LTE_eNB_UE_stats UE_stats[NUMBER_OF_UE_MAX]; LTE_eNB_UE_stats *UE_stats_ptr[NUMBER_OF_UE_MAX]; @@ -339,6 +424,9 @@ typedef struct PHY_VARS_gNB_s { /// PDCCH DMRS sequence uint32_t ***nr_gold_pdcch_dmrs; + /// PDSCH DMRS sequence + uint32_t ****nr_gold_pdsch_dmrs; + /// Indicator set to 0 after first SR uint8_t first_sr[NUMBER_OF_UE_MAX]; @@ -401,7 +489,7 @@ typedef struct PHY_VARS_gNB_s { /// cba_last successful reception for each group, used for collision detection uint8_t cba_last_reception[4]; - // Pointers for active physicalConfigDedicated to be applied in current subframe + // Pointers for active physicalConfigDedicated to be applied in current slot struct PhysicalConfigDedicated *physicalConfigDedicated[NUMBER_OF_UE_MAX]; diff --git a/openair1/PHY/defs_nr_UE.h b/openair1/PHY/defs_nr_UE.h index 456ac0816f..54aed7adb1 100644 --- a/openair1/PHY/defs_nr_UE.h +++ b/openair1/PHY/defs_nr_UE.h @@ -57,7 +57,6 @@ # include "COMMON/ral_messages_types.h" # include "UTIL/queue.h" # endif -# include "common/utils/LOG/log.h" # define msg(aRGS...) LOG_D(PHY, ##aRGS) # else # define msg printf @@ -864,16 +863,16 @@ typedef struct { int32_t **rho; /// \brief Pointer to llrs, 4-bit resolution. /// - first index: ? [0..48*N_RB_DL[ - uint16_t *llr; + int16_t *llr; /// \brief Pointer to llrs, 16-bit resolution. /// - first index: ? [0..96*N_RB_DL[ - uint16_t *llr16; + int16_t *llr16; /// \brief \f$\overline{w}\f$ from 36-211. /// - first index: ? [0..48*N_RB_DL[ - uint16_t *wbar; + int16_t *wbar; /// \brief PDCCH/DCI e-sequence (input to rate matching). /// - first index: ? [0..96*N_RB_DL[ - int8_t *e_rx; + int16_t *e_rx; /// number of PDCCH symbols in current subframe uint8_t num_pdcch_symbols; /// Allocated CRNTI for UE @@ -893,6 +892,7 @@ typedef struct { //Check for specific DCIFormat and AgregationLevel uint8_t dciFormat; uint8_t agregationLevel; + t_nrPolar_paramsPtr nrPolar_params; #ifdef NR_PDCCH_DEFS_NR_UE int nb_searchSpaces; // CORESET structure, where maximum number of CORESETs to be handled is 3 (according to 38.331 V15.1.0) @@ -901,6 +901,7 @@ typedef struct { // Each SearchSpace is associated with one ControlResourceSet NR_UE_PDCCH_SEARCHSPACE searchSpace[NR_NBR_SEARCHSPACE_ACT_BWP]; + int n_RB_BWP[NR_NBR_SEARCHSPACE_ACT_BWP]; uint32_t nb_search_space; #endif } NR_UE_PDCCH; @@ -932,6 +933,8 @@ typedef struct { /// \brief Pointer to PBCH decoded output. /// - first index: ? [0..63] (hard coded) uint8_t *decoded_output; + /// polar decoder parameters + t_nrPolar_paramsPtr nrPolar_params; /// \brief Total number of PDU errors. uint32_t pdu_errors; /// \brief Total number of PDU errors 128 frames ago. @@ -980,6 +983,8 @@ typedef struct { int UE_scan; /// \brief Indicator that UE should perform coarse scanning around carrier int UE_scan_carrier; + /// \brief Indicator that UE should enable estimation and compensation of frequency offset + int UE_fo_compensation; /// \brief Indicator that UE is synchronized to an eNB int is_synchronized; /// Data structure for UE process scheduling @@ -1018,7 +1023,12 @@ typedef struct { nr_ue_if_module_t *if_inst; nr_downlink_indication_t dl_indication; nr_uplink_indication_t ul_indication; + /// UE FAPI DCI request + nr_dcireq_t dcireq; + + /// UE FAPI indication for DLSCH reception fapi_nr_rx_indication_t rx_ind; + /// UE FAPI indication for DCI reception fapi_nr_dci_indication_t dci_ind; // point to the current rxTx thread index @@ -1062,7 +1072,7 @@ typedef struct { uint32_t dmrs_pbch_bitmap_nr[DMRS_PBCH_I_SSB][DMRS_PBCH_N_HF][DMRS_BITMAP_SIZE]; #endif - t_nrPolar_params *nrPolar_params; + /// PBCH DMRS sequence uint32_t nr_gold_pbch[2][64][NR_PBCH_DMRS_LENGTH_DWORD]; diff --git a/openair1/PHY/defs_nr_common.h b/openair1/PHY/defs_nr_common.h index cc07173473..58a8833b18 100644 --- a/openair1/PHY/defs_nr_common.h +++ b/openair1/PHY/defs_nr_common.h @@ -40,6 +40,7 @@ #include "PHY/CODING/nrPolar_tools/nr_polar_defs.h" #define nr_subframe_t lte_subframe_t +#define nr_slot_t lte_subframe_t #define MAX_NUM_SUBCARRIER_SPACING 5 @@ -53,9 +54,12 @@ #define ONE_OVER_SQRT2_Q15 23170 #define ONE_OVER_TWO_Q15 16384 -#define NR_MOD_TABLE_SIZE_SHORT 14 +#define NR_MOD_TABLE_SIZE_SHORT 686 #define NR_MOD_TABLE_BPSK_OFFSET 1 #define NR_MOD_TABLE_QPSK_OFFSET 3 +#define NR_MOD_TABLE_QAM16_OFFSET 7 +#define NR_MOD_TABLE_QAM64_OFFSET 23 +#define NR_MOD_TABLE_QAM256_OFFSET 87 #define NR_PSS_LENGTH 127 #define NR_SSS_LENGTH 127 @@ -70,6 +74,9 @@ /*used for the resource mapping*/ #define NR_MAX_PDCCH_DMRS_LENGTH 576 // 16(L)*2(QPSK)*3(3 DMRS symbs per REG)*6(REG per CCE) +#define NR_MAX_PDSCH_DMRS_LENGTH 3300 //275*6(k)*2(QPSK) +#define NR_MAX_PDSCH_DMRS_INIT_LENGTH_DWORD 104 // ceil(NR_MAX_PDSCH_DMRS_LENGTH/32) + #define NR_MAX_DCI_PAYLOAD_SIZE 64 #define NR_MAX_DCI_SIZE 1728 //16(L)*2(QPSK)*9(12 RE per REG - 3(DMRS))*6(REG per CCE) #define NR_MAX_DCI_SIZE_DWORD 54 // ceil(NR_MAX_DCI_SIZE/32) @@ -79,6 +86,13 @@ #define NR_MAX_PDCCH_AGG_LEVEL 16 #define NR_MAX_CSET_DURATION 3 +#define NR_MAX_NB_RBG 18 +#define NR_MAX_NB_LAYERS 8 +#define NR_MAX_NB_CODEWORDS 2 +#define NR_MAX_NB_HARQ_PROCESSES 16 +#define NR_MAX_PDSCH_ENCODED_LENGTH 950984 +#define NR_MAX_PDSCH_TBS 3824 + typedef enum { NR_MU_0=0, NR_MU_1, @@ -108,6 +122,36 @@ typedef enum { nr_FR2 } nr_frequency_range_e; +typedef enum { + MOD_BPSK=0, + MOD_QPSK, + MOD_QAM16, + MOD_QAM64, + MOD_QAM256 +}nr_mod_t; + +typedef struct { + /// Size of first RBG + uint8_t start_size; + /// Nominal size + uint8_t P; + /// Size of last RBG + uint8_t end_size; + /// Number of RBG + uint8_t N_RBG; +}nr_rbg_parms_t; + +typedef struct { + /// Size of first PRG + uint8_t start_size; + /// Nominal size + uint8_t P_prime; + /// Size of last PRG + uint8_t end_size; + /// Number of PRG + uint8_t N_PRG; +} nr_prg_parms_t; + typedef struct NR_BWP_PARMS { /// BWP ID uint8_t bwp_id; @@ -121,6 +165,10 @@ typedef struct NR_BWP_PARMS { uint16_t ofdm_symbol_size; /// Cyclic prefix uint8_t cyclic_prefix; + /// RBG params + nr_rbg_parms_t rbg_parms; + /// PRG params + nr_prg_parms_t prg_parms; } NR_BWP_PARMS; typedef struct { @@ -209,15 +257,19 @@ typedef struct NR_DL_FRAME_PARMS { uint16_t slots_per_frame; /// Number of samples in a subframe uint32_t samples_per_subframe; + /// Number of samples in a slot + uint32_t samples_per_slot; /// Number of OFDM/SC-FDMA symbols in one subframe (to be modified to account for potential different in UL/DL) uint16_t symbols_per_tti; /// Number of samples in a radio frame uint32_t samples_per_frame; /// Number of samples in a subframe without CP uint32_t samples_per_subframe_wCP; + /// Number of samples in a slot without CP + uint32_t samples_per_slot_wCP; /// Number of samples in a radio frame without CP uint32_t samples_per_frame_wCP; - /// Number of samples in a tti (same as subrame in LTE, depending on numerology in NR) + /// Number of samples in a tti (same as subrame in LTE, slot in NR) uint32_t samples_per_tti; /// NR numerology index [0..5] as specified in 38.211 Section 4 (mu). 0=15khZ SCS, 1=30khZ, 2=60kHz, etc uint8_t numerology_index; diff --git a/openair1/PHY/phy_extern.h b/openair1/PHY/phy_extern.h index fe888c8d2e..623a4bb0c6 100644 --- a/openair1/PHY/phy_extern.h +++ b/openair1/PHY/phy_extern.h @@ -38,6 +38,7 @@ extern unsigned int TX_DMA_BUFFER[4][NB_ANTENNAS_TX]; extern unsigned int DAQ_MBOX; extern int number_of_cards; +extern short conjugate[8],conjugate2[8]; #ifndef OCP_FRAMEWORK //extern PHY_VARS_eNB ***PHY_vars_eNB_g; diff --git a/openair1/PHY/phy_extern_nr_ue.h b/openair1/PHY/phy_extern_nr_ue.h index 2fa9ef511e..66dbd82fc4 100644 --- a/openair1/PHY/phy_extern_nr_ue.h +++ b/openair1/PHY/phy_extern_nr_ue.h @@ -19,8 +19,8 @@ * contact@openairinterface.org */ -#ifndef __PHY_EXTERN_UE__H__ -#define __PHY_EXTERN_UE__H__ +#ifndef __PHY_EXTERN_NR_UE__H__ +#define __PHY_EXTERN_NR_UE__H__ #include "PHY/defs_nr_UE.h" //#include "common/ran_context.h" @@ -34,6 +34,7 @@ extern unsigned int TX_DMA_BUFFER[4][NB_ANTENNAS_TX]; //#include "PHY/LTE_TRANSPORT/transport_extern.h" +extern short conjugate[8],conjugate2[8]; extern int number_of_cards; diff --git a/openair1/PHY/phy_extern_ue.h b/openair1/PHY/phy_extern_ue.h index 8ae424f5f7..596686f261 100644 --- a/openair1/PHY/phy_extern_ue.h +++ b/openair1/PHY/phy_extern_ue.h @@ -36,6 +36,7 @@ extern unsigned int TX_DMA_BUFFER[4][NB_ANTENNAS_TX]; extern int number_of_cards; +extern short conjugate[8],conjugate2[8]; #ifndef OCP_FRAMEWORK extern PHY_VARS_UE ***PHY_vars_UE_g; diff --git a/openair1/PHY/phy_vars.h b/openair1/PHY/phy_vars.h index e8588ce0b8..897d826ae2 100644 --- a/openair1/PHY/phy_vars.h +++ b/openair1/PHY/phy_vars.h @@ -59,6 +59,8 @@ unsigned short rev[2048],rev_times4[8192],rev_half[1024]; unsigned short rev256[256],rev512[512],rev1024[1024],rev4096[4096],rev2048[2048],rev8192[8192]; +short conjugate[8]__attribute__((aligned(16))) = {-1,1,-1,1,-1,1,-1,1}; +short conjugate2[8]__attribute__((aligned(16))) = {1,-1,1,-1,1,-1,1,-1}; #include "SIMULATION/ETH_TRANSPORT/vars.h" diff --git a/openair1/PHY/phy_vars_nr_ue.h b/openair1/PHY/phy_vars_nr_ue.h index 93a489178a..0e7e4f79d1 100644 --- a/openair1/PHY/phy_vars_nr_ue.h +++ b/openair1/PHY/phy_vars_nr_ue.h @@ -19,8 +19,8 @@ * contact@openairinterface.org */ -#ifndef __PHY_VARS_NR_H__ -#define __PHY_VARS_NR_H__ +#ifndef __PHY_VARS_NR_UE_H__ +#define __PHY_VARS_NR_UE_H__ #include "PHY/types.h" #include "PHY/defs_nr_UE.h" @@ -55,6 +55,9 @@ unsigned short rev[2048],rev_times4[8192],rev_half[1024]; unsigned short rev256[256],rev512[512],rev1024[1024],rev4096[4096],rev2048[2048],rev8192[8192]; +short conjugate[8]__attribute__((aligned(16))) = {-1,1,-1,1,-1,1,-1,1}; +short conjugate2[8]__attribute__((aligned(16))) = {1,-1,1,-1,1,-1,1,-1}; + #include "SIMULATION/ETH_TRANSPORT/vars.h" @@ -143,5 +146,5 @@ int16_t unscrambling_lut[65536*16] __attribute__((aligned(32))); /// lookup table for scrambling in TX uint8_t scrambling_lut[65536*16] __attribute__((aligned(32))); -uint8_t max_turbo_iterations=4; +uint8_t max_ldpc_iterations=4; #endif /*__PHY_VARS_H__ */ diff --git a/openair1/PHY/phy_vars_ue.h b/openair1/PHY/phy_vars_ue.h index e00b302dd7..af3657ce2c 100644 --- a/openair1/PHY/phy_vars_ue.h +++ b/openair1/PHY/phy_vars_ue.h @@ -24,7 +24,6 @@ #include "PHY/types.h" #include "PHY/defs_UE.h" -#include "PHY/phy_vars_ue.h" #include "common/ran_context.h" @@ -54,6 +53,10 @@ unsigned short rev[2048],rev_times4[8192],rev_half[1024]; unsigned short rev256[256],rev512[512],rev1024[1024],rev4096[4096],rev2048[2048],rev8192[8192]; +short conjugate[8]__attribute__((aligned(16))) = {-1,1,-1,1,-1,1,-1,1}; +short conjugate2[8]__attribute__((aligned(16))) = {1,-1,1,-1,1,-1,1,-1}; + + #include "SIMULATION/ETH_TRANSPORT/vars.h" diff --git a/openair1/SCHED/ru_procedures.c b/openair1/SCHED/ru_procedures.c index 94882c64fd..f30f09859c 100644 --- a/openair1/SCHED/ru_procedures.c +++ b/openair1/SCHED/ru_procedures.c @@ -61,7 +61,7 @@ void feptx0(RU_t *ru,int slot) { unsigned int aa,slot_offset; int slot_sizeF = (fp->ofdm_symbol_size)* ((fp->Ncp==1) ? 6 : 7); - int subframe = ru->proc.subframe_tx; + int subframe = ru->proc.tti_tx; VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM+slot , 1 ); @@ -172,7 +172,7 @@ void feptx_ofdm_2thread(RU_t *ru) { LTE_DL_FRAME_PARMS *fp=ru->frame_parms; RU_proc_t *proc = &ru->proc; struct timespec wait; - int subframe = ru->proc.subframe_tx; + int subframe = ru->proc.tti_tx; wait.tv_sec=0; wait.tv_nsec=5000000L; @@ -239,7 +239,7 @@ void feptx_ofdm(RU_t *ru) { ((fp->Ncp==1) ? 6 : 7); int len,len2; int16_t *txdata; - int subframe = ru->proc.subframe_tx; + int subframe = ru->proc.tti_tx; // int CC_id = ru->proc.CC_id; @@ -374,7 +374,7 @@ void feptx_prec(RU_t *ru) { PHY_VARS_eNB **eNB_list = ru->eNB_list,*eNB; LTE_DL_FRAME_PARMS *fp; int32_t ***bw; - int subframe = ru->proc.subframe_tx; + int subframe = ru->proc.tti_tx; if (ru->num_eNB == 1) { eNB = eNB_list[0]; @@ -420,7 +420,7 @@ void fep0(RU_t *ru,int slot) { VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPRX+slot, 1); - remove_7_5_kHz(ru,(slot&1)+(proc->subframe_rx<<1)); + remove_7_5_kHz(ru,(slot&1)+(proc->tti_rx<<1)); for (l=0; lsymbols_per_tti/2; l++) { slot_fep_ul(ru, l, @@ -537,7 +537,7 @@ void ru_fep_full_2thread(RU_t *ru) { LTE_DL_FRAME_PARMS *fp=ru->frame_parms; if ((fp->frame_type == TDD) && - (subframe_select(fp,proc->subframe_rx) != SF_UL)) return; + (subframe_select(fp,proc->tti_rx) != SF_UL)) return; if (ru->idx == 0) VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPRX, 1 ); @@ -580,7 +580,7 @@ void ru_fep_full_2thread(RU_t *ru) { stop_meas(&ru->ofdm_demod_wait_stats); if(opp_enabled == 1 && ru->ofdm_demod_wakeup_stats.p_time>30*3000){ print_meas_now(&ru->ofdm_demod_wakeup_stats,"fep wakeup",stderr); - printf("delay in fep wait on codition in frame_rx: %d subframe_rx: %d \n",proc->frame_rx,proc->subframe_rx); + printf("delay in fep wait on codition in frame_rx: %d subframe_rx: %d \n",proc->frame_rx,proc->tti_rx); } stop_meas(&ru->ofdm_demod_stats); @@ -596,13 +596,13 @@ void fep_full(RU_t *ru) { LTE_DL_FRAME_PARMS *fp=ru->frame_parms; if ((fp->frame_type == TDD) && - (subframe_select(fp,proc->subframe_rx) != SF_UL)) return; + (subframe_select(fp,proc->tti_rx) != SF_UL)) return; start_meas(&ru->ofdm_demod_stats); if (ru->idx == 0) VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPRX, 1 ); - remove_7_5_kHz(ru,proc->subframe_rx<<1); - remove_7_5_kHz(ru,1+(proc->subframe_rx<<1)); + remove_7_5_kHz(ru,proc->tti_rx<<1); + remove_7_5_kHz(ru,1+(proc->tti_rx<<1)); for (l=0; lsymbols_per_tti/2; l++) { slot_fep_ul(ru, diff --git a/openair1/SCHED_NR/fapi_nr_l1.c b/openair1/SCHED_NR/fapi_nr_l1.c index 99c8ab3037..d663a95c35 100644 --- a/openair1/SCHED_NR/fapi_nr_l1.c +++ b/openair1/SCHED_NR/fapi_nr_l1.c @@ -55,19 +55,20 @@ void handle_nr_nfapi_bch_pdu(PHY_VARS_gNB *gNB, void handle_nfapi_nr_dci_dl_pdu(PHY_VARS_gNB *gNB, - int frame, int subframe, + int frame, int slot, gNB_L1_rxtx_proc_t *proc, - nfapi_nr_dl_config_request_pdu_t *dl_config_pdu) + nfapi_nr_dl_config_request_pdu_t *dl_config_pdu, + nfapi_nr_dl_config_request_pdu_t *dl_config_dlsch_pdu) { - int idx = subframe&1; + int idx = slot&1; NR_gNB_PDCCH *pdcch_vars = &gNB->pdcch_vars; - LOG_D(PHY,"Frame %d, Subframe %d: DCI processing - populating pdcch_vars->dci_alloc[%d] proc:subframe_tx:%d idx:%d pdcch_vars->num_dci:%d\n",frame,subframe, pdcch_vars->num_dci, proc->subframe_tx, idx, pdcch_vars->num_dci); + LOG_I(PHY,"Frame %d, Slot %d: DCI processing - populating pdcch_vars->dci_alloc[%d] proc:slot_tx:%d idx:%d pdcch_vars->num_dci:%d\n",frame,slot, pdcch_vars->num_dci, proc->slot_tx, idx, pdcch_vars->num_dci); // copy dci configuration into gNB structure - nr_fill_dci_and_dlsch(gNB,frame,subframe,proc,&pdcch_vars->dci_alloc[pdcch_vars->num_dci],dl_config_pdu); + nr_fill_dci_and_dlsch(gNB,frame,slot,proc,&pdcch_vars->dci_alloc[pdcch_vars->num_dci],&dl_config_pdu->dci_dl_pdu,&dl_config_dlsch_pdu->dlsch_pdu); - LOG_D(PHY,"Frame %d, Subframe %d: DCI processing - populated pdcch_vars->dci_alloc[%d] proc:subframe_tx:%d idx:%d pdcch_vars->num_dci:%d\n",proc->frame_tx,proc->subframe_tx, pdcch_vars->num_dci, proc->subframe_tx, idx, pdcch_vars->num_dci); + LOG_I(PHY,"Frame %d, Slot %d: DCI processing - populated pdcch_vars->dci_alloc[%d] proc:slot_tx:%d idx:%d pdcch_vars->num_dci:%d\n",proc->frame_tx,proc->slot_tx, pdcch_vars->num_dci, proc->slot_tx, idx, pdcch_vars->num_dci); } @@ -80,7 +81,7 @@ void nr_schedule_response(NR_Sched_Rsp_t *Sched_INFO){ nfapi_nr_dl_config_request_t *DL_req = Sched_INFO->DL_req; nfapi_tx_request_t *TX_req = Sched_INFO->TX_req; frame_t frame = Sched_INFO->frame; - sub_frame_t subframe = Sched_INFO->subframe; + sub_frame_t slot = Sched_INFO->slot; AssertFatal(RC.gNB!=NULL,"RC.gNB is null\n"); AssertFatal(RC.gNB[Mod_id]!=NULL,"RC.gNB[%d] is null\n",Mod_id); @@ -92,20 +93,26 @@ void nr_schedule_response(NR_Sched_Rsp_t *Sched_INFO){ uint8_t number_dl_pdu = DL_req->dl_config_request_body.number_pdu; nfapi_nr_dl_config_request_pdu_t *dl_config_pdu; + nfapi_nr_dl_config_request_pdu_t *dl_config_dlsch_pdu; int i; LOG_D(PHY,"NFAPI: Sched_INFO:SFN/SF:%04d%d DL_req:SFN/SF:%04d%d:dl_pdu:%d tx_req:SFN/SF:%04d%d:pdus:%d \n", - frame,subframe, + frame,slot, NFAPI_SFNSF2SFN(DL_req->sfn_sf),NFAPI_SFNSF2SF(DL_req->sfn_sf),number_dl_pdu, NFAPI_SFNSF2SFN(TX_req->sfn_sf),NFAPI_SFNSF2SF(TX_req->sfn_sf),TX_req->tx_request_body.number_of_pdus); int do_oai =0; int dont_send =0; + gNB->pdcch_vars.num_dci = 0; + gNB->pdcch_vars.num_pdsch_rnti = 0; + + gNB->pdcch_vars.num_dci=0; for (i=0;idl_config_request_body.dl_config_pdu_list[i]; - //LOG_D(PHY,"NFAPI: dl_pdu %d : type %d\n",i,dl_config_pdu->pdu_type); + LOG_D(PHY,"NFAPI: dl_pdu %d : type %d\n",i,dl_config_pdu->pdu_type); + printf("NFAPI: dl_pdu %d : type %d\n",i,dl_config_pdu->pdu_type); switch (dl_config_pdu->pdu_type) { case NFAPI_NR_DL_CONFIG_BCH_PDU_TYPE: AssertFatal(dl_config_pdu->bch_pdu_rel15.pdu_index < TX_req->tx_request_body.number_of_pdus, @@ -122,11 +129,14 @@ void nr_schedule_response(NR_Sched_Rsp_t *Sched_INFO){ break; case NFAPI_NR_DL_CONFIG_DCI_DL_PDU_TYPE: + dl_config_dlsch_pdu = &DL_req->dl_config_request_body.dl_config_pdu_list[++i]; handle_nfapi_nr_dci_dl_pdu(gNB, - frame, subframe, + frame, slot, proc, - dl_config_pdu); + dl_config_pdu, + dl_config_dlsch_pdu); gNB->pdcch_vars.num_dci++; + gNB->pdcch_vars.num_pdsch_rnti++; do_oai=1; break; } diff --git a/openair1/SCHED_NR/fapi_nr_l1.h b/openair1/SCHED_NR/fapi_nr_l1.h index 4c6b25e969..66eb0fada7 100644 --- a/openair1/SCHED_NR/fapi_nr_l1.h +++ b/openair1/SCHED_NR/fapi_nr_l1.h @@ -42,4 +42,5 @@ void nr_schedule_response(NR_Sched_Rsp_t *Sched_INFO); void handle_nfapi_nr_dci_dl_pdu(PHY_VARS_gNB *gNB, int frame, int subframe, gNB_L1_rxtx_proc_t *proc, - nfapi_nr_dl_config_request_pdu_t *dl_config_pdu); + nfapi_nr_dl_config_request_pdu_t *dl_config_pdu, + nfapi_nr_dl_config_request_pdu_t *dl_config_dlsch_pdu); diff --git a/openair1/SCHED_NR/nr_ru_procedures.c b/openair1/SCHED_NR/nr_ru_procedures.c index 89b2b4d136..65dfb64734 100644 --- a/openair1/SCHED_NR/nr_ru_procedures.c +++ b/openair1/SCHED_NR/nr_ru_procedures.c @@ -58,38 +58,58 @@ extern openair0_config_t openair0_cfg[MAX_CARDS]; extern int oai_exit; -void nr_feptx0(RU_t *ru,int slot) { +void nr_feptx0(RU_t *ru,int first_symbol, int num_symbols) { NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; - unsigned int aa,slot_offset; - int slot_sizeF = fp->ofdm_symbol_size * fp->symbols_per_slot; - int subframe = ru->proc.subframe_tx; + unsigned int aa,slot_offset,slot_offsetF; + int slot = ru->proc.tti_tx; - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM+slot , 1 ); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM+(first_symbol!=0?1:0) , 1 ); - slot_offset = subframe*fp->samples_per_subframe + (slot*(fp->samples_per_subframe / fp->slots_per_subframe)); + slot_offset = slot*fp->samples_per_slot; + slot_offsetF = first_symbol*fp->ofdm_symbol_size; - LOG_D(PHY,"SFN/SF:RU:TX:%d/%d Generating slot %d\n",ru->proc.frame_tx, ru->proc.subframe_tx,slot); + if (first_symbol>0) slot_offset += (fp->ofdm_symbol_size*first_symbol) + (fp->nb_prefix_samples0) + (fp->nb_prefix_samples*(first_symbol-1)); + + LOG_D(PHY,"SFN/SF:RU:TX:%d/%d Generating slot %d (first_symbol %d num_symbols %d)\n",ru->proc.frame_tx, ru->proc.tti_tx,slot,first_symbol,num_symbols); for (aa=0; aanb_tx; aa++) { if (fp->Ncp == 1) { - PHY_ofdm_mod(&ru->common.txdataF_BF[aa][slot*slot_sizeF], - (int*)&ru->common.txdata[aa][slot_offset], - fp->ofdm_symbol_size, - 12, - fp->nb_prefix_samples, - CYCLIC_PREFIX); + PHY_ofdm_mod(&ru->common.txdataF_BF[aa][slot_offsetF], + (int*)&ru->common.txdata[aa][slot_offset], + fp->ofdm_symbol_size, + num_symbols, + fp->nb_prefix_samples, + CYCLIC_PREFIX); } else { - nr_normal_prefix_mod(&ru->common.txdataF_BF[aa][slot*slot_sizeF], - (int*)&ru->common.txdata[aa][slot_offset], - fp->symbols_per_slot, - fp); + if (first_symbol==0) { + PHY_ofdm_mod(&ru->common.txdataF_BF[aa][slot_offsetF], + (int*)&ru->common.txdata[aa][slot_offset], + fp->ofdm_symbol_size, + 1, + fp->nb_prefix_samples0, + CYCLIC_PREFIX); + PHY_ofdm_mod(&ru->common.txdataF_BF[aa][slot_offsetF+fp->ofdm_symbol_size], + (int*)&ru->common.txdata[aa][slot_offset+fp->nb_prefix_samples0+fp->ofdm_symbol_size], + fp->ofdm_symbol_size, + num_symbols-1, + fp->nb_prefix_samples, + CYCLIC_PREFIX); + } + else { + PHY_ofdm_mod(&ru->common.txdataF_BF[aa][slot_offsetF], + (int*)&ru->common.txdata[aa][slot_offset], + fp->ofdm_symbol_size, + num_symbols, + fp->nb_prefix_samples, + CYCLIC_PREFIX); + } } } - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM+slot , 0); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM+(first_symbol!=0?1:0), 0); } void nr_feptx_ofdm_2thread(RU_t *ru) { @@ -98,24 +118,25 @@ void nr_feptx_ofdm_2thread(RU_t *ru) { nfapi_nr_config_request_t *cfg = &ru->gNB_list[0]->gNB_config; RU_proc_t *proc = &ru->proc; struct timespec wait; - int subframe = ru->proc.subframe_tx; + int slot = ru->proc.tti_tx; wait.tv_sec=0; wait.tv_nsec=5000000L; start_meas(&ru->ofdm_mod_stats); - if (nr_subframe_select(cfg,subframe) == SF_UL) return; + if (nr_slot_select(cfg,slot) == SF_UL) return; // this copy should be done in the precoding thread (currently inactive) for (int aa=0;aanb_tx;aa++) - memcpy((void*)ru->common.txdataF_BF[aa], - (void*)ru->gNB_list[0]->common_vars.txdataF[aa], fp->samples_per_subframe_wCP*sizeof(int32_t)); + memcpy((void*)ru->common.txdataF_BF[aa], + + (void*)ru->gNB_list[0]->common_vars.txdataF[aa], fp->samples_per_slot_wCP*sizeof(int32_t)); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM , 1 ); - if (nr_subframe_select(cfg,subframe)==SF_DL) { - // If this is not an S-subframe + if (nr_slot_select(cfg,slot)==SF_DL) { + // If this is not an S-tti if (pthread_mutex_timedlock(&proc->mutex_feptx,&wait) != 0) { printf("[RU] ERROR pthread_mutex_lock for feptx thread (IC %d)\n", proc->instance_cnt_feptx); exit_fun( "error locking mutex_feptx" ); @@ -141,8 +162,8 @@ void nr_feptx_ofdm_2thread(RU_t *ru) { pthread_mutex_unlock( &proc->mutex_feptx ); } - // call first slot in this thread - nr_feptx0(ru,0); + // call first half-slot in this thread + nr_feptx0(ru,0,fp->symbols_per_slot>>1); wait_on_busy_condition(&proc->mutex_feptx,&proc->cond_feptx,&proc->instance_cnt_feptx,"NR feptx thread"); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM , 0 ); @@ -163,7 +184,7 @@ static void *nr_feptx_thread(void *param) { while (!oai_exit) { if (wait_on_condition(&proc->mutex_feptx,&proc->cond_feptx,&proc->instance_cnt_feptx,"NR feptx thread")<0) break; - nr_feptx0(ru,1); + nr_feptx0(ru,ru->nr_frame_parms->symbols_per_slot>>1,ru->nr_frame_parms->symbols_per_slot>>1); if (release_thread(&proc->mutex_feptx,&proc->instance_cnt_feptx,"NR feptx thread")<0) break; if (pthread_cond_signal(&proc->cond_feptx) != 0) { @@ -196,11 +217,11 @@ void nr_feptx_ofdm(RU_t *ru) { NR_DL_FRAME_PARMS *fp=ru->nr_frame_parms; nfapi_nr_config_request_t *cfg = &ru->gNB_list[0]->gNB_config; - unsigned int aa=0,slot=0; + unsigned int aa=0; int slot_sizeF = (fp->ofdm_symbol_size)* ((cfg->subframe_config.dl_cyclic_prefix_type.value == 1) ? 12 : 14); - int subframe = ru->proc.subframe_tx; - int *txdata = &ru->common.txdata[aa][subframe*fp->samples_per_subframe]; + int slot = ru->proc.tti_tx; + int *txdata = &ru->common.txdata[aa][slot*fp->samples_per_slot]; VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM , 1 ); start_meas(&ru->ofdm_mod_stats); @@ -208,22 +229,21 @@ void nr_feptx_ofdm(RU_t *ru) { // this copy should be done in the precoding thread (currently inactive) for (int aa=0;aanb_tx;aa++) memcpy((void*)ru->common.txdataF_BF[aa], - (void*)ru->gNB_list[0]->common_vars.txdataF[aa], fp->samples_per_subframe_wCP*sizeof(int32_t)); + (void*)ru->gNB_list[0]->common_vars.txdataF[aa], fp->samples_per_slot_wCP*sizeof(int32_t)); - if ((nr_subframe_select(cfg,subframe)==SF_DL)|| - ((nr_subframe_select(cfg,subframe)==SF_S))) { + if ((nr_slot_select(cfg,slot)==SF_DL)|| + ((nr_slot_select(cfg,slot)==SF_S))) { // LOG_D(HW,"Frame %d: Generating slot %d\n",frame,next_slot); - for (slot=0; slotslots_per_subframe;slot++) - nr_feptx0(ru,slot); + nr_feptx0(ru,0,fp->symbols_per_slot); } VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_RU_FEPTX_OFDM , 0 ); stop_meas(&ru->ofdm_mod_stats); - LOG_D(PHY,"feptx_ofdm (TXPATH): frame %d, subframe %d: txp (time %p) %d dB, txp (freq) %d dB\n", - ru->proc.frame_tx,subframe,txdata,dB_fixed(signal_energy((int32_t*)txdata,fp->samples_per_subframe)), + LOG_D(PHY,"feptx_ofdm (TXPATH): frame %d, slot %d: txp (time %p) %d dB, txp (freq) %d dB\n", + ru->proc.frame_tx,slot,txdata,dB_fixed(signal_energy((int32_t*)txdata,fp->samples_per_slot)), dB_fixed(signal_energy_nodc(ru->common.txdataF_BF[aa],2*slot_sizeF))); } diff --git a/openair1/SCHED_NR/phy_procedures_nr_common.c b/openair1/SCHED_NR/phy_procedures_nr_common.c index 7c3f49a291..1a0a02f22d 100755 --- a/openair1/SCHED_NR/phy_procedures_nr_common.c +++ b/openair1/SCHED_NR/phy_procedures_nr_common.c @@ -32,243 +32,8 @@ #include "sched_nr.h" - -/// LUT for the number of symbols in the coreset indexed by coreset index (4 MSB rmsi_pdcch_config) -uint8_t nr_coreset_nsymb_pdcch_type_0_b40Mhz[16] = {2,2,2,2,2,3,3,3,3,3,1,1,1,2,2,2}; // below 40Mhz bw -uint8_t nr_coreset_nsymb_pdcch_type_0_a40Mhz[10] = {2,2,3,3,1,1,2,2,3,3}; // above 40Mhz bw -/// LUT for the number of RBs in the coreset indexed by coreset index -uint8_t nr_coreset_rb_offset_pdcch_type_0_b40Mhz[16] = {0,1,2,3,4,0,1,2,3,4,12,14,16,12,14,16}; -uint8_t nr_coreset_rb_offset_pdcch_type_0_a40Mhz[10] = {0,4,0,4,0,28,0,28,0,28}; -/// LUT for monitoring occasions param O indexed by ss index (4 LSB rmsi_pdcch_config) -uint8_t nr_ss_param_O_type_0_mux1_FR1[16] = {0,0,2,2,5,5,7,7,0,5,0,0,2,2,5,5}; -uint8_t nr_ss_param_O_type_0_mux1_FR2[14] = {0,0,2.5,2.5,5,5,0,2.5,5,7.5,7.5,7.5,0,5}; -/// LUT for number of SS sets per slot indexed by ss index -uint8_t nr_ss_sets_per_slot_type_0_FR1[16] = {1,2,1,2,1,2,1,2,1,1,1,1,1,1,1,1}; -uint8_t nr_ss_sets_per_slot_type_0_FR2[14] = {1,2,1,2,1,2,2,2,2,1,2,2,1,1}; -/// LUT for monitoring occasions param M indexed by ss index -uint8_t nr_ss_param_M_type_0_mux1_FR1[16] = {1,0.5,1,0.5,1,0.5,1,0.5,2,2,1,1,1,1,1,1}; -uint8_t nr_ss_param_M_type_0_mux1_FR2[14] = {1,0.5,1,0.5,1,0.5,0.5,0.5,0.5,1,0.5,0.5,2,2}; -/// LUT for SS first symbol index indexed by ss index -uint8_t nr_ss_first_symb_idx_type_0_mux1_FR1[8] = {0,0,1,2,1,2,1,2}; - - - -nr_subframe_t nr_subframe_select(nfapi_nr_config_request_t *cfg,unsigned char subframe) +nr_subframe_t nr_slot_select(nfapi_nr_config_request_t *cfg,unsigned char slot) { if (cfg->subframe_config.duplex_mode.value == FDD) return(SF_DL); } - - -void nr_configure_css_dci_from_mib(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, - nr_scs_e scs_common, - nr_scs_e pdcch_scs, - nr_frequency_range_e freq_range, - uint8_t rmsi_pdcch_config, - uint8_t ssb_idx, - uint16_t nb_slots_per_frame, - uint16_t N_RB) -{ - uint8_t O, M; - uint8_t ss_idx = rmsi_pdcch_config&0xf; - uint8_t cset_idx = (rmsi_pdcch_config>>4)&0xf; - uint8_t mu; - - /// Coreset params - switch(scs_common) { - - case kHz15: - mu = 0; - break; - - case kHz30: - mu = 1; - - if (N_RB < 106) { // Minimum 40Mhz bandwidth not satisfied - switch(pdcch_scs) { - case kHz15: - break; - - case kHz30: - pdcch_params->mux_pattern = NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1; - pdcch_params->n_rb = (cset_idx < 10)? 24 : 48; - pdcch_params->n_symb = nr_coreset_nsymb_pdcch_type_0_b40Mhz[cset_idx]; - pdcch_params->rb_offset = nr_coreset_rb_offset_pdcch_type_0_b40Mhz[cset_idx]; - break; - - default: - AssertFatal(1==0,"Invalid scs_common/pdcch_scs combination %d/%d \n", scs_common, pdcch_scs); - } - } - - else { - AssertFatal(ss_idx<10 ,"Invalid scs_common/pdcch_scs combination %d/%d \n", scs_common, pdcch_scs); - switch(pdcch_scs) { - case kHz15: - break; - - case kHz30: - pdcch_params->mux_pattern = NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1; - pdcch_params->n_rb = (cset_idx < 4)? 24 : 48; - pdcch_params->n_symb = nr_coreset_nsymb_pdcch_type_0_b40Mhz[cset_idx]; - pdcch_params->rb_offset = nr_coreset_rb_offset_pdcch_type_0_b40Mhz[cset_idx]; - break; - - default: - AssertFatal(1==0,"Invalid scs_common/pdcch_scs combination %d/%d \n", scs_common, pdcch_scs); - } - } - - case kHz60: - mu = 2; - break; - - case kHz120: - mu = 3; - break; - - default: - AssertFatal(1==0,"Invalid common subcarrier spacing %d\n", scs_common); - - } - - /// Search space params - switch(pdcch_params->mux_pattern) { - - case NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1: - if (freq_range == nr_FR1) { - O = nr_ss_param_O_type_0_mux1_FR1[ss_idx]; - pdcch_params->nb_ss_sets_per_slot = nr_ss_sets_per_slot_type_0_FR1[ss_idx]; - M = nr_ss_param_M_type_0_mux1_FR1[ss_idx]; - pdcch_params->first_symbol = (ss_idx < 8)? ( (ss_idx&1)? pdcch_params->n_symb : 0 ) : nr_ss_first_symb_idx_type_0_mux1_FR1[ss_idx - 8]; - } - - else { - AssertFatal(ss_idx<14 ,"Invalid search space index for multiplexing type 1 and FR2 %d\n", ss_idx); - O = nr_ss_param_O_type_0_mux1_FR2[ss_idx]; - pdcch_params->nb_ss_sets_per_slot = nr_ss_sets_per_slot_type_0_FR2[ss_idx]; - M = nr_ss_param_M_type_0_mux1_FR2[ss_idx]; - pdcch_params->first_symbol = (ss_idx < 12)? ( (ss_idx&1)? 7 : 0 ) : 0; - } - pdcch_params->nb_slots = 2; - pdcch_params->sfn_mod2 = ((uint8_t)(floor( (O*pow(2, mu) + floor(ssb_idx*M)) / nb_slots_per_frame )) & 1)? 1 : 0; - pdcch_params->first_slot = (uint8_t)(O*pow(2, mu) + floor(ssb_idx*M)) % nb_slots_per_frame; - - break; - - case NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE2: - break; - - case NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE3: - break; - - default: - AssertFatal(1==0, "Invalid SSB and coreset multiplexing pattern %d\n", pdcch_params->mux_pattern); - } - pdcch_params->config_type = NFAPI_NR_CSET_CONFIG_MIB_SIB1; - pdcch_params->cr_mapping_type = NFAPI_NR_CCE_REG_MAPPING_INTERLEAVED; - pdcch_params->precoder_granularity = NFAPI_NR_CSET_SAME_AS_REG_BUNDLE; - pdcch_params->reg_bundle_size = 6; - pdcch_params->interleaver_size = 2; -} - -void nr_configure_css_dci_from_pdcch_config(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, - nfapi_nr_coreset_t* coreset, - nfapi_nr_search_space_t* search_space) { -//coreset - - //ControlResourceSetId - pdcch_params->config_type = coreset->coreset_id; - - //frequencyDomainResources - pdcch_params->n_rb = coreset->frequency_domain_resources; - - //duration - pdcch_params->n_symb = coreset->duration; - - //cce-REG-MappingType - switch (coreset->cce_reg_mapping_type) { - case NFAPI_NR_CCE_REG_MAPPING_NON_INTERLEAVED: //Non_interleaved - pdcch_params->reg_bundle_size = 0; - pdcch_params->interleaver_size = 0; - pdcch_params->shift_index = 0; - break; - case NFAPI_NR_CCE_REG_MAPPING_INTERLEAVED: //interleaved - pdcch_params->reg_bundle_size = coreset->reg_bundle_size; - pdcch_params->interleaver_size = coreset->interleaver_size; - pdcch_params->shift_index = coreset->shift_index; - break; - } - //precoderGranularity - pdcch_params->precoder_granularity = coreset->precoder_granularity; -/* - if (tci_present_in_dci == tciPresentInDCI_t->tciPresentInDCI_enabled){ - //not sure information - } -*/ - //pdcch-DMRS-ScramblingID - pdcch_params->scrambling_id = coreset->dmrs_scrambling_id; - - - - -//SearchSpace 38.213, Reference defs_nr_UE.h - 755-840 code - //searchSpaceId - search_space->search_space_id = 0; - - //controlResourceSetId - if(coreset->coreset_id != NULL){ - search_space->coreset_id = coreset->coreset_id; //Value 0 or Value 1 - } - - //monitoringSlotPeriodicityAndOffset - //pdcch_params->nb_slots = search_space->slot_monitoring_offset; //output not sure. - pdcch_params->nb_slots = search_space->slot_monitoring_periodicity; //nr_sl1=1,nr_sl2=2,nr_sl4=4,nr_sl5=5,nr_sl8=8,nr_sl10=10,nr_sl16=16,nr_sl20=20,nr_sl40=40,nr_sl80=80,nr_sl160=160,nr_sl320=320,nr_sl640=640,nr_sl1280=1280,nr_sl2560=2560 - - //duration - search_space->duration = search_space->slot_monitoring_periodicity-1; //The maximum valid duration is periodicity-1 - - - //monitoringSymbolsWithinSlot - pdcch_params->first_symbol = search_space->monitoring_symbols_in_slot; //38.213 page.69 - - //nrofCandidates - //NFAPI_NR_MAX_NB_CCE_AGGREGATION_LEVELS = 5 {1,2,4,8,16} - pdcch_params->aggregation_level = search_space->number_of_candidates[NFAPI_NR_MAX_NB_CCE_AGGREGATION_LEVELS]; - - if (search_space->search_space_type == 0){ - //Common_CSS - pdcch_params->search_space_type = search_space->search_space_type; - if(search_space->css_formats_0_0_and_1_0 == 0){ //Type0-"SI-RNTI, Type0A-SI-RNTI, Type1-RA-RNTI or TC-RNTI, Type2-P-RNTI, Type3-C-RNTI or CS-RNTI" - pdcch_params->dci_format = search_space->css_formats_0_0_and_1_0; - } - if(search_space->css_format_2_0 == 2){ //Type3-"SFI-RNTI" - pdcch_params->dci_format = search_space->css_format_2_0; - //Not SFI-Aggregation-LeveL - //pdcch_params->aggregation_level = search_space->number_of_candidates[NFAPI_NR_MAX_NB_CCE_AGGREGATION_LEVELS]; - } - if(search_space->css_format_2_1 == 3){ //Type3-"INT-RNTI" - pdcch_params->dci_format = search_space->css_format_2_1; - } - if(search_space->css_format_2_2 == 4){ //Type3-"TPC-PUSCH_RNTI or TPC-PUCCH-RNTI" - pdcch_params->dci_format = search_space->css_format_2_2; - } - if(search_space->css_format_2_3 == 5){ //Type3-"TPC-SRS-RNTI" - pdcch_params->dci_format = search_space->css_format_2_3; - //Monitoring periodicity of SRS PDCCH in number of slots for DCI format 2-3 - pdcch_params->nb_slots = search_space->srs_monitoring_periodicity; //mp1=1,mp2=2,mp4=4,mp5=5,mp8=8,mp10=10,mp16=16,mp20=20 - } - } - if (search_space->search_space_type == 1){ - //Ue_Specific_USS - pdcch_params->search_space_type = search_space->search_space_type; - if(search_space->uss_dci_formats == 0){ //DCI formats 0-0 and 1-0 - pdcch_params->dci_format = search_space->uss_dci_formats; - } - if(search_space->uss_dci_formats == 1){ //DCI formats 0-1 and 1-1 - pdcch_params->dci_format = search_space->uss_dci_formats; - } - } - - -} diff --git a/openair1/SCHED_NR/phy_procedures_nr_gNB.c b/openair1/SCHED_NR/phy_procedures_nr_gNB.c index f3ed2153a2..a7f8ef6884 100644 --- a/openair1/SCHED_NR/phy_procedures_nr_gNB.c +++ b/openair1/SCHED_NR/phy_procedures_nr_gNB.c @@ -23,6 +23,7 @@ #include "PHY/defs_gNB.h" #include "sched_nr.h" #include "PHY/NR_TRANSPORT/nr_transport.h" +#include "PHY/NR_TRANSPORT/nr_dlsch.h" #include "SCHED/sched_eNB.h" #include "SCHED/sched_common_extern.h" #include "nfapi_interface.h" @@ -111,47 +112,48 @@ int nr_get_ssb_start_symbol(nfapi_nr_config_request_t *cfg, NR_DL_FRAME_PARMS *f void nr_set_ssb_first_subcarrier(nfapi_nr_config_request_t *cfg, NR_DL_FRAME_PARMS *fp) { - int start_rb = cfg->sch_config.n_ssb_crb.value / (1<subframe_config.numerology_index_mu.value); - fp->ssb_start_subcarrier = 12 * start_rb + cfg->sch_config.ssb_subcarrier_offset.value; - LOG_D(PHY, "SSB first subcarrier %d (%d,%d)\n", fp->ssb_start_subcarrier,start_rb,cfg->sch_config.ssb_subcarrier_offset.value); + fp->ssb_start_subcarrier = (12 * cfg->sch_config.n_ssb_crb.value + cfg->sch_config.ssb_subcarrier_offset.value)/(1<subframe_config.numerology_index_mu.value); + LOG_D(PHY, "SSB first subcarrier %d (%d,%d)\n", fp->ssb_start_subcarrier,cfg->sch_config.n_ssb_crb.value,cfg->sch_config.ssb_subcarrier_offset.value); } -void nr_common_signal_procedures (PHY_VARS_gNB *gNB,int frame, int subframe) { +void nr_common_signal_procedures (PHY_VARS_gNB *gNB,int frame, int slot) { NR_DL_FRAME_PARMS *fp=&gNB->frame_parms; nfapi_nr_config_request_t *cfg = &gNB->gNB_config; int **txdataF = gNB->common_vars.txdataF; uint8_t *pbch_pdu=&gNB->pbch_pdu[0]; - int ss_subframe = (cfg->sch_config.half_frame_index.value)? 5 : 0; + int ss_slot = (cfg->sch_config.half_frame_index.value)? 10 : 0; uint8_t Lmax, ssb_index=0, n_hf=0; - LOG_D(PHY,"common_signal_procedures: frame %d, subframe %d\n",frame,subframe); + LOG_D(PHY,"common_signal_procedures: frame %d, slot %d\n",frame,slot); int ssb_start_symbol = nr_get_ssb_start_symbol(cfg, fp); nr_set_ssb_first_subcarrier(cfg, fp); Lmax = (fp->dl_CarrierFreq < 3e9)? 4:8; - if (subframe == ss_subframe) + if (slot == ss_slot) { // Current implementation is based on SSB in first half frame, first candidate - LOG_D(PHY,"SS TX: frame %d, subframe %d, start_symbol %d\n",frame,subframe, ssb_start_symbol); + LOG_D(PHY,"SS TX: frame %d, slot %d, start_symbol %d\n",frame,slot, ssb_start_symbol); - nr_generate_pss(gNB->d_pss, txdataF, AMP, ssb_start_symbol, cfg, fp); - nr_generate_sss(gNB->d_sss, txdataF, AMP_OVER_2, ssb_start_symbol, cfg, fp); + nr_generate_pss(gNB->d_pss, txdataF[0], AMP, ssb_start_symbol, cfg, fp); + nr_generate_sss(gNB->d_sss, txdataF[0], AMP, ssb_start_symbol, cfg, fp); if (!(frame&7)){ - LOG_D(PHY,"%d.%d : pbch_configured %d\n",frame,subframe,gNB->pbch_configured); + LOG_D(PHY,"%d.%d : pbch_configured %d\n",frame,slot,gNB->pbch_configured); if (gNB->pbch_configured != 1)return; gNB->pbch_configured = 0; } - nr_generate_pbch_dmrs(gNB->nr_gold_pbch_dmrs[n_hf][ssb_index],txdataF, AMP_OVER_2, ssb_start_symbol, cfg, fp); + + nr_generate_pbch_dmrs(gNB->nr_gold_pbch_dmrs[n_hf][ssb_index],txdataF[0], AMP, ssb_start_symbol, cfg, fp); + nr_generate_pbch(&gNB->pbch, gNB->nrPolar_params, pbch_pdu, gNB->nr_pbch_interleaver, - txdataF, - AMP_OVER_2, + txdataF[0], + AMP, ssb_start_symbol, n_hf,Lmax,ssb_index, frame, cfg, fp); @@ -160,48 +162,59 @@ void nr_common_signal_procedures (PHY_VARS_gNB *gNB,int frame, int subframe) { } void phy_procedures_gNB_TX(PHY_VARS_gNB *gNB, - gNB_L1_rxtx_proc_t *proc, - int do_meas) + gNB_L1_rxtx_proc_t *proc, + int do_meas) { int aa; int frame=proc->frame_tx; - int subframe=proc->subframe_tx; - uint8_t num_dci=0; + + int slot=proc->slot_tx; + uint8_t num_dci=0,num_pdsch_rnti; NR_DL_FRAME_PARMS *fp=&gNB->frame_parms; nfapi_nr_config_request_t *cfg = &gNB->gNB_config; int offset = gNB->CC_id; - if ((cfg->subframe_config.duplex_mode.value == TDD) && (nr_subframe_select(cfg,subframe)==SF_UL)) return; + if ((cfg->subframe_config.duplex_mode.value == TDD) && (nr_slot_select(cfg,slot)==SF_UL)) return; VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_ENB_TX+offset,1); if (do_meas==1) start_meas(&gNB->phy_proc_tx); // clear the transmit data array for the current subframe - for (aa=0; aarf_config.tx_antenna_ports.value; aa++) { - memset(gNB->common_vars.txdataF[aa],0,fp->samples_per_subframe_wCP*sizeof(int32_t)); + for (aa=0; aa<1/*15*/; aa++) { + memset(gNB->common_vars.txdataF[aa],0,fp->samples_per_slot_wCP*sizeof(int32_t)); } if (nfapi_mode == 0 || nfapi_mode == 1) { - nr_common_signal_procedures(gNB,frame, subframe); + nr_common_signal_procedures(gNB,frame, slot); //if (frame == 9) //write_output("txdataF.m","txdataF",gNB->common_vars.txdataF[aa],fp->samples_per_frame_wCP, 1, 1); } num_dci = gNB->pdcch_vars.num_dci; + + num_pdsch_rnti = gNB->pdcch_vars.num_pdsch_rnti; if (num_dci) { - LOG_D(PHY, "[gNB %d] Frame %d subframe %d \ - Calling nr_generate_dci_top (number of DCI %d)\n", gNB->Mod_id, frame, subframe, num_dci); + LOG_I(PHY, "[gNB %d] Frame %d slot %d \ + Calling nr_generate_dci_top (number of DCI %d)\n", gNB->Mod_id, frame, slot, num_dci); - uint8_t slot_idx = gNB->pdcch_vars.dci_alloc[0].pdcch_params.first_slot; - - if (nfapi_mode == 0 || nfapi_mode == 1) + if (nfapi_mode == 0 || nfapi_mode == 1){ nr_generate_dci_top(gNB->pdcch_vars, &gNB->nrPolar_params, - gNB->nr_gold_pdcch_dmrs[slot_idx], - gNB->common_vars.txdataF, + gNB->nr_gold_pdcch_dmrs[slot], + gNB->common_vars.txdataF[0], AMP, *fp, *cfg); + if (num_pdsch_rnti) { + LOG_I(PHY, "PDSCH generation started (%d)\n", num_pdsch_rnti); + nr_generate_pdsch(*gNB->dlsch[0][0], + gNB->pdcch_vars.dci_alloc[0], + gNB->nr_gold_pdsch_dmrs[slot], + gNB->common_vars.txdataF, + AMP, slot, *fp, *cfg); + } + } } + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_ENB_TX+offset,0); } diff --git a/openair1/SCHED_NR/sched_nr.h b/openair1/SCHED_NR/sched_nr.h index 17354d4fb3..9b53dd050c 100644 --- a/openair1/SCHED_NR/sched_nr.h +++ b/openair1/SCHED_NR/sched_nr.h @@ -34,22 +34,22 @@ #include "PHY/NR_TRANSPORT/nr_dci.h" -lte_subframe_t nr_subframe_select (nfapi_nr_config_request_t *cfg, unsigned char subframe); +nr_slot_t nr_slot_select (nfapi_nr_config_request_t *cfg, unsigned char slot); void nr_set_ssb_first_subcarrier(nfapi_nr_config_request_t *cfg, NR_DL_FRAME_PARMS *fp); void phy_procedures_gNB_TX(PHY_VARS_gNB *gNB, gNB_L1_rxtx_proc_t *proc, int do_meas); -void nr_common_signal_procedures (PHY_VARS_gNB *gNB,int frame, int subframe); +void nr_common_signal_procedures (PHY_VARS_gNB *gNB,int frame, int slot); void nr_init_feptx_thread(RU_t *ru,pthread_attr_t *attr_feptx); void nr_feptx_ofdm(RU_t *ru); void nr_feptx_ofdm_2thread(RU_t *ru); -void nr_feptx0(RU_t *ru,int slot); +void nr_feptx0(RU_t *ru,int first_symbol, int num_symbols); -void nr_configure_css_dci_from_mib(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, - nr_scs_e scs_common, - nr_scs_e pdcch_scs, - nr_frequency_range_e freq_range, - uint8_t rmsi_pdcch_config, - uint8_t ssb_idx, - uint16_t nb_slots_per_frame, - uint16_t N_RB); +void nr_configure_css_dci_initial(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, + nr_scs_e scs_common, + nr_scs_e pdcch_scs, + nr_frequency_range_e freq_range, + uint8_t rmsi_pdcch_config, + uint8_t ssb_idx, + uint16_t nb_slots_per_frame, + uint16_t N_RB); #endif diff --git a/openair1/SCHED_NR_UE/defs.h b/openair1/SCHED_NR_UE/defs.h index 126fccb1b2..5a8dc96fab 100644 --- a/openair1/SCHED_NR_UE/defs.h +++ b/openair1/SCHED_NR_UE/defs.h @@ -112,7 +112,7 @@ int phy_procedures_RN_UE_RX(unsigned char last_slot, unsigned char next_slot, re @param mode calib/normal mode @param r_type indicates the relaying operation: 0: no_relaying, 1: unicast relaying type 1, 2: unicast relaying type 2, 3: multicast relaying */ -void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,runmode_t mode,relaying_type_t r_type); +void phy_procedures_nrUE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,runmode_t mode,relaying_type_t r_type); /*! \brief Scheduling for UE RX procedures in normal subframes. @param last_slot Index of last slot (0-19) @param phy_vars_ue Pointer to UE variables on which to act @@ -123,8 +123,9 @@ void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB @param r_type indicates the relaying operation: 0: no_relaying, 1: unicast relaying type 1, 2: unicast relaying type 2, 3: multicast relaying @param phy_vars_rn pointer to RN variables */ -int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,uint8_t do_pdcch_flag,runmode_t mode,relaying_type_t r_type); -int phy_procedures_slot_parallelization_UE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,uint8_t do_pdcch_flag,runmode_t mode,relaying_type_t r_type); +int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t do_pdcch_flag,runmode_t mode); +int phy_procedures_slot_parallelization_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,uint8_t do_pdcch_flag,runmode_t mode,relaying_type_t r_type); + #ifdef UE_SLOT_PARALLELISATION void *UE_thread_slot1_dl_processing(void *arg); @@ -136,7 +137,7 @@ void *UE_thread_slot1_dl_processing(void *arg); @param abstraction_flag Indicator of PHY abstraction @param r_type indicates the relaying operation: 0: no_relaying, 1: unicast relaying type 1, 2: unicast relaying type 2, 3: multicast relaying */ -void phy_procedures_UE_S_TX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstraction_flag,relaying_type_t r_type); +//void phy_procedures_UE_S_TX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstraction_flag,relaying_type_t r_type); /*! \brief Scheduling for UE RX procedures in TDD S-subframes. @param phy_vars_ue Pointer to UE variables on which to act @@ -144,7 +145,7 @@ void phy_procedures_UE_S_TX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstractio @param abstraction_flag Indicator of PHY abstraction @param r_type indicates the relaying operation: 0: no_relaying, 1: unicast relaying type 1, 2: unicast relaying type 2, 3: multicast relaying */ -void phy_procedures_UE_S_RX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstraction_flag, relaying_type_t r_type); +//void phy_procedures_UE_S_RX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstraction_flag, relaying_type_t r_type); /*! \brief Function to compute subframe type as a function of Frame type and TDD Configuration (implements Table 4.2.2 from 36.211, p.11 from version 8.6) and subframe index. @@ -154,7 +155,7 @@ void phy_procedures_UE_S_RX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstractio */ -nr_subframe_t nr_subframe_select(NR_DL_FRAME_PARMS *frame_parms,uint8_t subframe); +//nr_subframe_t nr_subframe_select(NR_DL_FRAME_PARMS *frame_parms,uint8_t subframe); /*! \brief Function to compute subframe type as a function of Frame type and TDD Configuration (implements Table 4.2.2 from 36.211, p.11 from version 8.6) and subframe index. Same as nr_subframe_select, except that it uses the Mod_id and is provided as a service to the MAC scheduler. @@ -368,6 +369,11 @@ int nr_is_srs_occasion_common(NR_DL_FRAME_PARMS *frame_parms,int frame_tx,int su void nr_compute_srs_pos(lte_frame_type_t frameType,uint16_t isrs,uint16_t *psrsPeriodicity,uint16_t *psrsOffset); +void set_tx_harq_id(NR_UE_ULSCH_t *ulsch, int harq_pid, int slot_tx); +int get_tx_harq_id(NR_UE_ULSCH_t *ulsch, int slot_tx); + + + /*@}*/ diff --git a/openair1/SCHED_NR_UE/fapi_nr_ue_l1.c b/openair1/SCHED_NR_UE/fapi_nr_ue_l1.c index 12a1cbaad1..1daaf90cd6 100644 --- a/openair1/SCHED_NR_UE/fapi_nr_ue_l1.c +++ b/openair1/SCHED_NR_UE/fapi_nr_ue_l1.c @@ -42,211 +42,216 @@ extern PHY_VARS_NR_UE ***PHY_vars_UE_g; int8_t nr_ue_scheduled_response(nr_scheduled_response_t *scheduled_response){ - /// module id - module_id_t module_id = scheduled_response->module_id; - /// component carrier id - uint8_t cc_id = scheduled_response->CC_id; - uint32_t i; - if(scheduled_response != NULL){ - NR_UE_PDCCH *pdcch_vars2 = PHY_vars_UE_g[module_id][cc_id]->pdcch_vars[0][0]; - NR_UE_DLSCH_t *dlsch0 = PHY_vars_UE_g[module_id][cc_id]->dlsch[0][0]; - NR_UE_ULSCH_t *ulsch0 = PHY_vars_UE_g[module_id][cc_id]->ulsch[0]; - NR_DL_FRAME_PARMS frame_parms = PHY_vars_UE_g[module_id][cc_id]->frame_parms; - PRACH_RESOURCES_t *prach_resources = PHY_vars_UE_g[module_id][cc_id]->prach_resources[0]; + /// module id + module_id_t module_id = scheduled_response->module_id; + /// component carrier id + uint8_t cc_id = scheduled_response->CC_id; + uint32_t i; + + if(scheduled_response != NULL){ + // Note: we have to handle the thread IDs for this. To be revisited completely. + NR_UE_PDCCH *pdcch_vars2 = PHY_vars_UE_g[module_id][cc_id]->pdcch_vars[0][0]; + NR_UE_DLSCH_t *dlsch0 = PHY_vars_UE_g[module_id][cc_id]->dlsch[0][0][0]; + NR_UE_ULSCH_t *ulsch0 = PHY_vars_UE_g[module_id][cc_id]->ulsch[0]; + NR_DL_FRAME_PARMS frame_parms = PHY_vars_UE_g[module_id][cc_id]->frame_parms; + PRACH_RESOURCES_t *prach_resources = PHY_vars_UE_g[module_id][cc_id]->prach_resources[0]; -// PUCCH_ConfigCommon_nr_t *pucch_config_common = PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0]; -// PUCCH_Config_t *pucch_config_dedicated = PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0]; + // PUCCH_ConfigCommon_nr_t *pucch_config_common = PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0]; + // PUCCH_Config_t *pucch_config_dedicated = PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0]; - if(scheduled_response->dl_config != NULL){ - fapi_nr_dl_config_request_t *dl_config = scheduled_response->dl_config; + if(scheduled_response->dl_config != NULL){ + fapi_nr_dl_config_request_t *dl_config = scheduled_response->dl_config; - for(i=0; inumber_pdus; ++i){ - if(dl_config->dl_config_list[i].pdu_type == FAPI_NR_DL_CONFIG_TYPE_DCI){ - pdcch_vars2->nb_search_space = pdcch_vars2->nb_search_space + 1; - fapi_nr_dl_config_dci_dl_pdu_rel15_t *dci_config = &dl_config->dl_config_list[i].dci_config_pdu.dci_config_rel15; + for(i=0; inumber_pdus; ++i){ + if(dl_config->dl_config_list[i].pdu_type == FAPI_NR_DL_CONFIG_TYPE_DCI){ + pdcch_vars2->nb_search_space = pdcch_vars2->nb_search_space + 1; + fapi_nr_dl_config_dci_dl_pdu_rel15_t *dci_config = &dl_config->dl_config_list[i].dci_config_pdu.dci_config_rel15; + + pdcch_vars2->n_RB_BWP[i] = dci_config->N_RB_BWP; + pdcch_vars2->searchSpace[i].monitoringSymbolWithinSlot = dci_config->monitoring_symbols_within_slot; - pdcch_vars2->searchSpace[i].monitoringSymbolWithinSlot = dci_config->monitoring_symbols_within_slot; + pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel1 = dci_config->number_of_candidates[0]; + pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel2 = dci_config->number_of_candidates[1]; + pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel4 = dci_config->number_of_candidates[2]; + pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel8 = dci_config->number_of_candidates[3]; + pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel16 = dci_config->number_of_candidates[4]; + + pdcch_vars2->coreset[i].duration = dci_config->coreset.duration; - pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel1 = dci_config->number_of_candidates[0]; - pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel2 = dci_config->number_of_candidates[1]; - pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel4 = dci_config->number_of_candidates[2]; - pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel8 = dci_config->number_of_candidates[3]; - pdcch_vars2->searchSpace[i].nrofCandidates_aggrlevel16 = dci_config->number_of_candidates[4]; + pdcch_vars2->coreset[i].frequencyDomainResources = dci_config->coreset.frequency_domain_resource; + pdcch_vars2->coreset[i].rb_offset = dci_config->coreset.rb_offset; - pdcch_vars2->coreset[i].duration = dci_config->coreset.duration; + if(dci_config->coreset.cce_reg_mapping_type == CCE_REG_MAPPING_TYPE_INTERLEAVED){ + pdcch_vars2->coreset[i].cce_reg_mappingType.shiftIndex = dci_config->coreset.cce_reg_interleaved_shift_index; + pdcch_vars2->coreset[i].cce_reg_mappingType.reg_bundlesize = dci_config->coreset.cce_reg_interleaved_reg_bundle_size; + pdcch_vars2->coreset[i].cce_reg_mappingType.interleaversize = dci_config->coreset.cce_reg_interleaved_interleaver_size; + }else{ //CCE_REG_MAPPING_TYPE_NON_INTERLEAVED + pdcch_vars2->coreset[i].cce_reg_mappingType.shiftIndex = 0; + pdcch_vars2->coreset[i].cce_reg_mappingType.reg_bundlesize = 0; + pdcch_vars2->coreset[i].cce_reg_mappingType.interleaversize = 0; + } - pdcch_vars2->coreset[i].frequencyDomainResources = dci_config->coreset.frequency_domain_resource; - pdcch_vars2->coreset[i].rb_offset = dci_config->coreset.rb_offset; + pdcch_vars2->coreset[i].precoderGranularity = dci_config->coreset.precoder_granularity; + //pdcch_vars2->coreset[i].tciStatesPDCCH; + //pdcch_vars2->coreset[i].tciPresentInDCI; + pdcch_vars2->coreset[i].pdcchDMRSScramblingID = dci_config->coreset.pdcch_dmrs_scrambling_id; - if(dci_config->coreset.cce_reg_mapping_type == CCE_REG_MAPPING_TYPE_INTERLEAVED){ - pdcch_vars2->coreset[i].cce_reg_mappingType.shiftIndex = dci_config->coreset.cce_reg_interleaved_shift_index; - pdcch_vars2->coreset[i].cce_reg_mappingType.reg_bundlesize = dci_config->coreset.cce_reg_interleaved_reg_bundle_size; - pdcch_vars2->coreset[i].cce_reg_mappingType.interleaversize = dci_config->coreset.cce_reg_interleaved_interleaver_size; - }else{ //CCE_REG_MAPPING_TYPE_NON_INTERLEAVED - pdcch_vars2->coreset[i].cce_reg_mappingType.shiftIndex = 0; - pdcch_vars2->coreset[i].cce_reg_mappingType.reg_bundlesize = 0; - pdcch_vars2->coreset[i].cce_reg_mappingType.interleaversize = 0; - } + }else{ //FAPI_NR_DL_CONFIG_TYPE_DLSCH + // dlsch config pdu + + fapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_config_pdu = &dl_config->dl_config_list[i].dlsch_config_pdu.dlsch_config_rel15; + uint8_t current_harq_pid = dlsch_config_pdu->harq_process_nbr; + dlsch0->current_harq_pid = current_harq_pid; + dlsch0->active = 1; - pdcch_vars2->coreset[i].precoderGranularity = dci_config->coreset.precoder_granularity; - //pdcch_vars2->coreset[i].tciStatesPDCCH; - //pdcch_vars2->coreset[i].tciPresentInDCI; - pdcch_vars2->coreset[i].pdcchDMRSScramblingID = dci_config->coreset.pdcch_dmrs_scrambling_id; - - }else{ //FAPI_NR_DL_CONFIG_TYPE_DLSCH - // dlsch config pdu - - fapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_config_pdu = &dl_config->dl_config_list[i].dlsch_config_pdu.dlsch_config_rel15; - uint8_t current_harq_pid = dlsch_config_pdu->harq_process_nbr; - dlsch0->current_harq_pid = current_harq_pid; - dlsch0->active = 1; + //dlsch0->harq_processes[0]->mcs = &dlsch_config_pdu->mcs; - //dlsch0->harq_processes[0]->mcs = &dlsch_config_pdu->mcs; + NR_DL_UE_HARQ_t dlsch0_harq = *(dlsch0->harq_processes[current_harq_pid]); + - NR_DL_UE_HARQ_t dlsch0_harq = dlsch0->harq_processes[current_harq_pid]; + //dlsch0->harq_processes[current_harq_pid]->nb_rb = dlsch_config_pdu->number_rbs; - //dlsch0->harq_processes[current_harq_pid]->nb_rb = dlsch_config_pdu->number_rbs; - - dlsch0_harq.nb_rb = dlsch_config_pdu->number_rbs; - dlsch0_harq.start_rb = dlsch_config_pdu->start_rb; - dlsch0_harq.nb_symbols = dlsch_config_pdu->number_symbols; - dlsch0_harq.start_symbol = dlsch_config_pdu->start_symbol; - dlsch0_harq.mcs = dlsch_config_pdu->mcs; - dlsch0_harq.DCINdi = dlsch_config_pdu->ndi; - dlsch0_harq.rvidx = dlsch_config_pdu->rv; - dlsch0->g_pucch = dlsch_config_pdu->accumulated_delta_PUCCH; - dlsch0_harq.harq_ack.pucch_resource_indicator = dlsch_config_pdu->pucch_resource_id; - dlsch0_harq.harq_ack.slot_for_feedback_ack = dlsch_config_pdu->pdsch_to_harq_feedback_time_ind; - printf(">>>> \tdlsch0->g_pucch=%d\tdlsch0_harq.mcs=%d\n",dlsch0->g_pucch,dlsch0_harq.mcs); - //for (int j = 0 ; j<1000; j++) printf("\nk = %d",j); + dlsch0_harq.nb_rb = dlsch_config_pdu->number_rbs; + dlsch0_harq.start_rb = dlsch_config_pdu->start_rb; + dlsch0_harq.nb_symbols = dlsch_config_pdu->number_symbols; + dlsch0_harq.start_symbol = dlsch_config_pdu->start_symbol; + dlsch0_harq.mcs = dlsch_config_pdu->mcs; + dlsch0_harq.DCINdi = dlsch_config_pdu->ndi; + dlsch0_harq.rvidx = dlsch_config_pdu->rv; + dlsch0->g_pucch = dlsch_config_pdu->accumulated_delta_PUCCH; + dlsch0_harq.harq_ack.pucch_resource_indicator = dlsch_config_pdu->pucch_resource_id; + dlsch0_harq.harq_ack.slot_for_feedback_ack = dlsch_config_pdu->pdsch_to_harq_feedback_time_ind; + printf(">>>> \tdlsch0->g_pucch=%d\tdlsch0_harq.mcs=%d\n",dlsch0->g_pucch,dlsch0_harq.mcs); + //for (int j = 0 ; j<1000; j++) printf("\nk = %d",j); - #if 0 - dlsch0->harq_processes[current_harq_pid]->mcs = dlsch_config_pdu->mcs; - dlsch0->g_pucch = dlsch_config_pdu->accumulated_delta_PUCCH; - //pdlsch0->rnti = rnti; - #endif - } - } - }else{ - pdcch_vars2->nb_search_space = 0; - } - - if(scheduled_response->ul_config != NULL){ - fapi_nr_ul_config_request_t *ul_config = scheduled_response->ul_config; - for(i=0; inumber_pdus; ++i){ - if(ul_config->ul_config_list[i].pdu_type == FAPI_NR_UL_CONFIG_TYPE_PUSCH){ - // pusch config pdu - fapi_nr_ul_config_pusch_pdu_rel15_t *pusch_config_pdu = &ul_config->ul_config_list[i].ulsch_config_pdu.ulsch_pdu_rel15; - uint8_t current_harq_pid = pusch_config_pdu->harq_process_nbr; - ulsch0->harq_processes[current_harq_pid]->nb_rb = pusch_config_pdu->number_rbs; - ulsch0->harq_processes[current_harq_pid]->first_rb = pusch_config_pdu->start_rb; - ulsch0->harq_processes[current_harq_pid]->nb_symbols = pusch_config_pdu->number_symbols; - ulsch0->harq_processes[current_harq_pid]->start_symbol = pusch_config_pdu->start_symbol; - ulsch0->harq_processes[current_harq_pid]->mcs = pusch_config_pdu->mcs; - ulsch0->harq_processes[current_harq_pid]->DCINdi = pusch_config_pdu->ndi; - ulsch0->harq_processes[current_harq_pid]->rvidx = pusch_config_pdu->rv; - ulsch0->f_pusch = pusch_config_pdu->absolute_delta_PUSCH; - } - if(ul_config->ul_config_list[i].pdu_type == FAPI_NR_UL_CONFIG_TYPE_PUCCH){ - // pucch config pdu - fapi_nr_ul_config_pucch_pdu *pucch_config_pdu = &ul_config->ul_config_list[i].pucch_config_pdu; - uint8_t pucch_resource_id = 0; //FIXME!!! - uint8_t format = 1; // FIXME!!! - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.initialCyclicShift = pucch_config_pdu->initialCyclicShift; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.nrofSymbols = pucch_config_pdu->nrofSymbols; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.startingSymbolIndex = pucch_config_pdu->startingSymbolIndex; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.nrofPRBs = pucch_config_pdu->nrofPRBs; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->startingPRB = pucch_config_pdu->startingPRB; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.timeDomainOCC = pucch_config_pdu->timeDomainOCC; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.occ_length = pucch_config_pdu->occ_length; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.occ_Index = pucch_config_pdu->occ_Index; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->intraSlotFrequencyHopping = pucch_config_pdu->intraSlotFrequencyHopping; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->secondHopPRB = pucch_config_pdu->secondHopPRB; // Not sure this parameter is used - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].formatConfig[format-1]->additionalDMRS = pucch_config_pdu->additionalDMRS; // At this point we need to know which format is going to be used - PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].formatConfig[format-1]->pi2PBSK = pucch_config_pdu->pi2PBSK; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0].pucch_GroupHopping = pucch_config_pdu->pucch_GroupHopping; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0].hoppingId = pucch_config_pdu->hoppingId; - PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0].p0_nominal = pucch_config_pdu->p0_nominal; -/* pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.initialCyclicShift = pucch_config_pdu->initialCyclicShift; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.nrofSymbols = pucch_config_pdu->nrofSymbols; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.startingSymbolIndex = pucch_config_pdu->startingSymbolIndex; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.nrofPRBs = pucch_config_pdu->nrofPRBs; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->startingPRB = pucch_config_pdu->startingPRB; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.timeDomainOCC = pucch_config_pdu->timeDomainOCC; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.occ_length = pucch_config_pdu->occ_length; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.occ_Index = pucch_config_pdu->occ_Index; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->intraSlotFrequencyHopping = pucch_config_pdu->intraSlotFrequencyHopping; - pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->secondHopPRB = pucch_config_pdu->secondHopPRB; // Not sure this parameter is used - pucch_config_dedicated->formatConfig[format-1]->additionalDMRS = pucch_config_pdu->additionalDMRS; // At this point we need to know which format is going to be used - pucch_config_dedicated->formatConfig[format-1]->pi2PBSK = pucch_config_pdu->pi2PBSK; - pucch_config_common->pucch_GroupHopping = pucch_config_pdu->pucch_GroupHopping; - pucch_config_common->hoppingId = pucch_config_pdu->hoppingId; - pucch_config_common->p0_nominal = pucch_config_pdu->p0_nominal;*/ - } - if(ul_config->ul_config_list[i].pdu_type == FAPI_NR_UL_CONFIG_TYPE_PRACH){ - // prach config pdu - fapi_nr_ul_config_prach_pdu *prach_config_pdu = &ul_config->ul_config_list[i].prach_config_pdu; - frame_parms.prach_config_common.rootSequenceIndex = prach_config_pdu->root_sequence_index; - frame_parms.prach_config_common.prach_ConfigInfo.prach_ConfigIndex = prach_config_pdu->prach_configuration_index; - frame_parms.prach_config_common.prach_ConfigInfo.zeroCorrelationZoneConfig = prach_config_pdu->zero_correlation_zone_config; - frame_parms.prach_config_common.prach_ConfigInfo.highSpeedFlag = prach_config_pdu->restrictedset_config; - frame_parms.prach_config_common.prach_ConfigInfo.prach_FreqOffset = prach_config_pdu->prach_freq_offset; - prach_resources->ra_PreambleIndex = prach_config_pdu->preamble_index; - } - } - }else{ - - } - - if(scheduled_response->tx_request != NULL){ - - }else{ - - } +#if 0 + dlsch0->harq_processes[current_harq_pid]->mcs = dlsch_config_pdu->mcs; + dlsch0->g_pucch = dlsch_config_pdu->accumulated_delta_PUCCH; + //pdlsch0->rnti = rnti; +#endif + } + } + }else{ + pdcch_vars2->nb_search_space = 0; } - return 0; + if(scheduled_response->ul_config != NULL){ + fapi_nr_ul_config_request_t *ul_config = scheduled_response->ul_config; + for(i=0; inumber_pdus; ++i){ + if(ul_config->ul_config_list[i].pdu_type == FAPI_NR_UL_CONFIG_TYPE_PUSCH){ + // pusch config pdu + fapi_nr_ul_config_pusch_pdu_rel15_t *pusch_config_pdu = &ul_config->ul_config_list[i].ulsch_config_pdu.ulsch_pdu_rel15; + uint8_t current_harq_pid = pusch_config_pdu->harq_process_nbr; + ulsch0->harq_processes[current_harq_pid]->nb_rb = pusch_config_pdu->number_rbs; + ulsch0->harq_processes[current_harq_pid]->first_rb = pusch_config_pdu->start_rb; + ulsch0->harq_processes[current_harq_pid]->nb_symbols = pusch_config_pdu->number_symbols; + ulsch0->harq_processes[current_harq_pid]->start_symbol = pusch_config_pdu->start_symbol; + ulsch0->harq_processes[current_harq_pid]->mcs = pusch_config_pdu->mcs; + ulsch0->harq_processes[current_harq_pid]->DCINdi = pusch_config_pdu->ndi; + ulsch0->harq_processes[current_harq_pid]->rvidx = pusch_config_pdu->rv; + ulsch0->f_pusch = pusch_config_pdu->absolute_delta_PUSCH; + } + if(ul_config->ul_config_list[i].pdu_type == FAPI_NR_UL_CONFIG_TYPE_PUCCH){ + // pucch config pdu + fapi_nr_ul_config_pucch_pdu *pucch_config_pdu = &ul_config->ul_config_list[i].pucch_config_pdu; + uint8_t pucch_resource_id = 0; //FIXME!!! + uint8_t format = 1; // FIXME!!! + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.initialCyclicShift = pucch_config_pdu->initialCyclicShift; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.nrofSymbols = pucch_config_pdu->nrofSymbols; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.startingSymbolIndex = pucch_config_pdu->startingSymbolIndex; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.nrofPRBs = pucch_config_pdu->nrofPRBs; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->startingPRB = pucch_config_pdu->startingPRB; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.timeDomainOCC = pucch_config_pdu->timeDomainOCC; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.occ_length = pucch_config_pdu->occ_length; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->format_parameters.occ_Index = pucch_config_pdu->occ_Index; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->intraSlotFrequencyHopping = pucch_config_pdu->intraSlotFrequencyHopping; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].PUCCH_Resource[pucch_resource_id]->secondHopPRB = pucch_config_pdu->secondHopPRB; // Not sure this parameter is used + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].formatConfig[format-1]->additionalDMRS = pucch_config_pdu->additionalDMRS; // At this point we need to know which format is going to be used + PHY_vars_UE_g[module_id][cc_id]->pucch_config_dedicated_nr[0].formatConfig[format-1]->pi2PBSK = pucch_config_pdu->pi2PBSK; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0].pucch_GroupHopping = pucch_config_pdu->pucch_GroupHopping; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0].hoppingId = pucch_config_pdu->hoppingId; + PHY_vars_UE_g[module_id][cc_id]->pucch_config_common_nr[0].p0_nominal = pucch_config_pdu->p0_nominal; + /* pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.initialCyclicShift = pucch_config_pdu->initialCyclicShift; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.nrofSymbols = pucch_config_pdu->nrofSymbols; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.startingSymbolIndex = pucch_config_pdu->startingSymbolIndex; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.nrofPRBs = pucch_config_pdu->nrofPRBs; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->startingPRB = pucch_config_pdu->startingPRB; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.timeDomainOCC = pucch_config_pdu->timeDomainOCC; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.occ_length = pucch_config_pdu->occ_length; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->format_parameters.occ_Index = pucch_config_pdu->occ_Index; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->intraSlotFrequencyHopping = pucch_config_pdu->intraSlotFrequencyHopping; + pucch_config_dedicated->PUCCH_Resource[pucch_resource_id]->secondHopPRB = pucch_config_pdu->secondHopPRB; // Not sure this parameter is used + pucch_config_dedicated->formatConfig[format-1]->additionalDMRS = pucch_config_pdu->additionalDMRS; // At this point we need to know which format is going to be used + pucch_config_dedicated->formatConfig[format-1]->pi2PBSK = pucch_config_pdu->pi2PBSK; + pucch_config_common->pucch_GroupHopping = pucch_config_pdu->pucch_GroupHopping; + pucch_config_common->hoppingId = pucch_config_pdu->hoppingId; + pucch_config_common->p0_nominal = pucch_config_pdu->p0_nominal;*/ + } + if(ul_config->ul_config_list[i].pdu_type == FAPI_NR_UL_CONFIG_TYPE_PRACH){ + // prach config pdu + fapi_nr_ul_config_prach_pdu *prach_config_pdu = &ul_config->ul_config_list[i].prach_config_pdu; + frame_parms.prach_config_common.rootSequenceIndex = prach_config_pdu->root_sequence_index; + frame_parms.prach_config_common.prach_ConfigInfo.prach_ConfigIndex = prach_config_pdu->prach_configuration_index; + frame_parms.prach_config_common.prach_ConfigInfo.zeroCorrelationZoneConfig = prach_config_pdu->zero_correlation_zone_config; + frame_parms.prach_config_common.prach_ConfigInfo.highSpeedFlag = prach_config_pdu->restrictedset_config; + frame_parms.prach_config_common.prach_ConfigInfo.prach_FreqOffset = prach_config_pdu->prach_freq_offset; + prach_resources->ra_PreambleIndex = prach_config_pdu->preamble_index; + } + } + }else{ + + } + + if(scheduled_response->tx_request != NULL){ + + }else{ + + } + } + + return 0; } int8_t nr_ue_phy_config_request(nr_phy_config_t *phy_config){ - if(phy_config != NULL){ - if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_PBCH){ - printf("[L1][IF module][PHY CONFIG]\n"); - printf("subcarrier spacing: %d\n", phy_config->config_req.pbch_config.subcarrier_spacing_common); - printf("ssb carrier offset: %d\n", phy_config->config_req.pbch_config.ssb_subcarrier_offset); - printf("dmrs type A position: %d\n", phy_config->config_req.pbch_config.dmrs_type_a_position); - printf("pdcch config sib1: %d\n", phy_config->config_req.pbch_config.pdcch_config_sib1); - printf("cell barred: %d\n", phy_config->config_req.pbch_config.cell_barred); - printf("intra frequency reselection: %d\n", phy_config->config_req.pbch_config.intra_frequency_reselection); - printf("system frame number: %d\n", phy_config->config_req.pbch_config.system_frame_number); - printf("ssb index: %d\n", phy_config->config_req.pbch_config.ssb_index); - printf("half frame bit: %d\n", phy_config->config_req.pbch_config.half_frame_bit); - printf("-------------------------------\n"); + if(phy_config != NULL){ + if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_PBCH){ + printf("[L1][IF module][PHY CONFIG]\n"); + printf("subcarrier spacing: %d\n", phy_config->config_req.pbch_config.subcarrier_spacing_common); + printf("ssb carrier offset: %d\n", phy_config->config_req.pbch_config.ssb_subcarrier_offset); + printf("dmrs type A position: %d\n", phy_config->config_req.pbch_config.dmrs_type_a_position); + printf("pdcch config sib1: %d\n", phy_config->config_req.pbch_config.pdcch_config_sib1); + printf("cell barred: %d\n", phy_config->config_req.pbch_config.cell_barred); + printf("intra frequency reselection: %d\n", phy_config->config_req.pbch_config.intra_frequency_reselection); + printf("system frame number: %d\n", phy_config->config_req.pbch_config.system_frame_number); + printf("ssb index: %d\n", phy_config->config_req.pbch_config.ssb_index); + printf("half frame bit: %d\n", phy_config->config_req.pbch_config.half_frame_bit); + printf("-------------------------------\n"); - PHY_vars_UE_g[0][0]->proc.proc_rxtx[0].frame_rx = phy_config->config_req.pbch_config.system_frame_number; - } - - if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_DL_BWP_COMMON){ - - } - - if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_UL_BWP_COMMON){ - - } - - if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_DL_BWP_DEDICATED){ - - } - - if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_UL_BWP_DEDICATED){ - - } + PHY_vars_UE_g[0][0]->proc.proc_rxtx[0].frame_rx = phy_config->config_req.pbch_config.system_frame_number; } + + if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_DL_BWP_COMMON){ + + } + + if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_UL_BWP_COMMON){ + + } + + if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_DL_BWP_DEDICATED){ + + } + + if(phy_config->config_req.config_mask & FAPI_NR_CONFIG_REQUEST_MASK_UL_BWP_DEDICATED){ + + } + } - return 0; + return 0; } + diff --git a/openair1/SCHED_NR_UE/harq_nr.c b/openair1/SCHED_NR_UE/harq_nr.c index 93a3f998cf..23d828ec93 100644 --- a/openair1/SCHED_NR_UE/harq_nr.c +++ b/openair1/SCHED_NR_UE/harq_nr.c @@ -120,7 +120,7 @@ void get_dci_info_for_harq(PHY_VARS_NR_UE *ue, NR_DCI_INFO_EXTRACTED_t *nr_dci_i dlsch[0]->current_harq_pid = nr_dci_info_extracted->harq_process_number; - NR_DL_UE_HARQ_t *dl_harq = &dlsch[0]->harq_processes[dlsch[0]->current_harq_pid]; + NR_DL_UE_HARQ_t *dl_harq = dlsch[0]->harq_processes[dlsch[0]->current_harq_pid]; dl_harq->harq_ack.vDAI_DL = nr_dci_info_extracted->dai+1; dl_harq->harq_ack.pucch_resource_indicator = nr_dci_info_extracted->pucch_resource_ind; @@ -401,9 +401,9 @@ void config_downlink_harq_process(PHY_VARS_NR_UE *ue, int gNB_id, int TB_id, int memset(&dlsch->harq_processes[harq_pid],0,sizeof(NR_DL_UE_HARQ_t)); - NR_DL_UE_HARQ_t dl_harq = dlsch->harq_processes[harq_pid]; + NR_DL_UE_HARQ_t *dl_harq = dlsch->harq_processes[harq_pid]; - init_downlink_harq_status(&dl_harq); + init_downlink_harq_status(dl_harq); } } diff --git a/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c b/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c index 44cae9679c..5eee7b9c0c 100644 --- a/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c +++ b/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c @@ -40,18 +40,20 @@ #include "PHY/phy_extern_nr_ue.h" #include "PHY/NR_REFSIG/refsig_defs_ue.h" #include "PHY/MODULATION/modulation_UE.h" -//#include "PHY/NR_UE_TRANSPORT/nr_transport_ue.h" +#include "PHY/NR_UE_TRANSPORT/nr_transport_ue.h" +#include "PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h" //#include "PHY/extern.h" #include "SCHED_NR_UE/defs.h" #include "SCHED_NR/extern.h" //#include //#include "targets/RT/USER/nr-softmodem.h" +#include "PHY/NR_UE_ESTIMATION/nr_estimation.h" #ifdef EMOS #include "SCHED/phy_procedures_emos.h" #endif -#define DEBUG_PHY_PROC +//#define DEBUG_PHY_PROC #define NR_PDCCH_SCHED #define NR_PDCCH_SCHED_DEBUG @@ -91,6 +93,7 @@ char mode_string[4][20] = {"NOT SYNCHED","PRACH","RAR","PUSCH"}; extern double cpuf; + int32_t nr_rx_pdcch(PHY_VARS_NR_UE *ue, uint32_t frame, uint8_t nr_tti_rx, @@ -117,25 +120,26 @@ uint8_t nr_dci_decoding_procedure(int s, uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES]); int nr_generate_ue_ul_dlsch_params_from_dci(PHY_VARS_NR_UE *ue, - uint8_t eNB_id, - int frame, - uint8_t nr_tti_rx, - uint32_t dci_pdu[4], - uint16_t rnti, - uint8_t dci_length, - NR_DCI_format_t dci_format, - NR_UE_PDCCH *pdcch_vars, - NR_UE_PDSCH *pdsch_vars, - NR_UE_DLSCH_t **dlsch, - NR_UE_ULSCH_t *ulsch, - NR_DL_FRAME_PARMS *frame_parms, - PDSCH_CONFIG_DEDICATED *pdsch_config_dedicated, - uint8_t beamforming_mode, - uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS][NBR_NR_FORMATS], - uint16_t n_RB_ULBWP, - uint16_t n_RB_DLBWP, - uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES], - NR_DCI_INFO_EXTRACTED_t *nr_dci_info_extracted); + uint8_t eNB_id, + int frame, + uint8_t nr_tti_rx, + uint32_t dci_pdu[4], + uint16_t rnti, + uint8_t dci_length, + NR_DCI_format_t dci_format, + NR_UE_PDCCH *pdcch_vars, + NR_UE_PDSCH *pdsch_vars, + NR_UE_DLSCH_t **dlsch, + NR_UE_ULSCH_t *ulsch, + NR_DL_FRAME_PARMS *frame_parms, + PDSCH_CONFIG_DEDICATED *pdsch_config_dedicated, + uint8_t beamforming_mode, + uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS][NBR_NR_FORMATS], + uint16_t n_RB_ULBWP, + uint16_t n_RB_DLBWP, + uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES], + NR_DCI_INFO_EXTRACTED_t *nr_dci_info_extracted); + #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) extern uint32_t downlink_frequency[MAX_NUM_CCs][4]; @@ -155,8 +159,8 @@ void nr_dump_dlsch(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uin ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_processes[harq_pid]->Nl, ue->pdcch_vars[0%RX_NB_TH][eNB_id]->num_pdcch_symbols, proc->frame_rx, - nr_tti_rx, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id]); + nr_tti_rx, + ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id]); write_output("rxsigF0.m","rxsF0", ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx]].rxdataF[0],2*nsymb*ue->frame_parms.ofdm_symbol_size,2,1); write_output("rxsigF0_ext.m","rxsF0_ext", ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][0]->rxdataF_ext[0],2*nsymb*ue->frame_parms.ofdm_symbol_size,1,1); @@ -186,12 +190,12 @@ void nr_dump_dlsch_SI(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id, 1, ue->pdcch_vars[0%RX_NB_TH][eNB_id]->num_pdcch_symbols, proc->frame_rx, - nr_tti_rx, - 0); + nr_tti_rx, + 0); LOG_D(PHY,"[UE %d] Dumping dlsch_SI : ofdm_symbol_size %d, nsymb %d, nb_rb %d, mcs %d, nb_rb %d, num_pdcch_symbols %d,G %d\n", ue->Mod_id, - ue->frame_parms.ofdm_symbol_size, - nsymb, + ue->frame_parms.ofdm_symbol_size, + nsymb, ue->dlsch_SI[eNB_id]->harq_processes[0]->nb_rb, ue->dlsch_SI[eNB_id]->harq_processes[0]->mcs, ue->dlsch_SI[eNB_id]->harq_processes[0]->nb_rb, @@ -291,8 +295,8 @@ void nr_dump_dlsch_ra(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id, 1, ue->pdcch_vars[0%RX_NB_TH][eNB_id]->num_pdcch_symbols, proc->frame_rx, - nr_tti_rx, - 0); + nr_tti_rx, + 0); LOG_D(PHY,"[UE %d] Dumping dlsch_ra : nb_rb %d, mcs %d, nb_rb %d, num_pdcch_symbols %d,G %d\n", ue->Mod_id, ue->dlsch_ra[eNB_id]->harq_processes[0]->nb_rb, @@ -328,37 +332,37 @@ void phy_reset_ue(uint8_t Mod_id,uint8_t CC_id,uint8_t eNB_index) //[NUMBER_OF_RX_THREAD=2][NUMBER_OF_CONNECTED_eNB_MAX][2]; for(int l=0; ldlsch[l][i][j]) { - for(k=0; kdlsch[l][i][j]->harq_processes[k]; k++) { - ue->dlsch[l][i][j]->harq_processes[k]->status = SCH_IDLE; - for (s=0; s<10; s++) { - // reset ACK/NACK bit to DTX for all nr_tti_rxs s = 0..9 - ue->dlsch[l][i][j]->harq_ack[s].ack = 2; - ue->dlsch[l][i][j]->harq_ack[s].send_harq_status = 0; - ue->dlsch[l][i][j]->harq_ack[s].vDAI_UL = 0xff; - ue->dlsch[l][i][j]->harq_ack[s].vDAI_DL = 0xff; - } - } - } - } - - //UL HARQ - if(ue->ulsch[i]) { - for(k=0; kulsch[i]->harq_processes[k]; k++) { - ue->ulsch[i]->harq_processes[k]->status = SCH_IDLE; - //Set NDIs for all UL HARQs to 0 - // ue->ulsch[i]->harq_processes[k]->Ndi = 0; - - } - } - - // flush Msg3 buffer - ue->ulsch_Msg3_active[i] = 0; - + for(i=0; idlsch[l][i][j]) { + for(k=0; kdlsch[l][i][j]->harq_processes[k]; k++) { + ue->dlsch[l][i][j]->harq_processes[k]->status = SCH_IDLE; + for (s=0; s<10; s++) { + // reset ACK/NACK bit to DTX for all nr_tti_rxs s = 0..9 + ue->dlsch[l][i][j]->harq_ack[s].ack = 2; + ue->dlsch[l][i][j]->harq_ack[s].send_harq_status = 0; + ue->dlsch[l][i][j]->harq_ack[s].vDAI_UL = 0xff; + ue->dlsch[l][i][j]->harq_ack[s].vDAI_DL = 0xff; + } + } + } } + + //UL HARQ + if(ue->ulsch[i]) { + for(k=0; kulsch[i]->harq_processes[k]; k++) { + ue->ulsch[i]->harq_processes[k]->status = SCH_IDLE; + //Set NDIs for all UL HARQs to 0 + // ue->ulsch[i]->harq_processes[k]->Ndi = 0; + + } + } + + // flush Msg3 buffer + ue->ulsch_Msg3_active[i] = 0; + + } } } @@ -512,14 +516,14 @@ void compute_cqi_ri_resources(PHY_VARS_NR_UE *ue, uint8_t cqi_status, uint8_t ri_status) { - //PHY_MEASUREMENTS *meas = &ue->measurements; - //uint8_t transmission_mode = ue->transmission_mode[eNB_id]; + //PHY_MEASUREMENTS *meas = &ue->measurements; + //uint8_t transmission_mode = ue->transmission_mode[eNB_id]; - //LOG_I(PHY,"compute_cqi_ri_resources O_RI %d O %d uci format %d \n",ulsch->O_RI,ulsch->O,ulsch->uci_format); - if (cqi_status == 1 || ri_status == 1) + //LOG_I(PHY,"compute_cqi_ri_resources O_RI %d O %d uci format %d \n",ulsch->O_RI,ulsch->O,ulsch->uci_format); + if (cqi_status == 1 || ri_status == 1) { - ulsch->O = 4; + ulsch->O = 4; } } @@ -541,77 +545,77 @@ void ue_compute_srs_occasion(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t if (isSubframeSRS) { LOG_D(PHY," SrsDedicatedSetup: %d \n",pSoundingrs_ul_config_dedicated->srsConfigDedicatedSetup); - if(pSoundingrs_ul_config_dedicated->srsConfigDedicatedSetup) + if(pSoundingrs_ul_config_dedicated->srsConfigDedicatedSetup) { - nr_compute_srs_pos(frame_parms->frame_type, pSoundingrs_ul_config_dedicated->srs_ConfigIndex, &srsPeriodicity, &srsOffset); + nr_compute_srs_pos(frame_parms->frame_type, pSoundingrs_ul_config_dedicated->srs_ConfigIndex, &srsPeriodicity, &srsOffset); - LOG_D(PHY," srsPeriodicity: %d srsOffset: %d isSubframeSRS %d \n",srsPeriodicity,srsOffset,isSubframeSRS); + LOG_D(PHY," srsPeriodicity: %d srsOffset: %d isSubframeSRS %d \n",srsPeriodicity,srsOffset,isSubframeSRS); - // transmit SRS if the four following constraints are respected: - // - UE is configured to transmit SRS - // - SRS are configured in current nr_tti_rx - // - UE is configured to send SRS in this nr_tti_tx + // transmit SRS if the four following constraints are respected: + // - UE is configured to transmit SRS + // - SRS are configured in current nr_tti_rx + // - UE is configured to send SRS in this nr_tti_tx - // 36.213 8.2 - // 1- A UE shall not transmit SRS whenever SRS and PUCCH format 2/2a/2b transmissions happen to coincide in the same nr_tti_rx - // 2- A UE shall not transmit SRS whenever SRS transmit - // on and PUCCH transmission carrying ACK/NACK and/or - // positive SR happen to coincide in the same nr_tti_rx if the parameter - // Simultaneous-AN-and-SRS is FALSE + // 36.213 8.2 + // 1- A UE shall not transmit SRS whenever SRS and PUCCH format 2/2a/2b transmissions happen to coincide in the same nr_tti_rx + // 2- A UE shall not transmit SRS whenever SRS transmit + // on and PUCCH transmission carrying ACK/NACK and/or + // positive SR happen to coincide in the same nr_tti_rx if the parameter + // Simultaneous-AN-and-SRS is FALSE - // check PUCCH format 2/2a/2b transmissions - is_pucch2_subframe = is_cqi_TXOp(ue,proc,eNB_id) && (ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.cqi_PMI_ConfigIndex>0); - is_pucch2_subframe = (is_ri_TXOp(ue,proc,eNB_id) && (ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.ri_ConfigIndex>0)) || is_pucch2_subframe; + // check PUCCH format 2/2a/2b transmissions + is_pucch2_subframe = is_cqi_TXOp(ue,proc,eNB_id) && (ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.cqi_PMI_ConfigIndex>0); + is_pucch2_subframe = (is_ri_TXOp(ue,proc,eNB_id) && (ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.ri_ConfigIndex>0)) || is_pucch2_subframe; - // check ACK/SR transmission - if(frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission == FALSE) + // check ACK/SR transmission + if(frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission == FALSE) { - if(nr_is_SR_TXOp(ue,proc,eNB_id)) + if(nr_is_SR_TXOp(ue,proc,eNB_id)) { - uint32_t SR_payload = 0; - if (ue->mac_enabled==1) + uint32_t SR_payload = 0; + if (ue->mac_enabled==1) { - int Mod_id = ue->Mod_id; - int CC_id = ue->CC_id; - SR_payload = mac_xface->ue_get_SR(Mod_id, - CC_id, - frame_tx, - eNB_id, - ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->crnti, - nr_tti_tx); // nr_tti_tx used for meas gap + int Mod_id = ue->Mod_id; + int CC_id = ue->CC_id; + SR_payload = mac_xface->ue_get_SR(Mod_id, + CC_id, + frame_tx, + eNB_id, + ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->crnti, + nr_tti_tx); // nr_tti_tx used for meas gap - if (SR_payload > 0) - is_sr_an_subframe = 1; + if (SR_payload > 0) + is_sr_an_subframe = 1; } } - uint8_t pucch_ack_payload[2]; - if (nr_get_ack(&ue->frame_parms, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, - nr_tti_tx,proc->nr_tti_rx,pucch_ack_payload,0) > 0) + uint8_t pucch_ack_payload[2]; + if (nr_get_ack(&ue->frame_parms, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, + nr_tti_tx,proc->nr_tti_rx,pucch_ack_payload,0) > 0) { - is_sr_an_subframe = 1; + is_sr_an_subframe = 1; } } - // check SRS UE opportunity - if( isSubframeSRS && - (((10*frame_tx+nr_tti_tx) % srsPeriodicity) == srsOffset) - ) + // check SRS UE opportunity + if( isSubframeSRS && + (((10*frame_tx+nr_tti_tx) % srsPeriodicity) == srsOffset) + ) { - if ((is_pucch2_subframe == 0) && (is_sr_an_subframe == 0)) + if ((is_pucch2_subframe == 0) && (is_sr_an_subframe == 0)) { - pSoundingrs_ul_config_dedicated->srsUeSubframe = 1; - ue->ulsch[eNB_id]->srs_active = 1; - ue->ulsch[eNB_id]->Nsymb_pusch = 12-(frame_parms->Ncp<<1)- ue->ulsch[eNB_id]->srs_active; + pSoundingrs_ul_config_dedicated->srsUeSubframe = 1; + ue->ulsch[eNB_id]->srs_active = 1; + ue->ulsch[eNB_id]->Nsymb_pusch = 12-(frame_parms->Ncp<<1)- ue->ulsch[eNB_id]->srs_active; } - else + else { - LOG_I(PHY,"DROP UE-SRS-TX for this nr_tti_tx %d.%d: collision with PUCCH2 or SR/AN: PUCCH2-occasion: %d, SR-AN-occasion[simSRS-SR-AN %d]: %d \n", frame_tx, nr_tti_tx, is_pucch2_subframe, frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, is_sr_an_subframe); + LOG_I(PHY,"DROP UE-SRS-TX for this nr_tti_tx %d.%d: collision with PUCCH2 or SR/AN: PUCCH2-occasion: %d, SR-AN-occasion[simSRS-SR-AN %d]: %d \n", frame_tx, nr_tti_tx, is_pucch2_subframe, frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, is_sr_an_subframe); } } } - LOG_D(PHY," srsCellSubframe: %d, srsUeSubframe: %d, Nsymb-pusch: %d \n", pSoundingrs_ul_config_dedicated->srsCellSubframe, pSoundingrs_ul_config_dedicated->srsUeSubframe, ue->ulsch[eNB_id]->Nsymb_pusch); + LOG_D(PHY," srsCellSubframe: %d, srsUeSubframe: %d, Nsymb-pusch: %d \n", pSoundingrs_ul_config_dedicated->srsCellSubframe, pSoundingrs_ul_config_dedicated->srsUeSubframe, ue->ulsch[eNB_id]->Nsymb_pusch); } } @@ -648,42 +652,42 @@ void nr_get_cqipmiri_params(PHY_VARS_NR_UE *ue,uint8_t eNB_id) else if (cqi_PMI_ConfigIndex > 317) { if (cqi_PMI_ConfigIndex <= 349) { // 32 ms CQI_PMI period - cqirep->Npd = 32; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-318; + cqirep->Npd = 32; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-318; } else if (cqi_PMI_ConfigIndex <= 413) { // 64 ms CQI_PMI period - cqirep->Npd = 64; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-350; + cqirep->Npd = 64; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-350; } else if (cqi_PMI_ConfigIndex <= 541) { // 128 ms CQI_PMI period - cqirep->Npd = 128; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-414; + cqirep->Npd = 128; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-414; } } } else { // TDD - if (cqi_PMI_ConfigIndex == 0) { // all UL subframes - cqirep->Npd = 1; - cqirep->N_OFFSET_CQI = 0; - } else if (cqi_PMI_ConfigIndex <= 6) { // 5 ms CQI_PMI period - cqirep->Npd = 5; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-1; - } else if (cqi_PMI_ConfigIndex <=16) { // 10ms CQI_PMI period - cqirep->Npd = 10; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-6; - } else if (cqi_PMI_ConfigIndex <= 36) { // 20 ms CQI_PMI period - cqirep->Npd = 20; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-16; - } else if (cqi_PMI_ConfigIndex <= 76) { // 40 ms CQI_PMI period - cqirep->Npd = 40; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-36; - } else if (cqi_PMI_ConfigIndex <= 156) { // 80 ms CQI_PMI period - cqirep->Npd = 80; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-76; - } else if (cqi_PMI_ConfigIndex <= 316) { // 160 ms CQI_PMI period - cqirep->Npd = 160; - cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-156; - } + if (cqi_PMI_ConfigIndex == 0) { // all UL subframes + cqirep->Npd = 1; + cqirep->N_OFFSET_CQI = 0; + } else if (cqi_PMI_ConfigIndex <= 6) { // 5 ms CQI_PMI period + cqirep->Npd = 5; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-1; + } else if (cqi_PMI_ConfigIndex <=16) { // 10ms CQI_PMI period + cqirep->Npd = 10; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-6; + } else if (cqi_PMI_ConfigIndex <= 36) { // 20 ms CQI_PMI period + cqirep->Npd = 20; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-16; + } else if (cqi_PMI_ConfigIndex <= 76) { // 40 ms CQI_PMI period + cqirep->Npd = 40; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-36; + } else if (cqi_PMI_ConfigIndex <= 156) { // 80 ms CQI_PMI period + cqirep->Npd = 80; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-76; + } else if (cqi_PMI_ConfigIndex <= 316) { // 160 ms CQI_PMI period + cqirep->Npd = 160; + cqirep->N_OFFSET_CQI = cqi_PMI_ConfigIndex-156; + } } } @@ -696,72 +700,72 @@ PUCCH_FMT_t get_pucch_format(lte_frame_type_t frame_type, uint8_t bundling_flag) { if((cqi_status == 0) && (ri_status==0)) - { + { // PUCCH Format 1 1a 1b // 1- SR only ==> PUCCH format 1 // 2- 1bit Ack/Nack with/without SR ==> PUCCH format 1a // 3- 2bits Ack/Nack with/without SR ==> PUCCH format 1b if((nb_cw == 1)&&(bundling_flag==bundling)) - { + { return pucch_format1a; - } + } if((nb_cw == 1)&&(bundling_flag==multiplexing)) - { + { return pucch_format1b; - } + } if(nb_cw == 2) - { + { return pucch_format1b; - } + } if(SR_payload == 1) - { + { return pucch_format1; /* - if (frame_type == FDD) { - return pucch_format1; - } else if (frame_type == TDD) { - return pucch_format1b; - } else { - AssertFatal(1==0,"Unknown frame_type"); - }*/ - } - } + if (frame_type == FDD) { + return pucch_format1; + } else if (frame_type == TDD) { + return pucch_format1b; + } else { + AssertFatal(1==0,"Unknown frame_type"); + }*/ + } + } else - { + { // PUCCH Format 2 2a 2b // 1- CQI only or RI only ==> PUCCH format 2 // 2- CQI or RI + 1bit Ack/Nack for normal CP ==> PUCCH format 2a // 3- CQI or RI + 2bits Ack/Nack for normal CP ==> PUCCH format 2b // 4- CQI or RI + Ack/Nack for extended CP ==> PUCCH format 2 if(nb_cw == 0) - { + { return pucch_format2; - } + } if(cyclic_prefix_type == NORMAL) - { + { if(nb_cw == 1) - { + { return pucch_format2a; - } + } if(nb_cw == 2) - { + { return pucch_format2b; - } - } + } + } else - { + { return pucch_format2; - } - } + } + } return pucch_format1a; } uint16_t nr_get_n1_pucch(PHY_VARS_NR_UE *ue, - UE_nr_rxtx_proc_t *proc, - nr_harq_status_t *harq_ack, - uint8_t eNB_id, - uint8_t *b, - uint8_t SR) + UE_nr_rxtx_proc_t *proc, + nr_harq_status_t *harq_ack, + uint8_t eNB_id, + uint8_t *b, + uint8_t SR) { NR_DL_FRAME_PARMS *frame_parms=&ue->frame_parms; @@ -841,9 +845,9 @@ uint16_t nr_get_n1_pucch(PHY_VARS_NR_UE *ue, } LOG_D(PHY,"SFN/SF %d/%d calculating n1_pucch0 from last_dl=%d\n", - proc->frame_tx%1024, - proc->nr_tti_tx, - last_dl); + proc->frame_tx%1024, + proc->nr_tti_tx, + last_dl); // i=0 nCCE0 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[last_dl]; @@ -856,11 +860,11 @@ uint16_t nr_get_n1_pucch(PHY_VARS_NR_UE *ue, if (SR == 0) { // last paragraph pg 68 from 36.213 (v8.6), m=0 b[0]=(M==2) ? 1-harq_ack0 : harq_ack0; b[1]=harq_ack0; // in case we use pucch format 1b (subframes 2,7) - ue->pucch_sel[nr_tti_tx] = 0; + ue->pucch_sel[nr_tti_tx] = 0; return(n1_pucch0); } else { // SR and only 0 or 1 ACKs (first 2 entries in Table 7.3-1 of 36.213) b[0]=harq_ack0; - return(ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex); + return(ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex); } } else { if (SR == 0) { @@ -902,10 +906,10 @@ uint16_t nr_get_n1_pucch(PHY_VARS_NR_UE *ue, harq_ack0 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[5+last_dl].ack; LOG_D(PHY,"SFN/SF %d/%d calculating n1_pucch cce0=%d n1_pucch0=%d cce1=%d n1_pucch1=%d\n", - proc->frame_tx%1024, - proc->nr_tti_tx, - nCCE0,n1_pucch0, - nCCE1,n1_pucch1); + proc->frame_tx%1024, + proc->nr_tti_tx, + nCCE0,n1_pucch0, + nCCE1,n1_pucch1); if (harq_ack1!=2) { // n-6 // nr_tti_tx 6,8,0 and maybe 5,7,9 is to be ACK/NAKed @@ -963,294 +967,294 @@ uint16_t nr_get_n1_pucch(PHY_VARS_NR_UE *ue, break; case 4: // DL:S:UL:UL:DL:DL:DL:DL:DL:DL - // in this configuration we have M=4 from pg 68 of 36.213 (v8.6) - // Note: this doesn't allow using nr_tti_tx 1 for PDSCH transmission!!! (i.e. SF 1 cannot be acked in SF 2) - // set ACK/NAKs to DTX - harq_ack3 = 2; // DTX - harq_ack2 = 2; // DTX - harq_ack1 = 2; // DTX - harq_ack0 = 2; // DTX - // This is the offset for a particular nr_tti_tx (2,3,4) => (0,2,4) - //last_dl = (nr_tti_tx-2)<<1; - if (nr_tti_tx == 2) { - // i=0 - //nCCE0 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[2+nr_tti_tx]; - nCCE0 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(8+nr_tti_tx)%10]; - n1_pucch0 = 2*nr_get_Np(frame_parms->N_RB_DL,nCCE0,0) + nCCE0+ frame_parms->pucch_config_common.n1PUCCH_AN; - // i=1 - nCCE1 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[2+nr_tti_tx]; - n1_pucch1 = nr_get_Np(frame_parms->N_RB_DL,nCCE1,0) + nr_get_Np(frame_parms->N_RB_DL,nCCE1,1) + nCCE1 + frame_parms->pucch_config_common.n1PUCCH_AN; - // i=2 - nCCE2 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(8+nr_tti_tx)%10]; + // in this configuration we have M=4 from pg 68 of 36.213 (v8.6) + // Note: this doesn't allow using nr_tti_tx 1 for PDSCH transmission!!! (i.e. SF 1 cannot be acked in SF 2) + // set ACK/NAKs to DTX + harq_ack3 = 2; // DTX + harq_ack2 = 2; // DTX + harq_ack1 = 2; // DTX + harq_ack0 = 2; // DTX + // This is the offset for a particular nr_tti_tx (2,3,4) => (0,2,4) + //last_dl = (nr_tti_tx-2)<<1; + if (nr_tti_tx == 2) { + // i=0 + //nCCE0 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[2+nr_tti_tx]; + nCCE0 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(8+nr_tti_tx)%10]; + n1_pucch0 = 2*nr_get_Np(frame_parms->N_RB_DL,nCCE0,0) + nCCE0+ frame_parms->pucch_config_common.n1PUCCH_AN; + // i=1 + nCCE1 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[2+nr_tti_tx]; + n1_pucch1 = nr_get_Np(frame_parms->N_RB_DL,nCCE1,0) + nr_get_Np(frame_parms->N_RB_DL,nCCE1,1) + nCCE1 + frame_parms->pucch_config_common.n1PUCCH_AN; + // i=2 + nCCE2 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(8+nr_tti_tx)%10]; - n1_pucch2 = 2*nr_get_Np(frame_parms->N_RB_DL,nCCE2,1) + nCCE2+ frame_parms->pucch_config_common.n1PUCCH_AN; - // i=3 - //nCCE3 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(9+nr_tti_tx)%10]; - //n1_pucch3 = nr_get_Np(frame_parms->N_RB_DL,nCCE3,1) + nCCE3 + frame_parms->pucch_config_common.n1PUCCH_AN; + n1_pucch2 = 2*nr_get_Np(frame_parms->N_RB_DL,nCCE2,1) + nCCE2+ frame_parms->pucch_config_common.n1PUCCH_AN; + // i=3 + //nCCE3 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(9+nr_tti_tx)%10]; + //n1_pucch3 = nr_get_Np(frame_parms->N_RB_DL,nCCE3,1) + nCCE3 + frame_parms->pucch_config_common.n1PUCCH_AN; - // set ACK/NAK to values if not DTX - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(8+nr_tti_tx)%10].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed - harq_ack0 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(8+nr_tti_tx)%10].ack; + // set ACK/NAK to values if not DTX + if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(8+nr_tti_tx)%10].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed + harq_ack0 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(8+nr_tti_tx)%10].ack; - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[2+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed - harq_ack1 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[2+nr_tti_tx].ack; + if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[2+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed + harq_ack1 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[2+nr_tti_tx].ack; - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[3+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed - harq_ack2 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[3+nr_tti_tx].ack; + if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[3+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed + harq_ack2 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[3+nr_tti_tx].ack; - //if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(9+nr_tti_tx)%10].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed - //harq_ack3 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(9+nr_tti_tx)%10].ack; - //LOG_I(PHY,"SFN/SF %d/%d calculating n1_pucch cce0=%d n1_pucch0=%d cce1=%d n1_pucch1=%d cce2=%d n1_pucch2=%d\n", - // proc->frame_tx%1024, - // proc->nr_tti_tx_tx, - // nCCE0,n1_pucch0, - // nCCE1,n1_pucch1, nCCE2, n1_pucch2); - }else if (nr_tti_tx == 3) { - // i=0 + //if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(9+nr_tti_tx)%10].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed + //harq_ack3 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(9+nr_tti_tx)%10].ack; + //LOG_I(PHY,"SFN/SF %d/%d calculating n1_pucch cce0=%d n1_pucch0=%d cce1=%d n1_pucch1=%d cce2=%d n1_pucch2=%d\n", + // proc->frame_tx%1024, + // proc->nr_tti_tx_tx, + // nCCE0,n1_pucch0, + // nCCE1,n1_pucch1, nCCE2, n1_pucch2); + }else if (nr_tti_tx == 3) { + // i=0 - nCCE0 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[4+nr_tti_tx]; - n1_pucch0 = 3*nr_get_Np(frame_parms->N_RB_DL,nCCE0,0) + nCCE0+ frame_parms->pucch_config_common.n1PUCCH_AN; - // i=1 - nCCE1 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[5+nr_tti_tx]; - n1_pucch1 = 2*nr_get_Np(frame_parms->N_RB_DL,nCCE1,0) + nr_get_Np(frame_parms->N_RB_DL,nCCE1,1) + nCCE1 + frame_parms->pucch_config_common.n1PUCCH_AN; - // i=2 - nCCE2 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(6+nr_tti_tx)]; - n1_pucch2 = nr_get_Np(frame_parms->N_RB_DL,nCCE2,0) + 2*nr_get_Np(frame_parms->N_RB_DL,nCCE2,1) + nCCE2+ frame_parms->pucch_config_common.n1PUCCH_AN; - // i=3 - nCCE3 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(3+nr_tti_tx)]; - n1_pucch3 = 3*nr_get_Np(frame_parms->N_RB_DL,nCCE3,1) + nCCE3 + frame_parms->pucch_config_common.n1PUCCH_AN; + nCCE0 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[4+nr_tti_tx]; + n1_pucch0 = 3*nr_get_Np(frame_parms->N_RB_DL,nCCE0,0) + nCCE0+ frame_parms->pucch_config_common.n1PUCCH_AN; + // i=1 + nCCE1 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[5+nr_tti_tx]; + n1_pucch1 = 2*nr_get_Np(frame_parms->N_RB_DL,nCCE1,0) + nr_get_Np(frame_parms->N_RB_DL,nCCE1,1) + nCCE1 + frame_parms->pucch_config_common.n1PUCCH_AN; + // i=2 + nCCE2 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(6+nr_tti_tx)]; + n1_pucch2 = nr_get_Np(frame_parms->N_RB_DL,nCCE2,0) + 2*nr_get_Np(frame_parms->N_RB_DL,nCCE2,1) + nCCE2+ frame_parms->pucch_config_common.n1PUCCH_AN; + // i=3 + nCCE3 = ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->nCCE[(3+nr_tti_tx)]; + n1_pucch3 = 3*nr_get_Np(frame_parms->N_RB_DL,nCCE3,1) + nCCE3 + frame_parms->pucch_config_common.n1PUCCH_AN; - // set ACK/NAK to values if not DTX - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[4+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed + // set ACK/NAK to values if not DTX + if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[4+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed harq_ack0 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[4+nr_tti_tx].ack; - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[5+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed + if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[5+nr_tti_tx].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed harq_ack1 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[5+nr_tti_tx].ack; - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(6+nr_tti_tx)].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed + if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(6+nr_tti_tx)].send_harq_status>0) // n-6 // nr_tti_tx 6 is to be ACK/NAKed harq_ack2 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(6+nr_tti_tx)].ack; - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(3+nr_tti_tx)].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed + if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(3+nr_tti_tx)].send_harq_status>0) // n-6 // nr_tti_tx 5 is to be ACK/NAKed harq_ack3 = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack[(3+nr_tti_tx)].ack; - } + } - //LOG_I(PHY,"SFN/SF %d/%d calculating n1_pucch cce0=%d n1_pucch0=%d harq_ack0=%d cce1=%d n1_pucch1=%d harq_ack1=%d cce2=%d n1_pucch2=%d harq_ack2=%d cce3=%d n1_pucch3=%d harq_ack3=%d bundling_flag=%d\n", - // proc->frame_tx%1024, - // proc->nr_tti_tx, - // nCCE0,n1_pucch0,harq_ack0, - // nCCE1,n1_pucch1,harq_ack1, nCCE2, n1_pucch2, harq_ack2, - // nCCE3, n1_pucch3, harq_ack3, bundling_flag); + //LOG_I(PHY,"SFN/SF %d/%d calculating n1_pucch cce0=%d n1_pucch0=%d harq_ack0=%d cce1=%d n1_pucch1=%d harq_ack1=%d cce2=%d n1_pucch2=%d harq_ack2=%d cce3=%d n1_pucch3=%d harq_ack3=%d bundling_flag=%d\n", + // proc->frame_tx%1024, + // proc->nr_tti_tx, + // nCCE0,n1_pucch0,harq_ack0, + // nCCE1,n1_pucch1,harq_ack1, nCCE2, n1_pucch2, harq_ack2, + // nCCE3, n1_pucch3, harq_ack3, bundling_flag); - if ((bundling_flag==bundling)&&(SR == 0)) { // This is for bundling without SR, - b[0] = 1; - ack_counter = 0; + if ((bundling_flag==bundling)&&(SR == 0)) { // This is for bundling without SR, + b[0] = 1; + ack_counter = 0; - if ((harq_ack3!=2) ) { - b[0] = b[0]&harq_ack3; - n1_pucch_inter = n1_pucch3; - ack_counter ++; - } - if ((harq_ack0!=2) ) { - b[0] = b[0]&harq_ack0; - n1_pucch_inter = n1_pucch0; - ack_counter ++; - } - if ((harq_ack1!=2) ) { - b[0] = b[0]&harq_ack1; - n1_pucch_inter = n1_pucch1; - ack_counter ++; - } - if ((harq_ack2!=2) ) { - b[0] = b[0]&harq_ack2; - n1_pucch_inter = n1_pucch2; - ack_counter ++; - } + if ((harq_ack3!=2) ) { + b[0] = b[0]&harq_ack3; + n1_pucch_inter = n1_pucch3; + ack_counter ++; + } + if ((harq_ack0!=2) ) { + b[0] = b[0]&harq_ack0; + n1_pucch_inter = n1_pucch0; + ack_counter ++; + } + if ((harq_ack1!=2) ) { + b[0] = b[0]&harq_ack1; + n1_pucch_inter = n1_pucch1; + ack_counter ++; + } + if ((harq_ack2!=2) ) { + b[0] = b[0]&harq_ack2; + n1_pucch_inter = n1_pucch2; + ack_counter ++; + } - if (ack_counter == 0) - b[0] = 0; + if (ack_counter == 0) + b[0] = 0; - /*if (nr_tti_tx == 3) { - n1_pucch_inter = n1_pucch2; - } else if (nr_tti_tx == 2) { - n1_pucch_inter = n1_pucch1; - }*/ + /*if (nr_tti_tx == 3) { + n1_pucch_inter = n1_pucch2; + } else if (nr_tti_tx == 2) { + n1_pucch_inter = n1_pucch1; + }*/ - //LOG_I(PHY,"SFN/SF %d/%d calculating n1_pucch n1_pucch_inter=%d b[0]=%d b[1]=%d \n", - // proc->frame_tx%1024, - // proc->nr_tti_tx,n1_pucch_inter, - // b[0],b[1]); + //LOG_I(PHY,"SFN/SF %d/%d calculating n1_pucch n1_pucch_inter=%d b[0]=%d b[1]=%d \n", + // proc->frame_tx%1024, + // proc->nr_tti_tx,n1_pucch_inter, + // b[0],b[1]); - return(n1_pucch_inter); + return(n1_pucch_inter); - } else if ((bundling_flag==multiplexing)&&(SR==0)) { // Table 10.1 + } else if ((bundling_flag==multiplexing)&&(SR==0)) { // Table 10.1 - if (nr_tti_tx == 3) { - LOG_I(PHY, "sbuframe=%d \n",nr_tti_tx); - if ((harq_ack0 == 1) && (harq_ack1 == 1) && (harq_ack2 == 1) && (harq_ack3 == 1)) { - b[0] = 1; - b[1] = 1; - return(n1_pucch1); - } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { - b[0] = 1; - b[1] = 0; - return(n1_pucch1); - } else if (((harq_ack0 == 0) || (harq_ack0 == 2)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 0) && (harq_ack3 == 2)) { - b[0] = 1; - b[1] = 1; - return(n1_pucch2); - } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 1)) { - b[0] = 1; - b[1] = 0; - return(n1_pucch1); - } else if ((harq_ack0 == 0) && (harq_ack1 == 2) && (harq_ack2 == 2) && (harq_ack3 == 2)) { - b[0] = 1; - b[1] = 0; - return(n1_pucch0); - } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { - b[0] = 1; - b[1] = 0; - return(n1_pucch1); - } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && (harq_ack3 == 1)) { - b[0] = 0; - b[1] = 1; - return(n1_pucch3); - } else if (((harq_ack0 == 0) || (harq_ack0 == 2)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 0)) { - b[0] = 1; - b[1] = 1; - return(n1_pucch3); - } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { - b[0] = 0; - b[1] = 1; - return(n1_pucch2); - } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 1)) { - b[0] = 0; - b[1] = 1; - return(n1_pucch0); - } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { - b[0] = 0; - b[1] = 1; - return(n1_pucch0); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && (harq_ack2 == 1) && (harq_ack3 == 1)) { - b[0] = 0; - b[1] = 1; - return(n1_pucch3); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 0) && (harq_ack2 == 2) && (harq_ack3 == 2)) { - b[0] = 0; - b[1] = 0; - return(n1_pucch1); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { - b[0] = 1; - b[1] = 0; - return(n1_pucch2); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 1)) { - b[0] = 1; - b[1] = 0; - return(n1_pucch3); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { - b[0] = 0; - b[1] = 1; - return(n1_pucch1); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && (harq_ack3 == 1)) { - b[0] = 0; - b[1] = 1; - return(n1_pucch3); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { - b[0] = 0; - b[1] = 0; - return(n1_pucch2); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack3 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { - b[0] = 0; - b[1] = 0; - return(n1_pucch3); - } - } else if (nr_tti_tx == 2) { - if ((harq_ack0 == 1) && (harq_ack1 == 1) && (harq_ack2 == 1)) { - b[0] = 1; - b[1] = 1; - return(n1_pucch2); - } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { - b[0] = 1; - b[1] = 1; - return(n1_pucch1); - } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1)) { - b[0] = 1; - b[1] = 1; - return(n1_pucch0); - } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { - b[0] = 0; - b[1] = 1; - return(n1_pucch0); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && (harq_ack2 == 1)) { - b[0] = 1; - b[1] = 0; - return(n1_pucch2); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { - b[1] = 0; - b[0] = 0; - return(n1_pucch1); - } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1)) { - b[0] = 0; - b[1] = 0; - return(n1_pucch2); - } else if ((harq_ack0 == 2) && (harq_ack1 == 2) && (harq_ack2 == 0)) { - b[0] = 0; - b[1] = 1; - return(n1_pucch2); - } else if ((harq_ack0 == 2) && (harq_ack1 == 0) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { - b[0] = 1; - b[1] = 0; - return(n1_pucch1); - } else if ((harq_ack0 == 0) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { - b[0] = 1; - b[1] = 0; - return(n1_pucch0); - } + if (nr_tti_tx == 3) { + LOG_I(PHY, "sbuframe=%d \n",nr_tti_tx); + if ((harq_ack0 == 1) && (harq_ack1 == 1) && (harq_ack2 == 1) && (harq_ack3 == 1)) { + b[0] = 1; + b[1] = 1; + return(n1_pucch1); + } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { + b[0] = 1; + b[1] = 0; + return(n1_pucch1); + } else if (((harq_ack0 == 0) || (harq_ack0 == 2)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 0) && (harq_ack3 == 2)) { + b[0] = 1; + b[1] = 1; + return(n1_pucch2); + } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 1)) { + b[0] = 1; + b[1] = 0; + return(n1_pucch1); + } else if ((harq_ack0 == 0) && (harq_ack1 == 2) && (harq_ack2 == 2) && (harq_ack3 == 2)) { + b[0] = 1; + b[1] = 0; + return(n1_pucch0); + } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { + b[0] = 1; + b[1] = 0; + return(n1_pucch1); + } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && (harq_ack3 == 1)) { + b[0] = 0; + b[1] = 1; + return(n1_pucch3); + } else if (((harq_ack0 == 0) || (harq_ack0 == 2)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 0)) { + b[0] = 1; + b[1] = 1; + return(n1_pucch3); + } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { + b[0] = 0; + b[1] = 1; + return(n1_pucch2); + } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 1)) { + b[0] = 0; + b[1] = 1; + return(n1_pucch0); + } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { + b[0] = 0; + b[1] = 1; + return(n1_pucch0); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && (harq_ack2 == 1) && (harq_ack3 == 1)) { + b[0] = 0; + b[1] = 1; + return(n1_pucch3); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 0) && (harq_ack2 == 2) && (harq_ack3 == 2)) { + b[0] = 0; + b[1] = 0; + return(n1_pucch1); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { + b[0] = 1; + b[1] = 0; + return(n1_pucch2); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && (harq_ack3 == 1)) { + b[0] = 1; + b[1] = 0; + return(n1_pucch3); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0)) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { + b[0] = 0; + b[1] = 1; + return(n1_pucch1); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && (harq_ack3 == 1)) { + b[0] = 0; + b[1] = 1; + return(n1_pucch3); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1) && ((harq_ack3 == 2) || (harq_ack3 == 0))) { + b[0] = 0; + b[1] = 0; + return(n1_pucch2); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack3 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { + b[0] = 0; + b[1] = 0; + return(n1_pucch3); + } + } else if (nr_tti_tx == 2) { + if ((harq_ack0 == 1) && (harq_ack1 == 1) && (harq_ack2 == 1)) { + b[0] = 1; + b[1] = 1; + return(n1_pucch2); + } else if ((harq_ack0 == 1) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { + b[0] = 1; + b[1] = 1; + return(n1_pucch1); + } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1)) { + b[0] = 1; + b[1] = 1; + return(n1_pucch0); + } else if ((harq_ack0 == 1) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { + b[0] = 0; + b[1] = 1; + return(n1_pucch0); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && (harq_ack2 == 1)) { + b[0] = 1; + b[1] = 0; + return(n1_pucch2); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && (harq_ack1 == 1) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { + b[1] = 0; + b[0] = 0; + return(n1_pucch1); + } else if (((harq_ack0 == 2) || (harq_ack0 == 0)) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && (harq_ack2 == 1)) { + b[0] = 0; + b[1] = 0; + return(n1_pucch2); + } else if ((harq_ack0 == 2) && (harq_ack1 == 2) && (harq_ack2 == 0)) { + b[0] = 0; + b[1] = 1; + return(n1_pucch2); + } else if ((harq_ack0 == 2) && (harq_ack1 == 0) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { + b[0] = 1; + b[1] = 0; + return(n1_pucch1); + } else if ((harq_ack0 == 0) && ((harq_ack1 == 2) || (harq_ack1 == 0)) && ((harq_ack2 == 2) || (harq_ack2 == 0))) { + b[0] = 1; + b[1] = 0; + return(n1_pucch0); + } - } - } else if (SR==1) { // SR and 0,1,or 2 ACKS, (first 3 entries in Table 7.3-1 of 36.213) - // this should be number of ACKs (including - ack_counter = 0; - if (harq_ack0==1) - ack_counter ++; - if (harq_ack1==1) - ack_counter ++; - if (harq_ack2==1) - ack_counter ++; - if (harq_ack3==1) - ack_counter ++; + } + } else if (SR==1) { // SR and 0,1,or 2 ACKS, (first 3 entries in Table 7.3-1 of 36.213) + // this should be number of ACKs (including + ack_counter = 0; + if (harq_ack0==1) + ack_counter ++; + if (harq_ack1==1) + ack_counter ++; + if (harq_ack2==1) + ack_counter ++; + if (harq_ack3==1) + ack_counter ++; - switch (ack_counter) { - case 0: - b[0] = 0; - b[1] = 0; - break; + switch (ack_counter) { + case 0: + b[0] = 0; + b[1] = 0; + break; - case 1: - b[0] = 1; - b[1] = 1; - break; + case 1: + b[0] = 1; + b[1] = 1; + break; - case 2: - b[0] = 1; - b[1] = 0; - break; + case 2: + b[0] = 1; + b[1] = 0; + break; - case 3: - b[0] = 0; - b[1] = 1; - break; + case 3: + b[0] = 0; + b[1] = 1; + break; - case 4: - b[0] = 1; - b[1] = 1; - break; - } + case 4: + b[0] = 1; + b[1] = 1; + break; + } - ack_counter = 0; - return(ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex); - } + ack_counter = 0; + return(ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex); + } - break; + break; } // switch tdd_config } @@ -1262,33 +1266,33 @@ uint16_t nr_get_n1_pucch(PHY_VARS_NR_UE *ue, #ifdef EMOS - void phy_procedures_emos_UE_TX(uint8_t next_slot,uint8_t eNB_id) { +void phy_procedures_emos_UE_TX(uint8_t next_slot,uint8_t eNB_id) { uint8_t harq_pid; if (next_slot%2==0) { - // get harq_pid from nr_tti_tx relationship - harq_pid = nr_subframe2harq_pid(&ue->frame_parms,ue->frame,(next_slot>>1)); - if (harq_pid==255) { - LOG_E(PHY,"[UE%d] Frame %d : FATAL ERROR: illegal harq_pid, returning\n", - 0,ue->frame); - return; - } + // get harq_pid from nr_tti_tx relationship + harq_pid = nr_subframe2harq_pid(&ue->frame_parms,ue->frame,(next_slot>>1)); + if (harq_pid==255) { + LOG_E(PHY,"[UE%d] Frame %d : FATAL ERROR: illegal harq_pid, returning\n", + 0,ue->frame); + return; + } - if (ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag == 1) { - emos_dump_UE.uci_cnt[next_slot>>1] = 1; - memcpy(emos_dump_UE.UCI_data[0][next_slot>>1].o,ulsch[eNB_id]->o,MAX_CQI_BITS*sizeof(char)); - emos_dump_UE.UCI_data[0][next_slot>>1].O = ulsch[eNB_id]->O; - memcpy(emos_dump_UE.UCI_data[0][next_slot>>1].o_RI,ulsch[eNB_id]->o_RI,2*sizeof(char)); - emos_dump_UE.UCI_data[0][next_slot>>1].O_RI = ulsch[eNB_id]->O_RI; - memcpy(emos_dump_UE.UCI_data[0][next_slot>>1].o_ACK,ulsch[eNB_id]->o_ACK,4*sizeof(char)); - emos_dump_UE.UCI_data[0][next_slot>>1].O_ACK = ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK; - } - else { - emos_dump_UE.uci_cnt[next_slot>>1] = 0; - } - } + if (ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag == 1) { + emos_dump_UE.uci_cnt[next_slot>>1] = 1; + memcpy(emos_dump_UE.UCI_data[0][next_slot>>1].o,ulsch[eNB_id]->o,MAX_CQI_BITS*sizeof(char)); + emos_dump_UE.UCI_data[0][next_slot>>1].O = ulsch[eNB_id]->O; + memcpy(emos_dump_UE.UCI_data[0][next_slot>>1].o_RI,ulsch[eNB_id]->o_RI,2*sizeof(char)); + emos_dump_UE.UCI_data[0][next_slot>>1].O_RI = ulsch[eNB_id]->O_RI; + memcpy(emos_dump_UE.UCI_data[0][next_slot>>1].o_ACK,ulsch[eNB_id]->o_ACK,4*sizeof(char)); + emos_dump_UE.UCI_data[0][next_slot>>1].O_ACK = ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK; + } + else { + emos_dump_UE.uci_cnt[next_slot>>1] = 0; + } } +} #endif @@ -1315,88 +1319,88 @@ void ulsch_common_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, uint8_ #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706)//this is the EXPRESS MIMO case ulsch_start = (ue->rx_offset+nr_tti_tx*frame_parms->samples_per_subframe- - ue->hw_timing_advance- - ue->timing_advance- - ue->N_TA_offset+5); + ue->hw_timing_advance- + ue->timing_advance- + ue->N_TA_offset+5); //LOG_E(PHY,"ul-signal [nr_tti_rx: %d, ulsch_start %d]\n",nr_tti_tx, ulsch_start); if(ulsch_start < 0) - ulsch_start = ulsch_start + (LTE_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe); + ulsch_start = ulsch_start + (LTE_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe); if (ulsch_start > (LTE_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe)) - ulsch_start = ulsch_start % (LTE_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe); + ulsch_start = ulsch_start % (LTE_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe); //LOG_E(PHY,"ul-signal [nr_tti_rx: %d, ulsch_start %d]\n",nr_tti_tx, ulsch_start); #else //this is the normal case ulsch_start = (frame_parms->samples_per_subframe*nr_tti_tx)-ue->N_TA_offset; //-ue->timing_advance; #endif //else EXMIMO -//#if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) + //#if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) if (empty_subframe) - { -//#if 1 + { + //#if 1 overflow = ulsch_start - 9*frame_parms->samples_per_subframe; for (aa=0; aanb_antennas_tx; aa++) { - if (overflow > 0) - { - memset(&ue->common_vars.txdata[aa][ulsch_start],0,4*(frame_parms->samples_per_subframe-overflow)); - memset(&ue->common_vars.txdata[aa][0],0,4*overflow); - } - else - { - memset(&ue->common_vars.txdata[aa][ulsch_start],0,4*frame_parms->samples_per_subframe); - } + if (overflow > 0) + { + memset(&ue->common_vars.txdata[aa][ulsch_start],0,4*(frame_parms->samples_per_subframe-overflow)); + memset(&ue->common_vars.txdata[aa][0],0,4*overflow); + } + else + { + memset(&ue->common_vars.txdata[aa][ulsch_start],0,4*frame_parms->samples_per_subframe); + } } -/*#else - overflow = ulsch_start - 9*frame_parms->samples_per_subframe; - for (aa=0; aanb_antennas_tx; aa++) { - for (k=ulsch_start; ksamples_per_subframe*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME,ulsch_start+frame_parms->samples_per_subframe); k++) { - ((short*)ue->common_vars.txdata[aa])[2*k] = 0; - ((short*)ue->common_vars.txdata[aa])[2*k+1] = 0; - } + /*#else + overflow = ulsch_start - 9*frame_parms->samples_per_subframe; + for (aa=0; aanb_antennas_tx; aa++) { + for (k=ulsch_start; ksamples_per_subframe*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME,ulsch_start+frame_parms->samples_per_subframe); k++) { + ((short*)ue->common_vars.txdata[aa])[2*k] = 0; + ((short*)ue->common_vars.txdata[aa])[2*k+1] = 0; + } - for (k=0; kcommon_vars.txdata[aa])[2*k] = 0; - ((short*)ue->common_vars.txdata[aa])[2*k+1] = 0; - } - } -endif*/ + for (k=0; kcommon_vars.txdata[aa])[2*k] = 0; + ((short*)ue->common_vars.txdata[aa])[2*k+1] = 0; + } + } + endif*/ return; - } + } if ((frame_tx%100) == 0) LOG_D(PHY,"[UE %d] Frame %d, nr_tti_rx %d: ulsch_start = %d (rxoff %d, HW TA %d, timing advance %d, TA_offset %d\n", - ue->Mod_id,frame_tx,nr_tti_tx, - ulsch_start, - ue->rx_offset, - ue->hw_timing_advance, - ue->timing_advance, - ue->N_TA_offset); + ue->Mod_id,frame_tx,nr_tti_tx, + ulsch_start, + ue->rx_offset, + ue->hw_timing_advance, + ue->timing_advance, + ue->N_TA_offset); for (aa=0; aanb_antennas_tx; aa++) { if (frame_parms->Ncp == 1) PHY_ofdm_mod(&ue->common_vars.txdataF[aa][nr_tti_tx*nsymb*frame_parms->ofdm_symbol_size], #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) - dummy_tx_buffer, + dummy_tx_buffer, #else - &ue->common_vars.txdata[aa][ulsch_start], + &ue->common_vars.txdata[aa][ulsch_start], #endif - frame_parms->ofdm_symbol_size, - nsymb, - frame_parms->nb_prefix_samples, - CYCLIC_PREFIX); + frame_parms->ofdm_symbol_size, + nsymb, + frame_parms->nb_prefix_samples, + CYCLIC_PREFIX); else normal_prefix_mod(&ue->common_vars.txdataF[aa][nr_tti_tx*nsymb*frame_parms->ofdm_symbol_size], #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) - dummy_tx_buffer, + dummy_tx_buffer, #else - &ue->common_vars.txdata[aa][ulsch_start], + &ue->common_vars.txdata[aa][ulsch_start], #endif - nsymb, - &ue->frame_parms); + nsymb, + &ue->frame_parms); #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) @@ -1425,170 +1429,34 @@ endif*/ // handle switch before 1st TX nr_tti_rx, guarantee that the slot prior to transmission is switch on for (k=ulsch_start - (frame_parms->samples_per_subframe>>1) ; kcommon_vars.txdata[aa][k+frame_parms->samples_per_subframe*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME] &= 0xFFFEFFFE; + ue->common_vars.txdata[aa][k+frame_parms->samples_per_subframe*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME] &= 0xFFFEFFFE; else if (k>(frame_parms->samples_per_subframe*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME)) - ue->common_vars.txdata[aa][k-frame_parms->samples_per_subframe*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME] &= 0xFFFEFFFE; + ue->common_vars.txdata[aa][k-frame_parms->samples_per_subframe*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME] &= 0xFFFEFFFE; else - ue->common_vars.txdata[aa][k] &= 0xFFFEFFFE; + ue->common_vars.txdata[aa][k] &= 0xFFFEFFFE; } #endif #endif /* - only for debug - LOG_I(PHY,"ul-signal [nr_tti_rx: %d, ulsch_start %d, TA: %d, rxOffset: %d, timing_advance: %d, hw_timing_advance: %d]\n",nr_tti_tx, ulsch_start, ue->N_TA_offset, ue->rx_offset, ue->timing_advance, ue->hw_timing_advance); - if( (crash == 1) && (nr_tti_tx == 0) ) - { + only for debug + LOG_I(PHY,"ul-signal [nr_tti_rx: %d, ulsch_start %d, TA: %d, rxOffset: %d, timing_advance: %d, hw_timing_advance: %d]\n",nr_tti_tx, ulsch_start, ue->N_TA_offset, ue->rx_offset, ue->timing_advance, ue->hw_timing_advance); + if( (crash == 1) && (nr_tti_tx == 0) ) + { LOG_E(PHY,"***** DUMP TX Signal [ulsch_start %d] *****\n",ulsch_start); write_output("txBuff.m","txSignal",&ue->common_vars.txdata[aa][ulsch_start],frame_parms->samples_per_subframe,1,1); - } + } */ } //nb_antennas_tx #if UE_TIMING_TRACE - stop_meas(&ue->ofdm_mod_stats); + stop_meas(&ue->ofdm_mod_stats); #endif VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_ULSCH_COMMON,VCD_FUNCTION_OUT); } -#endif //0 -void ue_nr_prach_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,runmode_t mode) { - - int frame_tx = proc->frame_tx; - int nr_tti_tx = proc->nr_tti_tx; - int prach_power; - PRACH_RESOURCES_t prach_resources_local; - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_PRACH, VCD_FUNCTION_IN); - - ue->generate_prach=0; - - if (ue->mac_enabled==0){ - ue->prach_resources[eNB_id] = &prach_resources_local; - prach_resources_local.ra_RNTI = 0xbeef; - prach_resources_local.ra_PreambleIndex = 0; - } - - if (ue->mac_enabled==1){ - // ask L2 for RACH transport - if ((mode != rx_calib_ue) && (mode != rx_calib_ue_med) && (mode != rx_calib_ue_byp) && (mode != no_L2_connect) ) { - LOG_D(PHY,"Getting PRACH resources\n"); - //ue->prach_resources[eNB_id] = mac_xface->ue_get_rach(ue->Mod_id, - // ue->CC_id, - // frame_tx, - // eNB_id, - // nr_tti_tx); - //LOG_D(PHY,"Got prach_resources for eNB %d address %p, RRCCommon %p\n",eNB_id,ue->prach_resources[eNB_id],UE_mac_inst[ue->Mod_id].radioResourceConfigCommon); - LOG_D(PHY,"Prach resources %p\n",ue->prach_resources[eNB_id]); - } - } - - if (ue->prach_resources[eNB_id]!=NULL) { - - ue->generate_prach=1; - ue->prach_cnt=0; -#ifdef SMBV - ue->prach_resources[eNB_id]->ra_PreambleIndex = 19; -#endif -#ifdef OAI_EMU - ue->prach_PreambleIndex=ue->prach_resources[eNB_id]->ra_PreambleIndex; -#endif - - if (abstraction_flag == 0) { - - LOG_I(PHY,"mode %d\n",mode); - - if ((ue->mac_enabled==1) && (mode != calib_prach_tx)) { - ue->tx_power_dBm[nr_tti_tx] = ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER+get_PL(ue->Mod_id,ue->CC_id,eNB_id); - } - else { - ue->tx_power_dBm[nr_tti_tx] = ue->tx_power_max_dBm; - ue->prach_resources[eNB_id]->ra_PreambleIndex = 19; - } - - LOG_I(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d : Generating PRACH, preamble %d,PL %d, P0_PRACH %d, TARGET_RECEIVED_POWER %d dBm, PRACH TDD Resource index %d, RA-RNTI %d\n", - ue->Mod_id, - frame_tx, - nr_tti_tx, - ue->prach_resources[eNB_id]->ra_PreambleIndex, - get_PL(ue->Mod_id,ue->CC_id,eNB_id), - ue->tx_power_dBm[nr_tti_tx], - ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER, - ue->prach_resources[eNB_id]->ra_TDD_map_index, - ue->prach_resources[eNB_id]->ra_RNTI); - - ue->tx_total_RE[nr_tti_tx] = 96; - -#if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) - ue->prach_vars[eNB_id]->amp = get_tx_amp(ue->tx_power_dBm[nr_tti_tx], - ue->tx_power_max_dBm, - ue->frame_parms.N_RB_UL, - 6); -#else - ue->prach_vars[eNB_id]->amp = AMP; -#endif - if ((mode == calib_prach_tx) && (((proc->frame_tx&0xfffe)%100)==0)) - LOG_D(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d : PRACH TX power %d dBm, amp %d\n", - ue->Mod_id, - proc->frame_rx, - proc->nr_tti_tx, - ue->tx_power_dBm[nr_tti_tx], - ue->prach_vars[eNB_id]->amp); - - - // start_meas(&ue->tx_prach); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GENERATE_PRACH, VCD_FUNCTION_IN); - - //prach_power = generate_prach(ue,eNB_id,nr_tti_tx,frame_tx); - prach_power = generate_nr_prach(ue,eNB_id,nr_tti_tx,frame_tx); - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GENERATE_PRACH, VCD_FUNCTION_OUT); - // stop_meas(&ue->tx_prach); - LOG_D(PHY,"[UE %d][RAPROC] PRACH PL %d dB, power %d dBm, digital power %d dB (amp %d)\n", - ue->Mod_id, - get_PL(ue->Mod_id,ue->CC_id,eNB_id), - ue->tx_power_dBm[nr_tti_tx], - dB_fixed(prach_power), - ue->prach_vars[eNB_id]->amp); - } else { - //UE_transport_info[ue->Mod_id][ue->CC_id].cntl.prach_flag=1; - //UE_transport_info[ue->Mod_id][ue->CC_id].cntl.prach_id=ue->prach_resources[eNB_id]->ra_PreambleIndex; - } - - if (ue->mac_enabled==1){ - //mac_xface->Msg1_transmitted(ue->Mod_id, - // ue->CC_id, - // frame_tx, - // eNB_id); - } - - LOG_I(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d: Generating PRACH (eNB %d) preamble index %d for UL, TX power %d dBm (PL %d dB), l3msg \n", - ue->Mod_id,frame_tx,nr_tti_tx,eNB_id, - ue->prach_resources[eNB_id]->ra_PreambleIndex, - ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER+get_PL(ue->Mod_id,ue->CC_id,eNB_id), - get_PL(ue->Mod_id,ue->CC_id,eNB_id)); - - } - - - // if we're calibrating the PRACH kill the pointer to its resources so that the RA protocol doesn't continue - if (mode == calib_prach_tx) - ue->prach_resources[eNB_id]=NULL; - - LOG_D(PHY,"[UE %d] frame %d nr_tti_rx %d : generate_prach %d, prach_cnt %d\n", - ue->Mod_id,frame_tx,nr_tti_tx,ue->generate_prach,ue->prach_cnt); - - ue->prach_cnt++; - - if (ue->prach_cnt==3) - ue->generate_prach=0; - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_PRACH, VCD_FUNCTION_OUT); -} - -#if 0 void ue_prach_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,runmode_t mode) { int frame_tx = proc->frame_tx; @@ -1611,113 +1479,113 @@ void ue_prach_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_ if ((mode != rx_calib_ue) && (mode != rx_calib_ue_med) && (mode != rx_calib_ue_byp) && (mode != no_L2_connect) ) { LOG_D(PHY,"Getting PRACH resources\n"); //ue->prach_resources[eNB_id] = mac_xface->ue_get_rach(ue->Mod_id, - ue->CC_id, - frame_tx, - eNB_id, - nr_tti_tx); - LOG_D(PHY,"Got prach_resources for eNB %d address %p, RRCCommon %p\n",eNB_id,ue->prach_resources[eNB_id],UE_mac_inst[ue->Mod_id].radioResourceConfigCommon); - LOG_D(PHY,"Prach resources %p\n",ue->prach_resources[eNB_id]); - } + ue->CC_id, + frame_tx, + eNB_id, + nr_tti_tx); + LOG_D(PHY,"Got prach_resources for eNB %d address %p, RRCCommon %p\n",eNB_id,ue->prach_resources[eNB_id],UE_mac_inst[ue->Mod_id].radioResourceConfigCommon); + LOG_D(PHY,"Prach resources %p\n",ue->prach_resources[eNB_id]); } +} - if (ue->prach_resources[eNB_id]!=NULL) { +if (ue->prach_resources[eNB_id]!=NULL) { - ue->generate_prach=1; - ue->prach_cnt=0; + ue->generate_prach=1; + ue->prach_cnt=0; #ifdef SMBV - ue->prach_resources[eNB_id]->ra_PreambleIndex = 19; + ue->prach_resources[eNB_id]->ra_PreambleIndex = 19; #endif #ifdef OAI_EMU - ue->prach_PreambleIndex=ue->prach_resources[eNB_id]->ra_PreambleIndex; + ue->prach_PreambleIndex=ue->prach_resources[eNB_id]->ra_PreambleIndex; #endif - if (abstraction_flag == 0) { + if (abstraction_flag == 0) { - LOG_I(PHY,"mode %d\n",mode); + LOG_I(PHY,"mode %d\n",mode); - if ((ue->mac_enabled==1) && (mode != calib_prach_tx)) { - ue->tx_power_dBm[nr_tti_tx] = ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER+get_PL(ue->Mod_id,ue->CC_id,eNB_id); - } - else { - ue->tx_power_dBm[nr_tti_tx] = ue->tx_power_max_dBm; - ue->prach_resources[eNB_id]->ra_PreambleIndex = 19; - } + if ((ue->mac_enabled==1) && (mode != calib_prach_tx)) { + ue->tx_power_dBm[nr_tti_tx] = ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER+get_PL(ue->Mod_id,ue->CC_id,eNB_id); + } + else { + ue->tx_power_dBm[nr_tti_tx] = ue->tx_power_max_dBm; + ue->prach_resources[eNB_id]->ra_PreambleIndex = 19; + } - LOG_I(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d : Generating PRACH, preamble %d,PL %d, P0_PRACH %d, TARGET_RECEIVED_POWER %d dBm, PRACH TDD Resource index %d, RA-RNTI %d\n", - ue->Mod_id, - frame_tx, - nr_tti_tx, - ue->prach_resources[eNB_id]->ra_PreambleIndex, - get_PL(ue->Mod_id,ue->CC_id,eNB_id), - ue->tx_power_dBm[nr_tti_tx], - ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER, - ue->prach_resources[eNB_id]->ra_TDD_map_index, - ue->prach_resources[eNB_id]->ra_RNTI); + LOG_I(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d : Generating PRACH, preamble %d,PL %d, P0_PRACH %d, TARGET_RECEIVED_POWER %d dBm, PRACH TDD Resource index %d, RA-RNTI %d\n", + ue->Mod_id, + frame_tx, + nr_tti_tx, + ue->prach_resources[eNB_id]->ra_PreambleIndex, + get_PL(ue->Mod_id,ue->CC_id,eNB_id), + ue->tx_power_dBm[nr_tti_tx], + ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER, + ue->prach_resources[eNB_id]->ra_TDD_map_index, + ue->prach_resources[eNB_id]->ra_RNTI); - ue->tx_total_RE[nr_tti_tx] = 96; + ue->tx_total_RE[nr_tti_tx] = 96; #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) - ue->prach_vars[eNB_id]->amp = get_tx_amp(ue->tx_power_dBm[nr_tti_tx], - ue->tx_power_max_dBm, - ue->frame_parms.N_RB_UL, - 6); + ue->prach_vars[eNB_id]->amp = get_tx_amp(ue->tx_power_dBm[nr_tti_tx], + ue->tx_power_max_dBm, + ue->frame_parms.N_RB_UL, + 6); #else - ue->prach_vars[eNB_id]->amp = AMP; + ue->prach_vars[eNB_id]->amp = AMP; #endif - if ((mode == calib_prach_tx) && (((proc->frame_tx&0xfffe)%100)==0)) - LOG_D(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d : PRACH TX power %d dBm, amp %d\n", - ue->Mod_id, - proc->frame_rx, - proc->nr_tti_tx, - ue->tx_power_dBm[nr_tti_tx], - ue->prach_vars[eNB_id]->amp); + if ((mode == calib_prach_tx) && (((proc->frame_tx&0xfffe)%100)==0)) + LOG_D(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d : PRACH TX power %d dBm, amp %d\n", + ue->Mod_id, + proc->frame_rx, + proc->nr_tti_tx, + ue->tx_power_dBm[nr_tti_tx], + ue->prach_vars[eNB_id]->amp); - // start_meas(&ue->tx_prach); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GENERATE_PRACH, VCD_FUNCTION_IN); - prach_power = generate_prach(ue,eNB_id,nr_tti_tx,frame_tx); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GENERATE_PRACH, VCD_FUNCTION_OUT); - // stop_meas(&ue->tx_prach); - LOG_D(PHY,"[UE %d][RAPROC] PRACH PL %d dB, power %d dBm, digital power %d dB (amp %d)\n", - ue->Mod_id, - get_PL(ue->Mod_id,ue->CC_id,eNB_id), - ue->tx_power_dBm[nr_tti_tx], - dB_fixed(prach_power), - ue->prach_vars[eNB_id]->amp); - } else { - UE_transport_info[ue->Mod_id][ue->CC_id].cntl.prach_flag=1; - UE_transport_info[ue->Mod_id][ue->CC_id].cntl.prach_id=ue->prach_resources[eNB_id]->ra_PreambleIndex; - } - - if (ue->mac_enabled==1){ - //mac_xface->Msg1_transmitted(ue->Mod_id, - ue->CC_id, - frame_tx, - eNB_id); - } - - LOG_I(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d: Generating PRACH (eNB %d) preamble index %d for UL, TX power %d dBm (PL %d dB), l3msg \n", - ue->Mod_id,frame_tx,nr_tti_tx,eNB_id, - ue->prach_resources[eNB_id]->ra_PreambleIndex, - ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER+get_PL(ue->Mod_id,ue->CC_id,eNB_id), - get_PL(ue->Mod_id,ue->CC_id,eNB_id)); - + // start_meas(&ue->tx_prach); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GENERATE_PRACH, VCD_FUNCTION_IN); + prach_power = generate_prach(ue,eNB_id,nr_tti_tx,frame_tx); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GENERATE_PRACH, VCD_FUNCTION_OUT); + // stop_meas(&ue->tx_prach); + LOG_D(PHY,"[UE %d][RAPROC] PRACH PL %d dB, power %d dBm, digital power %d dB (amp %d)\n", + ue->Mod_id, + get_PL(ue->Mod_id,ue->CC_id,eNB_id), + ue->tx_power_dBm[nr_tti_tx], + dB_fixed(prach_power), + ue->prach_vars[eNB_id]->amp); + } else { + UE_transport_info[ue->Mod_id][ue->CC_id].cntl.prach_flag=1; + UE_transport_info[ue->Mod_id][ue->CC_id].cntl.prach_id=ue->prach_resources[eNB_id]->ra_PreambleIndex; } + if (ue->mac_enabled==1){ + //mac_xface->Msg1_transmitted(ue->Mod_id, + ue->CC_id, + frame_tx, + eNB_id); + } - // if we're calibrating the PRACH kill the pointer to its resources so that the RA protocol doesn't continue - if (mode == calib_prach_tx) - ue->prach_resources[eNB_id]=NULL; +LOG_I(PHY,"[UE %d][RAPROC] Frame %d, nr_tti_rx %d: Generating PRACH (eNB %d) preamble index %d for UL, TX power %d dBm (PL %d dB), l3msg \n", + ue->Mod_id,frame_tx,nr_tti_tx,eNB_id, + ue->prach_resources[eNB_id]->ra_PreambleIndex, + ue->prach_resources[eNB_id]->ra_PREAMBLE_RECEIVED_TARGET_POWER+get_PL(ue->Mod_id,ue->CC_id,eNB_id), + get_PL(ue->Mod_id,ue->CC_id,eNB_id)); - LOG_D(PHY,"[UE %d] frame %d nr_tti_rx %d : generate_prach %d, prach_cnt %d\n", - ue->Mod_id,frame_tx,nr_tti_tx,ue->generate_prach,ue->prach_cnt); +} - ue->prach_cnt++; - if (ue->prach_cnt==3) - ue->generate_prach=0; +// if we're calibrating the PRACH kill the pointer to its resources so that the RA protocol doesn't continue +if (mode == calib_prach_tx) + ue->prach_resources[eNB_id]=NULL; - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_PRACH, VCD_FUNCTION_OUT); +LOG_D(PHY,"[UE %d] frame %d nr_tti_rx %d : generate_prach %d, prach_cnt %d\n", + ue->Mod_id,frame_tx,nr_tti_tx,ue->generate_prach,ue->prach_cnt); + +ue->prach_cnt++; + +if (ue->prach_cnt==3) + ue->generate_prach=0; + +VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_PRACH, VCD_FUNCTION_OUT); } #endif @@ -1755,33 +1623,33 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 if (ue->mac_enabled == 1) { if ((ue->ulsch_Msg3_active[eNB_id] == 1) && - (ue->ulsch_Msg3_frame[eNB_id] == frame_tx) && - (ue->ulsch_Msg3_subframe[eNB_id] == nr_tti_tx)) { // Initial Transmission of Msg3 + (ue->ulsch_Msg3_frame[eNB_id] == frame_tx) && + (ue->ulsch_Msg3_subframe[eNB_id] == nr_tti_tx)) { // Initial Transmission of Msg3 ue->ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag = 1; if (ue->ulsch[eNB_id]->harq_processes[harq_pid]->round==0) - generate_ue_ulsch_params_from_rar(ue, - proc, - eNB_id); + generate_ue_ulsch_params_from_rar(ue, + proc, + eNB_id); ue->ulsch[eNB_id]->power_offset = 14; LOG_D(PHY,"[UE %d][RAPROC] Frame %d: Setting Msg3_flag in nr_tti_rx %d, for harq_pid %d\n", - Mod_id, - frame_tx, - nr_tti_tx, - harq_pid); + Mod_id, + frame_tx, + nr_tti_tx, + harq_pid); Msg3_flag = 1; } else { - /* no pusch has been scheduled on this transmit slot */ + /* no pusch has been scheduled on this transmit slot */ if (harq_pid == NR_MAX_HARQ_PROCESSES) { - LOG_E(PHY,"[UE%d] Frame %d nr_tti_rx %d ulsch_decoding.c: FATAL ERROR: illegal harq_pid, returning\n", + LOG_E(PHY,"[UE%d] Frame %d nr_tti_rx %d ulsch_decoding.c: FATAL ERROR: illegal harq_pid, returning\n", Mod_id,frame_tx, nr_tti_tx); //mac_xface->macphy_exit("Error in ulsch_decoding"); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX, VCD_FUNCTION_OUT); #if UE_TIMING_TRACE - stop_meas(&ue->phy_proc_tx); + stop_meas(&ue->phy_proc_tx); #endif return; } @@ -1795,31 +1663,31 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 uint8_t isBad = 0; if (ue->frame_parms.N_RB_UL <= ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb) { LOG_D(PHY,"Invalid PUSCH first_RB=%d for N_RB_UL=%d\n", - ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb, - ue->frame_parms.N_RB_UL); + ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb, + ue->frame_parms.N_RB_UL); isBad = 1; } if (ue->frame_parms.N_RB_UL < ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb) { LOG_D(PHY,"Invalid PUSCH num_RB=%d for N_RB_UL=%d\n", - ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb, - ue->frame_parms.N_RB_UL); + ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb, + ue->frame_parms.N_RB_UL); isBad = 1; } if (0 > ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb) { LOG_D(PHY,"Invalid PUSCH first_RB=%d\n", - ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb); + ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb); isBad = 1; } if (0 >= ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb) { LOG_D(PHY,"Invalid PUSCH num_RB=%d\n", - ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb); + ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb); isBad = 1; } if (ue->frame_parms.N_RB_UL < (ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb + ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb)) { LOG_D(PHY,"Invalid PUSCH num_RB=%d + first_RB=%d for N_RB_UL=%d\n", - ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb, - ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb, - ue->frame_parms.N_RB_UL); + ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb, + ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb, + ue->frame_parms.N_RB_UL); isBad = 1; } if ((0 > ue->ulsch[eNB_id]->harq_processes[harq_pid]->rvidx) || @@ -1829,8 +1697,8 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 } if (20 < ue->ulsch[eNB_id]->harq_processes[harq_pid]->mcs) { - LOG_D(PHY,"Not supported MCS in OAI mcs=%d\n", ue->ulsch[eNB_id]->harq_processes[harq_pid]->mcs); - isBad = 1; + LOG_D(PHY,"Not supported MCS in OAI mcs=%d\n", ue->ulsch[eNB_id]->harq_processes[harq_pid]->mcs); + isBad = 1; } if (isBad) { @@ -1846,26 +1714,26 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 // ue->ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag = 0; LOG_D(PHY,"Generating PUSCH (Abssubframe: %d.%d): harq-Id: %d, round: %d, MaxReTrans: %d \n",frame_tx,nr_tti_tx,harq_pid,ue->ulsch[eNB_id]->harq_processes[harq_pid]->round,ue->ulsch[eNB_id]->Mlimit); if (ue->ulsch[eNB_id]->harq_processes[harq_pid]->round >= (ue->ulsch[eNB_id]->Mlimit - 1)) - { + { LOG_D(PHY,"PUSCH MAX Retransmission achieved ==> send last pusch\n"); ue->ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag = 0; ue->ulsch[eNB_id]->harq_processes[harq_pid]->round = 0; - } + } ack_status_cw0 = nr_reset_ack(&ue->frame_parms, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &Nbundled, - 0); + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + ue->ulsch[eNB_id]->o_ACK, + &Nbundled, + 0); ack_status_cw1 = nr_reset_ack(&ue->frame_parms, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][1]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &NbundledCw1, - 1); + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][1]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + ue->ulsch[eNB_id]->o_ACK, + &NbundledCw1, + 1); //Nbundled = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack; //ue->ulsch[eNB_id]->bundling = Nbundled; @@ -1876,10 +1744,10 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 // check Periodic CQI/RI reporting cqi_status = ((ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.cqi_PMI_ConfigIndex>0)&& - (is_cqi_TXOp(ue,proc,eNB_id)==1)); + (is_cqi_TXOp(ue,proc,eNB_id)==1)); ri_status = ((ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.ri_ConfigIndex>0) && - (is_ri_TXOp(ue,proc,eNB_id)==1)); + (is_ri_TXOp(ue,proc,eNB_id)==1)); // compute CQI/RI resources compute_cqi_ri_resources(ue, ue->ulsch[eNB_id], eNB_id, ue->ulsch[eNB_id]->rnti, P_RNTI, CBA_RNTI, cqi_status, ri_status); @@ -1889,44 +1757,44 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 // check if we received a PDSCH at nr_tti_tx - 4 // ==> send ACK/NACK on PUSCH if (ue->frame_parms.frame_type == FDD) - { - ue->ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK = ack_status_cw0 + ack_status_cw1; - } + { + ue->ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK = ack_status_cw0 + ack_status_cw1; + } #if T_TRACER - if(ue->ulsch[eNB_id]->o_ACK[0]) - { - LOG_I(PHY,"PUSCH ACK\n"); - T(T_UE_PHY_DLSCH_UE_ACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx), T_INT(Mod_id), T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti), - T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); - } - else - { - LOG_I(PHY,"PUSCH NACK\n"); - T(T_UE_PHY_DLSCH_UE_NACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx), T_INT(Mod_id), T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti), - T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); - } + if(ue->ulsch[eNB_id]->o_ACK[0]) + { + LOG_I(PHY,"PUSCH ACK\n"); + T(T_UE_PHY_DLSCH_UE_ACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx), T_INT(Mod_id), T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti), + T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); + } + else + { + LOG_I(PHY,"PUSCH NACK\n"); + T(T_UE_PHY_DLSCH_UE_NACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx), T_INT(Mod_id), T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti), + T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); + } #endif #ifdef UE_DEBUG_TRACE LOG_I(PHY,"[UE %d][PDSCH %x] AbsSubFrame %d.%d Generating ACK (%d,%d) for %d bits on PUSCH\n", - Mod_id, - ue->ulsch[eNB_id]->rnti, - frame_tx%1024,nr_tti_tx, - ue->ulsch[eNB_id]->o_ACK[0],ue->ulsch[eNB_id]->o_ACK[1], - ue->ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK); + Mod_id, + ue->ulsch[eNB_id]->rnti, + frame_tx%1024,nr_tti_tx, + ue->ulsch[eNB_id]->o_ACK[0],ue->ulsch[eNB_id]->o_ACK[1], + ue->ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK); #endif } //#ifdef UE_DEBUG_TRACE - LOG_I(PHY, - "[UE %d][PUSCH %d] AbsSubframe %d.%d Generating PUSCH : first_rb %d, nb_rb %d, round %d, mcs %d, tbs %d, rv %d, " - "cyclic_shift %d (cyclic_shift_common %d,n_DMRS2 %d,n_PRS %d), ACK (%d,%d), O_ACK %d, ack_status_cw0 %d ack_status_cw1 %d bundling %d, Nbundled %d, CQI %d, RI %d\n", + LOG_I(PHY, + "[UE %d][PUSCH %d] AbsSubframe %d.%d Generating PUSCH : first_rb %d, nb_rb %d, round %d, mcs %d, tbs %d, rv %d, " + "cyclic_shift %d (cyclic_shift_common %d,n_DMRS2 %d,n_PRS %d), ACK (%d,%d), O_ACK %d, ack_status_cw0 %d ack_status_cw1 %d bundling %d, Nbundled %d, CQI %d, RI %d\n", Mod_id,harq_pid,frame_tx%1024,nr_tti_tx, first_rb,nb_rb, ue->ulsch[eNB_id]->harq_processes[harq_pid]->round, ue->ulsch[eNB_id]->harq_processes[harq_pid]->mcs, - ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS, + ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS, ue->ulsch[eNB_id]->harq_processes[harq_pid]->rvidx, (ue->frame_parms.pusch_config_common.ul_ReferenceSignalsPUSCH.cyclicShift+ ue->ulsch[eNB_id]->harq_processes[harq_pid]->n_DMRS2+ @@ -1941,7 +1809,7 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 ue->ulsch[eNB_id]->bundling, Nbundled, cqi_status, ri_status); - //#endif + //#endif @@ -1996,149 +1864,149 @@ void ue_ulsch_uespec_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8 stop_meas(&ue->ulsch_encoding_stats); #endif if (ue->mac_enabled == 1) { - // signal MAC that Msg3 was sent - //mac_xface->Msg3_transmitted(Mod_id, - CC_id, - frame_tx, - eNB_id); - } - } // Msg3_flag==1 - else { - input_buffer_length = ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS/8; + // signal MAC that Msg3 was sent + //mac_xface->Msg3_transmitted(Mod_id, + CC_id, + frame_tx, + eNB_id); + } + } // Msg3_flag==1 + else { + input_buffer_length = ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS/8; - if (ue->mac_enabled==1) { - // LOG_D(PHY,"[UE %d] ULSCH : Searching for MAC SDUs\n",Mod_id); - if (ue->ulsch[eNB_id]->harq_processes[harq_pid]->round==0) { - //if (ue->ulsch[eNB_id]->harq_processes[harq_pid]->calibration_flag == 0) { - access_mode=SCHEDULED_ACCESS; - //mac_xface->ue_get_sdu(Mod_id, + if (ue->mac_enabled==1) { + // LOG_D(PHY,"[UE %d] ULSCH : Searching for MAC SDUs\n",Mod_id); + if (ue->ulsch[eNB_id]->harq_processes[harq_pid]->round==0) { + //if (ue->ulsch[eNB_id]->harq_processes[harq_pid]->calibration_flag == 0) { + access_mode=SCHEDULED_ACCESS; + //mac_xface->ue_get_sdu(Mod_id, CC_id, - frame_tx, - proc->subframe_tx, - nr_tti_tx%(ue->frame_parms.ttis_per_subframe), - eNB_id, - ulsch_input_buffer, - input_buffer_length, - &access_mode); - } + frame_tx, + proc->subframe_tx, + nr_tti_tx%(ue->frame_parms.ttis_per_subframe), + eNB_id, + ulsch_input_buffer, + input_buffer_length, + &access_mode); + } #ifdef DEBUG_PHY_PROC #ifdef DEBUG_ULSCH - LOG_D(PHY,"[UE] Frame %d, nr_tti_rx %d : ULSCH SDU (TX harq_pid %d) (%d bytes) : \n",frame_tx,nr_tti_tx,harq_pid, ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS>>3); + LOG_D(PHY,"[UE] Frame %d, nr_tti_rx %d : ULSCH SDU (TX harq_pid %d) (%d bytes) : \n",frame_tx,nr_tti_tx,harq_pid, ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS>>3); - for (i=0; iulsch[eNB_id]->harq_processes[harq_pid]->TBS>>3; i++) - LOG_T(PHY,"%x.",ulsch_input_buffer[i]); + for (i=0; iulsch[eNB_id]->harq_processes[harq_pid]->TBS>>3; i++) + LOG_T(PHY,"%x.",ulsch_input_buffer[i]); - LOG_T(PHY,"\n"); + LOG_T(PHY,"\n"); #endif #endif - } - else { - unsigned int taus(void); + } + else { + unsigned int taus(void); - for (i=0; iulsch_encoding_stats); + start_meas(&ue->ulsch_encoding_stats); #endif - if (abstraction_flag==0) { + if (abstraction_flag==0) { - if (ulsch_encoding(ulsch_input_buffer, - ue, - harq_pid, - eNB_id, - proc->nr_tti_rx, - ue->transmission_mode[eNB_id],0, - Nbundled)!=0) { - LOG_E(PHY,"ulsch_coding.c: FATAL ERROR: returning\n"); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX, VCD_FUNCTION_OUT); + if (ulsch_encoding(ulsch_input_buffer, + ue, + harq_pid, + eNB_id, + proc->nr_tti_rx, + ue->transmission_mode[eNB_id],0, + Nbundled)!=0) { + LOG_E(PHY,"ulsch_coding.c: FATAL ERROR: returning\n"); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX, VCD_FUNCTION_OUT); #if UE_TIMING_TRACE stop_meas(&ue->phy_proc_tx); #endif - return; - } - } + return; + } + } #ifdef PHY_ABSTRACTION - else { - ulsch_encoding_emul(ulsch_input_buffer,ue,eNB_id,proc->nr_tti_rx,harq_pid,0); - } + else { + ulsch_encoding_emul(ulsch_input_buffer,ue,eNB_id,proc->nr_tti_rx,harq_pid,0); + } #endif #if UE_TIMING_TRACE - stop_meas(&ue->ulsch_encoding_stats); + stop_meas(&ue->ulsch_encoding_stats); #endif - } +} - if (abstraction_flag == 0) { - if (ue->mac_enabled==1) { - nr_pusch_power_cntl(ue,proc,eNB_id,1, abstraction_flag); - ue->tx_power_dBm[nr_tti_tx] = ue->ulsch[eNB_id]->Po_PUSCH; - } - else { - ue->tx_power_dBm[nr_tti_tx] = ue->tx_power_max_dBm; - } - ue->tx_total_RE[nr_tti_tx] = nb_rb*12; +if (abstraction_flag == 0) { + if (ue->mac_enabled==1) { + nr_pusch_power_cntl(ue,proc,eNB_id,1, abstraction_flag); + ue->tx_power_dBm[nr_tti_tx] = ue->ulsch[eNB_id]->Po_PUSCH; + } + else { + ue->tx_power_dBm[nr_tti_tx] = ue->tx_power_max_dBm; + } + ue->tx_total_RE[nr_tti_tx] = nb_rb*12; #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) - tx_amp = get_tx_amp(ue->tx_power_dBm[nr_tti_tx], - ue->tx_power_max_dBm, - ue->frame_parms.N_RB_UL, - nb_rb); + tx_amp = get_tx_amp(ue->tx_power_dBm[nr_tti_tx], + ue->tx_power_max_dBm, + ue->frame_parms.N_RB_UL, + nb_rb); #else - tx_amp = AMP; + tx_amp = AMP; #endif #if T_TRACER - T(T_UE_PHY_PUSCH_TX_POWER, T_INT(eNB_id),T_INT(Mod_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx),T_INT(ue->tx_power_dBm[nr_tti_tx]), - T_INT(tx_amp),T_INT(ue->ulsch[eNB_id]->f_pusch),T_INT(get_PL(Mod_id,0,eNB_id)),T_INT(nb_rb)); + T(T_UE_PHY_PUSCH_TX_POWER, T_INT(eNB_id),T_INT(Mod_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx),T_INT(ue->tx_power_dBm[nr_tti_tx]), + T_INT(tx_amp),T_INT(ue->ulsch[eNB_id]->f_pusch),T_INT(get_PL(Mod_id,0,eNB_id)),T_INT(nb_rb)); #endif #ifdef UE_DEBUG_TRACE - LOG_I(PHY,"[UE %d][PUSCH %d] AbsSubFrame %d.%d, generating PUSCH, Po_PUSCH: %d dBm (max %d dBm), amp %d\n", - Mod_id,harq_pid,frame_tx%1024,nr_tti_tx,ue->tx_power_dBm[nr_tti_tx],ue->tx_power_max_dBm, tx_amp); + LOG_I(PHY,"[UE %d][PUSCH %d] AbsSubFrame %d.%d, generating PUSCH, Po_PUSCH: %d dBm (max %d dBm), amp %d\n", + Mod_id,harq_pid,frame_tx%1024,nr_tti_tx,ue->tx_power_dBm[nr_tti_tx],ue->tx_power_max_dBm, tx_amp); #endif - if (tx_amp>100) - tx_amp =100; + if (tx_amp>100) + tx_amp =100; - //LOG_I(PHY,"[UE %d][PUSCH %d] after AbsSubFrame %d.%d, generating PUSCH, Po_PUSCH: %d dBm (max %d dBm), amp %d\n", - // Mod_id,harq_pid,frame_tx%1024,nr_tti_tx,ue->tx_power_dBm[nr_tti_tx],ue->tx_power_max_dBm, tx_amp); + //LOG_I(PHY,"[UE %d][PUSCH %d] after AbsSubFrame %d.%d, generating PUSCH, Po_PUSCH: %d dBm (max %d dBm), amp %d\n", + // Mod_id,harq_pid,frame_tx%1024,nr_tti_tx,ue->tx_power_dBm[nr_tti_tx],ue->tx_power_max_dBm, tx_amp); #if UE_TIMING_TRACE - start_meas(&ue->ulsch_modulation_stats); + start_meas(&ue->ulsch_modulation_stats); #endif - ulsch_modulation(ue->common_vars.txdataF, - tx_amp, - frame_tx, - nr_tti_tx, - &ue->frame_parms, - ue->ulsch[eNB_id]); - for (aa=0; aa<1/*frame_parms->nb_antennas_tx*/; aa++) - generate_drs_pusch(ue, - proc, - eNB_id, - tx_amp, - nr_tti_tx, - first_rb, - nb_rb, - aa); + ulsch_modulation(ue->common_vars.txdataF, + tx_amp, + frame_tx, + nr_tti_tx, + &ue->frame_parms, + ue->ulsch[eNB_id]); + for (aa=0; aa<1/*frame_parms->nb_antennas_tx*/; aa++) + generate_drs_pusch(ue, + proc, + eNB_id, + tx_amp, + nr_tti_tx, + first_rb, + nb_rb, + aa); #if UE_TIMING_TRACE - stop_meas(&ue->ulsch_modulation_stats); + stop_meas(&ue->ulsch_modulation_stats); #endif - } + } - if (abstraction_flag==1) { - // clear SR - ue->sr[nr_tti_tx]=0; - } - } // subframe_scheduling_flag==1 +if (abstraction_flag==1) { + // clear SR + ue->sr[nr_tti_tx]=0; + } +} // subframe_scheduling_flag==1 - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_ULSCH_UESPEC,VCD_FUNCTION_OUT); +VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_ULSCH_UESPEC,VCD_FUNCTION_OUT); #endif @@ -2160,47 +2028,47 @@ void ue_srs_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id uint8_t isSrsTxOccasion = pSoundingrs_ul_config_dedicated->srsUeSubframe; if(isSrsTxOccasion) - { - ue->generate_ul_signal[eNB_id] = 1; - if (ue->mac_enabled==1) { - srs_power_cntl(ue,proc,eNB_id, (uint8_t*)(&nb_rb_srs), abstraction_flag); - Po_SRS = ue->ulsch[eNB_id]->Po_SRS; - } - else - { - Po_SRS = ue->tx_power_max_dBm; - } + ue->generate_ul_signal[eNB_id] = 1; + if (ue->mac_enabled==1) + { + srs_power_cntl(ue,proc,eNB_id, (uint8_t*)(&nb_rb_srs), abstraction_flag); + Po_SRS = ue->ulsch[eNB_id]->Po_SRS; + } + else + { + Po_SRS = ue->tx_power_max_dBm; + } #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) - if (ue->mac_enabled==1) - { - tx_amp = get_tx_amp(Po_SRS, - ue->tx_power_max_dBm, - ue->frame_parms.N_RB_UL, - nb_rb_srs); - } - else - { - tx_amp = AMP; - } + if (ue->mac_enabled==1) + { + tx_amp = get_tx_amp(Po_SRS, + ue->tx_power_max_dBm, + ue->frame_parms.N_RB_UL, + nb_rb_srs); + } + else + { + tx_amp = AMP; + } #else tx_amp = AMP; #endif - LOG_D(PHY,"SRS PROC; TX_MAX_POWER %d, Po_SRS %d, NB_RB_UL %d, NB_RB_SRS %d TX_AMPL %d\n",ue->tx_power_max_dBm, + LOG_D(PHY,"SRS PROC; TX_MAX_POWER %d, Po_SRS %d, NB_RB_UL %d, NB_RB_SRS %d TX_AMPL %d\n",ue->tx_power_max_dBm, Po_SRS, ue->frame_parms.N_RB_UL, nb_rb_srs, tx_amp); - uint16_t nsymb = (ue->frame_parms.Ncp==0) ? 14:12; - uint16_t symbol_offset = (int)ue->frame_parms.ofdm_symbol_size*((nr_tti_tx*nsymb)+(nsymb-1)); - generate_srs(&ue->frame_parms, - &ue->soundingrs_ul_config_dedicated[eNB_id], - &ue->common_vars.txdataF[eNB_id][symbol_offset], - tx_amp, - nr_tti_tx); - } + uint16_t nsymb = (ue->frame_parms.Ncp==0) ? 14:12; + uint16_t symbol_offset = (int)ue->frame_parms.ofdm_symbol_size*((nr_tti_tx*nsymb)+(nsymb-1)); + generate_srs(&ue->frame_parms, + &ue->soundingrs_ul_config_dedicated[eNB_id], + &ue->common_vars.txdataF[eNB_id][symbol_offset], + tx_amp, + nr_tti_tx); + } } int16_t get_pucch2_cqi(PHY_VARS_NR_UE *ue,int eNB_id,int *len) { @@ -2208,16 +2076,16 @@ int16_t get_pucch2_cqi(PHY_VARS_NR_UE *ue,int eNB_id,int *len) { if ((ue->transmission_mode[eNB_id]<4)|| (ue->transmission_mode[eNB_id]==7)) { // Mode 1-0 feedback // 4-bit CQI message - /*LOG_I(PHY,"compute CQI value, TM %d, length 4, Cqi Avg %d, value %d \n", ue->transmission_mode[eNB_id], - ue->measurements.wideband_cqi_avg[eNB_id], - sinr2cqi((double)ue->measurements.wideband_cqi_avg[eNB_id], - ue->transmission_mode[eNB_id]));*/ + /*LOG_I(PHY,"compute CQI value, TM %d, length 4, Cqi Avg %d, value %d \n", ue->transmission_mode[eNB_id], + ue->measurements.wideband_cqi_avg[eNB_id], + sinr2cqi((double)ue->measurements.wideband_cqi_avg[eNB_id], + ue->transmission_mode[eNB_id]));*/ *len=4; return(sinr2cqi((double)ue->measurements.wideband_cqi_avg[eNB_id], - ue->transmission_mode[eNB_id])); + ue->transmission_mode[eNB_id])); } else { // Mode 1-1 feedback, later - //LOG_I(PHY,"compute CQI value, TM %d, length 0, Cqi Avg 0 \n", ue->transmission_mode[eNB_id]); + //LOG_I(PHY,"compute CQI value, TM %d, length 0, Cqi Avg 0 \n", ue->transmission_mode[eNB_id]); *len=0; // 2-antenna ports RI=1, 6 bits (2 PMI, 4 CQI) @@ -2245,54 +2113,54 @@ void get_pucch_param(PHY_VARS_NR_UE *ue, uint16_t *plength) { - switch (format) { - case pucch_format1: + switch (format) { + case pucch_format1: { - pucch_resource[0] = ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex; - pucch_payload[0] = 0; // payload is ignored in case of format1 - pucch_payload[1] = 0; // payload is ignored in case of format1 + pucch_resource[0] = ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex; + pucch_payload[0] = 0; // payload is ignored in case of format1 + pucch_payload[1] = 0; // payload is ignored in case of format1 } break; - case pucch_format1a: - case pucch_format1b: + case pucch_format1a: + case pucch_format1b: { - pucch_resource[0] = nr_get_n1_pucch(ue, - proc, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, - eNB_id, - ack_payload, - SR); - pucch_payload[0] = ack_payload[0]; - pucch_payload[1] = ack_payload[1]; - //pucch_payload[1] = 1; + pucch_resource[0] = nr_get_n1_pucch(ue, + proc, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, + eNB_id, + ack_payload, + SR); + pucch_payload[0] = ack_payload[0]; + pucch_payload[1] = ack_payload[1]; + //pucch_payload[1] = 1; } break; - case pucch_format2: + case pucch_format2: { - pucch_resource[0] = ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.cqi_PUCCH_ResourceIndex; - if(cqi_report) + pucch_resource[0] = ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.cqi_PUCCH_ResourceIndex; + if(cqi_report) { - pucch_payload[0] = get_pucch2_cqi(ue,eNB_id,(int*)plength); + pucch_payload[0] = get_pucch2_cqi(ue,eNB_id,(int*)plength); } - else + else { - *plength = 1; - pucch_payload[0] = get_pucch2_ri(ue,eNB_id); + *plength = 1; + pucch_payload[0] = get_pucch2_ri(ue,eNB_id); } } break; - case pucch_format2a: - case pucch_format2b: - LOG_E(PHY,"NO Resource available for PUCCH 2a/2b \n"); + case pucch_format2a: + case pucch_format2b: + LOG_E(PHY,"NO Resource available for PUCCH 2a/2b \n"); break; - case pucch_format3: - fprintf(stderr, "PUCCH format 3 not handled\n"); - abort(); - } + case pucch_format3: + fprintf(stderr, "PUCCH format 3 not handled\n"); + abort(); + } } #ifdef NR_PUCCH_SCHED @@ -2336,15 +2204,15 @@ void ue_pucch_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_ */ int harq_pid = nr_subframe2harq_pid(&ue->frame_parms, - frame_tx, - nr_tti_tx); + frame_tx, + nr_tti_tx); if(ue->ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag) - { + { LOG_D(PHY,"PUSCH is programmed on this nr_tti_rx [pid %d] AbsSuframe %d.%d ==> Skip PUCCH transmission \n",harq_pid,frame_tx,nr_tti_tx); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_PUCCH,VCD_FUNCTION_OUT); return; - } + } uint8_t isShortenPucch = (pSoundingrs_ul_config_dedicated->srsCellSubframe && frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission); @@ -2369,50 +2237,50 @@ void ue_pucch_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_ SR_payload = 0; if (nr_is_SR_TXOp(ue,proc,eNB_id)==1) - { + { if (ue->mac_enabled==1) { - SR_payload = mac_xface->ue_get_SR(Mod_id, - CC_id, - frame_tx, - eNB_id, - ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->crnti, - nr_tti_tx); // nr_tti_rx used for meas gap + SR_payload = mac_xface->ue_get_SR(Mod_id, + CC_id, + frame_tx, + eNB_id, + ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->crnti, + nr_tti_tx); // nr_tti_rx used for meas gap } else { - SR_payload = 1; + SR_payload = 1; } - } + } ack_status_cw0 = nr_get_ack(&ue->frame_parms, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - pucch_ack_payload, - 0); + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + pucch_ack_payload, + 0); ack_status_cw1 = nr_get_ack(&ue->frame_parms, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][1]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - pucch_ack_payload, - 1); + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][1]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + pucch_ack_payload, + 1); nb_cw = ( (ack_status_cw0 != 0) ? 1:0) + ( (ack_status_cw1 != 0) ? 1:0); cqi_status = ((ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.cqi_PMI_ConfigIndex>0)&& - (is_cqi_TXOp(ue,proc,eNB_id)==1)); + (is_cqi_TXOp(ue,proc,eNB_id)==1)); ri_status = ((ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.ri_ConfigIndex>0) && - (is_ri_TXOp(ue,proc,eNB_id)==1)); + (is_ri_TXOp(ue,proc,eNB_id)==1)); // Part - II // if nothing to report ==> exit function if( (nb_cw==0) && (SR_payload==0) && (cqi_status==0) && (ri_status==0) ) - { + { LOG_D(PHY,"PUCCH No feedback AbsSubframe %d.%d SR_payload %d nb_cw %d pucch_ack_payload[0] %d pucch_ack_payload[1] %d cqi_status %d Return \n", frame_tx%1024, nr_tti_tx, SR_payload, nb_cw, pucch_ack_payload[0], pucch_ack_payload[1], cqi_status); return; - } + } // Part - III // Decide which PUCCH format should be used if needed @@ -2447,368 +2315,181 @@ void ue_pucch_procedures(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_ case pucch_format1: case pucch_format1a: case pucch_format1b: - { + { if (ue->mac_enabled == 1) { - Po_PUCCH = nr_pucch_power_cntl(ue,proc,nr_tti_tx,eNB_id,format); + Po_PUCCH = nr_pucch_power_cntl(ue,proc,nr_tti_tx,eNB_id,format); } else { - Po_PUCCH = ue->tx_power_max_dBm; + Po_PUCCH = ue->tx_power_max_dBm; } ue->tx_power_dBm[nr_tti_tx] = Po_PUCCH; ue->tx_total_RE[nr_tti_tx] = 12; #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) tx_amp = get_tx_amp(Po_PUCCH, - ue->tx_power_max_dBm, - ue->frame_parms.N_RB_UL, - 1); + ue->tx_power_max_dBm, + ue->frame_parms.N_RB_UL, + 1); #else tx_amp = AMP; #endif #if T_TRACER T(T_UE_PHY_PUCCH_TX_POWER, T_INT(eNB_id),T_INT(Mod_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx),T_INT(ue->tx_power_dBm[nr_tti_tx]), - T_INT(tx_amp),T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id))); + T_INT(tx_amp),T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id))); #endif #ifdef UE_DEBUG_TRACE if(format == pucch_format1) - { + { LOG_I(PHY,"[UE %d][SR %x] AbsSubframe %d.%d Generating PUCCH 1 (SR for PUSCH), an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, Po_PUCCH %d\n", - Mod_id, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, - frame_tx%1024, nr_tti_tx, - frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, - isShortenPucch, - ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex, - Po_PUCCH); - } + Mod_id, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, + frame_tx%1024, nr_tti_tx, + frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, + isShortenPucch, + ue->scheduling_request_config[eNB_id].sr_PUCCH_ResourceIndex, + Po_PUCCH); + } else - { + { if (SR_payload>0) { - LOG_I(PHY,"[UE %d][SR %x] AbsSubFrame %d.%d Generating PUCCH %s payload %d,%d (with SR for PUSCH), an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, Po_PUCCH %d, amp %d\n", - Mod_id, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, - frame_tx % 1024, nr_tti_tx, - (format == pucch_format1a? "1a": ( - format == pucch_format1b? "1b" : "??")), - pucch_ack_payload[0],pucch_ack_payload[1], - frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, - isShortenPucch, - pucch_resource, - Po_PUCCH, - tx_amp); + LOG_I(PHY,"[UE %d][SR %x] AbsSubFrame %d.%d Generating PUCCH %s payload %d,%d (with SR for PUSCH), an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, Po_PUCCH %d, amp %d\n", + Mod_id, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, + frame_tx % 1024, nr_tti_tx, + (format == pucch_format1a? "1a": ( + format == pucch_format1b? "1b" : "??")), + pucch_ack_payload[0],pucch_ack_payload[1], + frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, + isShortenPucch, + pucch_resource, + Po_PUCCH, + tx_amp); } else { - LOG_I(PHY,"[UE %d][PDSCH %x] AbsSubFrame %d.%d rx_offset_diff: %d, Generating PUCCH %s, an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, b[0]=%d,b[1]=%d (SR_Payload %d), Po_PUCCH %d, amp %d\n", - Mod_id, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, - frame_tx%1024, nr_tti_tx,ue->rx_offset_diff, - (format == pucch_format1a? "1a": ( - format == pucch_format1b? "1b" : "??")), - frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, - isShortenPucch, - pucch_resource,pucch_payload[0],pucch_payload[1],SR_payload, - Po_PUCCH, - tx_amp); + LOG_I(PHY,"[UE %d][PDSCH %x] AbsSubFrame %d.%d rx_offset_diff: %d, Generating PUCCH %s, an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, b[0]=%d,b[1]=%d (SR_Payload %d), Po_PUCCH %d, amp %d\n", + Mod_id, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, + frame_tx%1024, nr_tti_tx,ue->rx_offset_diff, + (format == pucch_format1a? "1a": ( + format == pucch_format1b? "1b" : "??")), + frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission, + isShortenPucch, + pucch_resource,pucch_payload[0],pucch_payload[1],SR_payload, + Po_PUCCH, + tx_amp); } - } + } #endif #if T_TRACER if(pucch_payload[0]) - { + { T(T_UE_PHY_DLSCH_UE_ACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx), T_INT(Mod_id), T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti), - T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); - } + T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); + } else - { + { T(T_UE_PHY_DLSCH_UE_NACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx), T_INT(Mod_id), T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti), - T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); - } + T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid)); + } #endif if (abstraction_flag == 0) { - generate_pucch1x(ue->common_vars.txdataF, - &ue->frame_parms, - ue->ncs_cell, - format, - &ue->pucch_config_dedicated[eNB_id], - pucch_resource, - isShortenPucch, // shortened format - pucch_payload, - tx_amp, - nr_tti_tx); + generate_pucch1x(ue->common_vars.txdataF, + &ue->frame_parms, + ue->ncs_cell, + format, + &ue->pucch_config_dedicated[eNB_id], + pucch_resource, + isShortenPucch, // shortened format + pucch_payload, + tx_amp, + nr_tti_tx); } else { #ifdef PHY_ABSTRACTION - LOG_D(PHY,"Calling generate_pucch_emul ... (ACK %d %d, SR %d)\n",pucch_ack_payload[0],pucch_ack_payload[1],SR_payload); - generate_pucch_emul(ue, - proc, - format, - ue->frame_parms.pucch_config_common.nCS_AN, - pucch_payload, - SR_payload); + LOG_D(PHY,"Calling generate_pucch_emul ... (ACK %d %d, SR %d)\n",pucch_ack_payload[0],pucch_ack_payload[1],SR_payload); + generate_pucch_emul(ue, + proc, + format, + ue->frame_parms.pucch_config_common.nCS_AN, + pucch_payload, + SR_payload); #endif } - } - break; + } + break; case pucch_format2: - { + { if (ue->mac_enabled == 1) { - Po_PUCCH = nr_pucch_power_cntl(ue,proc,nr_tti_tx,eNB_id,format); + Po_PUCCH = nr_pucch_power_cntl(ue,proc,nr_tti_tx,eNB_id,format); } else { - Po_PUCCH = ue->tx_power_max_dBm; + Po_PUCCH = ue->tx_power_max_dBm; } ue->tx_power_dBm[nr_tti_tx] = Po_PUCCH; ue->tx_total_RE[nr_tti_tx] = 12; #if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) || defined(OAI_ADRV9371_ZC706) tx_amp = get_tx_amp(Po_PUCCH, - ue->tx_power_max_dBm, - ue->frame_parms.N_RB_UL, - 1); + ue->tx_power_max_dBm, + ue->frame_parms.N_RB_UL, + 1); #else tx_amp = AMP; #endif #if T_TRACER T(T_UE_PHY_PUCCH_TX_POWER, T_INT(eNB_id),T_INT(Mod_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx),T_INT(ue->tx_power_dBm[nr_tti_tx]), - T_INT(tx_amp),T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id))); + T_INT(tx_amp),T_INT(ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id))); #endif #ifdef UE_DEBUG_TRACE LOG_I(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2 (RI or CQI), Po_PUCCH %d, isShortenPucch %d, amp %d\n", - Mod_id, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, - frame_tx%1024, nr_tti_tx, - Po_PUCCH, - isShortenPucch, - tx_amp); + Mod_id, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, + frame_tx%1024, nr_tti_tx, + Po_PUCCH, + isShortenPucch, + tx_amp); #endif generate_pucch2x(ue->common_vars.txdataF, - &ue->frame_parms, - ue->ncs_cell, - format, - &ue->pucch_config_dedicated[eNB_id], - pucch_resource, - pucch_payload, - len, // A - 0, // B2 not needed - tx_amp, - nr_tti_tx, - ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->crnti); - } - break; + &ue->frame_parms, + ue->ncs_cell, + format, + &ue->pucch_config_dedicated[eNB_id], + pucch_resource, + pucch_payload, + len, // A + 0, // B2 not needed + tx_amp, + nr_tti_tx, + ue->pdcch_vars[ue->current_thread_id[proc->nr_tti_rx]][eNB_id]->crnti); + } + break; case pucch_format2a: - LOG_D(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2a (RI or CQI) Ack/Nack 1bit \n", - Mod_id, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, - frame_tx%1024, nr_tti_tx); - break; + LOG_D(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2a (RI or CQI) Ack/Nack 1bit \n", + Mod_id, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, + frame_tx%1024, nr_tti_tx); + break; case pucch_format2b: - LOG_D(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2b (RI or CQI) Ack/Nack 2bits\n", - Mod_id, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, - frame_tx%1024, nr_tti_tx); - break; + LOG_D(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2b (RI or CQI) Ack/Nack 2bits\n", + Mod_id, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->rnti, + frame_tx%1024, nr_tti_tx); + break; default: - break; + break; } VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_PUCCH,VCD_FUNCTION_OUT); } -#endif //0 - -void phy_procedures_nrUE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,runmode_t mode,relaying_type_t r_type) { - NR_DL_FRAME_PARMS *frame_parms=&ue->frame_parms; - //int32_t ulsch_start=0; - int nr_tti_tx = proc->nr_tti_tx; - int frame_tx = proc->frame_tx; - unsigned int aa; - uint8_t isSubframeSRS; - - uint8_t next1_thread_id = ue->current_thread_id[proc->nr_tti_rx]== (RX_NB_TH-1) ? 0:(ue->current_thread_id[proc->nr_tti_rx]+1); - uint8_t next2_thread_id = next1_thread_id== (RX_NB_TH-1) ? 0:(next1_thread_id+1); - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX,VCD_FUNCTION_IN); - - LOG_D(PHY,"****** start TX-Chain for AbsSubframe %d.%d ******\n", frame_tx, nr_tti_tx); -#if T_TRACER - T(T_UE_PHY_UL_TICK, T_INT(ue->Mod_id), T_INT(frame_tx%1024), T_INT(nr_tti_tx)); -#endif - - ue->generate_ul_signal[eNB_id] = 0; -#if UE_TIMING_TRACE - start_meas(&ue->phy_proc_tx); -#endif - -#ifdef EMOS - //phy_procedures_emos_UE_TX(next_slot); -#endif - - ue->tx_power_dBm[nr_tti_tx]=-127; - - if (abstraction_flag==0) { - for (aa=0; aanb_antennas_tx; aa++) { - memset(&ue->common_vars.txdataF[aa][nr_tti_tx*frame_parms->ofdm_symbol_size*frame_parms->symbols_per_tti], - 0, - frame_parms->ofdm_symbol_size*frame_parms->symbols_per_tti*sizeof(int32_t)); - } - } - - if (ue->UE_mode[eNB_id] != PRACH) { - // check cell srs nr_tti_rx and ue srs nr_tti_rx. This has an impact on pusch encoding - //isSubframeSRS = nr_is_srs_occasion_common(&ue->frame_parms,proc->frame_tx,proc->nr_tti_tx); - - //ue_compute_srs_occasion(ue,proc,eNB_id,isSubframeSRS); - - //ue_ulsch_uespec_procedures(ue,proc,eNB_id,abstraction_flag); - - } - - if (ue->UE_mode[eNB_id] == PUSCH) { - // check if we need to use PUCCH 1a/1b - //ue_pucch_procedures(ue,proc,eNB_id,abstraction_flag); - // check if we need to use SRS - //ue_srs_procedures(ue,proc,eNB_id,abstraction_flag); - } // UE_mode==PUSCH - - -#ifdef CBA - - if ((ue->ulsch[eNB_id]->harq_processes[harq_pid]->subframe_cba_scheduling_flag >= 1) && - (ue->ulsch[eNB_id]->harq_processes[harq_pid]->status == CBA_ACTIVE)) { - ue->ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag=0; //-=1 - // ue->ulsch[eNB_id]->harq_processes[harq_pid]->status= IDLE; - first_rb = ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb; - nb_rb = ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb; - //cba_mcs=ue->ulsch[eNB_id]->harq_processes[harq_pid]->mcs; - input_buffer_length = ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS/8; - access_mode=CBA_ACCESS; - - LOG_D(PHY,"[UE %d] Frame %d, nr_tti_rx %d: CBA num dci %d\n", - Mod_id,frame_tx,nr_tti_tx, - ue->ulsch[eNB_id]->num_cba_dci[nr_tti_tx]); - - /*mac_xface->ue_get_sdu(Mod_id, - CC_id, - frame_tx, - proc->subframe_tx, - nr_tti_tx%(ue->frame_parms.ttis_per_subframe), - eNB_id, - ulsch_input_buffer, - input_buffer_length, - &access_mode);*/ - - ue->ulsch[eNB_id]->num_cba_dci[nr_tti_tx]=0; - - if (access_mode > UNKNOWN_ACCESS) { - - if (abstraction_flag==0) { - if (ulsch_encoding(ulsch_input_buffer, - ue, - harq_pid, - eNB_id, - proc->nr_tti_rx, - ue->transmission_mode[eNB_id],0, - 0)!=0) { // Nbundled, to be updated!!!! - LOG_E(PHY,"ulsch_coding.c: FATAL ERROR: returning\n"); - return; - } - } - -#ifdef PHY_ABSTRACTION - else { - ulsch_encoding_emul(ulsch_input_buffer,ue,eNB_id,proc->nr_tti_rx,harq_pid,0); - } - -#endif - } else { - ue->ulsch[eNB_id]->harq_processes[harq_pid]->status= IDLE; - //reset_cba_uci(ue->ulsch[eNB_id]->o); - LOG_N(PHY,"[UE %d] Frame %d, nr_tti_rx %d: CBA transmission cancelled or postponed\n", - Mod_id, frame_tx,nr_tti_tx); - } - } - -#endif // end CBA - - - if (abstraction_flag == 0) { - //ulsch_common_procedures(ue,proc, (ue->generate_ul_signal[eNB_id] == 0)); - } // mode != PRACH - - - if ((ue->UE_mode[eNB_id] == PRACH) && - (ue->frame_parms.prach_config_common.prach_Config_enabled==1)) { - - // check if we have PRACH opportunity - - if (is_nr_prach_subframe(&ue->frame_parms,frame_tx,nr_tti_tx)) { - - ue_nr_prach_procedures(ue,proc,eNB_id,abstraction_flag,mode); - } - } // mode is PRACH - else { - ue->generate_prach=0; - } - - // reset DL ACK/NACK status - uint8_t N_bundled = 0; - uint8_t current_pid = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->current_harq_pid; - if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0] != NULL) - { -/* - nr_reset_ack(&ue->frame_parms, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_processes[current_pid]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); - nr_reset_ack(&ue->frame_parms, - ue->dlsch[next1_thread_id][eNB_id][0]->harq_processes[current_pid]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); - nr_reset_ack(&ue->frame_parms, - ue->dlsch[next2_thread_id][eNB_id][0]->harq_processes[current_pid]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); -*/ - } - - if (ue->dlsch_SI[eNB_id] != NULL) -/* - nr_reset_ack(&ue->frame_parms, - ue->dlsch_SI[eNB_id]->harq_processes[current_pid]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); -*/ - - LOG_D(PHY,"****** end TX-Chain for AbsSubframe %d.%d ******\n", frame_tx, nr_tti_tx); - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX, VCD_FUNCTION_OUT); -#if UE_TIMING_TRACE - stop_meas(&ue->phy_proc_tx); -#endif -} - - -#if 0 void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t abstraction_flag,runmode_t mode,relaying_type_t r_type) { @@ -2843,8 +2524,8 @@ void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB if (abstraction_flag==0) { for (aa=0; aanb_antennas_tx; aa++) { memset(&ue->common_vars.txdataF[aa][nr_tti_tx*frame_parms->ofdm_symbol_size*frame_parms->symbols_per_tti], - 0, - frame_parms->ofdm_symbol_size*frame_parms->symbols_per_tti*sizeof(int32_t)); + 0, + frame_parms->ofdm_symbol_size*frame_parms->symbols_per_tti*sizeof(int32_t)); } } @@ -2859,10 +2540,10 @@ void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB } if (ue->UE_mode[eNB_id] == PUSCH) { - // check if we need to use PUCCH 1a/1b - ue_pucch_procedures(ue,proc,eNB_id,abstraction_flag); - // check if we need to use SRS - ue_srs_procedures(ue,proc,eNB_id,abstraction_flag); + // check if we need to use PUCCH 1a/1b + ue_pucch_procedures(ue,proc,eNB_id,abstraction_flag); + // check if we need to use SRS + ue_srs_procedures(ue,proc,eNB_id,abstraction_flag); } // UE_mode==PUSCH @@ -2879,18 +2560,18 @@ void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB access_mode=CBA_ACCESS; LOG_D(PHY,"[UE %d] Frame %d, nr_tti_rx %d: CBA num dci %d\n", - Mod_id,frame_tx,nr_tti_tx, - ue->ulsch[eNB_id]->num_cba_dci[nr_tti_tx]); + Mod_id,frame_tx,nr_tti_tx, + ue->ulsch[eNB_id]->num_cba_dci[nr_tti_tx]); /*mac_xface->ue_get_sdu(Mod_id, - CC_id, - frame_tx, - proc->subframe_tx, - nr_tti_tx%(ue->frame_parms.ttis_per_subframe), - eNB_id, - ulsch_input_buffer, - input_buffer_length, - &access_mode);*/ + CC_id, + frame_tx, + proc->subframe_tx, + nr_tti_tx%(ue->frame_parms.ttis_per_subframe), + eNB_id, + ulsch_input_buffer, + input_buffer_length, + &access_mode);*/ ue->ulsch[eNB_id]->num_cba_dci[nr_tti_tx]=0; @@ -2919,7 +2600,7 @@ void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB ue->ulsch[eNB_id]->harq_processes[harq_pid]->status= IDLE; //reset_cba_uci(ue->ulsch[eNB_id]->o); LOG_N(PHY,"[UE %d] Frame %d, nr_tti_rx %d: CBA transmission cancelled or postponed\n", - Mod_id, frame_tx,nr_tti_tx); + Mod_id, frame_tx,nr_tti_tx); } } @@ -2948,38 +2629,38 @@ void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB // reset DL ACK/NACK status uint8_t N_bundled = 0; if (ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0] != NULL) - { - nr_reset_ack(&ue->frame_parms, - ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); - nr_reset_ack(&ue->frame_parms, - ue->dlsch[next1_thread_id][eNB_id][0]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); - nr_reset_ack(&ue->frame_parms, - ue->dlsch[next2_thread_id][eNB_id][0]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); - } + { + nr_reset_ack(&ue->frame_parms, + ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][eNB_id][0]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + ue->ulsch[eNB_id]->o_ACK, + &N_bundled, + 0); + nr_reset_ack(&ue->frame_parms, + ue->dlsch[next1_thread_id][eNB_id][0]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + ue->ulsch[eNB_id]->o_ACK, + &N_bundled, + 0); + nr_reset_ack(&ue->frame_parms, + ue->dlsch[next2_thread_id][eNB_id][0]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + ue->ulsch[eNB_id]->o_ACK, + &N_bundled, + 0); + } if (ue->dlsch_SI[eNB_id] != NULL) nr_reset_ack(&ue->frame_parms, - ue->dlsch_SI[eNB_id]->harq_ack, - nr_tti_tx, - proc->nr_tti_rx, - ue->ulsch[eNB_id]->o_ACK, - &N_bundled, - 0); + ue->dlsch_SI[eNB_id]->harq_ack, + nr_tti_tx, + proc->nr_tti_rx, + ue->ulsch[eNB_id]->o_ACK, + &N_bundled, + 0); LOG_D(PHY,"****** end TX-Chain for AbsSubframe %d.%d ******\n", frame_tx, nr_tti_tx); @@ -2989,9 +2670,7 @@ void phy_procedures_UE_TX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB stop_meas(&ue->phy_proc_tx); #endif } -#endif //0 -#if 0 void phy_procedures_UE_S_TX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstraction_flag,relaying_type_t r_type) { int aa;//i,aa; @@ -3004,30 +2683,35 @@ void phy_procedures_UE_S_TX(PHY_VARS_NR_UE *ue,uint8_t eNB_id,uint8_t abstractio int i; // set the whole tx buffer to RX for (i=0; isamples_per_subframe; i++) - ue->common_vars.txdata[aa][i] = 0x00010001; + ue->common_vars.txdata[aa][i] = 0x00010001; #else //this is the normal case memset(&ue->common_vars.txdata[aa][0],0, - (LTE_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe)*sizeof(int32_t)); + (LTE_NUMBER_OF_SUBFRAMES_PER_FRAME*frame_parms->samples_per_subframe)*sizeof(int32_t)); #endif //else EXMIMO } } } -void ue_measurement_procedures( - uint16_t l, // symbol index of each slot [0..6] - PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc, uint8_t eNB_id, - uint16_t slot, // slot index of each radio frame [0..19] - uint8_t abstraction_flag,runmode_t mode) +#endif + +void nr_ue_measurement_procedures( + uint16_t l, // symbol index of each slot [0..6] + PHY_VARS_NR_UE *ue, + UE_nr_rxtx_proc_t *proc, + uint8_t eNB_id, + uint16_t slot, // slot index of each radio frame [0..19] + runmode_t mode) { - //LOG_I(PHY,"ue_measurement_procedures l %d Ncp %d\n",l,ue->frame_parms.Ncp); + LOG_D(PHY,"ue_measurement_procedures l %u Ncp %d\n",l,ue->frame_parms.Ncp); NR_DL_FRAME_PARMS *frame_parms=&ue->frame_parms; int nr_tti_rx = proc->nr_tti_rx; +#if 0 VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_MEASUREMENT_PROCEDURES, VCD_FUNCTION_IN); if (l==0) { @@ -3036,53 +2720,51 @@ void ue_measurement_procedures( LOG_D(PHY,"Calling measurements nr_tti_rx %d, rxdata %p\n",nr_tti_rx,ue->common_vars.rxdata); lte_ue_measurements(ue, - (nr_tti_rx*frame_parms->samples_per_tti+ue->rx_offset)%(frame_parms->samples_per_tti*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME), - (nr_tti_rx == 1) ? 1 : 0, - 0, - 0, - nr_tti_rx); + (nr_tti_rx*frame_parms->samples_per_tti+ue->rx_offset)%(frame_parms->samples_per_tti*LTE_NUMBER_OF_SUBFRAMES_PER_FRAME), + (nr_tti_rx == 1) ? 1 : 0, + 0, + 0, + nr_tti_rx); } else { lte_ue_measurements(ue, - 0, - 0, - 1, - 0, - nr_tti_rx); + 0, + 0, + 1, + 0, + nr_tti_rx); } #if T_TRACER if(slot == 0) T(T_UE_PHY_MEAS, T_INT(eNB_id), T_INT(ue->Mod_id), T_INT(proc->frame_rx%1024), T_INT(proc->nr_tti_rx), - T_INT((int)(10*log10(ue->measurements.rsrp[0])-ue->rx_total_gain_dB)), - T_INT((int)ue->measurements.rx_rssi_dBm[0]), - T_INT((int)(ue->measurements.rx_power_avg_dB[0] - ue->measurements.n0_power_avg_dB)), - T_INT((int)ue->measurements.rx_power_avg_dB[0]), - T_INT((int)ue->measurements.n0_power_avg_dB), - T_INT((int)ue->measurements.wideband_cqi_avg[0]), - T_INT((int)ue->common_vars.freq_offset)); + T_INT((int)(10*log10(ue->measurements.rsrp[0])-ue->rx_total_gain_dB)), + T_INT((int)ue->measurements.rx_rssi_dBm[0]), + T_INT((int)(ue->measurements.rx_power_avg_dB[0] - ue->measurements.n0_power_avg_dB)), + T_INT((int)ue->measurements.rx_power_avg_dB[0]), + T_INT((int)ue->measurements.n0_power_avg_dB), + T_INT((int)ue->measurements.wideband_cqi_avg[0]), + T_INT((int)ue->common_vars.freq_offset)); #endif } if (l==(6-ue->frame_parms.Ncp)) { // make sure we have signal from PSS/SSS for N0 measurement - // LOG_I(PHY," l==(6-ue->frame_parms.Ncp) ue_rrc_measurements\n"); + // LOG_I(PHY," l==(6-ue->frame_parms.Ncp) ue_rrc_measurements\n"); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_RRC_MEASUREMENTS, VCD_FUNCTION_IN); ue_rrc_measurements(ue, - slot, - abstraction_flag); + slot, + abstraction_flag); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_RRC_MEASUREMENTS, VCD_FUNCTION_OUT); - if (abstraction_flag==1) - ue->sinr_eff = sinr_eff_cqi_calc(ue, 0, nr_tti_rx); } - +#endif // accumulate and filter timing offset estimation every nr_tti_rx (instead of every frame) - if (( (slot%2) == 0) && (l==(4-frame_parms->Ncp))) { + if (( slot == 2) && (l==(1-frame_parms->Ncp))) { // AGC - +/* VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GAIN_CONTROL, VCD_FUNCTION_IN); #ifndef OAI_USRP @@ -3096,24 +2778,24 @@ void ue_measurement_procedures( #endif VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_GAIN_CONTROL, VCD_FUNCTION_OUT); - +*/ eNB_id = 0; - if (abstraction_flag == 0) { + LOG_D(PHY,"start adjust sync l = %d slot = %d no timing %d\n",l, slot, ue->no_timing_correction); if (ue->no_timing_correction==0) - lte_adjust_synch(&ue->frame_parms, - ue, - eNB_id, - nr_tti_rx, - 0, - 16384); - } - + nr_adjust_synch_ue(&ue->frame_parms, + ue, + eNB_id, + nr_tti_rx, + 0, + 16384); + } VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_MEASUREMENT_PROCEDURES, VCD_FUNCTION_OUT); } + #ifdef EMOS void phy_procedures_emos_UE_RX(PHY_VARS_NR_UE *ue,uint8_t last_slot,uint8_t eNB_id) { @@ -3139,14 +2821,14 @@ void phy_procedures_emos_UE_RX(PHY_VARS_NR_UE *ue,uint8_t last_slot,uint8_t eNB_ if ((last_slot==10) || (last_slot==11)) { for (i=0; iframe_parms.nb_antennas_rx; i++) for (j=0; jframe_parms.nb_antennas_tx; j++) { - // first OFDM symbol with pilots - memcpy(&emos_dump_UE.channel[i][j][(last_slot%2)*2*ue->frame_parms.ofdm_symbol_size], - &ue->common_vars.dl_ch_estimates[eNB_id][(j<<1) + i][0], - ue->frame_parms.ofdm_symbol_size*sizeof(int)); - // second OFDM symbol with pilots - memcpy(&emos_dump_UE.channel[i][j][((last_slot%2)*2+1)*ue->frame_parms.ofdm_symbol_size], - &ue->common_vars.dl_ch_estimates[eNB_id][(j<<1) + i][(ue->frame_parms.Ncp == 0 ? 4 : 3)*ue->frame_parms.ofdm_symbol_size], - ue->frame_parms.ofdm_symbol_size*sizeof(int)); + // first OFDM symbol with pilots + memcpy(&emos_dump_UE.channel[i][j][(last_slot%2)*2*ue->frame_parms.ofdm_symbol_size], + &ue->common_vars.dl_ch_estimates[eNB_id][(j<<1) + i][0], + ue->frame_parms.ofdm_symbol_size*sizeof(int)); + // second OFDM symbol with pilots + memcpy(&emos_dump_UE.channel[i][j][((last_slot%2)*2+1)*ue->frame_parms.ofdm_symbol_size], + &ue->common_vars.dl_ch_estimates[eNB_id][(j<<1) + i][(ue->frame_parms.Ncp == 0 ? 4 : 3)*ue->frame_parms.ofdm_symbol_size], + ue->frame_parms.ofdm_symbol_size*sizeof(int)); } } @@ -3194,7 +2876,7 @@ void phy_procedures_emos_UE_RX(PHY_VARS_NR_UE *ue,uint8_t last_slot,uint8_t eNB_ LOG_W(PHY,"[UE %d] frame %d, slot %d, Problem writing EMOS data to FIFO\n",Mod_id,proc->frame_rx, last_slot); } else { if (proc->frame_rx%100==0) { - LOG_I(PHY,"[UE %d] frame %d, slot %d, Writing %d bytes EMOS data to FIFO\n",Mod_id,proc->frame_rx, last_slot, bytes); + LOG_I(PHY,"[UE %d] frame %d, slot %d, Writing %d bytes EMOS data to FIFO\n",Mod_id,proc->frame_rx, last_slot, bytes); } } } @@ -3202,7 +2884,7 @@ void phy_procedures_emos_UE_RX(PHY_VARS_NR_UE *ue,uint8_t last_slot,uint8_t eNB_ } #endif - +#if 0 void restart_phy(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc, uint8_t eNB_id,uint8_t abstraction_flag) { @@ -3266,14 +2948,15 @@ void restart_phy(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc, uint8_t eNB_id,uint //ue->bitrate[eNB_id] = 0; //ue->total_received_bits[eNB_id] = 0; } +#endif //(0) - -void ue_pbch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc, uint8_t abstraction_flag) +void nr_ue_pbch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc, uint8_t abstraction_flag) { // int i; - int pbch_tx_ant=0; - uint8_t pbch_phase; + //int pbch_tx_ant=0; + //uint8_t pbch_phase; + int ret = 0; uint16_t frame_tx; static uint8_t first_run = 1; uint8_t pbch_trials = 0; @@ -3285,172 +2968,51 @@ void ue_pbch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *pro VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PBCH_PROCEDURES, VCD_FUNCTION_IN); - pbch_phase=(frame_rx%4); - - if (pbch_phase>=4) - pbch_phase=0; - - for (pbch_trials=0; pbch_trials<4; pbch_trials++) { - //for (pbch_phase=0;pbch_phase<4;pbch_phase++) { - //LOG_I(PHY,"[UE %d] Frame %d, Trying PBCH %d (NidCell %d, eNB_id %d)\n",ue->Mod_id,frame_rx,pbch_phase,ue->frame_parms.Nid_cell,eNB_id); - if (abstraction_flag == 0) { - pbch_tx_ant = rx_pbch(ue, proc, - ue->pbch_vars[eNB_id], - &ue->frame_parms, - eNB_id, - ue->frame_parms.mode1_flag==1?SISO:ALAMOUTI, - ue->high_speed_flag, - pbch_phase); - - - - } - -#ifdef PHY_ABSTRACTION - else { - pbch_tx_ant = rx_pbch_emul(ue, - eNB_id, - pbch_phase); - } - -#endif - - if ((pbch_tx_ant>0) && (pbch_tx_ant<=4)) { - break; - } - - pbch_phase++; - - if (pbch_phase>=4) - pbch_phase=0; - } - - - - if ((pbch_tx_ant>0) && (pbch_tx_ant<=4)) { - - if (opt_enabled) { - static uint8_t dummy[3]; - dummy[0] = ue->pbch_vars[eNB_id]->decoded_output[2]; - dummy[1] = ue->pbch_vars[eNB_id]->decoded_output[1]; - dummy[2] = ue->pbch_vars[eNB_id]->decoded_output[0]; - trace_pdu(1, dummy, 3, ue->Mod_id, 0, 0, - frame_rx, nr_tti_rx, 0, 0); - LOG_D(OPT,"[UE %d][PBCH] Frame %d trace pdu for PBCH\n", - ue->Mod_id, nr_tti_rx); - } - - if (pbch_tx_ant>2) { - LOG_W(PHY,"[openair][SCHED][SYNCH] PBCH decoding: pbch_tx_ant>2 not supported\n"); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PBCH_PROCEDURES, VCD_FUNCTION_OUT); - return; - } - + //LOG_I(PHY,"[UE %d] Frame %d, Trying PBCH %d (NidCell %d, eNB_id %d)\n",ue->Mod_id,frame_rx,pbch_phase,ue->frame_parms.Nid_cell,eNB_id); + ret = nr_rx_pbch(ue, proc, + ue->pbch_vars[eNB_id], + &ue->frame_parms, + eNB_id, + SISO, + ue->high_speed_flag); + if (ret==0) { ue->pbch_vars[eNB_id]->pdu_errors_conseq = 0; - frame_tx = (((int)(ue->pbch_vars[eNB_id]->decoded_output[2]&0x03))<<8); - frame_tx += ((int)(ue->pbch_vars[eNB_id]->decoded_output[1]&0xfc)); - frame_tx += pbch_phase; - //if (ue->mac_enabled==1) { - //mac_xface->dl_phy_sync_success(ue->Mod_id,frame_rx,eNB_id,first_run); - //} - -#ifdef EMOS - //emos_dump_UE.frame_tx = frame_tx; - //emos_dump_UE.mimo_mode = ue->pbch_vars[eNB_id]->decoded_output[1]; -#endif - - if (first_run) { - first_run = 0; - - proc->frame_rx = (proc->frame_rx & 0xFFFFFC00) | (frame_tx & 0x000003FF); - proc->frame_tx = proc->frame_rx; - for(int th_id=0; th_idproc.proc_rxtx[th_id].frame_rx = proc->frame_rx; - ue->proc.proc_rxtx[th_id].frame_tx = proc->frame_tx; - - printf("[UE %d] frame %d, nr_tti_rx %d: Adjusting frame counter (PBCH ant_tx=%d, frame_tx=%d, phase %d, rx_offset %d) => new frame %d\n", - ue->Mod_id, - ue->proc.proc_rxtx[th_id].frame_rx, - nr_tti_rx, - pbch_tx_ant, - frame_tx, - pbch_phase, - ue->rx_offset, - proc->frame_rx); - } - - - frame_rx = proc->frame_rx; - - } else if (((frame_tx & 0x03FF) != (proc->frame_rx & 0x03FF))) { - //(pbch_tx_ant != ue->frame_parms.nb_antennas_tx)) { - LOG_D(PHY,"[UE %d] frame %d, nr_tti_rx %d: Re-adjusting frame counter (PBCH ant_tx=%d, frame_rx=%d, frame%%1024=%d, phase %d).\n", - ue->Mod_id, - proc->frame_rx, - nr_tti_rx, - pbch_tx_ant, - frame_tx, - frame_rx & 0x03FF, - pbch_phase); - - proc->frame_rx = (proc->frame_rx & 0xFFFFFC00) | (frame_tx & 0x000003FF); - proc->frame_tx = proc->frame_rx; - frame_rx = proc->frame_rx; - - for(int th_id=0; th_idproc.proc_rxtx[th_id].frame_rx = (proc->frame_rx & 0xFFFFFC00) | (frame_tx & 0x000003FF); - ue->proc.proc_rxtx[th_id].frame_tx = proc->frame_rx; - } - - } #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d] frame %d, nr_tti_rx %d, Received PBCH (MIB): mode1_flag %d, tx_ant %d, frame_tx %d. N_RB_DL %d, phich_duration %d, phich_resource %d/6!\n", + LOG_D(PHY,"[UE %d] frame %d, nr_tti_rx %d, Received PBCH (MIB): frame_tx %d. N_RB_DL %d\n", ue->Mod_id, frame_rx, nr_tti_rx, - ue->frame_parms.mode1_flag, - pbch_tx_ant, frame_tx, - ue->frame_parms.N_RB_DL, - ue->frame_parms.phich_config_common.phich_duration, - ue->frame_parms.phich_config_common.phich_resource); + ue->frame_parms.N_RB_DL); #endif } else { - /* LOG_E(PHY,"[UE %d] frame %d, nr_tti_rx %d, Error decoding PBCH!\n", - ue->Mod_id,frame_rx, nr_tti_rx); - - LOG_I(PHY,"[UE %d] rx_offset %d\n",ue->Mod_id,ue->rx_offset); - + ue->Mod_id,frame_rx, nr_tti_rx); + /* write_output("rxsig0.m","rxs0", ue->common_vars.rxdata[0],ue->frame_parms.samples_per_subframe,1,1); - write_output("H00.m","h00",&(ue->common_vars.dl_ch_estimates[0][0][0]),((ue->frame_parms.Ncp==0)?7:6)*(ue->frame_parms.ofdm_symbol_size),1,1); - write_output("H10.m","h10",&(ue->common_vars.dl_ch_estimates[0][2][0]),((ue->frame_parms.Ncp==0)?7:6)*(ue->frame_parms.ofdm_symbol_size),1,1); - write_output("rxsigF0.m","rxsF0", ue->common_vars.rxdataF[0],8*ue->frame_parms.ofdm_symbol_size,1,1); - write_output("PBCH_rxF0_ext.m","pbch0_ext",ue->pbch_vars[0]->rxdataF_ext[0],12*4*6,1,1); - write_output("PBCH_rxF0_comp.m","pbch0_comp",ue->pbch_vars[0]->rxdataF_comp[0],12*4*6,1,1); - write_output("PBCH_rxF_llr.m","pbch_llr",ue->pbch_vars[0]->llr,(ue->frame_parms.Ncp==0) ? 1920 : 1728,1,4); - exit(-1); + write_output("H00.m","h00",&(ue->common_vars.dl_ch_estimates[0][0][0]),((ue->frame_parms.Ncp==0)?7:6)*(ue->frame_parms.ofdm_symbol_size),1,1); + write_output("H10.m","h10",&(ue->common_vars.dl_ch_estimates[0][2][0]),((ue->frame_parms.Ncp==0)?7:6)*(ue->frame_parms.ofdm_symbol_size),1,1); + + write_output("rxsigF0.m","rxsF0", ue->common_vars.rxdataF[0],8*ue->frame_parms.ofdm_symbol_size,1,1); + write_output("PBCH_rxF0_ext.m","pbch0_ext",ue->pbch_vars[0]->rxdataF_ext[0],12*4*6,1,1); + write_output("PBCH_rxF0_comp.m","pbch0_comp",ue->pbch_vars[0]->rxdataF_comp[0],12*4*6,1,1); + write_output("PBCH_rxF_llr.m","pbch_llr",ue->pbch_vars[0]->llr,(ue->frame_parms.Ncp==0) ? 1920 : 1728,1,4); + exit(-1); */ ue->pbch_vars[eNB_id]->pdu_errors_conseq++; ue->pbch_vars[eNB_id]->pdu_errors++; - if (ue->mac_enabled == 1) { - //mac_xface->out_of_sync_ind(ue->Mod_id,frame_rx,eNB_id); - } - else{ - if (ue->pbch_vars[eNB_id]->pdu_errors_conseq>=100) { - LOG_E(PHY,"More that 100 consecutive PBCH errors! Exiting!\n"); - //mac_xface->macphy_exit("More that 100 consecutive PBCH errors!"); - } + + if (ue->pbch_vars[eNB_id]->pdu_errors_conseq>=100) { + LOG_E(PHY,"More that 100 consecutive PBCH errors! Exiting!\n"); + exit_fun("More that 100 consecutive PBCH errors! Exiting!\n"); } } @@ -3461,15 +3023,14 @@ void ue_pbch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *pro #ifdef DEBUG_PHY_PROC LOG_D(PHY,"[UE %d] frame %d, slot %d, PBCH errors = %d, consecutive errors = %d!\n", - ue->Mod_id,frame_rx, nr_tti_rx, - ue->pbch_vars[eNB_id]->pdu_errors, - ue->pbch_vars[eNB_id]->pdu_errors_conseq); + ue->Mod_id,frame_rx, nr_tti_rx, + ue->pbch_vars[eNB_id]->pdu_errors, + ue->pbch_vars[eNB_id]->pdu_errors_conseq); #endif VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PBCH_PROCEDURES, VCD_FUNCTION_OUT); } -#endif //(0) unsigned int get_tx_amp(int power_dBm, int power_max_dBm, int N_RB_UL, int nb_rb) { @@ -3490,13 +3051,13 @@ unsigned int get_tx_amp(int power_dBm, int power_max_dBm, int N_RB_UL, int nb_rb #ifdef NR_PDCCH_SCHED -int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t abstraction_flag) +int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc) { // unsigned int dci_cnt=0, i; //removed for nr_ue_pdcch_procedures and added in the loop for nb_coreset_active - #ifdef NR_PDCCH_SCHED_DEBUG - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_ue_pdcch_procedures() \n"); - #endif +#ifdef NR_PDCCH_SCHED_DEBUG + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_ue_pdcch_procedures() \n"); +#endif int frame_rx = proc->frame_rx; int nr_tti_rx = proc->nr_tti_rx; @@ -3508,47 +3069,41 @@ int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t * // each element represents the size in bits for each dci field, for each decoded dci -> [dci_cnt-1] // each time a dci is decode at dci_cnt, the values of the table dci_fields_sizes[i][j] will be copied at table dci_fields_sizes_cnt[dci_cnt-1][i][j] // table dci_fields_sizes_cnt[dci_cnt-1][i][j] will then be used in function nr_extract_dci_info - uint8_t dci_fields_sizes_cnt[MAX_NR_DCI_DECODED_SLOT][NBR_NR_DCI_FIELDS][NBR_NR_FORMATS] = {0}; + uint8_t dci_fields_sizes_cnt[MAX_NR_DCI_DECODED_SLOT][NBR_NR_DCI_FIELDS][NBR_NR_FORMATS]; int nb_searchspace_active=0; NR_UE_PDCCH **pdcch_vars = ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]]; NR_UE_PDCCH *pdcch_vars2 = ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]; - - uint16_t c_rnti=pdcch_vars[eNB_id]->crnti; - uint16_t cs_rnti=1,new_rnti=1,tc_rnti=1; - uint16_t p_rnti=P_RNTI; - uint16_t si_rnti=SI_RNTI; - uint16_t ra_rnti=3; - uint16_t sp_csi_rnti=1,sfi_rnti=1,int_rnti=1,tpc_pusch_rnti=1,tpc_pucch_rnti=1,tpc_srs_rnti=1; //FIXME - uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES] = - {c_rnti,cs_rnti,new_rnti,tc_rnti,p_rnti,si_rnti,ra_rnti,sp_csi_rnti,sfi_rnti,int_rnti,tpc_pusch_rnti,tpc_pucch_rnti,tpc_srs_rnti}; - // s in TS 38.212 Subclause 10.1, for each active BWP the UE can deal with 10 different search spaces // Higher layers have updated the number of searchSpaces with are active in the current slot and this value is stored in variable nb_searchspace_total int nb_searchspace_total = pdcch_vars2->nb_search_space; + + uint16_t c_rnti=pdcch_vars[eNB_id]->crnti; + uint16_t cs_rnti,new_rnti,tc_rnti; + uint16_t p_rnti=P_RNTI; + uint16_t si_rnti=SI_RNTI; + uint16_t ra_rnti=99; + uint16_t sp_csi_rnti,sfi_rnti,int_rnti,tpc_pusch_rnti,tpc_pucch_rnti,tpc_srs_rnti; //FIXME + uint16_t crc_scrambled_values[TOTAL_NBR_SCRAMBLED_VALUES] = + {c_rnti,cs_rnti,new_rnti,tc_rnti,p_rnti,si_rnti,ra_rnti,sp_csi_rnti,sfi_rnti,int_rnti,tpc_pusch_rnti,tpc_pucch_rnti,tpc_srs_rnti}; #ifdef NR_PDCCH_SCHED_DEBUG - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> there is a bug in FAPI to calculate nb_searchspace_total=%d\n",nb_searchspace_total); + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> there is a bug in FAPI to calculate nb_searchspace_total=%d\n",nb_searchspace_total); #endif if (nb_searchspace_total>1) nb_searchspace_total=1; // to be removed when fixing bug in FAPI #ifdef NR_PDCCH_SCHED_DEBUG - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> there is a bug in FAPI to calculate nb_searchspace_total so we set it to 1...\n"); - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> the number of searchSpaces active in the current slot(%d) is %d) \n", - nr_tti_rx,nb_searchspace_total); + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> there is a bug in FAPI to calculate nb_searchspace_total so we set it to 1...\n"); + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> the number of searchSpaces active in the current slot(%d) is %d) \n", + nr_tti_rx,nb_searchspace_total); #endif + // p in TS 38.212 Subclause 10.1, for each active BWP the UE can deal with 3 different CORESETs (including coresetId 0 for common search space) - //int nb_coreset_total = NR_NBR_CORESET_ACT_BWP; - uint8_t dci_cnt=0; + int nb_coreset_total = NR_NBR_CORESET_ACT_BWP; + unsigned int dci_cnt=0; // this table contains 56 (NBR_NR_DCI_FIELDS) elements for each dci field and format described in TS 38.212. Each element represents the size in bits for each dci field - //uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS][NBR_NR_FORMATS] = {0}; + uint8_t dci_fields_sizes[NBR_NR_DCI_FIELDS][NBR_NR_FORMATS] = {{0}}; // this is the UL bandwidth part. FIXME! To be defined where this value comes from - uint16_t n_RB_ULBWP = 106; + // uint16_t n_RB_ULBWP = 106; // this is the DL bandwidth part. FIXME! To be defined where this value comes from - uint16_t n_RB_DLBWP = 106; - #ifdef NR_PDCCH_SCHED_DEBUG - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> n_RB_ULBWP=%d n_RB_DLBWP=%d\n", - n_RB_ULBWP, - n_RB_DLBWP); - #endif // First we have to identify each searchSpace active at a time and do PDCCH monitoring corresponding to current searchSpace // Up to 10 searchSpaces can be configured to UE (s<=10) @@ -3558,40 +3113,23 @@ int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t * /* * The following code has been removed as it is handled by higher layers (fapi) * - // Verify that monitoring is required at the slot nr_tti_rx. We will run pdcch procedure only if do_pdcch_monitoring_current_slot=1 - // For Type0-PDCCH searchspace, we need to calculate the monitoring slot from Tables 13-1 .. 13-15 in TS 38.213 Subsection 13 - NR_UE_SLOT_PERIOD_OFFSET_t sl_period_offset_mon = pdcch_vars2->searchSpace[nb_searchspace_active].monitoringSlotPeriodicityAndOffset; - if (sl_period_offset_mon == nr_sl1) { - do_pdcch_monitoring_current_slot=1; // PDCCH monitoring in every slot - } else if (nr_tti_rx%(uint16_t)sl_period_offset_mon == pdcch_vars2->searchSpace[nb_searchspace_active].monitoringSlotPeriodicityAndOffset_offset) { - do_pdcch_monitoring_current_slot=1; // PDCCH monitoring in every monitoringSlotPeriodicityAndOffset slot with offset - }*/ + // Verify that monitoring is required at the slot nr_tti_rx. We will run pdcch procedure only if do_pdcch_monitoring_current_slot=1 + // For Type0-PDCCH searchspace, we need to calculate the monitoring slot from Tables 13-1 .. 13-15 in TS 38.213 Subsection 13 + NR_UE_SLOT_PERIOD_OFFSET_t sl_period_offset_mon = pdcch_vars2->searchSpace[nb_searchspace_active].monitoringSlotPeriodicityAndOffset; + if (sl_period_offset_mon == nr_sl1) { + do_pdcch_monitoring_current_slot=1; // PDCCH monitoring in every slot + } else if (nr_tti_rx%(uint16_t)sl_period_offset_mon == pdcch_vars2->searchSpace[nb_searchspace_active].monitoringSlotPeriodicityAndOffset_offset) { + do_pdcch_monitoring_current_slot=1; // PDCCH monitoring in every monitoringSlotPeriodicityAndOffset slot with offset + }*/ /* * FIXME * For PDCCH monitoring when overlap with SS/PBCH according to 38.213 v15.1.0 Section 10 * To be implemented LATER !!! */ //int _offset,_index,_M; - //int searchSpace_id = pdcch_vars2->searchSpace[nb_searchspace_active].searchSpaceId; - /* - * The following code has been removed as it is handled by higher layers (fapi) - if (searchSpace_id == 0){ // Implementing TS 38.213 subclause 13, UE procedure for monitoring Type0-PDCCH common search space - * - * according to TS 38.213 subclause 13 - * For the SS/PBCH block and control resource set (CORESET) multiplexing pattern 1, - * a UE monitors PDCCH in the Type0-PDCCH common search space over two consecutive slots starting from slot n0 - * - if (frame_rx%2 == 0) { - if ((((_offset*2+((_index*_M)/20))%2) != 0) || ((((_offset*2)+(_index*_M))%20) != nr_tti_rx) || ((((_offset*2)+(_index*_M))%20) != nr_tti_rx-1)){ - do_pdcch_monitoring_current_slot = 0; - } - } - if (frame_rx%2 == 1) { - if ((((_offset*2+((_index*_M)/20))%2) != 1) || ((((_offset*2)+(_index*_M))%20) != nr_tti_rx) || ((((_offset*2)+(_index*_M))%20) != nr_tti_rx-1)){ - do_pdcch_monitoring_current_slot = 0; - } - } - }*/ + int searchSpace_id = pdcch_vars2->searchSpace[nb_searchspace_active].searchSpaceId; + + #ifdef NR_PDCCH_SCHED_DEBUG printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> nb_searchspace_active=%d do_pdcch_monitoring_current_slot=%d (to be removed)\n", nb_searchspace_active, @@ -3602,16 +3140,16 @@ int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t * // the searchSpace indicates that we need to monitor PDCCH in current nr_tti_rx // get the parameters describing the current SEARCHSPACE // the CORESET id applicable to the current SearchSpace - //int searchSpace_coreset_id = pdcch_vars2->searchSpace[nb_searchspace_active].controlResourceSetId; + int searchSpace_coreset_id = pdcch_vars2->searchSpace[nb_searchspace_active].controlResourceSetId; // FIXME this variable is a bit string (14 bits) identifying every OFDM symbol in a slot. // at the moment we will not take into consideration this variable and we will consider that the OFDM symbol offset is always the first OFDM in a symbol uint16_t symbol_within_slot_mon = pdcch_vars2->searchSpace[nb_searchspace_active].monitoringSymbolWithinSlot; // get the remaining parameters describing the current SEARCHSPACE: // FIXME! To be defined where we get this information from - //NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L1 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel1; - //NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L2 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel2; - //NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L4 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel4; - //NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L8 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel8; - //NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L16 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel16; + NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L1 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel1; + NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L2 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel2; + NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L4 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel4; + NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L8 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel8; + NR_UE_SEARCHSPACE_nbrCAND_t num_cand_L16 = pdcch_vars2->searchSpace[nb_searchspace_active].nrofCandidates_aggrlevel16; // FIXME! A table of five enum elements // searchSpaceType indicates whether this is a common search space or a UE-specific search space //int searchSpaceType = pdcch_vars2->searchSpace[nb_searchspace_active].searchSpaceType.type; @@ -3627,64 +3165,50 @@ int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t * if (nb_coreset_active >= nb_coreset_total) return 0; // the coreset_id could not be found. There is a problem }*/ - /* - * we do not need these parameters yet - * - // get the parameters describing the current CORESET - int coreset_duration = pdcch_vars2->coreset[nb_coreset_active].duration; - uint64_t coreset_freq_dom = pdcch_vars2->coreset[nb_coreset_active].frequencyDomainResources; - int coreset_shift_index = pdcch_vars2->coreset[nb_coreset_active].cce_reg_mappingType.shiftIndex; - NR_UE_CORESET_REG_bundlesize_t coreset_bundlesize = pdcch_vars2->coreset[nb_coreset_active].cce_reg_mappingType.reg_bundlesize; - NR_UE_CORESET_interleaversize_t coreset_interleaversize = pdcch_vars2->coreset[nb_coreset_active].cce_reg_mappingType.interleaversize; - NR_UE_CORESET_precoder_granularity_t precoder_granularity = pdcch_vars2->coreset[nb_coreset_active].precoderGranularity; - int tci_statesPDCCH = pdcch_vars2->coreset[nb_coreset_active].tciStatesPDCCH; - int tci_present = pdcch_vars2->coreset[nb_coreset_active].tciPresentInDCI; - uint16_t pdcch_DMRS_scrambling_id = pdcch_vars2->coreset[nb_coreset_active].pdcchDMRSScramblingID; - */ - // A set of PDCCH candidates for a UE to monitor is defined in terms of PDCCH search spaces. - // Searchspace types: - // Type0-PDCCH common search space for a DCI format with CRC scrambled by a SI-RNTI - // number of consecutive resource blocks and a number of consecutive symbols for - // the control resource set of the Type0-PDCCH common search space from - // the four most significant bits of RMSI-PDCCH-Config as described in Tables 13-1 through 13-10 - // and determines PDCCH monitoring occasions - // from the four least significant bits of RMSI-PDCCH-Config, - // included in MasterInformationBlock, as described in Tables 13-11 through 13-15 - // Type0A-PDCCH common search space for a DCI format with CRC scrambled by a SI-RNTI - // Type1-PDCCH common search space for a DCI format with CRC scrambled by a RA-RNTI, or a TC-RNTI, or a C-RNTI - // Type2-PDCCH common search space for a DCI format with CRC scrambled by a P-RNTI - // Type3-PDCCH common search space for a DCI format with CRC scrambled by INT-RNTI, or SFI-RNTI, - // or TPC-PUSCH-RNTI, or TPC-PUCCH-RNTI, or TPC-SRS-RNTI, or C-RNTI, or CS-RNTI(s), or SP-CSI-RNTI + /* + * we do not need these parameters yet + * + // get the parameters describing the current CORESET + int coreset_duration = pdcch_vars2->coreset[nb_coreset_active].duration; + uint64_t coreset_freq_dom = pdcch_vars2->coreset[nb_coreset_active].frequencyDomainResources; + int coreset_shift_index = pdcch_vars2->coreset[nb_coreset_active].cce_reg_mappingType.shiftIndex; + NR_UE_CORESET_REG_bundlesize_t coreset_bundlesize = pdcch_vars2->coreset[nb_coreset_active].cce_reg_mappingType.reg_bundlesize; + NR_UE_CORESET_interleaversize_t coreset_interleaversize = pdcch_vars2->coreset[nb_coreset_active].cce_reg_mappingType.interleaversize; + NR_UE_CORESET_precoder_granularity_t precoder_granularity = pdcch_vars2->coreset[nb_coreset_active].precoderGranularity; + int tci_statesPDCCH = pdcch_vars2->coreset[nb_coreset_active].tciStatesPDCCH; + int tci_present = pdcch_vars2->coreset[nb_coreset_active].tciPresentInDCI; + uint16_t pdcch_DMRS_scrambling_id = pdcch_vars2->coreset[nb_coreset_active].pdcchDMRSScramblingID; + */ + + // A set of PDCCH candidates for a UE to monitor is defined in terms of PDCCH search spaces. + // Searchspace types: + // Type0-PDCCH common search space for a DCI format with CRC scrambled by a SI-RNTI + // number of consecutive resource blocks and a number of consecutive symbols for + // the control resource set of the Type0-PDCCH common search space from + // the four most significant bits of RMSI-PDCCH-Config as described in Tables 13-1 through 13-10 + // and determines PDCCH monitoring occasions + // from the four least significant bits of RMSI-PDCCH-Config, + // included in MasterInformationBlock, as described in Tables 13-11 through 13-15 + // Type0A-PDCCH common search space for a DCI format with CRC scrambled by a SI-RNTI + // Type1-PDCCH common search space for a DCI format with CRC scrambled by a RA-RNTI, or a TC-RNTI, or a C-RNTI + // Type2-PDCCH common search space for a DCI format with CRC scrambled by a P-RNTI + // Type3-PDCCH common search space for a DCI format with CRC scrambled by INT-RNTI, or SFI-RNTI, + // or TPC-PUSCH-RNTI, or TPC-PUCCH-RNTI, or TPC-SRS-RNTI, or C-RNTI, or CS-RNTI(s), or SP-CSI-RNTI -#ifdef PHY_ABSTRACTION - int CC_id; - int UE_id; - uint8_t harq_pid; -#endif - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PDCCH_PROCEDURES, VCD_FUNCTION_IN); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PDCCH_PROCEDURES, VCD_FUNCTION_IN); #if UE_TIMING_TRACE - start_meas(&ue->dlsch_rx_pdcch_stats); + start_meas(&ue->dlsch_rx_pdcch_stats); #endif - // if (nr_tti_rx != 5) - // return 0; - if (abstraction_flag == 0) { - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_RX_PDCCH, VCD_FUNCTION_IN); -/* rx_pdcch(ue, - proc->frame_rx, - nr_tti_rx, - eNB_id, - (ue->frame_parms.mode1_flag == 1) ? SISO : ALAMOUTI, - ue->high_speed_flag, - ue->is_secondary_ue);*/ //removed for nr_ue_pdcch_procedures - #ifdef NR_PDCCH_SCHED_DEBUG - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_rx_pdcch with eNB_id=%d (nb_coreset_active=%d, (symbol_within_slot_mon&0x3FFF)=%d, searchSpaceType=%d)\n", + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_RX_PDCCH, VCD_FUNCTION_IN); +#ifdef NR_PDCCH_SCHED_DEBUG + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_rx_pdcch with eNB_id=%d (nb_coreset_active=%d, (symbol_within_slot_mon&0x3FFF)=%d, searchSpaceType=%d)\n", eNB_id,nb_coreset_active,(symbol_within_slot_mon&0x3FFF), searchSpaceType); - #endif +#endif nr_rx_pdcch(ue, proc->frame_rx, nr_tti_rx, @@ -3696,169 +3220,87 @@ int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t * nb_coreset_active, (symbol_within_slot_mon&0x3FFF), searchSpaceType); - #ifdef NR_PDCCH_SCHED_DEBUG +#ifdef NR_PDCCH_SCHED_DEBUG printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Ending function nr_rx_pdcch(nb_coreset_active=%d, (symbol_within_slot_mon&0x3FFF)=%d, searchSpaceType=%d)\n", nb_coreset_active,(symbol_within_slot_mon&0x3FFF), searchSpaceType); - #endif +#endif VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_RX_PDCCH, VCD_FUNCTION_OUT); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_DCI_DECODING, VCD_FUNCTION_IN); -/* - //printf("Decode SIB frame param agregation + DCI %d %d \n",agregationLevel,dciFormat); - - //agregation level == FF means no configuration on - if(ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->agregationLevel == 0xFF || ue->decode_SIB) - { - // search all possible dcis - dci_cnt = dci_decoding_procedure(ue, - dci_alloc_rx, - (ue->UE_mode[eNB_id] < PUSCH)? 1 : 0, // if we're in PUSCH don't listen to common search space, - // later when we need paging or RA during connection, update this ... - eNB_id,nr_tti_rx); - } - else - { - // search only preconfigured dcis - // search C RNTI dci - dci_cnt = dci_CRNTI_decoding_procedure(ue, - dci_alloc_rx, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dciFormat, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->agregationLevel, - eNB_id, - nr_tti_rx); - } -*/ //removed for nr_ue_pdcch_procedures crc_scrambled_t crc_scrambled; format_found_t format_found=255; - #ifdef NR_PDCCH_SCHED_DEBUG - if (searchSpaceType == common) { // search all possible dci's for COMMON SEARCH SPACES according to the current SEARCHSPACE configuration - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_dci_decoding_procedure with eNB_id=%d (n_RB_ULBWP=%d, n_RB_DLBWP=%d, searchSpaceType=%d, nb_searchspace_active=%d, nb_coreset_active=%d) -> dci_cnt=%d\n", - eNB_id,n_RB_ULBWP,n_RB_DLBWP,searchSpaceType, - nb_searchspace_active, - nb_coreset_active, - dci_cnt); - } - if (searchSpaceType == ue_specific){// search all possible dci's for UE-SPECIFIC SEARCH SPACES according to the current SEARCHSPACE configuration - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_dci_decoding_procedure with eNB_id=%d (n_RB_ULBWP=%d, n_RB_DLBWP=%d, searchSpaceType=%d, nb_searchspace_active=%d, nb_coreset_active=%d) -> dci_cnt=%d\n", - eNB_id,n_RB_ULBWP,n_RB_DLBWP,searchSpaceType, - nb_searchspace_active, - nb_coreset_active, - dci_cnt); - } - #endif - - dci_cnt += nr_dci_decoding_procedure(nb_searchspace_active, - nb_coreset_active, - ue, - &dci_alloc_rx[dci_cnt], - searchSpaceType, - eNB_id, - nr_tti_rx, - dci_fields_sizes_cnt, - n_RB_ULBWP, - n_RB_DLBWP, - &crc_scrambled, - &format_found, - crc_scrambled_values); - #ifdef NR_PDCCH_SCHED_DEBUG - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Ending function nr_dci_decoding_procedure() -> dci_cnt=%d\n",dci_cnt); - #endif +#ifdef NR_PDCCH_SCHED_DEBUG + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_dci_decoding_procedure with eNB_id=%d (n_RB_ULBWP=%d, n_RB_DLBWP=%d, searchSpaceType=%d, nb_searchspace_active=%d, nb_coreset_active=%d)\n", + eNB_id, + pdcch_vars2->n_RB_BWP[nb_searchspace_active], + pdcch_vars2->n_RB_BWP[nb_searchspace_active], + searchSpaceType, + nb_searchspace_active, + nb_coreset_active); + +#endif + + + dci_cnt += nr_dci_decoding_procedure(nb_searchspace_active, + nb_coreset_active, + ue, + &dci_alloc_rx[dci_cnt], + searchSpaceType, // if we're in PUSCH don't listen to common search space, + // later when we need paging or RA during connection, update this ... + eNB_id, + nr_tti_rx, + dci_fields_sizes_cnt, + pdcch_vars2->n_RB_BWP[nb_searchspace_active], + pdcch_vars2->n_RB_BWP[nb_searchspace_active], + &crc_scrambled, + &format_found, + crc_scrambled_values); +#ifdef NR_PDCCH_SCHED_DEBUG + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Ending function nr_dci_decoding_procedure() -> dci_cnt=%d\n",dci_cnt); +#endif VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_DCI_DECODING, VCD_FUNCTION_OUT); //LOG_D(PHY,"[UE %d][PUSCH] Frame %d nr_tti_rx %d PHICH RX\n",ue->Mod_id,frame_rx,nr_tti_rx); - /* if (is_phich_subframe(&ue->frame_parms,nr_tti_rx)) { - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_RX_PHICH, VCD_FUNCTION_IN); - rx_phich(ue,proc,nr_tti_rx,eNB_id); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_RX_PHICH, VCD_FUNCTION_OUT); - }*/ //removed for nr_ue_pdcch_procedures - } - #ifdef PHY_ABSTRACTION - else { - for (i=0; iframe_parms.Nid_cell == ue->frame_parms.Nid_cell) - break; + uint8_t *nCCE_current = &ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->nCCE[nr_tti_rx]; + uint8_t *nCCE_dest = &ue->pdcch_vars[next1_thread_id][eNB_id]->nCCE[nr_tti_rx]; + uint8_t *nCCE_dest1 = &ue->pdcch_vars[next2_thread_id][eNB_id]->nCCE[nr_tti_rx]; + memcpy(nCCE_dest, nCCE_current, sizeof(uint8_t)); + memcpy(nCCE_dest1, nCCE_current, sizeof(uint8_t)); - if (CC_id < MAX_NUM_CCs) - break; - } + LOG_D(PHY,"current_thread %d next1_thread %d next2_thread %d \n", ue->current_thread_id[nr_tti_rx], next1_thread_id, next2_thread_id); - if (i==NB_eNB_INST) { - LOG_E(PHY,"[UE %d] phy_procedures_lte_ue.c: FATAL : Could not find attached eNB for DCI emulation (Nid_cell %d)!!!!\n",ue->Mod_id,ue->frame_parms.Nid_cell); - mac_xface->macphy_exit("Could not find attached eNB for DCI emulation"); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PDCCH_PROCEDURES, VCD_FUNCTION_OUT); - return(-1); - } - - LOG_D(PHY,"Calling dci_decoding_proc_emul ...\n"); - dci_cnt = dci_decoding_procedure_emul(ue->pdcch_vars[nr_tti_rx&1], - PHY_vars_eNB_g[i][CC_id]->num_ue_spec_dci[nr_tti_rx&1], - PHY_vars_eNB_g[i][CC_id]->num_common_dci[nr_tti_rx&1], - PHY_vars_eNB_g[i][CC_id]->dci_alloc[nr_tti_rx&1], - dci_alloc_rx, - eNB_id); - // printf("DCI: dci_cnt %d\n",dci_cnt); - UE_id = (uint32_t)find_ue((int16_t)ue->pdcch_vars[nr_tti_rx&1][eNB_id]->crnti,PHY_vars_eNB_g[i][CC_id]); - - if (UE_id>=0) { - // printf("Checking PHICH for UE %d (eNB %d)\n",UE_id,i); - if (is_phich_subframe(&ue->frame_parms,nr_tti_rx)) { - harq_pid = phich_subframe_to_harq_pid(&ue->frame_parms,frame_rx,nr_tti_rx); - - if (ue->ulsch[eNB_id]->harq_processes[harq_pid]->status == ACTIVE) { - // ue->ulsch[eNB_id]->harq_processes[harq_pid]->phich_ACK=1; - ue->ulsch[eNB_id]->harq_processes[harq_pid]->subframe_scheduling_flag =0; - ue->ulsch[eNB_id]->harq_processes[harq_pid]->status = IDLE; - ue->ulsch_Msg3_active[eNB_id] = 0; - ue->ulsch[eNB_id]->harq_processes[harq_pid]->round = 0; - LOG_D(PHY,"Msg3 inactive\n"); - - } // harq_pid is ACTIVE - } // This is a PHICH nr_tti_rx - } // UE_id exists - } - - #endif - - uint8_t *nCCE_current = &ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->nCCE[nr_tti_rx]; - uint8_t *nCCE_dest = &ue->pdcch_vars[next1_thread_id][eNB_id]->nCCE[nr_tti_rx]; - uint8_t *nCCE_dest1 = &ue->pdcch_vars[next2_thread_id][eNB_id]->nCCE[nr_tti_rx]; - memcpy(nCCE_dest, nCCE_current, sizeof(uint8_t)); - memcpy(nCCE_dest1, nCCE_current, sizeof(uint8_t)); - - LOG_D(PHY,"current_thread %d next1_thread %d next2_thread %d \n", ue->current_thread_id[nr_tti_rx], next1_thread_id, next2_thread_id); - - LOG_D(PHY,"[UE %d] AbsSubFrame %d.%d, Mode %s: DCI found %i --> rnti %x / crnti %x : format %d\n", - ue->Mod_id,frame_rx%1024,nr_tti_rx,mode_string[ue->UE_mode[eNB_id]], - dci_cnt, - dci_alloc_rx[0].rnti, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, - dci_alloc_rx[0].format ); - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dci_received += dci_cnt; + LOG_D(PHY,"[UE %d] AbsSubFrame %d.%d, Mode %s: DCI found %i --> rnti %x / crnti %x : format %d\n", + ue->Mod_id,frame_rx%1024,nr_tti_rx,mode_string[ue->UE_mode[eNB_id]], + dci_cnt, + dci_alloc_rx[0].rnti, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, + dci_alloc_rx[0].format ); + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->dci_received += dci_cnt; #ifdef EMOS - //emos_dump_UE.dci_cnt[nr_tti_rx] = dci_cnt; + //emos_dump_UE.dci_cnt[nr_tti_rx] = dci_cnt; #endif - ue->dci_ind.number_of_dcis = dci_cnt; - ue->dl_indication.dci_ind = &ue->dci_ind; // hang on rx_ind instance + ue->dci_ind.number_of_dcis = dci_cnt; + ue->dl_indication.dci_ind = &ue->dci_ind; // hang on rx_ind instance - for (int i=0; idlsch_received[eNB_id]++; + if ((dci_alloc_rx[i].rnti == crc_scrambled_values[_SI_RNTI_]) && (dci_alloc_rx[i].format == format1_0)) ue->dlsch_SI_received[eNB_id]++; if ((dci_alloc_rx[i].rnti == crc_scrambled_values[_P_RNTI_]) && (dci_alloc_rx[i].format == format1_0)) ue->dlsch_p_received[eNB_id]++; if ((dci_alloc_rx[i].rnti == crc_scrambled_values[_RA_RNTI_]) && (dci_alloc_rx[i].format == format1_0)) ue->dlsch_ra_received[eNB_id]++; @@ -3875,324 +3317,181 @@ int nr_ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t * if ((dci_alloc_rx[i].format == format1_1)){ // This format not implemented at a first time. FIXME } NR_DCI_INFO_EXTRACTED_t nr_dci_info_extracted; - NR_DCI_INFO_EXTRACTED_t *ptr_nr_dci_info_extracted = &nr_dci_info_extracted; #ifdef NR_PDCCH_SCHED_DEBUG - printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_generate_ue_ul_dlsch_params_from_dci with eNB_id=%d dci_cnt=%d format=%d, rnti=%d\n", + printf("<-NR_PDCCH_PHY_PROCEDURES_LTE_UE (nr_ue_pdcch_procedures)-> Entering function nr_generate_ue_ul_dlsch_params_from_dci with eNB_id=%d dci_cnt=%d\n", eNB_id, - dci_cnt, - dci_alloc_rx[i].format, - dci_alloc_rx[i].rnti); + dci_cnt); #endif - nr_generate_ue_ul_dlsch_params_from_dci(ue, - eNB_id, - frame_rx, - nr_tti_rx, - dci_alloc_rx[i].dci_pdu, - dci_alloc_rx[i].rnti, - dci_alloc_rx[i].dci_length, - dci_alloc_rx[i].format, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->ulsch[eNB_id], - &ue->frame_parms, - ue->pdsch_config_dedicated, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id], - dci_fields_sizes_cnt[i], - n_RB_ULBWP, - n_RB_DLBWP, - crc_scrambled_values, - ptr_nr_dci_info_extracted);//&nr_dci_info_extracted); + eNB_id, + frame_rx, + nr_tti_rx, + dci_alloc_rx[i].dci_pdu, + dci_alloc_rx[i].rnti, + dci_alloc_rx[i].dci_length, + dci_alloc_rx[i].format, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], + ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id], + ue->ulsch[eNB_id], + &ue->frame_parms, + ue->pdsch_config_dedicated, + ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id], + dci_fields_sizes_cnt[i], + pdcch_vars2->n_RB_BWP[nb_searchspace_active], + pdcch_vars2->n_RB_BWP[nb_searchspace_active], + crc_scrambled_values, + &nr_dci_info_extracted); ue->dci_ind.dci_list[i].rnti = dci_alloc_rx[i].rnti; ue->dci_ind.dci_list[i].dci_format = dci_alloc_rx[i].format; ue->dci_ind.dci_list[i].n_CCE = dci_alloc_rx[i].firstCCE; ue->dci_ind.dci_list[i].N_CCE = (int)dci_alloc_rx[i].L; - ue->dci_ind.number_of_dcis = ue->dci_ind.number_of_dcis + 1; - memcpy(&ue->dci_ind.dci_list[i].dci, &nr_dci_info_extracted, sizeof(fapi_nr_dci_pdu_rel15_t) ); - - //printf(">>> example mcs=%d\n",nr_dci_info_extracted.mcs); - //printf(">>> calling MAC with dl_indication DCI_IND\n"); - // TODO: check where should we send up this message. - ue->dl_indication.dci_ind=&ue->dci_ind; - ue->if_inst->dl_indication(&ue->dl_indication); + memcpy(&ue->dci_ind.dci_list[i].dci, &nr_dci_info_extracted, sizeof(fapi_nr_dci_pdu_rel15_t) ); - /* - * This is the LTE part to be removed - - if ((ue->UE_mode[eNB_id]>PRACH) && (dci_alloc_rx[i].rnti == ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti) && (dci_alloc_rx[i].format != format0)) { - - LOG_D(PHY,"[UE %d][DCI][PDSCH %x] AbsSubframe %d.%d: format %d, num_pdcch_symbols %d, nCCE %d, total CCEs %d\n", - ue->Mod_id,dci_alloc_rx[i].rnti, - frame_rx%1024,nr_tti_rx, - dci_alloc_rx[i].format, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->nCCE[nr_tti_rx], - get_nCCE(3,&ue->frame_parms,get_mi(&ue->frame_parms,0))); - - //dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - - if ((ue->UE_mode[eNB_id] > PRACH) && - (generate_ue_dlsch_params_from_dci(frame_rx, - nr_tti_rx, - (void *)&dci_alloc_rx[i].dci_pdu, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, - dci_alloc_rx[i].format, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->pdsch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id], - &ue->frame_parms, - ue->pdsch_config_dedicated, - SI_RNTI, - 0, - P_RNTI, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id], - ue->pdcch_vars[0%RX_NB_TH][eNB_id]->crnti_is_temporary? ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti: 0)==0)) { - - // update TPC for PUCCH - if((dci_alloc_rx[i].format == format1) || - (dci_alloc_rx[i].format == format1A) || - (dci_alloc_rx[i].format == format1B) || - (dci_alloc_rx[i].format == format2) || - (dci_alloc_rx[i].format == format2A) || - (dci_alloc_rx[i].format == format2B)) - { - //ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->g_pucch += ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_processes[ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->current_harq_pid]->delta_PUCCH; - int32_t delta_pucch = ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_processes[ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->current_harq_pid]->delta_PUCCH; - for(int th_id=0; th_iddlsch[th_id][eNB_id][0]->g_pucch += delta_pucch; - } - LOG_D(PHY,"update TPC for PUCCH %d.%d / pid %d delta_PUCCH %d g_pucch %d %d \n",frame_rx, nr_tti_rx,ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->current_harq_pid, - delta_pucch, - ue->dlsch[0][eNB_id][0]->g_pucch, - ue->dlsch[1][eNB_id][0]->g_pucch - //ue->dlsch[2][eNB_id][0]->g_pucch - ); - } - - ue->dlsch_received[eNB_id]++; - - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d] Generated UE DLSCH C_RNTI format %d\n",ue->Mod_id,dci_alloc_rx[i].format); - dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - LOG_D(PHY,"[UE %d] *********** dlsch->active in nr_tti_rx %d=> %d\n",ue->Mod_id,nr_tti_rx,ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active); - #endif - - // we received a CRNTI, so we're in PUSCH - if (ue->UE_mode[eNB_id] != PUSCH) { - #ifdef DEBUG_PHY_PROC - LOG_I(PHY,"[UE %d] Frame %d, nr_tti_rx %d: Received DCI with CRNTI %x => Mode PUSCH\n",ue->Mod_id,frame_rx,nr_tti_rx,ue->pdcch_vars[nr_tti_rx&1][eNB_id]->crnti); - #endif - //dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - ue->UE_mode[eNB_id] = PUSCH; - //mac_xface->macphy_exit("Connected. Exiting\n"); - } - } else { - LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d: Problem in DCI!\n",ue->Mod_id,frame_rx,nr_tti_rx); - dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - } - } - - else if ((dci_alloc_rx[i].rnti == SI_RNTI) && - ((dci_alloc_rx[i].format == format1A) || (dci_alloc_rx[i].format == format1C))) { - - #ifdef DEBUG_PHY_PROC - LOG_I(PHY,"[UE %d] nr_tti_rx %d: Found rnti %x, format 1%s, dci_cnt %d\n",ue->Mod_id,nr_tti_rx,dci_alloc_rx[i].rnti,dci_alloc_rx[i].format==format1A?"A":"C",i); - #endif - - - if (generate_ue_dlsch_params_from_dci(frame_rx, - nr_tti_rx, - (void *)&dci_alloc_rx[i].dci_pdu, - SI_RNTI, - dci_alloc_rx[i].format, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->pdsch_vars_SI[eNB_id], - &ue->dlsch_SI[eNB_id], - &ue->frame_parms, - ue->pdsch_config_dedicated, - SI_RNTI, - 0, - P_RNTI, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id], - 0)==0) { - - ue->dlsch_SI_received[eNB_id]++; - - - LOG_I(PHY,"[UE %d] Frame %d, nr_tti_rx %d : Generate UE DLSCH SI_RNTI format 1%s\n",ue->Mod_id,frame_rx,nr_tti_rx,dci_alloc_rx[i].format==format1A?"A":"C"); - //dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - - } - } - - else if ((dci_alloc_rx[i].rnti == P_RNTI) && - ((dci_alloc_rx[i].format == format1A) || (dci_alloc_rx[i].format == format1C))) { - - #ifdef DEBUG_PHY_PROC - LOG_I(PHY,"[UE %d] nr_tti_rx %d: Found rnti %x, format 1%s, dci_cnt %d\n",ue->Mod_id,nr_tti_rx,dci_alloc_rx[i].rnti,dci_alloc_rx[i].format==format1A?"A":"C",i); - #endif - - - if (generate_ue_dlsch_params_from_dci(frame_rx, - nr_tti_rx, - (void *)&dci_alloc_rx[i].dci_pdu, - P_RNTI, - dci_alloc_rx[i].format, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->pdsch_vars_p[eNB_id], - &ue->dlsch_SI[eNB_id], - &ue->frame_parms, - ue->pdsch_config_dedicated, - SI_RNTI, - 0, - P_RNTI, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id], - 0)==0) { - - ue->dlsch_p_received[eNB_id]++; - LOG_D(PHY,"[UE %d] Frame %d, nr_tti_rx %d : Generate UE DLSCH P_RNTI format 1%s\n",ue->Mod_id,frame_rx,nr_tti_rx,dci_alloc_rx[i].format==format1A?"A":"C"); - //dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - - } - } - - else if ((ue->prach_resources[eNB_id]) && - (dci_alloc_rx[i].rnti == ue->prach_resources[eNB_id]->ra_RNTI) && - (dci_alloc_rx[i].format == format1A)) { - - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d][RAPROC] nr_tti_rx %d: Found RA rnti %x, format 1A, dci_cnt %d\n",ue->Mod_id,nr_tti_rx,dci_alloc_rx[i].rnti,i); - - //if (((frame_rx%100) == 0) || (frame_rx < 20)) - //dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - //mac_xface->macphy_exit("so far so good...\n"); - #endif - - - if (generate_ue_dlsch_params_from_dci(frame_rx, - nr_tti_rx, - (DCI1A_5MHz_TDD_1_6_t *)&dci_alloc_rx[i].dci_pdu, - ue->prach_resources[eNB_id]->ra_RNTI, - format1A, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], - ue->pdsch_vars_ra[eNB_id], - &ue->dlsch_ra[eNB_id], - &ue->frame_parms, - ue->pdsch_config_dedicated, - SI_RNTI, - ue->prach_resources[eNB_id]->ra_RNTI, - P_RNTI, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id], - 0)==0) { - - ue->dlsch_ra_received[eNB_id]++; - - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d] Generate UE DLSCH RA_RNTI format 1A, rb_alloc %x, dlsch_ra[eNB_id] %p\n", - ue->Mod_id,ue->dlsch_ra[eNB_id]->harq_processes[0]->rb_alloc_even[0],ue->dlsch_ra[eNB_id]); - #endif - } - } else if( (dci_alloc_rx[i].rnti == ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti) && - (dci_alloc_rx[i].format == format0)) { - - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d][PUSCH] Frame %d nr_tti_rx %d: Found rnti %x, format 0, dci_cnt %d\n", - ue->Mod_id,frame_rx,nr_tti_rx,dci_alloc_rx[i].rnti,i); - #endif - - ue->ulsch_no_allocation_counter[eNB_id] = 0; - //dump_dci(&ue->frame_parms,&dci_alloc_rx[i]); - - if ((ue->UE_mode[eNB_id] > PRACH) && - (generate_ue_ulsch_params_from_dci((void *)&dci_alloc_rx[i].dci_pdu, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, - nr_tti_rx, - format0, - ue, - proc, - SI_RNTI, - 0, - P_RNTI, - CBA_RNTI, - eNB_id, - 0)==0)) { - #if T_TRACER - NR_DL_FRAME_PARMS *frame_parms = &ue->frame_parms; - uint8_t harq_pid = subframe2harq_pid(frame_parms, - pdcch_alloc2ul_frame(frame_parms,proc->frame_rx,proc->nr_tti_rx), - pdcch_alloc2ul_subframe(frame_parms,proc->nr_tti_rx)); - - T(T_UE_PHY_ULSCH_UE_DCI, T_INT(eNB_id), T_INT(proc->frame_rx%1024), T_INT(proc->nr_tti_rx), T_INT(ue->Mod_id), - T_INT(dci_alloc_rx[i].rnti), T_INT(harq_pid), - T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->mcs), - T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->round), - T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb), - T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb), - T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS)); - #endif - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d] Generate UE ULSCH C_RNTI format 0 (nr_tti_rx %d)\n",ue->Mod_id,nr_tti_rx); - #endif - - } - } else if( (dci_alloc_rx[i].rnti == ue->ulsch[eNB_id]->cba_rnti[0]) && - (dci_alloc_rx[i].format == format0)) { - // UE could belong to more than one CBA group - // ue->Mod_id%ue->ulsch[eNB_id]->num_active_cba_groups] - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d][PUSCH] Frame %d nr_tti_rx %d: Found cba rnti %x, format 0, dci_cnt %d\n", - ue->Mod_id,frame_rx,nr_tti_rx,dci_alloc_rx[i].rnti,i); - - if (((frame_rx%100) == 0) || (frame_rx < 20)) - dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - - #endif - - ue->ulsch_no_allocation_counter[eNB_id] = 0; - //dump_dci(&ue->frame_parms,&dci_alloc_rx[i]); - - if ((ue->UE_mode[eNB_id] > PRACH) && - (generate_ue_ulsch_params_from_dci((void *)&dci_alloc_rx[i].dci_pdu, - ue->ulsch[eNB_id]->cba_rnti[0], - nr_tti_rx, - format0, - ue, - proc, - SI_RNTI, - 0, - P_RNTI, - CBA_RNTI, - eNB_id, - 0)==0)) { - - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d] Generate UE ULSCH CBA_RNTI format 0 (nr_tti_rx %d)\n",ue->Mod_id,nr_tti_rx); - #endif - ue->ulsch[eNB_id]->num_cba_dci[(nr_tti_rx+4)%10]++; - } - } - - else { - #ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d] frame %d, nr_tti_rx %d: received DCI %d with RNTI=%x (C-RNTI:%x, CBA_RNTI %x) and format %d!\n",ue->Mod_id,frame_rx,nr_tti_rx,i,dci_alloc_rx[i].rnti, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, - ue->ulsch[eNB_id]->cba_rnti[0], - dci_alloc_rx[i].format); - - // dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); - #endif - }*/ - } // end for loop dci_cnt + + // TODO: check where should we send up this message. + ue->if_inst->dl_indication(&ue->dl_indication); + + #if UE_TIMING_TRACE stop_meas(&ue->dlsch_rx_pdcch_stats); #endif VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PDCCH_PROCEDURES, VCD_FUNCTION_OUT); -// } // end if do_pdcch_monitoring_current_slot + + } // end for loop nb_searchspace_active + return(0); +} +#endif // NR_PDCCH_SCHED + + + + +#if 0 + + if (generate_ue_dlsch_params_from_dci(frame_rx, + nr_tti_rx, + (DCI1A_5MHz_TDD_1_6_t *)&dci_alloc_rx[i].dci_pdu, + ue->prach_resources[eNB_id]->ra_RNTI, + format1A, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id], + ue->pdsch_vars_ra[eNB_id], + &ue->dlsch_ra[eNB_id], + &ue->frame_parms, + ue->pdsch_config_dedicated, + SI_RNTI, + ue->prach_resources[eNB_id]->ra_RNTI, + P_RNTI, + ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id], + 0)==0) { + + ue->dlsch_ra_received[eNB_id]++; + + #ifdef DEBUG_PHY_PROC + LOG_D(PHY,"[UE %d] Generate UE DLSCH RA_RNTI format 1A, rb_alloc %x, dlsch_ra[eNB_id] %p\n", + ue->Mod_id,ue->dlsch_ra[eNB_id]->harq_processes[0]->rb_alloc_even[0],ue->dlsch_ra[eNB_id]); + #endif + } + } else if( (dci_alloc_rx[i].rnti == ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti) && + (dci_alloc_rx[i].format == format0)) { + + #ifdef DEBUG_PHY_PROC + LOG_D(PHY,"[UE %d][PUSCH] Frame %d nr_tti_rx %d: Found rnti %x, format 0, dci_cnt %d\n", + ue->Mod_id,frame_rx,nr_tti_rx,dci_alloc_rx[i].rnti,i); + #endif + + ue->ulsch_no_allocation_counter[eNB_id] = 0; + //dump_dci(&ue->frame_parms,&dci_alloc_rx[i]); + + if ((ue->UE_mode[eNB_id] > PRACH) && + (generate_ue_ulsch_params_from_dci((void *)&dci_alloc_rx[i].dci_pdu, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, + nr_tti_rx, + format0, + ue, + proc, + SI_RNTI, + 0, + P_RNTI, + CBA_RNTI, + eNB_id, + 0)==0)) { + #if T_TRACER + NR_DL_FRAME_PARMS *frame_parms = &ue->frame_parms; + uint8_t harq_pid = subframe2harq_pid(frame_parms, + pdcch_alloc2ul_frame(frame_parms,proc->frame_rx,proc->nr_tti_rx), + pdcch_alloc2ul_subframe(frame_parms,proc->nr_tti_rx)); + + T(T_UE_PHY_ULSCH_UE_DCI, T_INT(eNB_id), T_INT(proc->frame_rx%1024), T_INT(proc->nr_tti_rx), T_INT(ue->Mod_id), + T_INT(dci_alloc_rx[i].rnti), T_INT(harq_pid), + T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->mcs), + T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->round), + T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb), + T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb), + T_INT(ue->ulsch[eNB_id]->harq_processes[harq_pid]->TBS)); + #endif + #ifdef DEBUG_PHY_PROC + LOG_D(PHY,"[UE %d] Generate UE ULSCH C_RNTI format 0 (nr_tti_rx %d)\n",ue->Mod_id,nr_tti_rx); + #endif + + } + } else if( (dci_alloc_rx[i].rnti == ue->ulsch[eNB_id]->cba_rnti[0]) && + (dci_alloc_rx[i].format == format0)) { + // UE could belong to more than one CBA group + // ue->Mod_id%ue->ulsch[eNB_id]->num_active_cba_groups] + #ifdef DEBUG_PHY_PROC + LOG_D(PHY,"[UE %d][PUSCH] Frame %d nr_tti_rx %d: Found cba rnti %x, format 0, dci_cnt %d\n", + ue->Mod_id,frame_rx,nr_tti_rx,dci_alloc_rx[i].rnti,i); + + if (((frame_rx%100) == 0) || (frame_rx < 20)) + dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); + + #endif + + ue->ulsch_no_allocation_counter[eNB_id] = 0; + //dump_dci(&ue->frame_parms,&dci_alloc_rx[i]); + + if ((ue->UE_mode[eNB_id] > PRACH) && + (generate_ue_ulsch_params_from_dci((void *)&dci_alloc_rx[i].dci_pdu, + ue->ulsch[eNB_id]->cba_rnti[0], + nr_tti_rx, + format0, + ue, + proc, + SI_RNTI, + 0, + P_RNTI, + CBA_RNTI, + eNB_id, + 0)==0)) { + + #ifdef DEBUG_PHY_PROC + LOG_D(PHY,"[UE %d] Generate UE ULSCH CBA_RNTI format 0 (nr_tti_rx %d)\n",ue->Mod_id,nr_tti_rx); + #endif + ue->ulsch[eNB_id]->num_cba_dci[(nr_tti_rx+4)%10]++; + } + } + + else { + #ifdef DEBUG_PHY_PROC + LOG_D(PHY,"[UE %d] frame %d, nr_tti_rx %d: received DCI %d with RNTI=%x (C-RNTI:%x, CBA_RNTI %x) and format %d!\n",ue->Mod_id,frame_rx,nr_tti_rx,i,dci_alloc_rx[i].rnti, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, + ue->ulsch[eNB_id]->cba_rnti[0], + dci_alloc_rx[i].format); + + // dump_dci(&ue->frame_parms, &dci_alloc_rx[i]); + #endif + }*/ + + } // end for loop dci_cnt +#if UE_TIMING_TRACE + stop_meas(&ue->dlsch_rx_pdcch_stats); +#endif + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_PDCCH_PROCEDURES, VCD_FUNCTION_OUT); + + // } // end if do_pdcch_monitoring_current_slot } // end for loop nb_searchspace_active return(0); } @@ -4220,15 +3519,15 @@ void ue_pmch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc,int eNB_id,i LOG_D(PHY,"ue calling pmch nr_tti_rx ..\n "); LOG_D(PHY,"[UE %d] Frame %d, nr_tti_rx %d: Querying for PMCH demodulation\n", - ue->Mod_id,(nr_tti_rx==9?-1:0)+frame_rx,nr_tti_rx); + ue->Mod_id,(nr_tti_rx==9?-1:0)+frame_rx,nr_tti_rx); #if defined(Rel10) || defined(Rel14) /*pmch_mcs = mac_xface->ue_query_mch(ue->Mod_id, - CC_id, - frame_rx, - nr_tti_rx, - eNB_id, - &sync_area, - &mcch_active);*/ + CC_id, + frame_rx, + nr_tti_rx, + eNB_id, + &sync_area, + &mcch_active);*/ #else pmch_mcs=-1; @@ -4239,121 +3538,121 @@ void ue_pmch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc,int eNB_id,i fill_UE_dlsch_MCH(ue,pmch_mcs,1,0,0); if (abstraction_flag == 0 ) { - for (l=2; l<12; l++) { + for (l=2; l<12; l++) { - slot_fep_mbsfn(ue, - l, - nr_tti_rx, - 0,0);//ue->rx_offset,0); - } + slot_fep_mbsfn(ue, + l, + nr_tti_rx, + 0,0);//ue->rx_offset,0); + } - for (l=2; l<12; l++) { - rx_pmch(ue, - 0, - nr_tti_rx, - l); - } + for (l=2; l<12; l++) { + rx_pmch(ue, + 0, + nr_tti_rx, + l); + } - ue->dlsch_MCH[0]->harq_processes[0]->G = get_G(&ue->frame_parms, - ue->dlsch_MCH[0]->harq_processes[0]->nb_rb, - ue->dlsch_MCH[0]->harq_processes[0]->rb_alloc_even, - ue->dlsch_MCH[0]->harq_processes[0]->Qm, - 1, - 2, - frame_rx, - nr_tti_rx, - 0); + ue->dlsch_MCH[0]->harq_processes[0]->G = get_G(&ue->frame_parms, + ue->dlsch_MCH[0]->harq_processes[0]->nb_rb, + ue->dlsch_MCH[0]->harq_processes[0]->rb_alloc_even, + ue->dlsch_MCH[0]->harq_processes[0]->Qm, + 1, + 2, + frame_rx, + nr_tti_rx, + 0); - dlsch_unscrambling(&ue->frame_parms,1,ue->dlsch_MCH[0], - ue->dlsch_MCH[0]->harq_processes[0]->G, - ue->pdsch_vars_MCH[0]->llr[0],0,nr_tti_rx<<1); + dlsch_unscrambling(&ue->frame_parms,1,ue->dlsch_MCH[0], + ue->dlsch_MCH[0]->harq_processes[0]->G, + ue->pdsch_vars_MCH[0]->llr[0],0,nr_tti_rx<<1); #ifdef UE_DLSCH_PARALLELISATION - ret = dlsch_decoding_mthread(ue,proc, eNB_id, - ue->pdsch_vars_MCH[0]->llr[0], - &ue->frame_parms, - ue->dlsch_MCH[0], - ue->dlsch_MCH[0]->harq_processes[0], - frame_rx, - nr_tti_rx, - 0, - 0,1); + ret = dlsch_decoding_mthread(ue,proc, eNB_id, + ue->pdsch_vars_MCH[0]->llr[0], + &ue->frame_parms, + ue->dlsch_MCH[0], + ue->dlsch_MCH[0]->harq_processes[0], + frame_rx, + nr_tti_rx, + 0, + 0,1); #else - ret = dlsch_decoding(ue, - ue->pdsch_vars_MCH[0]->llr[0], - &ue->frame_parms, - ue->dlsch_MCH[0], - ue->dlsch_MCH[0]->harq_processes[0], - frame_rx, - nr_tti_rx, - 0, - 0,1); - printf("start pmch dlsch decoding\n"); + ret = dlsch_decoding(ue, + ue->pdsch_vars_MCH[0]->llr[0], + &ue->frame_parms, + ue->dlsch_MCH[0], + ue->dlsch_MCH[0]->harq_processes[0], + frame_rx, + nr_tti_rx, + 0, + 0,1); + printf("start pmch dlsch decoding\n"); #endif } else { // abstraction #ifdef PHY_ABSTRACTION - ret = dlsch_decoding_emul(ue, - nr_tti_rx, - 5, // PMCH - eNB_id); + ret = dlsch_decoding_emul(ue, + nr_tti_rx, + 5, // PMCH + eNB_id); #endif } if (mcch_active == 1) - ue->dlsch_mcch_trials[sync_area][0]++; + ue->dlsch_mcch_trials[sync_area][0]++; else - ue->dlsch_mtch_trials[sync_area][0]++; + ue->dlsch_mtch_trials[sync_area][0]++; if (ret == (1+ue->dlsch_MCH[0]->max_turbo_iterations)) { - if (mcch_active == 1) - ue->dlsch_mcch_errors[sync_area][0]++; - else - ue->dlsch_mtch_errors[sync_area][0]++; + if (mcch_active == 1) + ue->dlsch_mcch_errors[sync_area][0]++; + else + ue->dlsch_mtch_errors[sync_area][0]++; - LOG_D(PHY,"[UE %d] Frame %d, nr_tti_rx %d: PMCH in error (%d,%d), not passing to L2 (TBS %d, iter %d,G %d)\n", - ue->Mod_id, + LOG_D(PHY,"[UE %d] Frame %d, nr_tti_rx %d: PMCH in error (%d,%d), not passing to L2 (TBS %d, iter %d,G %d)\n", + ue->Mod_id, frame_rx,nr_tti_rx, - ue->dlsch_mcch_errors[sync_area][0], - ue->dlsch_mtch_errors[sync_area][0], - ue->dlsch_MCH[0]->harq_processes[0]->TBS>>3, - ue->dlsch_MCH[0]->max_turbo_iterations, - ue->dlsch_MCH[0]->harq_processes[0]->G); - dump_mch(ue,0,ue->dlsch_MCH[0]->harq_processes[0]->G,nr_tti_rx); + ue->dlsch_mcch_errors[sync_area][0], + ue->dlsch_mtch_errors[sync_area][0], + ue->dlsch_MCH[0]->harq_processes[0]->TBS>>3, + ue->dlsch_MCH[0]->max_turbo_iterations, + ue->dlsch_MCH[0]->harq_processes[0]->G); + dump_mch(ue,0,ue->dlsch_MCH[0]->harq_processes[0]->G,nr_tti_rx); #ifdef DEBUG_DLSCH - for (int i=0; idlsch_MCH[0]->harq_processes[0]->TBS>>3; i++) { - LOG_T(PHY,"%02x.",ue->dlsch_MCH[0]->harq_processes[0]->c[0][i]); - } + for (int i=0; idlsch_MCH[0]->harq_processes[0]->TBS>>3; i++) { + LOG_T(PHY,"%02x.",ue->dlsch_MCH[0]->harq_processes[0]->c[0][i]); + } - LOG_T(PHY,"\n"); + LOG_T(PHY,"\n"); #endif - if (nr_tti_rx==9) - //mac_xface->macphy_exit("Why are we exiting here?"); - } else { // decoding successful + if (nr_tti_rx==9) + //mac_xface->macphy_exit("Why are we exiting here?"); + } else { // decoding successful #if defined(Rel10) || defined(Rel14) - if (mcch_active == 1) { - /*mac_xface->ue_send_mch_sdu(ue->Mod_id, - CC_id, - frame_rx, - ue->dlsch_MCH[0]->harq_processes[0]->b, - ue->dlsch_MCH[0]->harq_processes[0]->TBS>>3, - eNB_id,// not relevant in eMBMS context - sync_area);*/ - ue->dlsch_mcch_received[sync_area][0]++; + if (mcch_active == 1) { + /*mac_xface->ue_send_mch_sdu(ue->Mod_id, + CC_id, + frame_rx, + ue->dlsch_MCH[0]->harq_processes[0]->b, + ue->dlsch_MCH[0]->harq_processes[0]->TBS>>3, + eNB_id,// not relevant in eMBMS context + sync_area);*/ + ue->dlsch_mcch_received[sync_area][0]++; - if (ue->dlsch_mch_received_sf[nr_tti_rx%5][0] == 1 ) { - ue->dlsch_mch_received_sf[nr_tti_rx%5][0]=0; - } else { - ue->dlsch_mch_received[0]+=1; - ue->dlsch_mch_received_sf[nr_tti_rx][0]=1; - } + if (ue->dlsch_mch_received_sf[nr_tti_rx%5][0] == 1 ) { + ue->dlsch_mch_received_sf[nr_tti_rx%5][0]=0; + } else { + ue->dlsch_mch_received[0]+=1; + ue->dlsch_mch_received_sf[nr_tti_rx][0]=1; + } - } + } #endif // Rel10 || Rel14 } // decoding sucessful } // pmch_mcs>=0 @@ -4363,48 +3662,46 @@ void ue_pmch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc,int eNB_id,i void copy_harq_proc_struct(NR_DL_UE_HARQ_t *harq_processes_dest, NR_DL_UE_HARQ_t *current_harq_processes) { - harq_processes_dest->B = current_harq_processes->B ; - harq_processes_dest->C = current_harq_processes->C ; - harq_processes_dest->Cminus = current_harq_processes->Cminus ; - harq_processes_dest->Cplus = current_harq_processes->Cplus ; - harq_processes_dest->DCINdi = current_harq_processes->DCINdi ; - harq_processes_dest->F = current_harq_processes->F ; - harq_processes_dest->G = current_harq_processes->G ; - harq_processes_dest->Kminus = current_harq_processes->Kminus ; - harq_processes_dest->Kplus = current_harq_processes->Kplus ; - harq_processes_dest->Nl = current_harq_processes->Nl ; - harq_processes_dest->Qm = current_harq_processes->Qm ; - harq_processes_dest->TBS = current_harq_processes->TBS ; - harq_processes_dest->b = current_harq_processes->b ; - harq_processes_dest->codeword = current_harq_processes->codeword ; - harq_processes_dest->delta_PUCCH = current_harq_processes->delta_PUCCH ; - harq_processes_dest->dl_power_off = current_harq_processes->dl_power_off ; - harq_processes_dest->first_tx = current_harq_processes->first_tx ; - harq_processes_dest->mcs = current_harq_processes->mcs ; - harq_processes_dest->mimo_mode = current_harq_processes->mimo_mode ; - harq_processes_dest->nb_rb = current_harq_processes->nb_rb ; - harq_processes_dest->pmi_alloc = current_harq_processes->pmi_alloc ; - harq_processes_dest->rb_alloc_even[0] = current_harq_processes->rb_alloc_even[0] ; - harq_processes_dest->rb_alloc_even[1] = current_harq_processes->rb_alloc_even[1] ; - harq_processes_dest->rb_alloc_even[2] = current_harq_processes->rb_alloc_even[2] ; - harq_processes_dest->rb_alloc_even[3] = current_harq_processes->rb_alloc_even[3] ; - harq_processes_dest->rb_alloc_odd[0] = current_harq_processes->rb_alloc_odd[0] ; - harq_processes_dest->rb_alloc_odd[1] = current_harq_processes->rb_alloc_odd[1] ; - harq_processes_dest->rb_alloc_odd[2] = current_harq_processes->rb_alloc_odd[2] ; - harq_processes_dest->rb_alloc_odd[3] = current_harq_processes->rb_alloc_odd[3] ; - harq_processes_dest->round = current_harq_processes->round ; - harq_processes_dest->rvidx = current_harq_processes->rvidx ; - harq_processes_dest->status = current_harq_processes->status ; - harq_processes_dest->vrb_type = current_harq_processes->vrb_type ; + + harq_processes_dest->B = current_harq_processes->B ; + harq_processes_dest->C = current_harq_processes->C ; + harq_processes_dest->DCINdi = current_harq_processes->DCINdi ; + harq_processes_dest->F = current_harq_processes->F ; + harq_processes_dest->G = current_harq_processes->G ; + harq_processes_dest->K = current_harq_processes->K ; + harq_processes_dest->Nl = current_harq_processes->Nl ; + harq_processes_dest->Qm = current_harq_processes->Qm ; + harq_processes_dest->TBS = current_harq_processes->TBS ; + harq_processes_dest->b = current_harq_processes->b ; + harq_processes_dest->codeword = current_harq_processes->codeword ; + harq_processes_dest->delta_PUCCH = current_harq_processes->delta_PUCCH ; + harq_processes_dest->dl_power_off = current_harq_processes->dl_power_off ; + harq_processes_dest->first_tx = current_harq_processes->first_tx ; + harq_processes_dest->mcs = current_harq_processes->mcs ; + harq_processes_dest->mimo_mode = current_harq_processes->mimo_mode ; + harq_processes_dest->nb_rb = current_harq_processes->nb_rb ; + harq_processes_dest->pmi_alloc = current_harq_processes->pmi_alloc ; + harq_processes_dest->rb_alloc_even[0] = current_harq_processes->rb_alloc_even[0] ; + harq_processes_dest->rb_alloc_even[1] = current_harq_processes->rb_alloc_even[1] ; + harq_processes_dest->rb_alloc_even[2] = current_harq_processes->rb_alloc_even[2] ; + harq_processes_dest->rb_alloc_even[3] = current_harq_processes->rb_alloc_even[3] ; + harq_processes_dest->rb_alloc_odd[0] = current_harq_processes->rb_alloc_odd[0] ; + harq_processes_dest->rb_alloc_odd[1] = current_harq_processes->rb_alloc_odd[1] ; + harq_processes_dest->rb_alloc_odd[2] = current_harq_processes->rb_alloc_odd[2] ; + harq_processes_dest->rb_alloc_odd[3] = current_harq_processes->rb_alloc_odd[3] ; + harq_processes_dest->round = current_harq_processes->round ; + harq_processes_dest->rvidx = current_harq_processes->rvidx ; + harq_processes_dest->status = current_harq_processes->status ; + harq_processes_dest->vrb_type = current_harq_processes->vrb_type ; + } /*void copy_ack_struct(nr_harq_status_t *harq_ack_dest, nr_harq_status_t *current_harq_ack) -{ - memcpy(harq_ack_dest, current_harq_ack, sizeof(nr_harq_status_t)); -}*/ + { + memcpy(harq_ack_dest, current_harq_ack, sizeof(nr_harq_status_t)); + }*/ -#if 0 -void ue_pdsch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, PDSCH_t pdsch, NR_UE_DLSCH_t *dlsch0, NR_UE_DLSCH_t *dlsch1, int s0, int s1, int abstraction_flag) { +void nr_ue_pdsch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, PDSCH_t pdsch, NR_UE_DLSCH_t *dlsch0, NR_UE_DLSCH_t *dlsch1, int s0, int s1) { int nr_tti_rx = proc->nr_tti_rx; int m; @@ -4419,23 +3716,23 @@ void ue_pdsch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id if (dlsch0 && (!dlsch1)) { harq_pid = dlsch0->current_harq_pid; - LOG_D(PHY,"[UE %d] PDSCH active in nr_tti_rx %d, harq_pid %d Symbol %d\n",ue->Mod_id,nr_tti_rx,harq_pid,m); + LOG_D(PHY,"[UE %d] PDSCH type %d active in nr_tti_rx %d, harq_pid %d Symbol %d\n",ue->Mod_id,pdsch,nr_tti_rx,harq_pid,m); if ((pdsch==PDSCH) && (ue->transmission_mode[eNB_id] == 5) && - (dlsch0->harq_processes[harq_pid].dl_power_off==0) && + (dlsch0->harq_processes[harq_pid]->dl_power_off==0) && (ue->use_ia_receiver ==1)) { dual_stream_UE = 1; eNB_id_i = ue->n_connected_eNB; - i_mod = dlsch0->harq_processes[harq_pid].Qm; + i_mod = dlsch0->harq_processes[harq_pid]->Qm; } else if((pdsch==PDSCH) && (ue->transmission_mode[eNB_id]==3)) - { + { dual_stream_UE = rx_IC_dual_stream; eNB_id_i = eNB_id; i_mod = 0; - } + } else { dual_stream_UE = 0; eNB_id_i = eNB_id+1; @@ -4446,17 +3743,17 @@ void ue_pdsch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id if (ue->transmission_mode[eNB_id]==7) { if (ue->frame_parms.Ncp==0) { //if ((m==3) || (m==6) || (m==9) || (m==12)) - //LOG_D(PHY,"[UE %d] dlsch->active in nr_tti_rx %d => %d, l=%d\n",phy_vars_ue->Mod_id,nr_tti_rx,phy_vars_ue->dlsch_ue[eNB_id][0]->active, l); - //lte_dl_bf_channel_estimation(ue,eNB_id,0,nr_tti_rx*2+(m>6?1:0),5,m); + //LOG_D(PHY,"[UE %d] dlsch->active in nr_tti_rx %d => %d, l=%d\n",phy_vars_ue->Mod_id,nr_tti_rx,phy_vars_ue->dlsch_ue[eNB_id][0]->active, l); + //lte_dl_bf_channel_estimation(ue,eNB_id,0,nr_tti_rx*2+(m>6?1:0),5,m); } else { LOG_E(PHY,"[UE %d]Beamforming channel estimation not supported yet for TM7 extented CP.\n",ue->Mod_id); } } - if ((m==s0) && (m<4)) - first_symbol_flag = 1; + if ((m==s0) && (m<3)) + first_symbol_flag = 1; else - first_symbol_flag = 0; + first_symbol_flag = 0; #if UE_TIMING_TRACE uint8_t slot = 0; if(m >= ue->frame_parms.symbols_per_tti>>1) @@ -4464,7 +3761,7 @@ void ue_pdsch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id start_meas(&ue->dlsch_llr_stats_parallelization[ue->current_thread_id[nr_tti_rx]][slot]); #endif // process DLSCH received in first slot - rx_pdsch(ue, + nr_rx_pdsch(ue, pdsch, eNB_id, eNB_id_i, @@ -4478,22 +3775,20 @@ void ue_pdsch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id #if UE_TIMING_TRACE stop_meas(&ue->dlsch_llr_stats_parallelization[ue->current_thread_id[nr_tti_rx]][slot]); #if DISABLE_LOG_X - printf("[AbsSFN %d.%d] LLR Computation Symbol %d %5.2f \n",proc->frame_rx,nr_tti_rx,m,ue->dlsch_llr_stats_parallelization[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); + printf("[AbsSFN %d.%d] LLR Computation Symbol %d %5.2f \n",proc->frame_rx,nr_tti_rx,m,ue->dlsch_llr_stats_parallelization[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); #else - LOG_D(PHY, "[AbsSFN %d.%d] LLR Computation Symbol %d %5.2f \n",proc->frame_rx,nr_tti_rx,m,ue->dlsch_llr_stats_parallelization[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); + LOG_D(PHY, "[AbsSFN %d.%d] LLR Computation Symbol %d %5.2f \n",proc->frame_rx,nr_tti_rx,m,ue->dlsch_llr_stats_parallelization[ue->current_thread_id[nr_tti_rx]][slot].p_time/(cpuf*1000.0)); #endif #endif if(first_symbol_flag) - { + { proc->first_symbol_available = 1; - } + } } // CRNTI active } } -#endif - #if 0 void process_rar(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, runmode_t mode, int abstraction_flag) { @@ -4507,43 +3802,43 @@ void process_rar(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, runmod uint8_t next2_thread_id = next1_thread_id== (RX_NB_TH-1) ? 0:(next1_thread_id+1); LOG_D(PHY,"[UE %d][RAPROC] Frame %d nr_tti_rx %d Received RAR mode %d\n", - ue->Mod_id, - frame_rx, - nr_tti_rx, ue->UE_mode[eNB_id]); + ue->Mod_id, + frame_rx, + nr_tti_rx, ue->UE_mode[eNB_id]); if (ue->mac_enabled == 1) { if ((ue->UE_mode[eNB_id] != PUSCH) && - (ue->prach_resources[eNB_id]->Msg3!=NULL)) { + (ue->prach_resources[eNB_id]->Msg3!=NULL)) { LOG_D(PHY,"[UE %d][RAPROC] Frame %d nr_tti_rx %d Invoking MAC for RAR (current preamble %d)\n", ue->Mod_id,frame_rx, nr_tti_rx, ue->prach_resources[eNB_id]->ra_PreambleIndex); -/* timing_advance = mac_xface->ue_process_rar(ue->Mod_id, - ue->CC_id, - frame_rx, - ue->prach_resources[eNB_id]->ra_RNTI, - dlsch0->harq_processes[0]->b, - &ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, - ue->prach_resources[eNB_id]->ra_PreambleIndex, - dlsch0->harq_processes[0]->b); // alter the 'b' buffer so it contains only the selected RAR header and RAR payload -*/ + /* timing_advance = mac_xface->ue_process_rar(ue->Mod_id, + ue->CC_id, + frame_rx, + ue->prach_resources[eNB_id]->ra_RNTI, + dlsch0->harq_processes[0]->b, + &ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, + ue->prach_resources[eNB_id]->ra_PreambleIndex, + dlsch0->harq_processes[0]->b); // alter the 'b' buffer so it contains only the selected RAR header and RAR payload + */ ue->pdcch_vars[next1_thread_id][eNB_id]->crnti = ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti; ue->pdcch_vars[next2_thread_id][eNB_id]->crnti = ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti; if (timing_advance!=0xffff) { - LOG_D(PHY,"[UE %d][RAPROC] Frame %d nr_tti_rx %d Got rnti %x and timing advance %d from RAR\n", + LOG_D(PHY,"[UE %d][RAPROC] Frame %d nr_tti_rx %d Got rnti %x and timing advance %d from RAR\n", ue->Mod_id, frame_rx, nr_tti_rx, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti, timing_advance); - // remember this c-rnti is still a tc-rnti + // remember this c-rnti is still a tc-rnti - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti_is_temporary = 1; + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->crnti_is_temporary = 1; //timing_advance = 0; nr_process_timing_advance_rar(ue,proc,timing_advance); @@ -4551,10 +3846,10 @@ void process_rar(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, runmod if (mode!=debug_prach) { ue->ulsch_Msg3_active[eNB_id]=1; nr_get_Msg3_alloc(&ue->frame_parms, - nr_tti_rx, - frame_rx, - &ue->ulsch_Msg3_frame[eNB_id], - &ue->ulsch_Msg3_subframe[eNB_id]); + nr_tti_rx, + frame_rx, + &ue->ulsch_Msg3_frame[eNB_id], + &ue->ulsch_Msg3_subframe[eNB_id]); LOG_D(PHY,"[UE %d][RAPROC] Got Msg3_alloc Frame %d nr_tti_rx %d: Msg3_frame %d, Msg3_subframe %d\n", ue->Mod_id, @@ -4563,8 +3858,8 @@ void process_rar(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, runmod ue->ulsch_Msg3_frame[eNB_id], ue->ulsch_Msg3_subframe[eNB_id]); harq_pid = nr_subframe2harq_pid(&ue->frame_parms, - ue->ulsch_Msg3_frame[eNB_id], - ue->ulsch_Msg3_subframe[eNB_id]); + ue->ulsch_Msg3_frame[eNB_id], + ue->ulsch_Msg3_subframe[eNB_id]); ue->ulsch[eNB_id]->harq_processes[harq_pid]->round = 0; ue->UE_mode[eNB_id] = RA_RESPONSE; @@ -4573,9 +3868,9 @@ void process_rar(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, runmod ue->ulsch_no_allocation_counter[eNB_id] = 0; } } else { // PRACH preamble doesn't match RAR - LOG_W(PHY,"[UE %d][RAPROC] Received RAR preamble (%d) doesn't match !!!\n", - ue->Mod_id, - ue->prach_resources[eNB_id]->ra_PreambleIndex); + LOG_W(PHY,"[UE %d][RAPROC] Received RAR preamble (%d) doesn't match !!!\n", + ue->Mod_id, + ue->prach_resources[eNB_id]->ra_PreambleIndex); } } // mode != PUSCH } @@ -4586,16 +3881,16 @@ void process_rar(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, runmod } } +#endif -void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, +void nr_ue_dlsch_procedures(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, int eNB_id, PDSCH_t pdsch, NR_UE_DLSCH_t *dlsch0, NR_UE_DLSCH_t *dlsch1, int *dlsch_errors, - runmode_t mode, - int abstraction_flag) { + runmode_t mode) { int harq_pid; int frame_rx = proc->frame_rx; @@ -4605,9 +3900,13 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, NR_UE_PDSCH *pdsch_vars; uint8_t is_cw0_active = 0; uint8_t is_cw1_active = 0; + // to be updated by higher layer + uint8_t nb_re_dmrs = 6; + uint16_t length_dmrs = 1; + uint16_t nb_symb_sch = 8; if (dlsch0==NULL) - AssertFatal(0,"dlsch0 should be defined at this level \n"); + AssertFatal(0,"dlsch0 should be defined at this level \n"); harq_pid = dlsch0->current_harq_pid; is_cw0_active = dlsch0->harq_processes[harq_pid]->status; @@ -4615,18 +3914,18 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, if(dlsch1) is_cw1_active = dlsch1->harq_processes[harq_pid]->status; - LOG_D(PHY,"AbsSubframe %d.%d Start Turbo Decoder for CW0 [harq_pid %d] ? %d \n", frame_rx%1024, nr_tti_rx, harq_pid, is_cw0_active); - LOG_D(PHY,"AbsSubframe %d.%d Start Turbo Decoder for CW1 [harq_pid %d] ? %d \n", frame_rx%1024, nr_tti_rx, harq_pid, is_cw1_active); + LOG_D(PHY,"AbsSubframe %d.%d Start LDPC Decoder for CW0 [harq_pid %d] ? %d \n", frame_rx%1024, nr_tti_rx, harq_pid, is_cw0_active); + LOG_D(PHY,"AbsSubframe %d.%d Start LDPC Decoder for CW1 [harq_pid %d] ? %d \n", frame_rx%1024, nr_tti_rx, harq_pid, is_cw1_active); if(is_cw0_active && is_cw1_active) - { + { dlsch0->Kmimo = 2; dlsch1->Kmimo = 2; - } + } else - { + { dlsch0->Kmimo = 1; - } + } if (1) { switch (pdsch) { case SI_PDSCH: @@ -4658,49 +3957,45 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, if (pdsch==RA_PDSCH) { if (ue->prach_resources[eNB_id]!=NULL) - dlsch0->rnti = ue->prach_resources[eNB_id]->ra_RNTI; + dlsch0->rnti = ue->prach_resources[eNB_id]->ra_RNTI; else { - LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d: FATAL, prach_resources is NULL\n",ue->Mod_id,frame_rx,nr_tti_rx); - //mac_xface->macphy_exit("prach_resources is NULL"); - return; + LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d: FATAL, prach_resources is NULL\n",ue->Mod_id,frame_rx,nr_tti_rx); + //mac_xface->macphy_exit("prach_resources is NULL"); + return; } } - if (abstraction_flag == 0) { - // start turbo decode for CW 0 - dlsch0->harq_processes[harq_pid]->G = get_G(&ue->frame_parms, - dlsch0->harq_processes[harq_pid]->nb_rb, - dlsch0->harq_processes[harq_pid]->rb_alloc_even, - dlsch0->harq_processes[harq_pid]->Qm, - dlsch0->harq_processes[harq_pid]->Nl, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - frame_rx, - nr_tti_rx, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id]); + // start ldpc decode for CW 0 + dlsch0->harq_processes[harq_pid]->G = nr_get_G(dlsch0->harq_processes[harq_pid]->nb_rb, + nb_symb_sch, + nb_re_dmrs, + length_dmrs, + dlsch0->harq_processes[harq_pid]->Qm, + dlsch0->harq_processes[harq_pid]->Nl); #if UE_TIMING_TRACE start_meas(&ue->dlsch_unscrambling_stats); #endif - dlsch_unscrambling(&ue->frame_parms, - 0, - dlsch0, - dlsch0->harq_processes[harq_pid]->G, - pdsch_vars->llr[0], - 0, - nr_tti_rx<<1); + nr_dlsch_unscrambling(pdsch_vars->llr[0], + dlsch0->harq_processes[harq_pid]->G, + 0, + ue->frame_parms.Nid_cell, + dlsch0->rnti); + + #if UE_TIMING_TRACE stop_meas(&ue->dlsch_unscrambling_stats); #endif #if 0 - LOG_I(PHY," ------ start turbo decoder for AbsSubframe %d.%d / %d ------ \n", frame_rx, nr_tti_rx, harq_pid); - LOG_I(PHY,"start turbo decode for CW 0 for AbsSubframe %d.%d / %d --> nb_rb %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->nb_rb); - LOG_I(PHY,"start turbo decode for CW 0 for AbsSubframe %d.%d / %d --> rb_alloc_even %x \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->rb_alloc_even); - LOG_I(PHY,"start turbo decode for CW 0 for AbsSubframe %d.%d / %d --> Qm %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->Qm); - LOG_I(PHY,"start turbo decode for CW 0 for AbsSubframe %d.%d / %d --> Nl %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->Nl); - LOG_I(PHY,"start turbo decode for CW 0 for AbsSubframe %d.%d / %d --> G %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->G); - LOG_I(PHY,"start turbo decode for CW 0 for AbsSubframe %d.%d / %d --> Kmimo %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->Kmimo); - LOG_I(PHY,"start turbo decode for CW 0 for AbsSubframe %d.%d / %d --> Pdcch Sym %d \n", frame_rx, nr_tti_rx, harq_pid, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols); + LOG_I(PHY," ------ start ldpc decoder for AbsSubframe %d.%d / %d ------ \n", frame_rx, nr_tti_rx, harq_pid); + LOG_I(PHY,"start ldpc decode for CW 0 for AbsSubframe %d.%d / %d --> nb_rb %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->nb_rb); + LOG_I(PHY,"start ldpc decode for CW 0 for AbsSubframe %d.%d / %d --> rb_alloc_even %x \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->rb_alloc_even); + LOG_I(PHY,"start ldpc decode for CW 0 for AbsSubframe %d.%d / %d --> Qm %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->Qm); + LOG_I(PHY,"start ldpc decode for CW 0 for AbsSubframe %d.%d / %d --> Nl %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->Nl); + LOG_I(PHY,"start ldpc decode for CW 0 for AbsSubframe %d.%d / %d --> G %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->harq_processes[harq_pid]->G); + LOG_I(PHY,"start ldpc decode for CW 0 for AbsSubframe %d.%d / %d --> Kmimo %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch0->Kmimo); + LOG_I(PHY,"start ldpc decode for CW 0 for AbsSubframe %d.%d / %d --> Pdcch Sym %d \n", frame_rx, nr_tti_rx, harq_pid, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols); #endif #if UE_TIMING_TRACE @@ -4708,23 +4003,25 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, #endif #ifdef UE_DLSCH_PARALLELISATION - ret = dlsch_decoding_mthread(ue,proc,eNB_id, + ret = nr_dlsch_decoding_mthread(ue,proc,eNB_id, pdsch_vars->llr[0], &ue->frame_parms, dlsch0, dlsch0->harq_processes[harq_pid], frame_rx, + nb_symb_sch, nr_tti_rx, harq_pid, pdsch==PDSCH?1:0, dlsch0->harq_processes[harq_pid]->TBS>256?1:0); #else - ret = dlsch_decoding(ue, + ret = nr_dlsch_decoding(ue, pdsch_vars->llr[0], &ue->frame_parms, dlsch0, dlsch0->harq_processes[harq_pid], frame_rx, + nb_symb_sch, nr_tti_rx, harq_pid, pdsch==PDSCH?1:0,//proc->decoder_switch, @@ -4737,50 +4034,45 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, #if DISABLE_LOG_X printf(" --> Unscrambling for CW0 %5.3f\n", (ue->dlsch_unscrambling_stats.p_time)/(cpuf*1000.0)); - printf("AbsSubframe %d.%d --> Turbo Decoding for CW0 %5.3f\n", + printf("AbsSubframe %d.%d --> LDPC Decoding for CW0 %5.3f\n", frame_rx%1024, nr_tti_rx,(ue->dlsch_decoding_stats[ue->current_thread_id[nr_tti_rx]].p_time)/(cpuf*1000.0)); #else LOG_I(PHY, " --> Unscrambling for CW0 %5.3f\n", (ue->dlsch_unscrambling_stats.p_time)/(cpuf*1000.0)); - LOG_I(PHY, "AbsSubframe %d.%d --> Turbo Decoding for CW0 %5.3f\n", + LOG_I(PHY, "AbsSubframe %d.%d --> LDPC Decoding for CW0 %5.3f\n", frame_rx%1024, nr_tti_rx,(ue->dlsch_decoding_stats[ue->current_thread_id[nr_tti_rx]].p_time)/(cpuf*1000.0)); #endif #endif if(is_cw1_active) { - // start turbo decode for CW 1 - dlsch1->harq_processes[harq_pid]->G = get_G(&ue->frame_parms, - dlsch1->harq_processes[harq_pid]->nb_rb, - dlsch1->harq_processes[harq_pid]->rb_alloc_even, - dlsch1->harq_processes[harq_pid]->Qm, - dlsch1->harq_processes[harq_pid]->Nl, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - frame_rx, - nr_tti_rx, - ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id]); + // start ldpc decode for CW 1 + dlsch1->harq_processes[harq_pid]->G = nr_get_G(dlsch1->harq_processes[harq_pid]->nb_rb, + nb_symb_sch, + nb_re_dmrs, + length_dmrs, + dlsch1->harq_processes[harq_pid]->Qm, + dlsch1->harq_processes[harq_pid]->Nl); #if UE_TIMING_TRACE start_meas(&ue->dlsch_unscrambling_stats); #endif - dlsch_unscrambling(&ue->frame_parms, - 0, - dlsch1, - dlsch1->harq_processes[harq_pid]->G, - pdsch_vars->llr[1], - 1, - nr_tti_rx<<1); + nr_dlsch_unscrambling(pdsch_vars->llr[1], + dlsch1->harq_processes[harq_pid]->G, + 0, + ue->frame_parms.Nid_cell, + dlsch1->rnti); #if UE_TIMING_TRACE stop_meas(&ue->dlsch_unscrambling_stats); #endif #if 0 - LOG_I(PHY,"start turbo decode for CW 1 for AbsSubframe %d.%d / %d --> nb_rb %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->nb_rb); - LOG_I(PHY,"start turbo decode for CW 1 for AbsSubframe %d.%d / %d --> rb_alloc_even %x \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->rb_alloc_even); - LOG_I(PHY,"start turbo decode for CW 1 for AbsSubframe %d.%d / %d --> Qm %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->Qm); - LOG_I(PHY,"start turbo decode for CW 1 for AbsSubframe %d.%d / %d --> Nl %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->Nl); - LOG_I(PHY,"start turbo decode for CW 1 for AbsSubframe %d.%d / %d --> G %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->G); - LOG_I(PHY,"start turbo decode for CW 1 for AbsSubframe %d.%d / %d --> Kmimo %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->Kmimo); - LOG_I(PHY,"start turbo decode for CW 1 for AbsSubframe %d.%d / %d --> Pdcch Sym %d \n", frame_rx, nr_tti_rx, harq_pid, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols); + LOG_I(PHY,"start ldpc decode for CW 1 for AbsSubframe %d.%d / %d --> nb_rb %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->nb_rb); + LOG_I(PHY,"start ldpc decode for CW 1 for AbsSubframe %d.%d / %d --> rb_alloc_even %x \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->rb_alloc_even); + LOG_I(PHY,"start ldpc decode for CW 1 for AbsSubframe %d.%d / %d --> Qm %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->Qm); + LOG_I(PHY,"start ldpc decode for CW 1 for AbsSubframe %d.%d / %d --> Nl %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->Nl); + LOG_I(PHY,"start ldpc decode for CW 1 for AbsSubframe %d.%d / %d --> G %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->harq_processes[harq_pid]->G); + LOG_I(PHY,"start ldpc decode for CW 1 for AbsSubframe %d.%d / %d --> Kmimo %d \n", frame_rx, nr_tti_rx, harq_pid, dlsch1->Kmimo); + LOG_I(PHY,"start ldpc decode for CW 1 for AbsSubframe %d.%d / %d --> Pdcch Sym %d \n", frame_rx, nr_tti_rx, harq_pid, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols); #endif #if UE_TIMING_TRACE @@ -4788,24 +4080,26 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, #endif #ifdef UE_DLSCH_PARALLELISATION - ret1 = dlsch_decoding_mthread(ue,proc, eNB_id, + ret1 = nr_dlsch_decoding_mthread(ue,proc, eNB_id, pdsch_vars->llr[1], &ue->frame_parms, dlsch1, dlsch1->harq_processes[harq_pid], frame_rx, + nb_symb_sch, nr_tti_rx, harq_pid, pdsch==PDSCH?1:0, dlsch1->harq_processes[harq_pid]->TBS>256?1:0); #else - ret1 = dlsch_decoding(ue, + ret1 = nr_dlsch_decoding(ue, pdsch_vars->llr[1], &ue->frame_parms, dlsch1, dlsch1->harq_processes[harq_pid], frame_rx, + nb_symb_sch, nr_tti_rx, harq_pid, pdsch==PDSCH?1:0,//proc->decoder_switch, @@ -4818,189 +4112,28 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, #if DISABLE_LOG_X printf(" --> Unscrambling for CW1 %5.3f\n", (ue->dlsch_unscrambling_stats.p_time)/(cpuf*1000.0)); - printf("AbsSubframe %d.%d --> Turbo Decoding for CW1 %5.3f\n", + printf("AbsSubframe %d.%d --> ldpc Decoding for CW1 %5.3f\n", frame_rx%1024, nr_tti_rx,(ue->dlsch_decoding_stats[ue->current_thread_id[nr_tti_rx]].p_time)/(cpuf*1000.0)); #else LOG_D(PHY, " --> Unscrambling for CW1 %5.3f\n", (ue->dlsch_unscrambling_stats.p_time)/(cpuf*1000.0)); - LOG_D(PHY, "AbsSubframe %d.%d --> Turbo Decoding for CW1 %5.3f\n", + LOG_D(PHY, "AbsSubframe %d.%d --> ldpc Decoding for CW1 %5.3f\n", frame_rx%1024, nr_tti_rx,(ue->dlsch_decoding_stats[ue->current_thread_id[nr_tti_rx]].p_time)/(cpuf*1000.0)); #endif #endif - LOG_I(PHY,"AbsSubframe %d.%d --> Turbo Decoding for CW1 %5.3f\n", + LOG_I(PHY,"AbsSubframe %d.%d --> ldpc Decoding for CW1 %5.3f\n", frame_rx%1024, nr_tti_rx,(ue->dlsch_decoding_stats[ue->current_thread_id[nr_tti_rx]].p_time)/(cpuf*1000.0)); } - LOG_D(PHY," ------ end turbo decoder for AbsSubframe %d.%d ------ \n", frame_rx, nr_tti_rx); + LOG_D(PHY," ------ end ldpc decoder for AbsSubframe %d.%d ------ \n", frame_rx, nr_tti_rx); } - else { - LOG_D(PHY,"Calling dlsch_decoding_emul ...\n"); -#ifdef PHY_ABSTRACTION - ret = dlsch_decoding_emul(ue, - nr_tti_rx, - pdsch, - eNB_id); -#endif - } - // Check CRC for CW 0 - if (ret == (1+dlsch0->max_turbo_iterations)) { - *dlsch_errors=*dlsch_errors+1; - - if(dlsch0->rnti != 0xffff) - { - LOG_D(PHY,"[UE %d][PDSCH %x/%d] AbsSubframe %d.%d : DLSCH CW0 in error (rv %d,round %d, mcs %d,TBS %d)\n", - ue->Mod_id,dlsch0->rnti, - harq_pid,frame_rx,nr_tti_rx, - dlsch0->harq_processes[harq_pid]->rvidx, - dlsch0->harq_processes[harq_pid]->round, - dlsch0->harq_processes[harq_pid]->mcs, - dlsch0->harq_processes[harq_pid]->TBS); - } - - - } else { - if(dlsch0->rnti != 0xffff) - { - LOG_D(PHY,"[UE %d][PDSCH %x/%d] AbsSubframe %d.%d : Received DLSCH CW0 (rv %d,round %d, mcs %d,TBS %d)\n", - ue->Mod_id,dlsch0->rnti, - harq_pid,frame_rx,nr_tti_rx, - dlsch0->harq_processes[harq_pid]->rvidx, - dlsch0->harq_processes[harq_pid]->round, - dlsch0->harq_processes[harq_pid]->mcs, - dlsch0->harq_processes[harq_pid]->TBS); - } - -#ifdef DEBUG_DLSCH - int j; - LOG_D(PHY,"dlsch harq_pid %d (rx): \n",dlsch0->current_harq_pid); - - for (j=0; jharq_processes[dlsch0->current_harq_pid]->TBS>>3; j++) - LOG_T(PHY,"%x.",dlsch0->harq_processes[dlsch0->current_harq_pid]->b[j]); - - LOG_T(PHY,"\n"); -#endif - - - switch (pdsch) { - case PDSCH: - /* if (ue->mac_enabled == 1) - mac_xface->ue_send_sdu(ue->Mod_id, - CC_id, - frame_rx, - nr_tti_rx, - dlsch0->harq_processes[dlsch0->current_harq_pid]->b, - dlsch0->harq_processes[dlsch0->current_harq_pid]->TBS>>3, - eNB_id);*/ - break; - case SI_PDSCH: - /*mac_xface->ue_decode_si(ue->Mod_id, - CC_id, - frame_rx, - eNB_id, - ue->dlsch_SI[eNB_id]->harq_processes[0]->b, - ue->dlsch_SI[eNB_id]->harq_processes[0]->TBS>>3);*/ - break; - case P_PDSCH: - /* if (ue->mac_enabled == 1) - mac_xface->ue_decode_p(ue->Mod_id, - CC_id, - frame_rx, - eNB_id, - ue->dlsch_SI[eNB_id]->harq_processes[0]->b, - ue->dlsch_SI[eNB_id]->harq_processes[0]->TBS>>3);*/ - break; - case RA_PDSCH: - if (ue->mac_enabled == 1) - process_rar(ue,proc,eNB_id,mode,abstraction_flag); - break; - case PDSCH1: - LOG_E(PHY,"Shouldn't have PDSCH1 yet, come back later\n"); - AssertFatal(1==0,"exiting"); - break; - case PMCH: - LOG_E(PHY,"Shouldn't have PMCH here\n"); - AssertFatal(1==0,"exiting"); - break; - - } - ue->total_TBS[eNB_id] = ue->total_TBS[eNB_id] + - dlsch0->harq_processes[dlsch0->current_harq_pid]->TBS; - ue->total_received_bits[eNB_id] = ue->total_TBS[eNB_id] + - dlsch0->harq_processes[dlsch0->current_harq_pid]->TBS; - } - // Check CRC for CW 1 - if(is_cw1_active) - { - if (ret1 == (1+dlsch0->max_turbo_iterations)) { - LOG_I(PHY,"[UE %d][PDSCH %x/%d] Frame %d nr_tti_rx %d DLSCH CW1 in error (rv %d,mcs %d,TBS %d)\n", - ue->Mod_id,dlsch0->rnti, - harq_pid,frame_rx,nr_tti_rx, - dlsch0->harq_processes[harq_pid]->rvidx, - dlsch0->harq_processes[harq_pid]->mcs, - dlsch0->harq_processes[harq_pid]->TBS); - - } else { - LOG_I(PHY,"[UE %d][PDSCH %x/%d] Frame %d nr_tti_rx %d: Received DLSCH CW1 (rv %d,mcs %d,TBS %d)\n", - ue->Mod_id,dlsch0->rnti, - harq_pid,frame_rx,nr_tti_rx, - dlsch0->harq_processes[harq_pid]->rvidx, - dlsch0->harq_processes[harq_pid]->mcs, - dlsch0->harq_processes[harq_pid]->TBS); - - - if (ue->mac_enabled == 1) { - switch (pdsch) { - case PDSCH: - if(is_cw1_active) - /*mac_xface->ue_send_sdu(ue->Mod_id, - CC_id, - frame_rx, - nr_tti_rx, - dlsch1->harq_processes[dlsch1->current_harq_pid]->b, - dlsch1->harq_processes[dlsch1->current_harq_pid]->TBS>>3, - eNB_id);*/ - break; - case SI_PDSCH: - case P_PDSCH: - case RA_PDSCH: - case PDSCH1: - case PMCH: - AssertFatal(0,"exiting"); - break; - } - } - } - } - -#ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[UE %d][PDSCH %x/%d] Frame %d nr_tti_rx %d: PDSCH/DLSCH decoding iter %d (mcs %d, rv %d, TBS %d)\n", - ue->Mod_id, - dlsch0->rnti,harq_pid, - frame_rx,nr_tti_rx,ret, - dlsch0->harq_processes[harq_pid]->mcs, - dlsch0->harq_processes[harq_pid]->rvidx, - dlsch0->harq_processes[harq_pid]->TBS); - - if (frame_rx%100==0) { - LOG_D(PHY,"[UE %d][PDSCH %x] Frame %d nr_tti_rx %d dlsch_errors %d, dlsch_received %d, dlsch_fer %d, current_dlsch_cqi %d\n", - ue->Mod_id,dlsch0->rnti, - frame_rx,nr_tti_rx, - ue->dlsch_errors[eNB_id], - ue->dlsch_received[eNB_id], - ue->dlsch_fer[eNB_id], - ue->measurements.wideband_cqi_tot[eNB_id]); - } - -#endif - - } } -#endif + /*! * \brief This is the UE synchronize thread. @@ -5012,174 +4145,174 @@ void ue_dlsch_procedures(PHY_VARS_NR_UE *ue, #define FIFO_PRIORITY 40 void *UE_thread_slot1_dl_processing(void *arg) { - static __thread int UE_dl_slot1_processing_retval; - struct rx_tx_thread_data *rtd = arg; - UE_nr_rxtx_proc_t *proc = rtd->proc; - PHY_VARS_NR_UE *ue = rtd->UE; + static __thread int UE_dl_slot1_processing_retval; + struct rx_tx_thread_data *rtd = arg; + UE_nr_rxtx_proc_t *proc = rtd->proc; + PHY_VARS_NR_UE *ue = rtd->UE; - int frame_rx; - uint8_t nr_tti_rx; - uint8_t pilot0; - uint8_t pilot1; - uint8_t slot1; + int frame_rx; + uint8_t nr_tti_rx; + uint8_t pilot0; + uint8_t pilot1; + uint8_t slot1; - uint8_t next_nr_tti_rx; - uint8_t next_subframe_slot0; + uint8_t next_nr_tti_rx; + uint8_t next_subframe_slot0; - proc->instance_cnt_slot1_dl_processing=-1; - proc->nr_tti_rx=proc->sub_frame_start; + proc->instance_cnt_slot1_dl_processing=-1; + proc->nr_tti_rx=proc->sub_frame_start; - char threadname[256]; - sprintf(threadname,"UE_thread_slot1_dl_processing_%d", proc->sub_frame_start); + char threadname[256]; + sprintf(threadname,"UE_thread_slot1_dl_processing_%d", proc->sub_frame_start); - cpu_set_t cpuset; - CPU_ZERO(&cpuset); - if ( (proc->sub_frame_start+1)%RX_NB_TH == 0 && threads.slot1_proc_one != -1 ) - CPU_SET(threads.slot1_proc_one, &cpuset); - if ( (proc->sub_frame_start+1)%RX_NB_TH == 1 && threads.slot1_proc_two != -1 ) - CPU_SET(threads.slot1_proc_two, &cpuset); - if ( (proc->sub_frame_start+1)%RX_NB_TH == 2 && threads.slot1_proc_three != -1 ) - CPU_SET(threads.slot1_proc_three, &cpuset); + cpu_set_t cpuset; + CPU_ZERO(&cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 0 && threads.slot1_proc_one != -1 ) + CPU_SET(threads.slot1_proc_one, &cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 1 && threads.slot1_proc_two != -1 ) + CPU_SET(threads.slot1_proc_two, &cpuset); + if ( (proc->sub_frame_start+1)%RX_NB_TH == 2 && threads.slot1_proc_three != -1 ) + CPU_SET(threads.slot1_proc_three, &cpuset); - init_thread(900000,1000000 , FIFO_PRIORITY-1, &cpuset, - threadname); + init_thread(900000,1000000 , FIFO_PRIORITY-1, &cpuset, + threadname); - while (!oai_exit) { - if (pthread_mutex_lock(&proc->mutex_slot1_dl_processing) != 0) { - LOG_E( PHY, "[SCHED][UE] error locking mutex for UE slot1 dl processing\n" ); - exit_fun("nothing to add"); - } - while (proc->instance_cnt_slot1_dl_processing < 0) { - // most of the time, the thread is waiting here - pthread_cond_wait( &proc->cond_slot1_dl_processing, &proc->mutex_slot1_dl_processing ); - } - if (pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) != 0) { - LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE slot1 dl processing \n" ); - exit_fun("nothing to add"); - } + while (!oai_exit) { + if (pthread_mutex_lock(&proc->mutex_slot1_dl_processing) != 0) { + LOG_E( PHY, "[SCHED][UE] error locking mutex for UE slot1 dl processing\n" ); + exit_fun("nothing to add"); + } + while (proc->instance_cnt_slot1_dl_processing < 0) { + // most of the time, the thread is waiting here + pthread_cond_wait( &proc->cond_slot1_dl_processing, &proc->mutex_slot1_dl_processing ); + } + if (pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) != 0) { + LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE slot1 dl processing \n" ); + exit_fun("nothing to add"); + } - /*for(int th_idx=0; th_idx< RX_NB_TH; th_idx++) - { - frame_rx = ue->proc.proc_rxtx[0].frame_rx; - nr_tti_rx = ue->proc.proc_rxtx[0].nr_tti_rx; - printf("AbsSubframe %d.%d execute dl slot1 processing \n", frame_rx, nr_tti_rx); - }*/ - frame_rx = proc->frame_rx; - nr_tti_rx = proc->nr_tti_rx; - next_nr_tti_rx = (1+nr_tti_rx)%10; - next_subframe_slot0 = next_nr_tti_rx<<1; + /*for(int th_idx=0; th_idx< RX_NB_TH; th_idx++) + { + frame_rx = ue->proc.proc_rxtx[0].frame_rx; + nr_tti_rx = ue->proc.proc_rxtx[0].nr_tti_rx; + printf("AbsSubframe %d.%d execute dl slot1 processing \n", frame_rx, nr_tti_rx); + }*/ + frame_rx = proc->frame_rx; + nr_tti_rx = proc->nr_tti_rx; + next_nr_tti_rx = (1+nr_tti_rx)%10; + next_subframe_slot0 = next_nr_tti_rx<<1; - slot1 = (nr_tti_rx<<1) + 1; - pilot0 = 0; + slot1 = (nr_tti_rx<<1) + 1; + pilot0 = 0; - //printf("AbsSubframe %d.%d execute dl slot1 processing \n", frame_rx, nr_tti_rx); + //printf("AbsSubframe %d.%d execute dl slot1 processing \n", frame_rx, nr_tti_rx); - if (ue->frame_parms.Ncp == 0) { // normal prefix - pilot1 = 4; - } else { // extended prefix - pilot1 = 3; - } + if (ue->frame_parms.Ncp == 0) { // normal prefix + pilot1 = 4; + } else { // extended prefix + pilot1 = 3; + } - /**** Slot1 FE Processing ****/ + /**** Slot1 FE Processing ****/ #if UE_TIMING_TRACE - start_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1]); + start_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1]); #endif - // I- start dl slot1 processing - // do first symbol of next downlink nr_tti_rx for channel estimation - /* - // 1- perform FFT for pilot ofdm symbols first (ofdmSym0 next nr_tti_rx ofdmSym11) - if (nr_subframe_select(&ue->frame_parms,next_nr_tti_rx) != SF_UL) - { - front_end_fft(ue, - pilot0, - next_subframe_slot0, - 0, - 0); - } + // I- start dl slot1 processing + // do first symbol of next downlink nr_tti_rx for channel estimation + /* + // 1- perform FFT for pilot ofdm symbols first (ofdmSym0 next nr_tti_rx ofdmSym11) + if (nr_subframe_select(&ue->frame_parms,next_nr_tti_rx) != SF_UL) + { + front_end_fft(ue, + pilot0, + next_subframe_slot0, + 0, + 0); + } - front_end_fft(ue, - pilot1, - slot1, - 0, - 0); - */ - // 1- perform FFT - for (int l=1; lframe_parms.symbols_per_tti>>1; l++) - { - //if( (l != pilot0) && (l != pilot1)) - { + front_end_fft(ue, + pilot1, + slot1, + 0, + 0); + */ + // 1- perform FFT + for (int l=1; lframe_parms.symbols_per_tti>>1; l++) + { + //if( (l != pilot0) && (l != pilot1)) + { #if UE_TIMING_TRACE - start_meas(&ue->ofdm_demod_stats); + start_meas(&ue->ofdm_demod_stats); #endif - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_IN); - //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,l); - front_end_fft(ue, + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_IN); + //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,l); + front_end_fft(ue, l, slot1, 0, 0); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_OUT); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_OUT); #if UE_TIMING_TRACE - stop_meas(&ue->ofdm_demod_stats); + stop_meas(&ue->ofdm_demod_stats); #endif - } - } // for l=1..l2 + } + } // for l=1..l2 - if (nr_subframe_select(&ue->frame_parms,next_nr_tti_rx) != SF_UL) - { - //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,next_subframe_slot0,pilot0); - front_end_fft(ue, - pilot0, - next_subframe_slot0, - 0, - 0); - } + if (nr_subframe_select(&ue->frame_parms,next_nr_tti_rx) != SF_UL) + { + //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,next_subframe_slot0,pilot0); + front_end_fft(ue, + pilot0, + next_subframe_slot0, + 0, + 0); + } - // 2- perform Channel Estimation for slot1 - for (int l=1; lframe_parms.symbols_per_tti>>1; l++) - { - if(l == pilot1) - { - //wait until channel estimation for pilot0/slot1 is available - uint32_t wait = 0; - while(proc->chan_est_pilot0_slot1_available == 0) - { - usleep(1); - wait++; - } - //printf("[slot1 dl processing] ChanEst symbol %d slot %d wait%d\n",l,slot1,wait); - } - //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,l); - front_end_chanEst(ue, - l, - slot1, - 0); - ue_measurement_procedures(l-1,ue,proc,0,1+(nr_tti_rx<<1),0,ue->mode); - } - //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,next_subframe_slot0,pilot0); - front_end_chanEst(ue, - pilot0, - next_subframe_slot0, - 0); + // 2- perform Channel Estimation for slot1 + for (int l=1; lframe_parms.symbols_per_tti>>1; l++) + { + if(l == pilot1) + { + //wait until channel estimation for pilot0/slot1 is available + uint32_t wait = 0; + while(proc->chan_est_pilot0_slot1_available == 0) + { + usleep(1); + wait++; + } + //printf("[slot1 dl processing] ChanEst symbol %d slot %d wait%d\n",l,slot1,wait); + } + //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,l); + front_end_chanEst(ue, + l, + slot1, + 0); + ue_measurement_procedures(l-1,ue,proc,0,1+(nr_tti_rx<<1),0,ue->mode); + } + //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,next_subframe_slot0,pilot0); + front_end_chanEst(ue, + pilot0, + next_subframe_slot0, + 0); - if ( (nr_tti_rx == 0) && (ue->decode_MIB == 1)) - { - ue_pbch_procedures(0,ue,proc,0); - } + if ( (nr_tti_rx == 0) && (ue->decode_MIB == 1)) + { + ue_pbch_procedures(0,ue,proc,0); + } - proc->chan_est_slot1_available = 1; - //printf("Set available slot 1channelEst to 1 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); - //printf(" [slot1 dl processing] ==> FFT/CHanEst Done for AbsSubframe %d.%d \n", proc->frame_rx, proc->nr_tti_rx); + proc->chan_est_slot1_available = 1; + //printf("Set available slot 1channelEst to 1 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); + //printf(" [slot1 dl processing] ==> FFT/CHanEst Done for AbsSubframe %d.%d \n", proc->frame_rx, proc->nr_tti_rx); - //printf(" [slot1 dl processing] ==> Start LLR Comuptation slot1 for AbsSubframe %d.%d \n", proc->frame_rx, proc->nr_tti_rx); + //printf(" [slot1 dl processing] ==> Start LLR Comuptation slot1 for AbsSubframe %d.%d \n", proc->frame_rx, proc->nr_tti_rx); #if UE_TIMING_TRACE - stop_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1]); + stop_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1]); #if DISABLE_LOG_X - printf("[AbsSFN %d.%d] Slot1: FFT + Channel Estimate + Pdsch Proc Slot0 %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1].p_time/(cpuf*1000.0)); + printf("[AbsSFN %d.%d] Slot1: FFT + Channel Estimate + Pdsch Proc Slot0 %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1].p_time/(cpuf*1000.0)); #else - LOG_D(PHY, "[AbsSFN %d.%d] Slot1: FFT + Channel Estimate + Pdsch Proc Slot0 %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1].p_time/(cpuf*1000.0)); + LOG_D(PHY, "[AbsSFN %d.%d] Slot1: FFT + Channel Estimate + Pdsch Proc Slot0 %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][1].p_time/(cpuf*1000.0)); #endif #endif @@ -5187,10 +4320,10 @@ void *UE_thread_slot1_dl_processing(void *arg) { //wait until pdcch is decoded uint32_t wait = 0; while(proc->dci_slot0_available == 0) - { + { usleep(1); wait++; - } + } //printf("[slot1 dl processing] AbsSubframe %d.%d LLR Computation Start wait DCI %d\n",frame_rx,nr_tti_rx,wait); @@ -5206,67 +4339,67 @@ void *UE_thread_slot1_dl_processing(void *arg) { // start slave thread for Pdsch Procedure (slot1) // do procedures for C-RNTI uint8_t eNB_id = 0; - uint8_t abstraction_flag = 0; - if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { - //wait until first ofdm symbol is processed - //wait = 0; - //while(proc->first_symbol_available == 0) - //{ - // usleep(1); - // wait++; - //} - //printf("[slot1 dl processing] AbsSubframe %d.%d LLR Computation Start wait First Ofdm Sym %d\n",frame_rx,nr_tti_rx,wait); - //VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); - ue_pdsch_procedures(ue, - proc, - eNB_id, - PDSCH, - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], - NULL, - (ue->frame_parms.symbols_per_tti>>1), - ue->frame_parms.symbols_per_tti-1, - abstraction_flag); - LOG_D(PHY," ------ end PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); - LOG_D(PHY," ------ --> PDSCH Turbo Decoder slot 0/1: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); + if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { + //wait until first ofdm symbol is processed + //wait = 0; + //while(proc->first_symbol_available == 0) + //{ + // usleep(1); + // wait++; + //} + //printf("[slot1 dl processing] AbsSubframe %d.%d LLR Computation Start wait First Ofdm Sym %d\n",frame_rx,nr_tti_rx,wait); + + //VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); + ue_pdsch_procedures(ue, + proc, + eNB_id, + PDSCH, + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], + NULL, + (ue->frame_parms.symbols_per_tti>>1), + ue->frame_parms.symbols_per_tti-1, + abstraction_flag); + LOG_D(PHY," ------ end PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); + LOG_D(PHY," ------ --> PDSCH Turbo Decoder slot 0/1: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); } // do procedures for SI-RNTI if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) { - ue_pdsch_procedures(ue, - proc, - eNB_id, - SI_PDSCH, - ue->dlsch_SI[eNB_id], - NULL, - (ue->frame_parms.symbols_per_tti>>1), - ue->frame_parms.symbols_per_tti-1, - abstraction_flag); + ue_pdsch_procedures(ue, + proc, + eNB_id, + SI_PDSCH, + ue->dlsch_SI[eNB_id], + NULL, + (ue->frame_parms.symbols_per_tti>>1), + ue->frame_parms.symbols_per_tti-1, + abstraction_flag); } // do procedures for P-RNTI if ((ue->dlsch_p[eNB_id]) && (ue->dlsch_p[eNB_id]->active == 1)) { - ue_pdsch_procedures(ue, - proc, - eNB_id, - P_PDSCH, - ue->dlsch_p[eNB_id], - NULL, - (ue->frame_parms.symbols_per_tti>>1), - ue->frame_parms.symbols_per_tti-1, - abstraction_flag); + ue_pdsch_procedures(ue, + proc, + eNB_id, + P_PDSCH, + ue->dlsch_p[eNB_id], + NULL, + (ue->frame_parms.symbols_per_tti>>1), + ue->frame_parms.symbols_per_tti-1, + abstraction_flag); } // do procedures for RA-RNTI if ((ue->dlsch_ra[eNB_id]) && (ue->dlsch_ra[eNB_id]->active == 1)) { - ue_pdsch_procedures(ue, - proc, - eNB_id, - RA_PDSCH, - ue->dlsch_ra[eNB_id], - NULL, - (ue->frame_parms.symbols_per_tti>>1), - ue->frame_parms.symbols_per_tti-1, - abstraction_flag); + ue_pdsch_procedures(ue, + proc, + eNB_id, + RA_PDSCH, + ue->dlsch_ra[eNB_id], + NULL, + (ue->frame_parms.symbols_per_tti>>1), + ue->frame_parms.symbols_per_tti-1, + abstraction_flag); } proc->llr_slot1_available=1; @@ -5282,561 +4415,37 @@ void *UE_thread_slot1_dl_processing(void *arg) { #endif - if (pthread_mutex_lock(&proc->mutex_slot1_dl_processing) != 0) { - LOG_E( PHY, "[SCHED][UE] error locking mutex for UE RXTX\n" ); - exit_fun("noting to add"); - } - proc->instance_cnt_slot1_dl_processing--; - if (pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) != 0) { - LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE FEP Slo1\n" ); - exit_fun("noting to add"); - } + if (pthread_mutex_lock(&proc->mutex_slot1_dl_processing) != 0) { + LOG_E( PHY, "[SCHED][UE] error locking mutex for UE RXTX\n" ); + exit_fun("noting to add"); } - // thread finished - free(arg); - return &UE_dl_slot1_processing_retval; + proc->instance_cnt_slot1_dl_processing--; + if (pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) != 0) { + LOG_E( PHY, "[SCHED][UE] error unlocking mutex for UE FEP Slo1\n" ); + exit_fun("noting to add"); + } + } + // thread finished + free(arg); + return &UE_dl_slot1_processing_retval; } #endif #ifdef UE_SLOT_PARALLELISATION -int phy_procedures_slot_parallelization_UE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id, +int phy_procedures_slot_parallelization_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id, uint8_t abstraction_flag,uint8_t do_pdcch_flag,runmode_t mode, relaying_type_t r_type) { - int l,l2; - int pmch_flag=0; - int frame_rx = proc->frame_rx; - int nr_tti_rx = proc->nr_tti_rx; - uint8_t pilot0; - uint8_t pilot1; - uint8_t slot0; - uint8_t slot1; - uint8_t first_ofdm_sym; - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_IN); - -#if T_TRACER - T(T_UE_PHY_DL_TICK, T_INT(ue->Mod_id), T_INT(frame_rx%1024), T_INT(nr_tti_rx)); - - T(T_UE_PHY_INPUT_SIGNAL, T_INT(ue->Mod_id), T_INT(frame_rx%1024), T_INT(nr_tti_rx), T_INT(0), - T_BUFFER(&ue->common_vars.rxdata[0][nr_tti_rx*ue->frame_parms.samples_per_subframe], - ue->frame_parms.samples_per_subframe * 4)); -#endif - - // start timers -#ifdef UE_DEBUG_TRACE - LOG_I(PHY," ****** start RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); -#endif - -#if UE_TIMING_TRACE - start_meas(&ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]]); - start_meas(&ue->ue_front_end_stat[ue->current_thread_id[nr_tti_rx]]); -#endif - - pmch_flag = is_pmch_subframe(frame_rx,nr_tti_rx,&ue->frame_parms) ? 1 : 0; - - if (do_pdcch_flag) { - // deactivate reception until we scan pdcch - if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]) - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active = 0; - if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]) - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]->active = 0; - - if (ue->dlsch_SI[eNB_id]) - ue->dlsch_SI[eNB_id]->active = 0; - if (ue->dlsch_p[eNB_id]) - ue->dlsch_p[eNB_id]->active = 0; - if (ue->dlsch_ra[eNB_id]) - ue->dlsch_ra[eNB_id]->active = 0; - } - -#ifdef DEBUG_PHY_PROC - LOG_D(PHY,"[%s %d] Frame %d nr_tti_rx %d: Doing phy_procedures_UE_RX\n", - (r_type == multicast_relay) ? "RN/UE" : "UE", - ue->Mod_id,frame_rx, nr_tti_rx); -#endif - - - - - if (nr_subframe_select(&ue->frame_parms,nr_tti_rx) == SF_S) { // S-subframe, do first 5 symbols only - l2 = 5; - } else if (pmch_flag == 1) { // do first 2 symbols only - l2 = 1; - } else { // normal nr_tti_rx, last symbol to be processed is the first of the second slot - l2 = (ue->frame_parms.symbols_per_tti/2)-1; - } - - int prev_nr_tti_rx = (nr_tti_rx - 1)<0? 9: (nr_tti_rx - 1); - if (nr_subframe_select(&ue->frame_parms,prev_nr_tti_rx) != SF_DL) { - //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ - // RX processing of symbols l=0...l2 - //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ - first_ofdm_sym = 0; - } else { - //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ - // RX processing of symbols l=1...l2 (l=0 is done in last scheduling epoch) - //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ - first_ofdm_sym = 1; - } - slot0 = (nr_tti_rx<<1); - slot1 = (nr_tti_rx<<1) + 1; - pilot0 = 0; - if (ue->frame_parms.Ncp == 0) { // normal prefix - pilot1 = 4; - } else { // extended prefix - pilot1 = 3; - } - - //LOG_I(PHY,"Set available channelEst to 0 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); - //LOG_I(PHY,"Set available llrs slot1 to 0 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); - //LOG_I(PHY,"Set available dci info slot0 to 0 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); - proc->chan_est_pilot0_slot1_available=0; - proc->llr_slot1_available=0; - proc->dci_slot0_available=0; - proc->first_symbol_available=0; - proc->chan_est_slot1_available=0; - //proc->channel_level=0; - - if (pthread_mutex_lock(&proc->mutex_slot1_dl_processing) != 0) { - LOG_E( PHY, "[SCHED][UE %d][Slot0] error locking mutex for UE slot1 dl processing\n",ue->Mod_id ); - exit_fun("nothing to add"); - } - - proc->instance_cnt_slot1_dl_processing++; - if (proc->instance_cnt_slot1_dl_processing == 0) - { - LOG_D(PHY,"unblock slot1 dl processing thread blocked on instance_cnt_slot1_dl_processing : %d \n", proc->instance_cnt_slot1_dl_processing ); - if (pthread_cond_signal(&proc->cond_slot1_dl_processing) != 0) { - LOG_E( PHY, "[SCHED][UE %d][Slot0] ERROR pthread_cond_signal for UE slot1 processing thread\n", ue->Mod_id); - exit_fun("nothing to add"); - } - if (pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) != 0) { - LOG_E( PHY, "[SCHED][UE %d][Slot0] error unlocking mutex for UE slot1 dl processing \n",ue->Mod_id ); - exit_fun("nothing to add"); - } - - } else - { - LOG_E( PHY, "[SCHED][UE %d] UE RX thread busy (IC %d)!!\n", ue->Mod_id, proc->instance_cnt_slot1_dl_processing); - if (proc->instance_cnt_slot1_dl_processing > 2) - exit_fun("instance_cnt_slot1_dl_processing > 2"); - } - //AssertFatal(pthread_cond_signal(&proc->cond_slot1_dl_processing) ==0 ,""); - AssertFatal(pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) ==0,""); - - - /**** Slot0 FE Processing ****/ - // I- start main thread for FFT/ChanEst symbol: 0/1 --> 7 -#if UE_TIMING_TRACE - start_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); -#endif - // 1- perform FFT for pilot ofdm symbols first (ofdmSym7 ofdmSym4 or (ofdmSym6 ofdmSym3)) - //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,pilot0); - front_end_fft(ue, - pilot0, - slot1, - 0, - 0); - //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,pilot1); - front_end_fft(ue, - pilot1, - slot0, - 0, - 0); - //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,pilot1); - front_end_chanEst(ue, - pilot1, - slot0, - 0); - //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,pilot0); - front_end_chanEst(ue, - pilot0, - slot1, - 0); - proc->chan_est_pilot0_slot1_available = 1; - //printf("Set available channelEst to 1 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); - - // 2- perform FFT for other ofdm symbols other than pilots - for (l=first_ofdm_sym; l<=l2; l++) - { - if( (l != pilot0) && (l != pilot1)) - { - //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,l); -#if UE_TIMING_TRACE - start_meas(&ue->ofdm_demod_stats); -#endif - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_IN); - front_end_fft(ue, - l, - slot0, - 0, - 0); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_OUT); -#if UE_TIMING_TRACE - stop_meas(&ue->ofdm_demod_stats); -#endif - } - } // for l=1..l2 - - // 3- perform Channel Estimation for slot0 - for (l=first_ofdm_sym; l<=l2; l++) - { - if( (l != pilot0) && (l != pilot1)) - { - //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,l); - front_end_chanEst(ue, - l, - slot0, - 0); - } - ue_measurement_procedures(l-1,ue,proc,eNB_id,(nr_tti_rx<<1),abstraction_flag,mode); - } - - if (do_pdcch_flag) { -#if UE_TIMING_TRACE - start_meas(&ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); -#endif - if (ue_pdcch_procedures(eNB_id,ue,proc,abstraction_flag) == -1) { - LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d: Error in pdcch procedures\n",ue->Mod_id,frame_rx,nr_tti_rx); -#if UE_TIMING_TRACE - stop_meas(&ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); -#if DISABLE_LOG_X - printf("[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#endif -#endif - //proc->dci_slot0_available = 1; - return(-1); - } - //proc->dci_slot0_available=1; -#if UE_TIMING_TRACE - stop_meas(&ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); -#if DISABLE_LOG_X - printf("[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#endif -#endif - } - - //printf("num_pdcch_symbols %d\n",ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols); - - // first slot has been processed (FFTs + Channel Estimation, PCFICH/PHICH/PDCCH) -#if UE_TIMING_TRACE - stop_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); -#if DISABLE_LOG_X - printf("[AbsSFN %d.%d] Slot0: FFT + Channel Estimate + PCFICH/PHICH/PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "[AbsSFN %d.%d] Slot0: FFT + Channel Estimate + PCFICH/PHICH/PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); -#endif -#endif - - //wait until slot1 FE is done - uint32_t wait = 0; - while(proc->chan_est_slot1_available == 0) - { - usleep(1); - wait++; - } - -#if UE_TIMING_TRACE - stop_meas(&ue->ue_front_end_stat[ue->current_thread_id[nr_tti_rx]]); -#if DISABLE_LOG_X - printf("[AbsSFN %d.%d] FULL FE Processing %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "[AbsSFN %d.%d] FULL FE Processing %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); -#endif -#endif - /**** End nr_tti_rx FE Processing ****/ - - -#if 0 - //Trigger LLR parallelized for Slot 1 - proc->dci_slot0_available=1; - printf("Set available dci slot0 to 1 AbsSubframe %d.%d \n",frame_rx%1024,nr_tti_rx); -#endif - - /**** Pdsch Procedure Slot0 ****/ - // start main thread for Pdsch Procedure (slot0) - // do procedures for C-RNTI - //printf("AbsSubframe %d.%d Pdsch Procedure (slot0)\n",frame_rx%1024,nr_tti_rx); - //printf("AbsSubframe %d.%d Pdsch Procedure PDSCH Active %d \n",frame_rx%1024,nr_tti_rx, ue->dlsch[ue->current_thread_id[nr_tti_rx]][0][0]->active); - -#if UE_TIMING_TRACE - start_meas(&ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); -#endif - -#if UE_TIMING_TRACE - start_meas(&ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); -#endif - if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); - ue_pdsch_procedures(ue, - proc, - eNB_id, - PDSCH, - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], - NULL, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - (ue->frame_parms.symbols_per_tti>>1)-1, - abstraction_flag); - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT); - } - - // do procedures for SI-RNTI - if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) { - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_SI, VCD_FUNCTION_IN); - ue_pdsch_procedures(ue, - proc, - eNB_id, - SI_PDSCH, - ue->dlsch_SI[eNB_id], - NULL, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - (ue->frame_parms.symbols_per_tti>>1)-1, - abstraction_flag); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_SI, VCD_FUNCTION_OUT); - } - - // do procedures for SI-RNTI - if ((ue->dlsch_p[eNB_id]) && (ue->dlsch_p[eNB_id]->active == 1)) { - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_P, VCD_FUNCTION_IN); - ue_pdsch_procedures(ue, - proc, - eNB_id, - P_PDSCH, - ue->dlsch_p[eNB_id], - NULL, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - (ue->frame_parms.symbols_per_tti>>1)-1, - abstraction_flag); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_P, VCD_FUNCTION_OUT); - } - - // do procedures for RA-RNTI - if ((ue->dlsch_ra[eNB_id]) && (ue->dlsch_ra[eNB_id]->active == 1)) { - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_RA, VCD_FUNCTION_IN); - ue_pdsch_procedures(ue, - proc, - eNB_id, - RA_PDSCH, - ue->dlsch_ra[eNB_id], - NULL, - ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - (ue->frame_parms.symbols_per_tti>>1)-1, - abstraction_flag); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_RA, VCD_FUNCTION_OUT); - } - -#if 1 - // LLR linear - proc->dci_slot0_available=1; - //printf("Set available dci slot0 to 1 AbsSubframe %d.%d \n",frame_rx%1024,nr_tti_rx); -#endif - -#if UE_TIMING_TRACE - stop_meas(&ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); -#if DISABLE_LOG_X - printf("[AbsSFN %d.%d] Slot0: LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "[AbsSFN %d.%d] Slot0: LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); -#endif -#endif - - - //wait until LLR Slot1 is done - wait = 0; - while(proc->llr_slot1_available == 0) - { - usleep(1); - wait++; - } - - - -#if UE_TIMING_TRACE - stop_meas(&ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); -#if DISABLE_LOG_X - printf("[AbsSFN %d.%d] Full LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "[AbsSFN %d.%d] Full LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#endif -#endif - //printf("[slot0 dl processing] AbsSubframe %d.%d Channel Decoder Start wait %d\n",frame_rx,nr_tti_rx,wait); - - - //=====================================================================// -#if UE_TIMING_TRACE - start_meas(&ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); -#endif - - LOG_D(PHY,"==> Start Turbo Decoder active dlsch %d SI %d RA %d \n",ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active, - ue->dlsch_SI[eNB_id]->active, - //ue->dlsch_p[eNB_id]->active, - ue->dlsch_ra[eNB_id]->active); - // Start Turbo decoder - if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { - //VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); - ue_dlsch_procedures(ue, - proc, - eNB_id, - PDSCH, - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1], - &ue->dlsch_errors[eNB_id], - mode, - abstraction_flag); - //VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT); - } - - // do procedures for SI-RNTI - if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) { - ue_dlsch_procedures(ue, - proc, - eNB_id, - SI_PDSCH, - ue->dlsch_SI[eNB_id], - NULL, - &ue->dlsch_SI_errors[eNB_id], - mode, - abstraction_flag); - ue->dlsch_SI[eNB_id]->active = 0; - } - - // do procedures for P-RNTI - if ((ue->dlsch_p[eNB_id]) && (ue->dlsch_p[eNB_id]->active == 1)) { - ue_dlsch_procedures(ue, - proc, - eNB_id, - P_PDSCH, - ue->dlsch_p[eNB_id], - NULL, - &ue->dlsch_p_errors[eNB_id], - mode, - abstraction_flag); - ue->dlsch_p[eNB_id]->active = 0; - } - // do procedures for RA-RNTI - if ((ue->dlsch_ra[eNB_id]) && (ue->dlsch_ra[eNB_id]->active == 1)) { - ue_dlsch_procedures(ue, - proc, - eNB_id, - RA_PDSCH, - ue->dlsch_ra[eNB_id], - NULL, - &ue->dlsch_ra_errors[eNB_id], - mode, - abstraction_flag); - ue->dlsch_ra[eNB_id]->active = 0; - } - -#if UE_TIMING_TRACE - stop_meas(&ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); -#if DISABLE_LOG_X - printf("[AbsSFN %d.%d] Channel Decoder: %5.2f \n",frame_rx,nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "[AbsSFN %d.%d] Channel Decoder: %5.2f \n",frame_rx,nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#endif - -#endif - - // duplicate harq structure - uint8_t current_harq_pid = ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->current_harq_pid; - NR_DL_UE_HARQ_t *current_harq_processes = ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_processes[current_harq_pid]; - nr_harq_status_t *current_harq_ack = &ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_ack[nr_tti_rx]; - - // For Debug parallelisation - //if (current_harq_ack->ack == 0) { - //printf("[slot0 dl processing][End of Channel Decoding] AbsSubframe %d.%d Decode Fail for HarqId%d Round%d\n",frame_rx,nr_tti_rx,current_harq_pid,current_harq_processes->round); - //} - for(uint8_t rx_th_idx=1; rx_th_idxdlsch[ue->current_thread_id[(nr_tti_rx+rx_th_idx)%10]][eNB_id][0]->harq_processes[current_harq_pid]; - nr_harq_status_t *harq_ack_dest = &ue->dlsch[ue->current_thread_id[(nr_tti_rx+rx_th_idx)%10]][eNB_id][0]->harq_ack[nr_tti_rx]; - - copy_harq_proc_struct(harq_processes_dest, current_harq_processes); - copy_ack_struct(harq_ack_dest, current_harq_ack); - - } - /* - NR_DL_UE_HARQ_t *harq_processes_dest = ue->dlsch[(nr_tti_rx+1)%RX_NB_TH][eNB_id][0]->harq_processes[current_harq_pid]; - NR_DL_UE_HARQ_t *harq_processes_dest1 = ue->dlsch[(nr_tti_rx+2)%RX_NB_TH][eNB_id][0]->harq_processes[current_harq_pid]; - - nr_harq_status_t *current_harq_ack = &ue->dlsch[nr_tti_rx%RX_NB_TH][eNB_id][0]->harq_ack[nr_tti_rx]; - nr_harq_status_t *harq_ack_dest = &ue->dlsch[(nr_tti_rx+1)%RX_NB_TH][eNB_id][0]->harq_ack[nr_tti_rx]; - nr_harq_status_t *harq_ack_dest1 = &ue->dlsch[(nr_tti_rx+2)%RX_NB_TH][eNB_id][0]->harq_ack[nr_tti_rx]; - - copy_harq_proc_struct(harq_processes_dest, current_harq_processes); - copy_ack_struct(harq_ack_dest, current_harq_ack); - - copy_harq_proc_struct(harq_processes_dest1, current_harq_processes); - copy_ack_struct(harq_ack_dest1, current_harq_ack); - */ - if (nr_tti_rx==9) { - if (frame_rx % 10 == 0) { - if ((ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]) != 0) - ue->dlsch_fer[eNB_id] = (100*(ue->dlsch_errors[eNB_id] - ue->dlsch_errors_last[eNB_id]))/(ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]); - - ue->dlsch_errors_last[eNB_id] = ue->dlsch_errors[eNB_id]; - ue->dlsch_received_last[eNB_id] = ue->dlsch_received[eNB_id]; - } - - - ue->bitrate[eNB_id] = (ue->total_TBS[eNB_id] - ue->total_TBS_last[eNB_id])*100; - ue->total_TBS_last[eNB_id] = ue->total_TBS[eNB_id]; - LOG_D(PHY,"[UE %d] Calculating bitrate Frame %d: total_TBS = %d, total_TBS_last = %d, bitrate %f kbits\n", - ue->Mod_id,frame_rx,ue->total_TBS[eNB_id], - ue->total_TBS_last[eNB_id],(float) ue->bitrate[eNB_id]/1000.0); - -#if UE_AUTOTEST_TRACE - if ((frame_rx % 100 == 0)) { - LOG_I(PHY,"[UE %d] AUTOTEST Metric : UE_DLSCH_BITRATE = %5.2f kbps (frame = %d) \n", ue->Mod_id, (float) ue->bitrate[eNB_id]/1000.0, frame_rx); - } -#endif - - } - -#ifdef EMOS - phy_procedures_emos_UE_RX(ue,slot,eNB_id); -#endif - - - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_OUT); - -#if UE_TIMING_TRACE - stop_meas(&ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]]); -#if DISABLE_LOG_X - printf("------FULL RX PROC [AbsSFN %d.%d]: %5.2f ------\n",frame_rx,nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#else - LOG_D(PHY, "------FULL RX PROC [AbsSFN %d.%d]: %5.2f ------\n",frame_rx,nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); -#endif -#endif - - LOG_D(PHY," ****** end RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); - return (0); -} -#endif - - -int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id, - uint8_t abstraction_flag,uint8_t do_pdcch_flag,runmode_t mode, - relaying_type_t r_type) { - - //int l,l2; - //int pilot1; - //int pmch_flag=0; + int l,l2; + int pmch_flag=0; int frame_rx = proc->frame_rx; int nr_tti_rx = proc->nr_tti_rx; - proc->decoder_switch = 0; - //int counter_decoder = 0; + uint8_t pilot0; + uint8_t pilot1; + uint8_t slot0; + uint8_t slot1; + uint8_t first_ofdm_sym; - //uint8_t next1_thread_id = ue->current_thread_id[nr_tti_rx]== (RX_NB_TH-1) ? 0:(ue->current_thread_id[nr_tti_rx]+1); - //uint8_t next2_thread_id = next1_thread_id== (RX_NB_TH-1) ? 0:(next1_thread_id+1); - -#if 0 VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_IN); #if T_TRACER @@ -5844,52 +4453,43 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN T(T_UE_PHY_INPUT_SIGNAL, T_INT(ue->Mod_id), T_INT(frame_rx%1024), T_INT(nr_tti_rx), T_INT(0), T_BUFFER(&ue->common_vars.rxdata[0][nr_tti_rx*ue->frame_parms.samples_per_subframe], - ue->frame_parms.samples_per_subframe * 4)); + ue->frame_parms.samples_per_subframe * 4)); #endif // start timers -//#ifdef UE_DEBUG_TRACE +#ifdef UE_DEBUG_TRACE LOG_I(PHY," ****** start RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); -//#endif +#endif #if UE_TIMING_TRACE start_meas(&ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]]); - start_meas(&ue->generic_stat); + start_meas(&ue->ue_front_end_stat[ue->current_thread_id[nr_tti_rx]]); #endif pmch_flag = is_pmch_subframe(frame_rx,nr_tti_rx,&ue->frame_parms) ? 1 : 0; - - - - if (do_pdcch_flag) { - // deactivate reception until we scan pdcch - if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]) - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active = 0; - if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]) - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]->active = 0; + // deactivate reception until we scan pdcch + if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]) + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active = 0; + if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]) + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]->active = 0; - if (ue->dlsch_SI[eNB_id]) - ue->dlsch_SI[eNB_id]->active = 0; - if (ue->dlsch_p[eNB_id]) - ue->dlsch_p[eNB_id]->active = 0; - if (ue->dlsch_ra[eNB_id]) - ue->dlsch_ra[eNB_id]->active = 0; + if (ue->dlsch_SI[eNB_id]) + ue->dlsch_SI[eNB_id]->active = 0; + if (ue->dlsch_p[eNB_id]) + ue->dlsch_p[eNB_id]->active = 0; + if (ue->dlsch_ra[eNB_id]) + ue->dlsch_ra[eNB_id]->active = 0; } #ifdef DEBUG_PHY_PROC LOG_D(PHY,"[%s %d] Frame %d nr_tti_rx %d: Doing phy_procedures_UE_RX\n", - (r_type == multicast_relay) ? "RN/UE" : "UE", - ue->Mod_id,frame_rx, nr_tti_rx); + (r_type == multicast_relay) ? "RN/UE" : "UE", + ue->Mod_id,frame_rx, nr_tti_rx); #endif - if (ue->frame_parms.Ncp == 0) { // normal prefix - pilot1 = 4; - } else { // extended prefix - pilot1 = 3; - } if (nr_subframe_select(&ue->frame_parms,nr_tti_rx) == SF_S) { // S-subframe, do first 5 symbols only @@ -5905,130 +4505,205 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // RX processing of symbols l=0...l2 //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ - l=0; + first_ofdm_sym = 0; } else { //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // RX processing of symbols l=1...l2 (l=0 is done in last scheduling epoch) //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ - l=1; + first_ofdm_sym = 1; + } + slot0 = (nr_tti_rx<<1); + slot1 = (nr_tti_rx<<1) + 1; + pilot0 = 0; + if (ue->frame_parms.Ncp == 0) { // normal prefix + pilot1 = 4; + } else { // extended prefix + pilot1 = 3; } - LOG_D(PHY," ------ slot 0 Processing: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); - LOG_D(PHY," ------ --> FFT/ChannelEst/PDCCH slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); -#endif + //LOG_I(PHY,"Set available channelEst to 0 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); + //LOG_I(PHY,"Set available llrs slot1 to 0 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); + //LOG_I(PHY,"Set available dci info slot0 to 0 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); + proc->chan_est_pilot0_slot1_available=0; + proc->llr_slot1_available=0; + proc->dci_slot0_available=0; + proc->first_symbol_available=0; + proc->chan_est_slot1_available=0; + //proc->channel_level=0; - //nr_gold_pdcch(ue,0, 2); + if (pthread_mutex_lock(&proc->mutex_slot1_dl_processing) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] error locking mutex for UE slot1 dl processing\n",ue->Mod_id ); + exit_fun("nothing to add"); + } - for (int l=0; l<2; l++) { - if (abstraction_flag == 0) { -#if UE_TIMING_TRACE - start_meas(&ue->ofdm_demod_stats); -#endif + proc->instance_cnt_slot1_dl_processing++; + if (proc->instance_cnt_slot1_dl_processing == 0) + { + LOG_D(PHY,"unblock slot1 dl processing thread blocked on instance_cnt_slot1_dl_processing : %d \n", proc->instance_cnt_slot1_dl_processing ); + if (pthread_cond_signal(&proc->cond_slot1_dl_processing) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] ERROR pthread_cond_signal for UE slot1 processing thread\n", ue->Mod_id); + exit_fun("nothing to add"); + } + if (pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) != 0) { + LOG_E( PHY, "[SCHED][UE %d][Slot0] error unlocking mutex for UE slot1 dl processing \n",ue->Mod_id ); + exit_fun("nothing to add"); + } - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_IN); - nr_slot_fep(ue, - l, - nr_tti_rx, - 0, - 0, - 1, - NR_PDCCH_EST); - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_OUT); -#if UE_TIMING_TRACE - stop_meas(&ue->ofdm_demod_stats); -#endif + } else + { + LOG_E( PHY, "[SCHED][UE %d] UE RX thread busy (IC %d)!!\n", ue->Mod_id, proc->instance_cnt_slot1_dl_processing); + if (proc->instance_cnt_slot1_dl_processing > 2) + exit_fun("instance_cnt_slot1_dl_processing > 2"); } - } + //AssertFatal(pthread_cond_signal(&proc->cond_slot1_dl_processing) ==0 ,""); + AssertFatal(pthread_mutex_unlock(&proc->mutex_slot1_dl_processing) ==0,""); - write_output("rxdataF.m","rxF",ue->common_vars.common_vars_rx_data_per_thread[ue->current_thread_id[nr_tti_rx>>1]].rxdataF[0],ue->frame_parms.ofdm_symbol_size*2,1,1); - //ue_measurement_procedures(l-1,ue,proc,eNB_id,(nr_tti_rx<<1),abstraction_flag,mode); - - //if (do_pdcch_flag) { - // if ((l==pilot1) || ((pmch_flag==1)&(l==l2))) { - // LOG_D(PHY,"[UE %d] Frame %d: Calling pdcch procedures (eNB %d)\n",ue->Mod_id,frame_rx,eNB_id); - - //start_meas(&ue->rx_pdcch_stats[ue->current_thread_id[nr_tti_rx]]); - - //printf(">>> at phy_procedures_nrUE_RX, nr_ue_pdcch_procedures init, dlsch->active=%d\n", - // ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active); - if (nr_ue_pdcch_procedures(eNB_id,ue,proc,abstraction_flag) == -1) { - LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d: Error in pdcch procedures\n",ue->Mod_id,frame_rx,nr_tti_rx); - return(-1); - } - //for (int m=0;m<1000;m++) printf("%d",m); - //printf("\n>>> at phy_procedures_nrUE_RX, nr_ue_pdcch_procedures end, dlsch->active=%d\n", - // ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active); - - // preparing for PDSCH procedures - // if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { - // ue_pdsch_procedures(ue, - // proc, - // eNB_id, - // PDSCH, - // ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], - // NULL, - // ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - // ue->frame_parms.symbols_per_tti>>1, - // abstraction_flag); - // } -#if 0 - //stop_meas(&ue->rx_pdcch_stats[ue->current_thread_id[nr_tti_rx]]); - //printf("nr_tti_rx %d n_pdcch_sym %d pdcch procedures %5.3f \n", - // nr_tti_rx, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - // (ue->rx_pdcch_stats[ue->current_thread_id[nr_tti_rx]].p_time)/(cpuf*1000.0)); - LOG_D(PHY,"num_pdcch_symbols %d\n",ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols); - // } - //} - - // } // for l=1..l2 - - - ue_measurement_procedures(l-1,ue,proc,eNB_id,(nr_tti_rx<<1),abstraction_flag,mode); - - LOG_D(PHY," ------ end FFT/ChannelEst/PDCCH slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); - // If this is PMCH, call procedures and return - if (pmch_flag == 1) { - ue_pmch_procedures(ue,proc,eNB_id,abstraction_flag); - return 0; - } - - nr_slot_fep(ue, - 0, - 1+(nr_tti_rx<<1), - 0, - 0, - 0, - NR_PDSCH_EST); + /**** Slot0 FE Processing ****/ + // I- start main thread for FFT/ChanEst symbol: 0/1 --> 7 +#if UE_TIMING_TRACE + start_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); #endif + // 1- perform FFT for pilot ofdm symbols first (ofdmSym7 ofdmSym4 or (ofdmSym6 ofdmSym3)) + //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,pilot0); + front_end_fft(ue, + pilot0, + slot1, + 0, + 0); + //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,pilot1); + front_end_fft(ue, + pilot1, + slot0, + 0, + 0); + //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,pilot1); + front_end_chanEst(ue, + pilot1, + slot0, + 0); + //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot1,pilot0); + front_end_chanEst(ue, + pilot0, + slot1, + 0); + proc->chan_est_pilot0_slot1_available = 1; + //printf("Set available channelEst to 1 AbsSubframe %d.%d \n",frame_rx,nr_tti_rx); + + // 2- perform FFT for other ofdm symbols other than pilots + for (l=first_ofdm_sym; l<=l2; l++) + { + if( (l != pilot0) && (l != pilot1)) + { + //printf("AbsSubframe %d.%d FFT slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,l); +#if UE_TIMING_TRACE + start_meas(&ue->ofdm_demod_stats); +#endif + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_IN); + front_end_fft(ue, + l, + slot0, + 0, + 0); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_OUT); +#if UE_TIMING_TRACE + stop_meas(&ue->ofdm_demod_stats); +#endif + } + } // for l=1..l2 + + // 3- perform Channel Estimation for slot0 + for (l=first_ofdm_sym; l<=l2; l++) + { + if( (l != pilot0) && (l != pilot1)) + { + //printf("AbsSubframe %d.%d ChanEst slot %d, symbol %d\n", frame_rx,nr_tti_rx,slot0,l); + front_end_chanEst(ue, + l, + slot0, + 0); + } + ue_measurement_procedures(l-1,ue,proc,eNB_id,(nr_tti_rx<<1),abstraction_flag,mode); + } + + if (do_pdcch_flag) { +#if UE_TIMING_TRACE + start_meas(&ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); +#endif + if (ue_pdcch_procedures(eNB_id,ue,proc,abstraction_flag) == -1) { + LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d: Error in pdcch procedures\n",ue->Mod_id,frame_rx,nr_tti_rx); +#if UE_TIMING_TRACE + stop_meas(&ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#else + LOG_D(PHY, "[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#endif +#endif + //proc->dci_slot0_available = 1; + return(-1); + } + //proc->dci_slot0_available=1; +#if UE_TIMING_TRACE + stop_meas(&ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#else + LOG_D(PHY, "[AbsSFN %d.%d] Slot0: PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->pdcch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#endif +#endif + } + + //printf("num_pdcch_symbols %d\n",ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols); // first slot has been processed (FFTs + Channel Estimation, PCFICH/PHICH/PDCCH) #if UE_TIMING_TRACE - stop_meas(&ue->generic_stat); + stop_meas(&ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); #if DISABLE_LOG_X - printf("[SFN %d] Slot0: FFT + Channel Estimate + PCFICH/PHICH/PDCCH %5.2f \n",nr_tti_rx,ue->generic_stat.p_time/(cpuf*1000.0)); + printf("[AbsSFN %d.%d] Slot0: FFT + Channel Estimate + PCFICH/PHICH/PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); #else - LOG_D(PHY, "[SFN %d] Slot0: FFT + Channel Estimate + PCFICH/PHICH/PDCCH %5.2f \n",nr_tti_rx,ue->generic_stat.p_time/(cpuf*1000.0)); + LOG_D(PHY, "[AbsSFN %d.%d] Slot0: FFT + Channel Estimate + PCFICH/PHICH/PDCCH %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); +#endif #endif -#endif -#if 0 - LOG_D(PHY," ------ --> PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); - //to update from pdsch config - nr_gold_pdsch(ue,0,0, 1); - - nr_slot_fep(ue, - 2, //to be updated from higher layer - (nr_tti_rx<<1), - 0, - 0, - 0, - NR_PDSCH_EST); + //wait until slot1 FE is done + uint32_t wait = 0; + while(proc->chan_est_slot1_available == 0) + { + usleep(1); + wait++; + } #if UE_TIMING_TRACE - start_meas(&ue->generic_stat); + stop_meas(&ue->ue_front_end_stat[ue->current_thread_id[nr_tti_rx]]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] FULL FE Processing %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); +#else + LOG_D(PHY, "[AbsSFN %d.%d] FULL FE Processing %5.2f \n",frame_rx,nr_tti_rx,ue->ue_front_end_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); #endif +#endif + /**** End nr_tti_rx FE Processing ****/ + + +#if 0 + //Trigger LLR parallelized for Slot 1 + proc->dci_slot0_available=1; + printf("Set available dci slot0 to 1 AbsSubframe %d.%d \n",frame_rx%1024,nr_tti_rx); +#endif + + /**** Pdsch Procedure Slot0 ****/ + // start main thread for Pdsch Procedure (slot0) // do procedures for C-RNTI + //printf("AbsSubframe %d.%d Pdsch Procedure (slot0)\n",frame_rx%1024,nr_tti_rx); + //printf("AbsSubframe %d.%d Pdsch Procedure PDSCH Active %d \n",frame_rx%1024,nr_tti_rx, ue->dlsch[ue->current_thread_id[nr_tti_rx]][0][0]->active); + +#if UE_TIMING_TRACE + start_meas(&ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); +#endif + +#if UE_TIMING_TRACE + start_meas(&ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); +#endif if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); ue_pdsch_procedures(ue, @@ -6038,13 +4713,12 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], NULL, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - ue->frame_parms.symbols_per_tti>>1, + (ue->frame_parms.symbols_per_tti>>1)-1, abstraction_flag); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT); } - LOG_D(PHY," ------ end PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); // do procedures for SI-RNTI if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) { VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_SI, VCD_FUNCTION_IN); @@ -6055,7 +4729,7 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN ue->dlsch_SI[eNB_id], NULL, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - ue->frame_parms.symbols_per_tti>>1, + (ue->frame_parms.symbols_per_tti>>1)-1, abstraction_flag); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_SI, VCD_FUNCTION_OUT); } @@ -6070,7 +4744,7 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN ue->dlsch_p[eNB_id], NULL, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - ue->frame_parms.symbols_per_tti>>1, + (ue->frame_parms.symbols_per_tti>>1)-1, abstraction_flag); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_P, VCD_FUNCTION_OUT); } @@ -6085,49 +4759,455 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN ue->dlsch_ra[eNB_id], NULL, ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, - ue->frame_parms.symbols_per_tti>>1, + (ue->frame_parms.symbols_per_tti>>1)-1, abstraction_flag); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_RA, VCD_FUNCTION_OUT); } +#if 1 + // LLR linear + proc->dci_slot0_available=1; + //printf("Set available dci slot0 to 1 AbsSubframe %d.%d \n",frame_rx%1024,nr_tti_rx); +#endif + +#if UE_TIMING_TRACE + stop_meas(&ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Slot0: LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); +#else + LOG_D(PHY, "[AbsSFN %d.%d] Slot0: LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_per_slot_stat[ue->current_thread_id[nr_tti_rx]][0].p_time/(cpuf*1000.0)); +#endif +#endif + + + //wait until LLR Slot1 is done + wait = 0; + while(proc->llr_slot1_available == 0) + { + usleep(1); + wait++; + } + + + +#if UE_TIMING_TRACE + stop_meas(&ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Full LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#else + LOG_D(PHY, "[AbsSFN %d.%d] Full LLR Computation %5.2f \n",frame_rx,nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#endif +#endif + //printf("[slot0 dl processing] AbsSubframe %d.%d Channel Decoder Start wait %d\n",frame_rx,nr_tti_rx,wait); + + + //=====================================================================// +#if UE_TIMING_TRACE + start_meas(&ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); +#endif + + LOG_D(PHY,"==> Start Turbo Decoder active dlsch %d SI %d RA %d \n",ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active, + ue->dlsch_SI[eNB_id]->active, + //ue->dlsch_p[eNB_id]->active, + ue->dlsch_ra[eNB_id]->active); + // Start Turbo decoder + if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { + //VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); + ue_dlsch_procedures(ue, + proc, + eNB_id, + PDSCH, + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1], + &ue->dlsch_errors[eNB_id], + mode, + abstraction_flag); + //VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT); + } + + // do procedures for SI-RNTI + if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) { + ue_dlsch_procedures(ue, + proc, + eNB_id, + SI_PDSCH, + ue->dlsch_SI[eNB_id], + NULL, + &ue->dlsch_SI_errors[eNB_id], + mode, + abstraction_flag); + ue->dlsch_SI[eNB_id]->active = 0; + } + + // do procedures for P-RNTI + if ((ue->dlsch_p[eNB_id]) && (ue->dlsch_p[eNB_id]->active == 1)) { + ue_dlsch_procedures(ue, + proc, + eNB_id, + P_PDSCH, + ue->dlsch_p[eNB_id], + NULL, + &ue->dlsch_p_errors[eNB_id], + mode, + abstraction_flag); + ue->dlsch_p[eNB_id]->active = 0; + } + // do procedures for RA-RNTI + if ((ue->dlsch_ra[eNB_id]) && (ue->dlsch_ra[eNB_id]->active == 1)) { + ue_dlsch_procedures(ue, + proc, + eNB_id, + RA_PDSCH, + ue->dlsch_ra[eNB_id], + NULL, + &ue->dlsch_ra_errors[eNB_id], + mode, + abstraction_flag); + ue->dlsch_ra[eNB_id]->active = 0; + } + +#if UE_TIMING_TRACE + stop_meas(&ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); +#if DISABLE_LOG_X + printf("[AbsSFN %d.%d] Channel Decoder: %5.2f \n",frame_rx,nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#else + LOG_D(PHY, "[AbsSFN %d.%d] Channel Decoder: %5.2f \n",frame_rx,nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#endif + +#endif + + // duplicate harq structure + uint8_t current_harq_pid = ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->current_harq_pid; + NR_DL_UE_HARQ_t *current_harq_processes = ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_processes[current_harq_pid]; + nr_harq_status_t *current_harq_ack = &ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_ack[nr_tti_rx]; + + // For Debug parallelisation + //if (current_harq_ack->ack == 0) { + //printf("[slot0 dl processing][End of Channel Decoding] AbsSubframe %d.%d Decode Fail for HarqId%d Round%d\n",frame_rx,nr_tti_rx,current_harq_pid,current_harq_processes->round); + //} + for(uint8_t rx_th_idx=1; rx_th_idxdlsch[ue->current_thread_id[(nr_tti_rx+rx_th_idx)%10]][eNB_id][0]->harq_processes[current_harq_pid]; + nr_harq_status_t *harq_ack_dest = &ue->dlsch[ue->current_thread_id[(nr_tti_rx+rx_th_idx)%10]][eNB_id][0]->harq_ack[nr_tti_rx]; + + copy_harq_proc_struct(harq_processes_dest, current_harq_processes); + copy_ack_struct(harq_ack_dest, current_harq_ack); + + } + /* + NR_DL_UE_HARQ_t *harq_processes_dest = ue->dlsch[(nr_tti_rx+1)%RX_NB_TH][eNB_id][0]->harq_processes[current_harq_pid]; + NR_DL_UE_HARQ_t *harq_processes_dest1 = ue->dlsch[(nr_tti_rx+2)%RX_NB_TH][eNB_id][0]->harq_processes[current_harq_pid]; + + nr_harq_status_t *current_harq_ack = &ue->dlsch[nr_tti_rx%RX_NB_TH][eNB_id][0]->harq_ack[nr_tti_rx]; + nr_harq_status_t *harq_ack_dest = &ue->dlsch[(nr_tti_rx+1)%RX_NB_TH][eNB_id][0]->harq_ack[nr_tti_rx]; + nr_harq_status_t *harq_ack_dest1 = &ue->dlsch[(nr_tti_rx+2)%RX_NB_TH][eNB_id][0]->harq_ack[nr_tti_rx]; + + copy_harq_proc_struct(harq_processes_dest, current_harq_processes); + copy_ack_struct(harq_ack_dest, current_harq_ack); + + copy_harq_proc_struct(harq_processes_dest1, current_harq_processes); + copy_ack_struct(harq_ack_dest1, current_harq_ack); + */ + if (nr_tti_rx==9) { + if (frame_rx % 10 == 0) { + if ((ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]) != 0) + ue->dlsch_fer[eNB_id] = (100*(ue->dlsch_errors[eNB_id] - ue->dlsch_errors_last[eNB_id]))/(ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]); + + ue->dlsch_errors_last[eNB_id] = ue->dlsch_errors[eNB_id]; + ue->dlsch_received_last[eNB_id] = ue->dlsch_received[eNB_id]; + } + + + ue->bitrate[eNB_id] = (ue->total_TBS[eNB_id] - ue->total_TBS_last[eNB_id])*100; + ue->total_TBS_last[eNB_id] = ue->total_TBS[eNB_id]; + LOG_D(PHY,"[UE %d] Calculating bitrate Frame %d: total_TBS = %d, total_TBS_last = %d, bitrate %f kbits\n", + ue->Mod_id,frame_rx,ue->total_TBS[eNB_id], + ue->total_TBS_last[eNB_id],(float) ue->bitrate[eNB_id]/1000.0); + +#if UE_AUTOTEST_TRACE + if ((frame_rx % 100 == 0)) { + LOG_I(PHY,"[UE %d] AUTOTEST Metric : UE_DLSCH_BITRATE = %5.2f kbps (frame = %d) \n", ue->Mod_id, (float) ue->bitrate[eNB_id]/1000.0, frame_rx); + } +#endif + + } + +#ifdef EMOS + phy_procedures_emos_UE_RX(ue,slot,eNB_id); +#endif + + + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_OUT); + +#if UE_TIMING_TRACE + stop_meas(&ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]]); +#if DISABLE_LOG_X + printf("------FULL RX PROC [AbsSFN %d.%d]: %5.2f ------\n",frame_rx,nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#else + LOG_D(PHY, "------FULL RX PROC [AbsSFN %d.%d]: %5.2f ------\n",frame_rx,nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +#endif +#endif + + LOG_D(PHY," ****** end RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); + return (0); +} +#endif + + +int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eNB_id, + uint8_t do_pdcch_flag,runmode_t mode) { + + + + int l,l2; + int pilot1; + + int frame_rx = proc->frame_rx; + int nr_tti_rx = proc->nr_tti_rx; + uint16_t nb_symb_sch = 8; // to be updated by higher layer + uint8_t nb_symb_pdcch =2; + //proc->decoder_switch = 0; + //int counter_decoder = 0; + + LOG_D(PHY," ****** start RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); + + uint8_t next1_thread_id = ue->current_thread_id[nr_tti_rx]== (RX_NB_TH-1) ? 0:(ue->current_thread_id[nr_tti_rx]+1); + uint8_t next2_thread_id = next1_thread_id== (RX_NB_TH-1) ? 0:(next1_thread_id+1); + +#if 0 + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_IN); + +#if T_TRACER + T(T_UE_PHY_DL_TICK, T_INT(ue->Mod_id), T_INT(frame_rx%1024), T_INT(nr_tti_rx)); + + T(T_UE_PHY_INPUT_SIGNAL, T_INT(ue->Mod_id), T_INT(frame_rx%1024), T_INT(nr_tti_rx), T_INT(0), + T_BUFFER(&ue->common_vars.rxdata[0][nr_tti_rx*ue->frame_parms.samples_per_subframe], + ue->frame_parms.samples_per_subframe * 4)); +#endif + + // start timers + //#ifdef UE_DEBUG_TRACE + LOG_I(PHY," ****** start RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); + //#endif + +#if UE_TIMING_TRACE + start_meas(&ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]]); + start_meas(&ue->generic_stat); +#endif + + if (do_pdcch_flag) { + // deactivate reception until we scan pdcch + if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]) + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active = 0; + if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]) + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1]->active = 0; + + if (ue->dlsch_SI[eNB_id]) + ue->dlsch_SI[eNB_id]->active = 0; + if (ue->dlsch_p[eNB_id]) + ue->dlsch_p[eNB_id]->active = 0; + if (ue->dlsch_ra[eNB_id]) + ue->dlsch_ra[eNB_id]->active = 0; + } + +#ifdef DEBUG_PHY_PROC + LOG_D(PHY,"[%s %d] Frame %d nr_tti_rx %d: Doing phy_procedures_UE_RX\n", + (r_type == multicast_relay) ? "RN/UE" : "UE", + ue->Mod_id,frame_rx, nr_tti_rx); +#endif + + + if (ue->frame_parms.Ncp == 0) { // normal prefix + pilot1 = 4; + } else { // extended prefix + pilot1 = 3; + } + + /* + if (nr_subframe_select(&ue->frame_parms,nr_tti_rx) == SF_S) { // S-subframe, do first 5 symbols only + l2 = 5; + } else */ + { // normal nr_tti_rx, last symbol to be processed is the first of the second slot + l2 = (ue->frame_parms.symbols_per_tti/2)-1; + } + + int prev_nr_tti_rx = (nr_tti_rx - 1)<0? 9: (nr_tti_rx - 1);/* + if (nr_subframe_select(&ue->frame_parms,prev_nr_tti_rx) != SF_DL) { + //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ + // RX processing of symbols l=0...l2 + //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ + l=0; + } else */ + { + //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ + // RX processing of symbols l=1...l2 (l=0 is done in last scheduling epoch) + //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ + l=1; + } + + LOG_D(PHY," ------ slot 0 Processing: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); + LOG_D(PHY," ------ --> FFT/ChannelEst/PDCCH slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); +#endif + +#ifdef NR_PDCCH_SCHED + //nr_gold_pdcch(ue,0, 2); + + if (nr_tti_rx==1){ + for (uint16_t l=0; lofdm_demod_stats); +#endif + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_IN); + nr_slot_fep(ue, + l, + nr_tti_rx<<1, + 0, + 0, + 1, + NR_PDCCH_EST); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_OUT); +#if UE_TIMING_TRACE + stop_meas(&ue->ofdm_demod_stats); +#endif + + //printf("phy procedure pdcch start measurement l =%d\n",l); + nr_ue_measurement_procedures(l,ue,proc,eNB_id,(nr_tti_rx<<1),mode); + + } + + if (nr_ue_pdcch_procedures(eNB_id,ue,proc) == -1) { + LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d: Error in pdcch procedures\n",ue->Mod_id,frame_rx,nr_tti_rx); + return(-1); + } + } +#endif //NR_PDCCH_SCHED + + LOG_D(PHY," ------ --> PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); + + if (nr_tti_rx==1){ + //to update from pdsch config + nr_gold_pdsch(ue,nb_symb_pdcch,0, 1); + + for (uint16_t m=nb_symb_pdcch;m<=(nb_symb_sch+nb_symb_pdcch-1) ; m++){ + nr_slot_fep(ue, + m, //to be updated from higher layer + nr_tti_rx<<1, + 0, + 0, + 1, + NR_PDSCH_EST); + + //printf("phy procedure pdsch start measurement\n"); + nr_ue_measurement_procedures(m,ue,proc,eNB_id,(nr_tti_rx<<1),mode); + + } + //set active for testing, to be removed + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active = 0; + } + else + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active = 0; + +#if UE_TIMING_TRACE + start_meas(&ue->generic_stat); +#endif + // do procedures for C-RNTI + if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); + nr_ue_pdsch_procedures(ue, + proc, + eNB_id, + PDSCH, + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], + NULL, + nb_symb_pdcch, //ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, + (nb_symb_sch+nb_symb_pdcch-1)); //ue->frame_parms.symbols_per_tti>>1, + + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT); + } + + LOG_D(PHY," ------ end PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); + // do procedures for SI-RNTI + if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) { + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_SI, VCD_FUNCTION_IN); + nr_ue_pdsch_procedures(ue, + proc, + eNB_id, + SI_PDSCH, + ue->dlsch_SI[eNB_id], + NULL, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, + ue->frame_parms.symbols_per_tti>>1); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_SI, VCD_FUNCTION_OUT); + } + + // do procedures for SI-RNTI + if ((ue->dlsch_p[eNB_id]) && (ue->dlsch_p[eNB_id]->active == 1)) { + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_P, VCD_FUNCTION_IN); + nr_ue_pdsch_procedures(ue, + proc, + eNB_id, + P_PDSCH, + ue->dlsch_p[eNB_id], + NULL, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, + ue->frame_parms.symbols_per_tti>>1); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_P, VCD_FUNCTION_OUT); + } + + // do procedures for RA-RNTI + if ((ue->dlsch_ra[eNB_id]) && (ue->dlsch_ra[eNB_id]->active == 1)) { + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_RA, VCD_FUNCTION_IN); + nr_ue_pdsch_procedures(ue, + proc, + eNB_id, + RA_PDSCH, + ue->dlsch_ra[eNB_id], + NULL, + ue->pdcch_vars[ue->current_thread_id[nr_tti_rx]][eNB_id]->num_pdcch_symbols, + ue->frame_parms.symbols_per_tti>>1); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC_RA, VCD_FUNCTION_OUT); + } +//#if 0 LOG_D(PHY," ------ slot 1 Processing: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); LOG_D(PHY," ------ --> FFT/ChannelEst/PDCCH slot 1: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); - if (nr_subframe_select(&ue->frame_parms,nr_tti_rx) != SF_S) { // do front-end processing for second slot, and first symbol of next nr_tti_rx + /*if (nr_subframe_select(&ue->frame_parms,nr_tti_rx) != SF_S)*/ + { // do front-end processing for second slot, and first symbol of next nr_tti_rx for (l=1; lframe_parms.symbols_per_tti>>1; l++) { - if (abstraction_flag == 0) { #if UE_TIMING_TRACE start_meas(&ue->ofdm_demod_stats); #endif VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_IN); - nr_slot_fep(ue, + /*nr_slot_fep(ue, l, 1+(nr_tti_rx<<1), 0, 0, 0, - NR_PDSCH_EST); + NR_PDSCH_EST);*/ VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_UE_SLOT_FEP, VCD_FUNCTION_OUT); + #if UE_TIMING_TRACE - stop_meas(&ue->ofdm_demod_stats); + stop_meas(&ue->ofdm_demod_stats); #endif - } + - //ue_measurement_procedures(l-1,ue,proc,eNB_id,1+(nr_tti_rx<<1),abstraction_flag,mode); + //ue_measurement_procedures(l-1,ue,proc,eNB_id,1+(nr_tti_rx<<1),abstraction_flag,mode); - } // for l=1..l2 + } // for l=1..l2 // do first symbol of next downlink nr_tti_rx for channel estimation - int next_nr_tti_rx = (1+nr_tti_rx)%10; - if (nr_subframe_select(&ue->frame_parms,next_nr_tti_rx) != SF_UL) + int next_nr_tti_rx = (1+nr_tti_rx)%10; + /* if (nr_subframe_select(&ue->frame_parms,next_nr_tti_rx) != SF_UL)*/ { - nr_slot_fep(ue, + /*nr_slot_fep(ue, 0, (next_nr_tti_rx<<1), 0, 0, 0, - NR_PDSCH_EST); + NR_PDSCH_EST);*/ } } // not an S-subframe #if UE_TIMING_TRACE @@ -6137,27 +5217,37 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN #else LOG_D(PHY, "[SFN %d] Slot1: FFT + Channel Estimate + Pdsch Proc Slot0 %5.2f \n",nr_tti_rx,ue->generic_stat.p_time/(cpuf*1000.0)); #endif - + #endif - LOG_D(PHY," ------ end FFT/ChannelEst/PDCCH slot 1: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); + //LOG_D(PHY," ------ end FFT/ChannelEst/PDCCH slot 1: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); if ( (nr_tti_rx == 0) && (ue->decode_MIB == 1)) - { - ue_pbch_procedures(eNB_id,ue,proc,abstraction_flag); - } - + { + for (int i=0; i<3; i++) + nr_slot_fep(ue, + (5+i), //mu=1 case B + nr_tti_rx, + 0, + 0, + 1, + NR_PBCH_EST); + + nr_ue_pbch_procedures(eNB_id,ue,proc,0); + } + // do procedures for C-RNTI LOG_D(PHY," ------ --> PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); if (ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->active == 1) { - + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN); #if UE_TIMING_TRACE + start_meas(&ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); #endif - ue_pdsch_procedures(ue, - proc, - eNB_id, + /*nr_ue_pdsch_procedures(ue, + proc, + eNB_id, PDSCH, ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], NULL, @@ -6165,40 +5255,45 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN ue->frame_parms.symbols_per_tti-1, abstraction_flag); LOG_D(PHY," ------ end PDSCH ChannelComp/LLR slot 0: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); - LOG_D(PHY," ------ --> PDSCH Turbo Decoder slot 0/1: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx); + LOG_D(PHY," ------ --> PDSCH Turbo Decoder slot 0/1: AbsSubframe %d.%d ------ \n", frame_rx%1024, nr_tti_rx);*/ #if UE_TIMING_TRACE stop_meas(&ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); start_meas(&ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); #endif - /*ue_dlsch_procedures(ue, - proc, - eNB_id, - PDSCH, - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], - ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1], - &ue->dlsch_errors[eNB_id], - mode, - abstraction_flag);*/ + +//#endif //slot 1 + + nr_ue_dlsch_procedures(ue, + proc, + eNB_id, + PDSCH, + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0], + ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][1], + &ue->dlsch_errors[eNB_id], + mode); + + #if UE_TIMING_TRACE - stop_meas(&ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); + stop_meas(&ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]]); #if DISABLE_LOG_X - printf("[SFN %d] Slot1: Pdsch Proc %5.2f\n",nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); - printf("[SFN %d] Slot0 Slot1: Dlsch Proc %5.2f\n",nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); + printf("[SFN %d] Slot1: Pdsch Proc %5.2f\n",nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); + printf("[SFN %d] Slot0 Slot1: Dlsch Proc %5.2f\n",nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); #else - LOG_D(PHY, "[SFN %d] Slot1: Pdsch Proc %5.2f\n",nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); - LOG_D(PHY, "[SFN %d] Slot0 Slot1: Dlsch Proc %5.2f\n",nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); + LOG_D(PHY, "[SFN %d] Slot1: Pdsch Proc %5.2f\n",nr_tti_rx,ue->pdsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); + LOG_D(PHY, "[SFN %d] Slot0 Slot1: Dlsch Proc %5.2f\n",nr_tti_rx,ue->dlsch_procedures_stat[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); #endif #endif - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT); + VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT); - } + } #if UE_TIMING_TRACE - start_meas(&ue->generic_stat); +start_meas(&ue->generic_stat); #endif #if 0 + if(nr_tti_rx==5 && ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_processes[ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->current_harq_pid]->nb_rb > 20){ //write_output("decoder_llr.m","decllr",dlsch_llr,G,1,0); //write_output("llr.m","llr", &ue->pdsch_vars[eNB_id]->llr[0][0],(14*nb_rb*12*dlsch1_harq->Qm) - 4*(nb_rb*4*dlsch1_harq->Qm),1,0); @@ -6220,15 +5315,14 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN // do procedures for SI-RNTI if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) { - ue_pdsch_procedures(ue, - proc, - eNB_id, - SI_PDSCH, - ue->dlsch_SI[eNB_id], - NULL, - 1+(ue->frame_parms.symbols_per_tti>>1), - ue->frame_parms.symbols_per_tti-1, - abstraction_flag); + nr_ue_pdsch_procedures(ue, + proc, + eNB_id, + SI_PDSCH, + ue->dlsch_SI[eNB_id], + NULL, + 1+(ue->frame_parms.symbols_per_tti>>1), + ue->frame_parms.symbols_per_tti-1); /*ue_dlsch_procedures(ue, proc, @@ -6244,15 +5338,15 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN // do procedures for P-RNTI if ((ue->dlsch_p[eNB_id]) && (ue->dlsch_p[eNB_id]->active == 1)) { - ue_pdsch_procedures(ue, - proc, - eNB_id, - P_PDSCH, - ue->dlsch_p[eNB_id], - NULL, - 1+(ue->frame_parms.symbols_per_tti>>1), - ue->frame_parms.symbols_per_tti-1, - abstraction_flag); + nr_ue_pdsch_procedures(ue, + proc, + eNB_id, + P_PDSCH, + ue->dlsch_p[eNB_id], + NULL, + 1+(ue->frame_parms.symbols_per_tti>>1), + ue->frame_parms.symbols_per_tti-1); + /*ue_dlsch_procedures(ue, proc, @@ -6267,15 +5361,15 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN } // do procedures for RA-RNTI if ((ue->dlsch_ra[eNB_id]) && (ue->dlsch_ra[eNB_id]->active == 1)) { - ue_pdsch_procedures(ue, - proc, - eNB_id, - RA_PDSCH, - ue->dlsch_ra[eNB_id], - NULL, - 1+(ue->frame_parms.symbols_per_tti>>1), - ue->frame_parms.symbols_per_tti-1, - abstraction_flag); + nr_ue_pdsch_procedures(ue, + proc, + eNB_id, + RA_PDSCH, + ue->dlsch_ra[eNB_id], + NULL, + 1+(ue->frame_parms.symbols_per_tti>>1), + ue->frame_parms.symbols_per_tti-1); + /*ue_dlsch_procedures(ue, proc, eNB_id, @@ -6298,63 +5392,63 @@ int phy_procedures_nrUE_RX(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN /*nr_harq_status_t *current_harq_ack = &ue->dlsch[ue->current_thread_id[nr_tti_rx]][eNB_id][0]->harq_ack[nr_tti_rx]; nr_harq_status_t *harq_ack_dest = &ue->dlsch[next1_thread_id][eNB_id][0]->harq_ack[nr_tti_rx]; nr_harq_status_t *harq_ack_dest1 = &ue->dlsch[next2_thread_id][eNB_id][0]->harq_ack[nr_tti_rx]; - */ +*/ - copy_harq_proc_struct(harq_processes_dest, current_harq_processes); - //copy_ack_struct(harq_ack_dest, current_harq_ack); +copy_harq_proc_struct(harq_processes_dest, current_harq_processes); +//copy_ack_struct(harq_ack_dest, current_harq_ack); - copy_harq_proc_struct(harq_processes_dest1, current_harq_processes); - //copy_ack_struct(harq_ack_dest1, current_harq_ack); +copy_harq_proc_struct(harq_processes_dest1, current_harq_processes); +//copy_ack_struct(harq_ack_dest1, current_harq_ack); - if (nr_tti_rx==9) { - if (frame_rx % 10 == 0) { - if ((ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]) != 0) - ue->dlsch_fer[eNB_id] = (100*(ue->dlsch_errors[eNB_id] - ue->dlsch_errors_last[eNB_id]))/(ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]); - - ue->dlsch_errors_last[eNB_id] = ue->dlsch_errors[eNB_id]; - ue->dlsch_received_last[eNB_id] = ue->dlsch_received[eNB_id]; - } - - - ue->bitrate[eNB_id] = (ue->total_TBS[eNB_id] - ue->total_TBS_last[eNB_id])*100; - ue->total_TBS_last[eNB_id] = ue->total_TBS[eNB_id]; - LOG_D(PHY,"[UE %d] Calculating bitrate Frame %d: total_TBS = %d, total_TBS_last = %d, bitrate %f kbits\n", - ue->Mod_id,frame_rx,ue->total_TBS[eNB_id], - ue->total_TBS_last[eNB_id],(float) ue->bitrate[eNB_id]/1000.0); - - #if UE_AUTOTEST_TRACE - if ((frame_rx % 100 == 0)) { - LOG_I(PHY,"[UE %d] AUTOTEST Metric : UE_DLSCH_BITRATE = %5.2f kbps (frame = %d) \n", ue->Mod_id, (float) ue->bitrate[eNB_id]/1000.0, frame_rx); - } - #endif +if (nr_tti_rx==9) { + if (frame_rx % 10 == 0) { + if ((ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]) != 0) + ue->dlsch_fer[eNB_id] = (100*(ue->dlsch_errors[eNB_id] - ue->dlsch_errors_last[eNB_id]))/(ue->dlsch_received[eNB_id] - ue->dlsch_received_last[eNB_id]); + ue->dlsch_errors_last[eNB_id] = ue->dlsch_errors[eNB_id]; + ue->dlsch_received_last[eNB_id] = ue->dlsch_received[eNB_id]; } + + ue->bitrate[eNB_id] = (ue->total_TBS[eNB_id] - ue->total_TBS_last[eNB_id])*100; + ue->total_TBS_last[eNB_id] = ue->total_TBS[eNB_id]; + LOG_D(PHY,"[UE %d] Calculating bitrate Frame %d: total_TBS = %d, total_TBS_last = %d, bitrate %f kbits\n", + ue->Mod_id,frame_rx,ue->total_TBS[eNB_id], + ue->total_TBS_last[eNB_id],(float) ue->bitrate[eNB_id]/1000.0); + +#if UE_AUTOTEST_TRACE + if ((frame_rx % 100 == 0)) { + LOG_I(PHY,"[UE %d] AUTOTEST Metric : UE_DLSCH_BITRATE = %5.2f kbps (frame = %d) \n", ue->Mod_id, (float) ue->bitrate[eNB_id]/1000.0, frame_rx); + } +#endif + + } + #if UE_TIMING_TRACE - stop_meas(&ue->generic_stat); - printf("after tubo until end of Rx %5.2f \n",ue->generic_stat.p_time/(cpuf*1000.0)); +stop_meas(&ue->generic_stat); +printf("after tubo until end of Rx %5.2f \n",ue->generic_stat.p_time/(cpuf*1000.0)); #endif #ifdef EMOS - phy_procedures_emos_UE_RX(ue,slot,eNB_id); +phy_procedures_emos_UE_RX(ue,slot,eNB_id); #endif - VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_OUT); +VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_OUT); #if UE_TIMING_TRACE - stop_meas(&ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]]); +stop_meas(&ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]]); #if DISABLE_LOG_X - printf("------FULL RX PROC [SFN %d]: %5.2f ------\n",nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +printf("------FULL RX PROC [SFN %d]: %5.2f ------\n",nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); #else - LOG_D(PHY, "------FULL RX PROC [SFN %d]: %5.2f ------\n",nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); +LOG_D(PHY, "------FULL RX PROC [SFN %d]: %5.2f ------\n",nr_tti_rx,ue->phy_proc_rx[ue->current_thread_id[nr_tti_rx]].p_time/(cpuf*1000.0)); #endif #endif -#endif //pdsch +//#endif //pdsch - LOG_D(PHY," ****** end RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); - return (0); +LOG_D(PHY," ****** end RX-Chain for AbsSubframe %d.%d ****** \n", frame_rx%1024, nr_tti_rx); +return (0); } @@ -6404,10 +5498,10 @@ void phy_procedures_UE_lte(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN int frame_tx = proc->frame_tx; int nr_tti_rx = proc->nr_tti_rx; int nr_tti_tx = proc->nr_tti_tx; -//#ifdef UE_NR_PHY_DEMO -// int nr_tti_rx = proc->nr_tti_rx; -// int nr_tti_tx = proc->nr_tti_tx; -//#endif + //#ifdef UE_NR_PHY_DEMO + // int nr_tti_rx = proc->nr_tti_rx; + // int nr_tti_tx = proc->nr_tti_tx; + //#endif #undef DEBUG_PHY_PROC @@ -6436,14 +5530,14 @@ void phy_procedures_UE_lte(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN switch (ITTI_MSG_ID(msg_p)) { case PHY_FIND_CELL_REQ: - LOG_I(PHY, "[UE %d] Received %s\n", Mod_id, msg_name); + LOG_I(PHY, "[UE %d] Received %s\n", Mod_id, msg_name); - /* TODO process the message */ - break; + /* TODO process the message */ + break; default: - LOG_E(PHY, "[UE %d] Received unexpected message %s\n", Mod_id, msg_name); - break; + LOG_E(PHY, "[UE %d] Received unexpected message %s\n", Mod_id, msg_name); + break; } result = itti_free (ITTI_MSG_ORIGIN_ID(msg_p), msg_p); @@ -6456,12 +5550,12 @@ void phy_procedures_UE_lte(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN for (slot=0;slot<2;slot++) { if ((nr_subframe_select(&ue->frame_parms,nr_tti_tx)==SF_UL)|| - (ue->frame_parms.frame_type == FDD)) { + (ue->frame_parms.frame_type == FDD)) { phy_procedures_UE_TX(ue,proc,eNB_id,abstraction_flag,mode,r_type); } if ((nr_subframe_select(&ue->frame_parms,nr_tti_rx)==SF_DL) || - (ue->frame_parms.frame_type == FDD)) { + (ue->frame_parms.frame_type == FDD)) { #if defined(Rel10) || defined(Rel14) if (phy_procedures_RN_UE_RX(nr_tti_rx, nr_tti_tx, r_type) != 0 ) @@ -6470,12 +5564,12 @@ void phy_procedures_UE_lte(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN } if ((nr_subframe_select(&ue->frame_parms,nr_tti_tx)==SF_S) && - (slot==1)) { + (slot==1)) { phy_procedures_UE_S_TX(ue,eNB_id,abstraction_flag,r_type); } if ((nr_subframe_select(&ue->frame_parms,nr_tti_rx)==SF_S) && - (slot==0)) { + (slot==0)) { #if defined(Rel10) || defined(Rel14) if (phy_procedures_RN_UE_RX(nr_tti_rx, nr_tti_tx, r_type) != 0 ) @@ -6485,44 +5579,44 @@ void phy_procedures_UE_lte(PHY_VARS_NR_UE *ue,UE_nr_rxtx_proc_t *proc,uint8_t eN if (ue->mac_enabled==1) { if (slot==0) { -//#ifdef UE_NR_PHY_DEMO + //#ifdef UE_NR_PHY_DEMO ret = mac_xface->ue_scheduler(ue->Mod_id, - frame_rx, - proc->subframe_rx, - nr_tti_rx%(ue->frame_parms.ttis_per_subframe), - frame_tx, - proc->subframe_tx, - nr_tti_tx%(ue->frame_parms.ttis_per_subframe), - nr_subframe_select(&ue->frame_parms,nr_tti_tx), - eNB_id, - 0/*FIXME CC_id*/); -//#else -/* ret = mac_xface->ue_scheduler(ue->Mod_id, - frame_rx, - nr_tti_rx, - frame_tx, - nr_tti_tx, - nr_subframe_select(&ue->frame_parms,nr_tti_tx), - eNB_id,*/ -// 0/*FIXME CC_id*/); -//#endif + frame_rx, + proc->subframe_rx, + nr_tti_rx%(ue->frame_parms.ttis_per_subframe), + frame_tx, + proc->subframe_tx, + nr_tti_tx%(ue->frame_parms.ttis_per_subframe), + nr_subframe_select(&ue->frame_parms,nr_tti_tx), + eNB_id, + 0/*FIXME CC_id*/); + //#else + /* ret = mac_xface->ue_scheduler(ue->Mod_id, + frame_rx, + nr_tti_rx, + frame_tx, + nr_tti_tx, + nr_subframe_select(&ue->frame_parms,nr_tti_tx), + eNB_id,*/ + // 0/*FIXME CC_id*/); + //#endif - if (ret == CONNECTION_LOST) { - LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d RRC Connection lost, returning to PRACH\n",ue->Mod_id, - frame_rx,nr_tti_tx); - ue->UE_mode[eNB_id] = PRACH; - // mac_xface->macphy_exit("Connection lost"); - } else if (ret == PHY_RESYNCH) { - LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d RRC Connection lost, trying to resynch\n", - ue->Mod_id, - frame_rx,nr_tti_tx); - ue->UE_mode[eNB_id] = RESYNCH; - // mac_xface->macphy_exit("Connection lost"); - } else if (ret == PHY_HO_PRACH) { - LOG_I(PHY,"[UE %d] Frame %d, nr_tti_rx %d, return to PRACH and perform a contention-free access\n", - ue->Mod_id,frame_rx,nr_tti_tx); - ue->UE_mode[eNB_id] = PRACH; - } + if (ret == CONNECTION_LOST) { + LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d RRC Connection lost, returning to PRACH\n",ue->Mod_id, + frame_rx,nr_tti_tx); + ue->UE_mode[eNB_id] = PRACH; + // mac_xface->macphy_exit("Connection lost"); + } else if (ret == PHY_RESYNCH) { + LOG_E(PHY,"[UE %d] Frame %d, nr_tti_rx %d RRC Connection lost, trying to resynch\n", + ue->Mod_id, + frame_rx,nr_tti_tx); + ue->UE_mode[eNB_id] = RESYNCH; + // mac_xface->macphy_exit("Connection lost"); + } else if (ret == PHY_HO_PRACH) { + LOG_I(PHY,"[UE %d] Frame %d, nr_tti_rx %d, return to PRACH and perform a contention-free access\n", + ue->Mod_id,frame_rx,nr_tti_tx); + ue->UE_mode[eNB_id] = PRACH; + } } } diff --git a/openair1/SCHED_NR_UE/pucch_uci_ue_nr.c b/openair1/SCHED_NR_UE/pucch_uci_ue_nr.c index bc23c5a974..fb38c80eb8 100644 --- a/openair1/SCHED_NR_UE/pucch_uci_ue_nr.c +++ b/openair1/SCHED_NR_UE/pucch_uci_ue_nr.c @@ -193,7 +193,7 @@ bool pucch_procedures_ue_nr(PHY_VARS_NR_UE *ue, uint8_t gNB_id, UE_nr_rxtx_proc_ if (dl_harq_pid < ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][gNB_id][0]->number_harq_processes_for_pdsch) { /* pucch indicator can be reseted in function get_downlink_ack so it should be get now */ - pucch_resource_indicator = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][gNB_id][0]->harq_processes[dl_harq_pid].harq_ack.pucch_resource_indicator; + pucch_resource_indicator = ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][gNB_id][0]->harq_processes[dl_harq_pid]->harq_ack.pucch_resource_indicator; } /* Part - I @@ -292,7 +292,7 @@ bool pucch_procedures_ue_nr(PHY_VARS_NR_UE *ue, uint8_t gNB_id, UE_nr_rxtx_proc_ int occ_length = 0; int occ_Index = 0; - NR_UE_HARQ_STATUS_t *harq_status = &ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][gNB_id][0]->harq_processes[dl_harq_pid].harq_ack; + NR_UE_HARQ_STATUS_t *harq_status = &ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][gNB_id][0]->harq_processes[dl_harq_pid]->harq_ack; if (select_pucch_resource(ue, gNB_id, N_UCI, pucch_resource_indicator, &initial_pucch_id, &pucch_resource_set, &pucch_resource_id, harq_status) == TRUE) { @@ -691,7 +691,7 @@ uint8_t get_downlink_ack(PHY_VARS_NR_UE *ue, uint8_t gNB_id, UE_nr_rxtx_proc_t for (int dl_harq_pid = 0; dl_harq_pid < number_pid_dl; dl_harq_pid++) { - harq_status = &ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][gNB_id][code_word]->harq_processes[dl_harq_pid].harq_ack; + harq_status = &ue->dlsch[ue->current_thread_id[proc->nr_tti_rx]][gNB_id][code_word]->harq_processes[dl_harq_pid]->harq_ack; /* check if current tx slot should transmit downlink acknowlegment */ if (harq_status->slot_for_feedback_ack == proc->nr_tti_tx) { diff --git a/openair1/SIMULATION/LTE_PHY/dlsim.c b/openair1/SIMULATION/LTE_PHY/dlsim.c index 63e6de2290..3807c51b96 100644 --- a/openair1/SIMULATION/LTE_PHY/dlsim.c +++ b/openair1/SIMULATION/LTE_PHY/dlsim.c @@ -1650,7 +1650,7 @@ int main(int argc, char **argv) start_meas(&eNB->ofdm_mod_stats); - ru->proc.subframe_tx=subframe; + ru->proc.tti_tx=subframe; memcpy((void*)ru->frame_parms,(void*)&eNB->frame_parms,sizeof(LTE_DL_FRAME_PARMS)); feptx_prec(ru); feptx_ofdm(ru); @@ -1670,7 +1670,7 @@ int main(int argc, char **argv) phy_procedures_eNB_TX(eNB,proc_eNB,0); - ru->proc.subframe_tx=(subframe+1)%10; + ru->proc.tti_tx=(subframe+1)%10; feptx_prec(ru); feptx_ofdm(ru); diff --git a/openair1/SIMULATION/LTE_PHY/ulsim.c b/openair1/SIMULATION/LTE_PHY/ulsim.c index 6bdec77b76..8d963b9186 100644 --- a/openair1/SIMULATION/LTE_PHY/ulsim.c +++ b/openair1/SIMULATION/LTE_PHY/ulsim.c @@ -1158,7 +1158,7 @@ int main(int argc, char **argv) { eNB->proc.frame_rx = 1; eNB->proc.subframe_rx = subframe; ru->proc.frame_rx = 1; - ru->proc.subframe_rx = subframe; + ru->proc.tti_rx = subframe; proc_rxtx_ue->frame_tx = proc_rxtx->frame_rx; proc_rxtx_ue->frame_rx = proc_rxtx->frame_tx; proc_rxtx_ue->subframe_tx = proc_rxtx->subframe_rx; diff --git a/openair1/SIMULATION/NR_PHY/dlschsim.c b/openair1/SIMULATION/NR_PHY/dlschsim.c new file mode 100644 index 0000000000..1c2b2305d6 --- /dev/null +++ b/openair1/SIMULATION/NR_PHY/dlschsim.c @@ -0,0 +1,668 @@ +/* + * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The OpenAirInterface Software Alliance licenses this file to You under + * the OAI Public License, Version 1.1 (the "License"); you may not use this file + * except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.openairinterface.org/?page_id=698 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + *------------------------------------------------------------------------------- + * For more information about the OpenAirInterface (OAI) Software Alliance: + * contact@openairinterface.org + */ + +#include +#include +#include +#include +#include +#include + +#include "common/config/config_userapi.h" +#include "common/utils/LOG/log.h" +#include "common/ran_context.h" + +#include "SIMULATION/TOOLS/sim.h" +#include "SIMULATION/RF/rf.h" +#include "PHY/types.h" +#include "PHY/defs_nr_common.h" +#include "PHY/defs_nr_UE.h" +#include "PHY/defs_gNB.h" +#include "PHY/INIT/phy_init.h" +#include "PHY/NR_REFSIG/refsig_defs_ue.h" +#include "PHY/NR_REFSIG/nr_mod_table.h" +#include "PHY/MODULATION/modulation_eNB.h" +#include "PHY/MODULATION/modulation_UE.h" +#include "PHY/NR_TRANSPORT/nr_transport.h" +#include "PHY/NR_TRANSPORT/nr_dlsch.h" +#include "PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h" + +#include "SCHED_NR/sched_nr.h" + +//#include "PHY/MODULATION/modulation_common.h" +//#include "common/config/config_load_configmodule.h" +//#include "UTIL/LISTS/list.h" +//#include "common/ran_context.h" + +//#define DEBUG_DLSCHSIM + +PHY_VARS_gNB *gNB; +PHY_VARS_NR_UE *UE; +RAN_CONTEXT_t RC; + +double cpuf; + +// dummy functions +int nfapi_mode = 0; +int oai_nfapi_hi_dci0_req(nfapi_hi_dci0_request_t *hi_dci0_req) { + return (0); +} +int oai_nfapi_tx_req(nfapi_tx_request_t *tx_req) { + return (0); +} + +int oai_nfapi_dl_config_req(nfapi_dl_config_request_t *dl_config_req) { + return (0); +} + +int oai_nfapi_ul_config_req(nfapi_ul_config_request_t *ul_config_req) { + return (0); +} + +int oai_nfapi_nr_dl_config_req(nfapi_nr_dl_config_request_t *dl_config_req) { + return (0); +} + +uint32_t from_nrarfcn(int nr_bandP, uint32_t dl_earfcn) { + return (0); +} +int32_t get_uldl_offset(int eutra_bandP) { + return (0); +} + +NR_IF_Module_t * +NR_IF_Module_init(int Mod_id) { + return (NULL); +} + +void exit_function(const char *file, const char *function, const int line, + const char *s) { + const char *msg = s == NULL ? "no comment" : s; + printf("Exiting at: %s:%d %s(), %s\n", file, line, function, msg); + exit(-1); +} + +// needed for some functions +PHY_VARS_NR_UE *PHY_vars_UE_g[1][1] = { { NULL } }; +uint16_t n_rnti = 0x1234; + +char quantize(double D, double x, unsigned char B) { + double qxd; + short maxlev; + qxd = floor(x / D); + maxlev = 1 << (B - 1); //(char)(pow(2,B-1)); + + //printf("x=%f,qxd=%f,maxlev=%d\n",x,qxd, maxlev); + + if (qxd <= -maxlev) + qxd = -maxlev; + else if (qxd >= maxlev) + qxd = maxlev - 1; + + return ((char) qxd); +} + +int main(int argc, char **argv) { + char c; + int i; //,j,l,aa; + double SNR, SNR_lin, snr0 = -2.0, snr1 = 2.0; + double snr_step = 0.1; + uint8_t snr1set = 0; + int **txdata; + double **s_re, **s_im, **r_re, **r_im; + // int sync_pos, sync_pos_slot; + // FILE *rx_frame_file; + FILE *output_fd = NULL; + uint8_t write_output_file = 0; + // int subframe_offset; + // char fname[40], vname[40]; + int trial, n_trials = 1, n_errors = 0, n_false_positive = 0; + uint8_t transmission_mode = 1, n_tx = 1, n_rx = 1; + uint16_t Nid_cell = 0; + channel_desc_t *gNB2UE; + uint8_t extended_prefix_flag = 0; + int8_t interf1 = -21, interf2 = -21; + FILE *input_fd = NULL, *pbch_file_fd = NULL; + //char input_val_str[50],input_val_str2[50]; + //uint16_t NB_RB=25; + SCM_t channel_model = AWGN; //Rayleigh1_anticorr; + uint8_t N_RB_DL = 106, mu = 1; + unsigned char frame_type = 0; + unsigned char pbch_phase = 0; + int frame = 0, subframe = 0; + int frame_length_complex_samples; + //int frame_length_complex_samples_no_prefix; + NR_DL_FRAME_PARMS *frame_parms; + //nfapi_nr_config_request_t *gNB_config; + uint8_t Kmimo = 0; + uint32_t Nsoft = 0; + double sigma; + unsigned char qbits = 8; + int ret; + //int run_initial_sync=0; + int loglvl = OAILOG_WARNING; + float target_error_rate = 0.01; + + uint16_t nb_symb_sch = 12; + uint16_t nb_rb = 50; + uint8_t Imcs = 9; + + cpuf = get_cpu_freq_GHz(); + + if (load_configmodule(argc, argv) == 0) { + exit_fun("[SOFTMODEM] Error, configuration module init failed\n"); + } + + //logInit(); + randominit(0); + + while ((c = getopt(argc, argv, "df:hpg:i:j:n:l:m:r:s:S:y:z:N:F:R:P:")) != -1) { + switch (c) { + case 'f': + write_output_file = 1; + output_fd = fopen(optarg, "w"); + + if (output_fd == NULL) { + printf("Error opening %s\n", optarg); + exit(-1); + } + + break; + + case 'd': + frame_type = 1; + break; + + case 'g': + switch ((char) *optarg) { + case 'A': + channel_model = SCM_A; + break; + + case 'B': + channel_model = SCM_B; + break; + + case 'C': + channel_model = SCM_C; + break; + + case 'D': + channel_model = SCM_D; + break; + + case 'E': + channel_model = EPA; + break; + + case 'F': + channel_model = EVA; + break; + + case 'G': + channel_model = ETU; + break; + + default: + msg("Unsupported channel model!\n"); + exit(-1); + } + + break; + + case 'i': + interf1 = atoi(optarg); + break; + + case 'j': + interf2 = atoi(optarg); + break; + + case 'n': + n_trials = atoi(optarg); + break; + + case 's': + snr0 = atof(optarg); + msg("Setting SNR0 to %f\n", snr0); + break; + + case 'S': + snr1 = atof(optarg); + snr1set = 1; + msg("Setting SNR1 to %f\n", snr1); + break; + + case 'p': + extended_prefix_flag = 1; + break; + + /* + case 'r': + ricean_factor = pow(10,-.1*atof(optarg)); + if (ricean_factor>1) { + printf("Ricean factor must be between 0 and 1\n"); + exit(-1); + } + break; + */ + + case 'y': + n_tx = atoi(optarg); + + if ((n_tx == 0) || (n_tx > 2)) { + msg("Unsupported number of tx antennas %d\n", n_tx); + exit(-1); + } + + break; + + case 'z': + n_rx = atoi(optarg); + + if ((n_rx == 0) || (n_rx > 2)) { + msg("Unsupported number of rx antennas %d\n", n_rx); + exit(-1); + } + + break; + + case 'N': + Nid_cell = atoi(optarg); + break; + + case 'R': + N_RB_DL = atoi(optarg); + break; + + case 'F': + input_fd = fopen(optarg, "r"); + + if (input_fd == NULL) { + printf("Problem with filename %s\n", optarg); + exit(-1); + } + + break; + + case 'P': + pbch_phase = atoi(optarg); + + if (pbch_phase > 3) + printf("Illegal PBCH phase (0-3) got %d\n", pbch_phase); + + break; + + case 'm': + Imcs = atoi(optarg); + break; + + case 'l': + nb_symb_sch = atoi(optarg); + break; + + case 'r': + nb_rb = atoi(optarg); + break; + + case 'x': + transmission_mode = atoi(optarg); + break; + + default: + case 'h': + printf("%s -h(elp) -p(extended_prefix) -N cell_id -f output_filename -F input_filename -g channel_model -n n_frames -t Delayspread -s snr0 -S snr1 -x transmission_mode -y TXant -z RXant -i Intefrence0 -j Interference1 -A interpolation_file -C(alibration offset dB) -N CellId\n", argv[0]); + printf("-h This message\n"); + printf("-p Use extended prefix mode\n"); + printf("-d Use TDD\n"); + printf("-n Number of frames to simulate\n"); + printf("-s Starting SNR, runs from SNR0 to SNR0 + 5 dB. If n_frames is 1 then just SNR is simulated\n"); + printf("-S Ending SNR, runs from SNR0 to SNR1\n"); + printf("-t Delay spread for multipath channel\n"); + printf("-g [A,B,C,D,E,F,G] Use 3GPP SCM (A,B,C,D) or 36-101 (E-EPA,F-EVA,G-ETU) models (ignores delay spread and Ricean factor)\n"); + printf("-x Transmission mode (1,2,6 for the moment)\n"); + printf("-y Number of TX antennas used in eNB\n"); + printf("-z Number of RX antennas used in UE\n"); + printf("-i Relative strength of first intefering eNB (in dB) - cell_id mod 3 = 1\n"); + printf("-j Relative strength of second intefering eNB (in dB) - cell_id mod 3 = 2\n"); + printf("-N Nid_cell\n"); + printf("-R N_RB_DL\n"); + printf("-O oversampling factor (1,2,4,8,16)\n"); + printf("-A Interpolation_filname Run with Abstraction to generate Scatter plot using interpolation polynomial in file\n"); + // printf("-C Generate Calibration information for Abstraction (effective SNR adjustment to remove Pe bias w.r.t. AWGN)\n"); + printf("-f Output filename (.txt format) for Pe/SNR results\n"); + printf("-F Input filename (.txt format) for RX conformance testing\n"); + exit(-1); + break; + } + } + + logInit(); + set_glog(loglvl); + T_stdout = 1; + + if (snr1set == 0) + snr1 = snr0 + 10; + + gNB2UE = new_channel_desc_scm(n_tx, n_rx, channel_model, 61.44e6, //N_RB2sampling_rate(N_RB_DL), + 40e6, //N_RB2channel_bandwidth(N_RB_DL), + 0, 0, 0); + + if (gNB2UE == NULL) { + msg("Problem generating channel model. Exiting.\n"); + exit(-1); + } + + RC.gNB = (PHY_VARS_gNB ** *) malloc(sizeof(PHY_VARS_gNB **)); + RC.gNB[0] = (PHY_VARS_gNB **) malloc(sizeof(PHY_VARS_gNB *)); + RC.gNB[0][0] = malloc(sizeof(PHY_VARS_gNB)); + gNB = RC.gNB[0][0]; + //gNB_config = &gNB->gNB_config; + frame_parms = &gNB->frame_parms; //to be initialized I suppose (maybe not necessary for PBCH) + frame_parms->nb_antennas_tx = n_tx; + frame_parms->nb_antennas_rx = n_rx; + frame_parms->N_RB_DL = N_RB_DL; + crcTableInit(); + nr_phy_config_request_sim(gNB, N_RB_DL, N_RB_DL, mu, Nid_cell); + phy_init_nr_gNB(gNB, 0, 0); + //init_eNB_afterRU(); + frame_length_complex_samples = frame_parms->samples_per_subframe; + //frame_length_complex_samples_no_prefix = frame_parms->samples_per_subframe_wCP; + s_re = malloc(2 * sizeof(double *)); + s_im = malloc(2 * sizeof(double *)); + r_re = malloc(2 * sizeof(double *)); + r_im = malloc(2 * sizeof(double *)); + txdata = malloc(2 * sizeof(int *)); + + for (i = 0; i < 2; i++) { + s_re[i] = malloc(frame_length_complex_samples * sizeof(double)); + bzero(s_re[i], frame_length_complex_samples * sizeof(double)); + s_im[i] = malloc(frame_length_complex_samples * sizeof(double)); + bzero(s_im[i], frame_length_complex_samples * sizeof(double)); + r_re[i] = malloc(frame_length_complex_samples * sizeof(double)); + bzero(r_re[i], frame_length_complex_samples * sizeof(double)); + r_im[i] = malloc(frame_length_complex_samples * sizeof(double)); + bzero(r_im[i], frame_length_complex_samples * sizeof(double)); + txdata[i] = malloc(frame_length_complex_samples * sizeof(int)); + bzero(r_re[i], frame_length_complex_samples * sizeof(int)); + } + + if (pbch_file_fd != NULL) { + load_pbch_desc(pbch_file_fd); + } + + /* for (int k=0; k<2; k++) { + // Create transport channel structures for 2 transport blocks (MIMO) + for (i=0; i<2; i++) { + gNB->dlsch[k][i] = new_gNB_dlsch(Kmimo,8,Nsoft,0,frame_parms,gNB_config); + + if (!gNB->dlsch[k][i]) { + printf("Can't get eNB dlsch structures\n"); + exit(-1); + } + gNB->dlsch[k][i]->Nsoft = 10; + gNB->dlsch[k][i]->rnti = n_rnti+k; + } + }*/ + //configure UE + UE = malloc(sizeof(PHY_VARS_NR_UE)); + memcpy(&UE->frame_parms, frame_parms, sizeof(NR_DL_FRAME_PARMS)); + + //phy_init_nr_top(frame_parms); + if (init_nr_ue_signal(UE, 1, 0) != 0) { + printf("Error at UE NR initialisation\n"); + exit(-1); + } + + //nr_init_frame_parms_ue(&UE->frame_parms); + //init_nr_ue_transport(UE, 0); + for (int sf = 0; sf < 2; sf++) { + for (i = 0; i < 2; i++) { + UE->dlsch[sf][0][i] = new_nr_ue_dlsch(Kmimo, 8, Nsoft, 5, N_RB_DL, + 0); + + if (!UE->dlsch[sf][0][i]) { + printf("Can't get ue dlsch structures\n"); + exit(-1); + } + + UE->dlsch[sf][0][i]->rnti = n_rnti; + } + } + + UE->dlsch_SI[0] = new_nr_ue_dlsch(1, 1, Nsoft, 5, N_RB_DL, 0); + UE->dlsch_ra[0] = new_nr_ue_dlsch(1, 1, Nsoft, 5, N_RB_DL, 0); + unsigned char harq_pid = 0; //dlsch->harq_ids[subframe]; + NR_gNB_DLSCH_t *dlsch = gNB->dlsch[0][0]; + nfapi_nr_dl_config_dlsch_pdu_rel15_t *rel15 = &dlsch->harq_processes[harq_pid]->dlsch_pdu.dlsch_pdu_rel15; + //time_stats_t *rm_stats, *te_stats, *i_stats; + uint8_t is_crnti = 0, llr8_flag = 0; + unsigned int TBS = 8424; + unsigned int available_bits; + uint8_t nb_re_dmrs = 6; + uint16_t length_dmrs = 1; + unsigned char mod_order; + uint8_t Nl = 1; + uint8_t rvidx = 0; + dlsch->rnti = 1; + /*dlsch->harq_processes[0]->mcs = Imcs; + dlsch->harq_processes[0]->rvidx = rvidx;*/ + //printf("dlschsim harqid %d nb_rb %d, mscs %d\n",dlsch->harq_ids[subframe], + // dlsch->harq_processes[0]->nb_rb,dlsch->harq_processes[0]->mcs,dlsch->harq_processes[0]->Nl); + mod_order = nr_get_Qm(Imcs, 1); + available_bits = nr_get_G(nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs, mod_order, 1); + TBS = nr_compute_tbs(Imcs, nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs, Nl); + printf("available bits %d TBS %d mod_order %d\n", available_bits, TBS, mod_order); + //dlsch->harq_ids[subframe]= 0; + rel15->n_prb = nb_rb; + rel15->nb_symbols = nb_symb_sch; + rel15->modulation_order = mod_order; + rel15->nb_layers = Nl; + rel15->nb_re_dmrs = nb_re_dmrs; + rel15->transport_block_size = TBS; + double *modulated_input = malloc16(sizeof(double) * 16 * 68 * 384); + short *channel_output_fixed = malloc16(sizeof(short) * 16 * 68 * 384); + short *channel_output_uncoded = malloc16(sizeof(unsigned short) * 16 * 68 * 384); + double errors_bit_uncoded = 0; + unsigned char *estimated_output; + unsigned char *estimated_output_bit; + unsigned char *test_input_bit; + unsigned int errors_bit = 0; + test_input_bit = (unsigned char *) malloc16(sizeof(unsigned char) * 16 * 68 * 384); + estimated_output = (unsigned char *) malloc16(sizeof(unsigned char) * 16 * 68 * 384); + estimated_output_bit = (unsigned char *) malloc16(sizeof(unsigned char) * 16 * 68 * 384); + NR_UE_DLSCH_t *dlsch0_ue = UE->dlsch[UE->current_thread_id[subframe]][0][0]; + NR_DL_UE_HARQ_t *harq_process = dlsch0_ue->harq_processes[harq_pid]; + harq_process->mcs = Imcs; + harq_process->Nl = Nl; + harq_process->nb_rb = nb_rb; + harq_process->Qm = mod_order; + harq_process->rvidx = rvidx; + printf("harq process ue mcs = %d Qm = %d, symb %d\n", harq_process->mcs, harq_process->Qm, nb_symb_sch); + unsigned char *test_input; + test_input = (unsigned char *) malloc16(sizeof(unsigned char) * TBS / 8); + + for (i = 0; i < TBS / 8; i++) + test_input[i] = (unsigned char) rand(); + + estimated_output = harq_process->b; + +#ifdef DEBUG_DLSCHSIM + for (i = 0; i < TBS / 8; i++) printf("test_input[i]=%d \n",test_input[i]); +#endif + + /*for (int i=0; i0x%08x\n",crc24c(test_input, 32)); + // generate signal + if (input_fd == NULL) { + nr_dlsch_encoding(test_input, subframe, dlsch, frame_parms); + } + + for (SNR = snr0; SNR < snr1; SNR += snr_step) { + n_errors = 0; + n_false_positive = 0; + + for (trial = 0; trial < n_trials; trial++) { + errors_bit_uncoded = 0; + for (i = 0; i < available_bits; i++) { +#ifdef DEBUG_CODER + if ((i&0xf)==0) + printf("\ne %d..%d: ",i,i+15); +#endif + + //if (i<16) + // printf("encoder output f[%d] = %d\n",i,dlsch->harq_processes[0]->f[i]); + if (dlsch->harq_processes[0]->f[i] == 0) + modulated_input[i] = 1.0; ///sqrt(2); //QPSK + else + modulated_input[i] = -1.0; ///sqrt(2); + + //if (i<16) printf("modulated_input[%d] = %d\n",i,modulated_input[i]); + //SNR =10; + SNR_lin = pow(10, SNR / 10.0); + sigma = 1.0 / sqrt(2 * SNR_lin); + channel_output_fixed[i] = (short) quantize(sigma / 4.0 / 4.0, + modulated_input[i] + sigma * gaussdouble(0.0, 1.0), + qbits); + //channel_output_fixed[i] = (char)quantize8bit(sigma/4.0,(2.0*modulated_input[i]) - 1.0 + sigma*gaussdouble(0.0,1.0)); + //printf("llr[%d]=%d\n",i,channel_output_fixed[i]); + //printf("channel_output_fixed[%d]: %d\n",i,channel_output_fixed[i]); + + //channel_output_fixed[i] = (char)quantize(1,channel_output_fixed[i],qbits); + + //if (i<16) printf("channel_output_fixed[%d] = %d\n",i,channel_output_fixed[i]); + + //Uncoded BER + if (channel_output_fixed[i] < 0) + channel_output_uncoded[i] = 1; //QPSK demod + else + channel_output_uncoded[i] = 0; + + if (channel_output_uncoded[i] != dlsch->harq_processes[harq_pid]->f[i]) + errors_bit_uncoded = errors_bit_uncoded + 1; + } + + //if (errors_bit_uncoded>10) + //printf("errors bits uncoded %f\n", errors_bit_uncoded); +#ifdef DEBUG_CODER + printf("\n"); + exit(-1); +#endif + ret = nr_dlsch_decoding(UE, channel_output_fixed, &UE->frame_parms, + dlsch0_ue, dlsch0_ue->harq_processes[0], frame, nb_symb_sch, + subframe, harq_pid, is_crnti, llr8_flag); + + if (ret > dlsch0_ue->max_ldpc_iterations) + n_errors++; + + //count errors + errors_bit = 0; + + for (i = 0; i < TBS; i++) { + estimated_output_bit[i] = (dlsch0_ue->harq_processes[0]->b[i / 8] & (1 << (i & 7))) >> (i & 7); + test_input_bit[i] = (test_input[i / 8] & (1 << (i & 7))) >> (i & 7); // Further correct for multiple segments + + if (estimated_output_bit[i] != test_input_bit[i]) { + errors_bit++; + //printf("estimated bits error occurs @%d ",i); + } + } + + if (errors_bit > 0) { + n_false_positive++; + if (n_trials == 1) + printf("errors_bit %d (trial %d)\n", errors_bit, trial); + } + } + + printf("SNR %f, BLER %f (false positive %f)\n", SNR, + (float) n_errors / (float) n_trials, + (float) n_false_positive / (float) n_trials); + + if ((float) n_errors / (float) n_trials < target_error_rate) + break; + } + + /*LOG_M("txsigF0.m","txsF0", gNB->common_vars.txdataF[0],frame_length_complex_samples_no_prefix,1,1); + if (gNB->frame_parms.nb_antennas_tx>1) + LOG_M("txsigF1.m","txsF1", gNB->common_vars.txdataF[1],frame_length_complex_samples_no_prefix,1,1);*/ + + //TODO: loop over slots + /*for (aa=0; aaframe_parms.nb_antennas_tx; aa++) { + if (gNB_config->subframe_config.dl_cyclic_prefix_type.value == 1) { + PHY_ofdm_mod(gNB->common_vars.txdataF[aa], + txdata[aa], + frame_parms->ofdm_symbol_size, + 12, + frame_parms->nb_prefix_samples, + CYCLIC_PREFIX); + } else { + nr_normal_prefix_mod(gNB->common_vars.txdataF[aa], + txdata[aa], + 14, + frame_parms); + } + } + + LOG_M("txsig0.m","txs0", txdata[0],frame_length_complex_samples,1,1); + if (gNB->frame_parms.nb_antennas_tx>1) + LOG_M("txsig1.m","txs1", txdata[1],frame_length_complex_samples,1,1); + + + for (i=0; inb_antennas_tx; aa++) { + r_re[aa][i] = ((double)(((short *)txdata[aa]))[(i<<1)]); + r_im[aa][i] = ((double)(((short *)txdata[aa]))[(i<<1)+1]); + } + }*/ + + for (i = 0; i < 2; i++) { + printf("gNB %d\n", i); + free_gNB_dlsch(gNB->dlsch[0][i]); + printf("UE %d\n", i); + free_nr_ue_dlsch(UE->dlsch[UE->current_thread_id[subframe]][0][i]); + } + + for (i = 0; i < 2; i++) { + free(s_re[i]); + free(s_im[i]); + free(r_re[i]); + free(r_im[i]); + free(txdata[i]); + } + + free(s_re); + free(s_im); + free(r_re); + free(r_im); + free(txdata); + + if (output_fd) + fclose(output_fd); + + if (input_fd) + fclose(input_fd); + + return (n_errors); +} + diff --git a/openair1/SIMULATION/NR_PHY/dlsim.c b/openair1/SIMULATION/NR_PHY/dlsim.c index cc89f245fc..1d440c5d0d 100644 --- a/openair1/SIMULATION/NR_PHY/dlsim.c +++ b/openair1/SIMULATION/NR_PHY/dlsim.c @@ -37,6 +37,7 @@ #include "PHY/defs_nr_UE.h" #include "PHY/defs_gNB.h" #include "PHY/NR_REFSIG/refsig_defs_ue.h" +#include "PHY/NR_REFSIG/nr_mod_table.h" #include "PHY/MODULATION/modulation_eNB.h" #include "PHY/MODULATION/modulation_UE.h" #include "PHY/INIT/phy_init.h" @@ -53,6 +54,7 @@ #include "NR_PHY_INTERFACE/NR_IF_Module.h" #include "NR_UE_PHY_INTERFACE/NR_IF_Module.h" + PHY_VARS_gNB *gNB; PHY_VARS_NR_UE *UE; RAN_CONTEXT_t RC; @@ -71,14 +73,15 @@ int oai_nfapi_ul_config_req(nfapi_ul_config_request_t *ul_config_req) { return(0 int oai_nfapi_nr_dl_config_req(nfapi_nr_dl_config_request_t *dl_config_req) {return(0);} -uint32_t from_earfcn(int eutra_bandP,uint32_t dl_earfcn) {return(0);} -int32_t get_uldl_offset(int eutra_bandP) {return(0);} +int32_t get_uldl_offset(int nr_bandP) {return(0);} NR_IF_Module_t *NR_IF_Module_init(int Mod_id){return(NULL);} int8_t dummy_nr_ue_dl_indication(nr_downlink_indication_t *dl_info){return(0);} int8_t dummy_nr_ue_ul_indication(nr_uplink_indication_t *ul_info){return(0);} +lte_subframe_t subframe_select(LTE_DL_FRAME_PARMS *frame_parms,unsigned char subframe) { return(SF_DL);} + void exit_function(const char* file, const char* function, const int line,const char *s) { const char * msg= s==NULL ? "no comment": s; printf("Exiting at: %s:%d %s(), %s\n", file, line, function, msg); @@ -96,6 +99,12 @@ int rlc_module_init (void) {return(0);} void pdcp_layer_init(void) {} int rrc_init_nr_global_param(void){return(0);} +void config_common(int Mod_idP, + int CC_idP, + int nr_bandP, + uint64_t dl_CarrierFreqP, + uint32_t dl_BandwidthP + ); // needed for some functions @@ -145,7 +154,7 @@ int main(int argc, char **argv) unsigned char frame_type = 0; unsigned char pbch_phase = 0; - int frame=0,subframe=1; + int frame=0,slot=1; int frame_length_complex_samples; int frame_length_complex_samples_no_prefix; NR_DL_FRAME_PARMS *frame_parms; @@ -375,6 +384,8 @@ int main(int argc, char **argv) RC.gNB = (PHY_VARS_gNB***) malloc(sizeof(PHY_VARS_gNB **)); RC.gNB[0] = (PHY_VARS_gNB**) malloc(sizeof(PHY_VARS_gNB *)); RC.gNB[0][0] = malloc(sizeof(PHY_VARS_gNB)); + memset(RC.gNB[0][0],0,sizeof(PHY_VARS_gNB)); + gNB = RC.gNB[0][0]; gNB_config = &gNB->gNB_config; frame_parms = &gNB->frame_parms; //to be initialized I suppose (maybe not necessary for PBCH) @@ -383,9 +394,14 @@ int main(int argc, char **argv) frame_parms->N_RB_DL = N_RB_DL; frame_parms->N_RB_UL = N_RB_DL; - nr_phy_config_request_sim(gNB,N_RB_DL,N_RB_DL,mu); + // stub to configure frame_parms + nr_phy_config_request_sim(gNB,N_RB_DL,N_RB_DL,mu,Nid_cell); + // call MAC to configure common parameters + phy_init_nr_gNB(gNB,0,0); + + double fs,bw; if (mu == 1 && N_RB_DL == 217) { @@ -454,13 +470,14 @@ int main(int argc, char **argv) //configure UE UE = malloc(sizeof(PHY_VARS_NR_UE)); + memset((void*)UE,0,sizeof(PHY_VARS_NR_UE)); PHY_vars_UE_g = malloc(sizeof(PHY_VARS_NR_UE**)); PHY_vars_UE_g[0] = malloc(sizeof(PHY_VARS_NR_UE*)); PHY_vars_UE_g[0][0] = UE; memcpy(&UE->frame_parms,frame_parms,sizeof(NR_DL_FRAME_PARMS)); - phy_init_nr_top(UE); + if (run_initial_sync==1) UE->is_synchronized = 0; - else UE->is_synchronized = 1; + else {UE->is_synchronized = 1; UE->UE_mode[0]=PUSCH;} UE->perfect_ce = 0; @@ -470,19 +487,25 @@ int main(int argc, char **argv) exit(-1); } + init_nr_ue_transport(UE,0); + nr_gold_pbch(UE); + nr_gold_pdcch(UE,0,2); RC.nb_nr_macrlc_inst = 1; mac_top_init_gNB(); gNB_mac = RC.nrmac[0]; + config_common(0,0,78,(uint64_t)3640000000L,N_RB_DL); + config_nr_mib(0,0,1,kHz30,0,0,0,0); + nr_l2_init_ue(); UE_mac = get_mac_inst(0); UE->if_inst = nr_ue_if_module_init(0); UE->if_inst->scheduled_response = nr_ue_scheduled_response; UE->if_inst->phy_config_request = nr_ue_phy_config_request; - UE->if_inst->dl_indication = dummy_nr_ue_dl_indication; + UE->if_inst->dl_indication = nr_ue_dl_indication; UE->if_inst->ul_indication = dummy_nr_ue_ul_indication; @@ -494,11 +517,11 @@ int main(int argc, char **argv) gNB->pbch_configured = 1; for (int i=0;i<4;i++) gNB->pbch_pdu[i]=i+1; - nr_schedule_css_dlsch_phytest(0,frame,subframe); + nr_schedule_css_dlsch_phytest(0,frame,slot); Sched_INFO.module_id = 0; Sched_INFO.CC_id = 0; Sched_INFO.frame = frame; - Sched_INFO.subframe = subframe; + Sched_INFO.slot = slot; Sched_INFO.DL_req = &gNB_mac->DL_req[0]; Sched_INFO.UL_req = NULL; Sched_INFO.HI_DCI0_req = NULL; @@ -506,7 +529,7 @@ int main(int argc, char **argv) nr_schedule_response(&Sched_INFO); gNB_proc.frame_tx = frame; - gNB_proc.subframe_tx = subframe; + gNB_proc.slot_tx = slot; phy_procedures_gNB_TX(gNB,&gNB_proc,0); //nr_common_signal_procedures (gNB,frame,subframe); @@ -562,7 +585,15 @@ int main(int argc, char **argv) } } + //Configure UE + rrc_gNB_carrier_data_t carrier; + uint32_t pdcch_ConfigSIB1 = 0; + uint32_t ssb_SubcarrierOffset = 0; + carrier.MIB = (uint8_t*) malloc(4); + carrier.sizeof_MIB = do_MIB_NR(&carrier,0,ssb_SubcarrierOffset,pdcch_ConfigSIB1,30,2); + + nr_rrc_mac_config_req_ue(0,0,0,carrier.mib.message.choice.mib,NULL,NULL,NULL); fapi_nr_dl_config_request_t dl_config; // Type0 PDCCH search space dl_config.number_pdus = 1; @@ -591,7 +622,7 @@ int main(int argc, char **argv) uint32_t number_of_search_space_per_slot=1; uint32_t first_symbol_index=0; - uint32_t search_space_duration; // element of search space + uint32_t search_space_duration=0; // element of search space uint32_t coreset_duration; // element of coreset coreset_duration = num_symbols * number_of_search_space_per_slot; @@ -603,8 +634,9 @@ int main(int argc, char **argv) dl_config.dl_config_list[0].dci_config_pdu.dci_config_rel15.number_of_candidates[4] = table_38213_10_1_1_c2[4]; // CCE aggregation level = 16 dl_config.dl_config_list[0].dci_config_pdu.dci_config_rel15.duration = search_space_duration; dl_config.dl_config_list[0].dci_config_pdu.dci_config_rel15.monitoring_symbols_within_slot = (0x3fff << first_symbol_index) & (0x3fff >> (14-coreset_duration-first_symbol_index)) & 0x3fff; - - + + dl_config.dl_config_list[0].dci_config_pdu.dci_config_rel15.N_RB_BWP = N_RB_DL; + for (SNR=snr0; SNRrx_offset=0; UE_proc.frame_rx = frame; - UE_proc.nr_tti_rx= subframe; - UE_proc.subframe_rx = subframe; + UE_proc.nr_tti_rx= slot; + UE_proc.subframe_rx = slot; UE_mac->scheduled_response.dl_config = &dl_config; nr_ue_scheduled_response(&UE_mac->scheduled_response); + printf("Running phy procedures UE RX %d.%d\n",frame,slot); + phy_procedures_nrUE_RX(UE, &UE_proc, 0, - 0, do_pdcch_flag, - normal_txrx, - no_relay); + normal_txrx); + if (UE->dci_ind.number_of_dcis==0) n_errors++; } diff --git a/openair1/SIMULATION/NR_PHY/pbchsim.c b/openair1/SIMULATION/NR_PHY/pbchsim.c index 52a9f48481..214a33fb73 100644 --- a/openair1/SIMULATION/NR_PHY/pbchsim.c +++ b/openair1/SIMULATION/NR_PHY/pbchsim.c @@ -37,6 +37,7 @@ #include "PHY/defs_nr_UE.h" #include "PHY/defs_gNB.h" #include "PHY/NR_REFSIG/refsig_defs_ue.h" +#include "PHY/NR_REFSIG/nr_mod_table.h" #include "PHY/MODULATION/modulation_eNB.h" #include "PHY/MODULATION/modulation_UE.h" #include "PHY/INIT/phy_init.h" @@ -65,8 +66,8 @@ int oai_nfapi_ul_config_req(nfapi_ul_config_request_t *ul_config_req) { return(0 int oai_nfapi_nr_dl_config_req(nfapi_nr_dl_config_request_t *dl_config_req) {return(0);} -uint32_t from_earfcn(int eutra_bandP,uint32_t dl_earfcn) {return(0);} -int32_t get_uldl_offset(int eutra_bandP) {return(0);} +uint32_t from_nrarfcn(int nr_bandP,uint32_t dl_nrarfcn) {return(0);} +int32_t get_uldl_offset(int nr_bandP) {return(0);} NR_IF_Module_t *NR_IF_Module_init(int Mod_id){return(NULL);} @@ -84,39 +85,43 @@ int main(int argc, char **argv) char c; - int i,l,aa; + int i,aa;//,l; double sigma2, sigma2_dB=10,SNR,snr0=-2.0,snr1=2.0; + double cfo=0; uint8_t snr1set=0; int **txdata; double **s_re,**s_im,**r_re,**r_im; double iqim = 0.0; + double ip =0.0; unsigned char pbch_pdu[6]; // int sync_pos, sync_pos_slot; // FILE *rx_frame_file; FILE *output_fd = NULL; uint8_t write_output_file=0; //int result; - int freq_offset; + //int freq_offset; // int subframe_offset; // char fname[40], vname[40]; - int trial,n_trials=1,n_errors,n_errors_payload; + int trial,n_trials=1,n_errors=0,n_errors_payload=0; uint8_t transmission_mode = 1,n_tx=1,n_rx=1; uint16_t Nid_cell=0; channel_desc_t *gNB2UE; - uint32_t nsymb,tx_lev,tx_lev1 = 0,tx_lev2 = 0; + uint8_t extended_prefix_flag=0; int8_t interf1=-21,interf2=-21; FILE *input_fd=NULL,*pbch_file_fd=NULL; - char input_val_str[50],input_val_str2[50]; - uint8_t frame_mod4,num_pdcch_symbols = 0; + //uint32_t nsymb,tx_lev,tx_lev1 = 0,tx_lev2 = 0; + //char input_val_str[50],input_val_str2[50]; + //uint8_t frame_mod4,num_pdcch_symbols = 0; + //double pbch_sinr; + //int pbch_tx_ant; SCM_t channel_model=AWGN;//Rayleigh1_anticorr; - double pbch_sinr; - int pbch_tx_ant; + int N_RB_DL=273,mu=1; unsigned char frame_type = 0; @@ -143,7 +148,7 @@ int main(int argc, char **argv) randominit(0); - while ((c = getopt (argc, argv, "f:hA:pf:g:i:j:n:s:S:t:x:y:z:N:F:GR:dP:IL:")) != -1) { + while ((c = getopt (argc, argv, "f:hA:pf:g:i:j:n:o:s:S:t:x:y:z:N:F:GR:dP:IL:")) != -1) { switch (c) { case 'f': write_output_file=1; @@ -209,6 +214,11 @@ int main(int argc, char **argv) n_trials = atoi(optarg); break; + case 'o': + cfo = atof(optarg); + msg("Setting CFO to %f Hz\n",cfo); + break; + case 's': snr0 = atof(optarg); msg("Setting SNR0 to %f\n",snr0); @@ -322,6 +332,7 @@ int main(int argc, char **argv) printf("-z Number of RX antennas used in UE\n"); printf("-i Relative strength of first intefering eNB (in dB) - cell_id mod 3 = 1\n"); printf("-j Relative strength of second intefering eNB (in dB) - cell_id mod 3 = 2\n"); + printf("-o Carrier frequency offset in Hz\n"); printf("-N Nid_cell\n"); printf("-R N_RB_DL\n"); printf("-O oversampling factor (1,2,4,8,16)\n"); @@ -353,30 +364,49 @@ int main(int argc, char **argv) frame_parms->nb_antennas_rx = n_rx; frame_parms->N_RB_DL = N_RB_DL; frame_parms->N_RB_UL = N_RB_DL; + frame_parms->Nid_cell = Nid_cell; - nr_phy_config_request_sim(gNB,N_RB_DL,N_RB_DL,mu); + nr_phy_config_request_sim(gNB,N_RB_DL,N_RB_DL,mu,Nid_cell); phy_init_nr_gNB(gNB,0,0); - double fs,bw; + double fs,bw,scs,eps; if (mu == 1 && N_RB_DL == 217) { fs = 122.88e6; bw = 80e6; + scs = 30000; } else if (mu == 1 && N_RB_DL == 245) { fs = 122.88e6; bw = 90e6; + scs = 30000; } else if (mu == 1 && N_RB_DL == 273) { fs = 122.88e6; bw = 100e6; + scs = 30000; } else if (mu == 1 && N_RB_DL == 106) { fs = 61.44e6; bw = 40e6; + scs = 30000; } else AssertFatal(1==0,"Unsupported numerology for mu %d, N_RB %d\n",mu, N_RB_DL); + // cfo with respect to sub-carrier spacing + eps = cfo/scs; + + // computation of integer and fractional FO to compare with estimation results + int IFO; + if(eps!=0.0){ + printf("Introducing a CFO of %lf relative to SCS of %d kHz\n",eps,(int)(scs/1000)); + if (eps>0) + IFO=(int)(eps+0.5); + else + IFO=(int)(eps-0.5); + printf("FFO = %lf; IFO = %d\n",eps-IFO,IFO); + } + gNB2UE = new_channel_desc_scm(n_tx, n_rx, channel_model, @@ -426,12 +456,15 @@ int main(int argc, char **argv) //configure UE UE = malloc(sizeof(PHY_VARS_NR_UE)); memcpy(&UE->frame_parms,frame_parms,sizeof(NR_DL_FRAME_PARMS)); - phy_init_nr_top(UE); + //phy_init_nr_top(UE); //called from init_nr_ue_signal if (run_initial_sync==1) UE->is_synchronized = 0; else UE->is_synchronized = 1; UE->perfect_ce = 0; + if(eps!=0.0) + UE->UE_fo_compensation = 1; // if a frequency offset is set then perform fo estimation and compensation + if (init_nr_ue_signal(UE, 1, 0) != 0) { printf("Error at UE NR initialisation\n"); @@ -473,7 +506,7 @@ int main(int argc, char **argv) frame_length_complex_samples, input_fd) != frame_length_complex_samples) { printf("error reading from file\n"); - exit(-1); + //exit(-1); } } @@ -489,6 +522,7 @@ int main(int argc, char **argv) // printf("txlev %d (%f)\n",txlev,10*log10(txlev)); + for (i=0; inb_antennas_tx; aa++) { r_re[aa][i] = ((double)(((short *)txdata[aa]))[(i<<1)]); @@ -507,10 +541,31 @@ int main(int argc, char **argv) //multipath_channel(gNB2UE,s_re,s_im,r_re,r_im,frame_length_complex_samples,0); //AWGN - sigma2_dB = 10*log10((double)txlev)-SNR; + sigma2_dB = 20*log10((double)AMP/4)-SNR; sigma2 = pow(10,sigma2_dB/10); - // printf("sigma2 %f (%f dB)\n",sigma2,sigma2_dB); + //printf("sigma2 %f (%f dB), tx_lev %f (%f dB)\n",sigma2,sigma2_dB,txlev,10*log10((double)txlev)); + if(eps!=0.0) + rf_rx(r_re, // real part of txdata + r_im, // imag part of txdata + NULL, // interference real part + NULL, // interference imag part + 0, // interference power + frame_parms->nb_antennas_rx, // number of rx antennas + frame_length_complex_samples, // number of samples in frame + 1.0e9/fs, //sampling time (ns) + cfo, // frequency offset in Hz + 0.0, // drift (not implemented) + 0.0, // noise figure (not implemented) + 0.0, // rx gain in dB ? + 200, // 3rd order non-linearity in dB ? + &ip, // initial phase + 30.0e3, // phase noise cutoff in kHz + -500.0, // phase noise amplitude in dBc + 0.0, // IQ imbalance (dB), + 0.0); // IQ phase imbalance (rad) + + for (i=0; inb_antennas_rx; aa++) { @@ -520,10 +575,11 @@ int main(int argc, char **argv) } if (n_trials==1) { - LOG_M("rxsig0.m","rxs0", UE->common_vars.rxdata[0],frame_length_complex_samples,1,1); + LOG_M("rxsig0.m","rxs0", UE->common_vars.rxdata[0],frame_parms->samples_per_subframe,1,1); if (gNB->frame_parms.nb_antennas_tx>1) - LOG_M("rxsig1.m","rxs1", UE->common_vars.rxdata[1],frame_length_complex_samples,1,1); + LOG_M("rxsig1.m","rxs1", UE->common_vars.rxdata[1],frame_parms->samples_per_subframe,1,1); } + if (UE->is_synchronized == 0) { ret = nr_initial_sync(UE, normal_txrx); printf("nr_initial_sync1 returns %d\n",ret); @@ -557,7 +613,7 @@ int main(int argc, char **argv) 0, 1, NR_PBCH_EST); - + ret = nr_rx_pbch(UE, &UE->proc.proc_rxtx[0], UE->pbch_vars[0], @@ -584,7 +640,7 @@ int main(int argc, char **argv) n_errors_payload++; } - if (ret<0) n_errors++; + if (ret!=0) n_errors++; } } //noise trials diff --git a/openair1/SIMULATION/NR_UE_PHY/unit_tests/src/dummy_functions.c b/openair1/SIMULATION/NR_UE_PHY/unit_tests/src/dummy_functions.c index 0fdd8ff383..fa243d1eb0 100644 --- a/openair1/SIMULATION/NR_UE_PHY/unit_tests/src/dummy_functions.c +++ b/openair1/SIMULATION/NR_UE_PHY/unit_tests/src/dummy_functions.c @@ -119,9 +119,9 @@ void dl_phy_sync_success(module_id_t module_idP, frame_t frameP, unsigned char eNB_index, uint8_t first_sync){} -uint32_t from_earfcn(int eutra_bandP, uint32_t dl_earfcn) { return(0);} +uint32_t from_nrarfcn(int nr_bandP, uint32_t dl_nrarfcn) { return(0);} -int32_t get_uldl_offset(int eutra_bandP) { return(0);} +int32_t get_uldl_offset(int nr_bandP) { return(0);} IF_Module_t *IF_Module_init(int Mod_id) { return(NULL);} diff --git a/openair1/SIMULATION/RF/rf.c b/openair1/SIMULATION/RF/rf.c index e317fbd154..bff3453de0 100644 --- a/openair1/SIMULATION/RF/rf.c +++ b/openair1/SIMULATION/RF/rf.c @@ -107,10 +107,10 @@ void rf_rx(double **r_re, exit(-1); } - if (fabs(f_off) > 10000.0) { +/* if (fabs(f_off) > 10000.0) { printf("rf.c: Illegal f_off %f\n",f_off); exit(-1); - } + }*/ if (fabs(drift) > 1000.0) { printf("rf.c: Illegal drift %f\n",drift); diff --git a/openair2/COMMON/rrc_messages_types.h b/openair2/COMMON/rrc_messages_types.h index 6d83aba907..922c949307 100644 --- a/openair2/COMMON/rrc_messages_types.h +++ b/openair2/COMMON/rrc_messages_types.h @@ -310,8 +310,8 @@ typedef struct NRRrcConfigurationReq_s { uint8_t tdd_config_s[MAX_NUM_CCs]; lte_prefix_type_t DL_prefix_type[MAX_NUM_CCs]; lte_prefix_type_t UL_prefix_type[MAX_NUM_CCs]; - int16_t eutra_band[MAX_NUM_CCs]; - uint32_t downlink_frequency[MAX_NUM_CCs]; + int16_t nr_band[MAX_NUM_CCs]; + uint64_t downlink_frequency[MAX_NUM_CCs]; int32_t uplink_frequency_offset[MAX_NUM_CCs]; int16_t Nid_cell[MAX_NUM_CCs];// for testing, change later int16_t N_RB_DL[MAX_NUM_CCs];// for testing, change later diff --git a/openair2/GNB_APP/RRC_nr_paramsvalues.h b/openair2/GNB_APP/RRC_nr_paramsvalues.h index a30e8c537d..ca7c659355 100644 --- a/openair2/GNB_APP/RRC_nr_paramsvalues.h +++ b/openair2/GNB_APP/RRC_nr_paramsvalues.h @@ -39,7 +39,7 @@ #define GNB_CONFIG_STRING_FRAME_TYPE "frame_type" #define GNB_CONFIG_STRING_DL_PREFIX_TYPE "DL_prefix_type" #define GNB_CONFIG_STRING_UL_PREFIX_TYPE "UL_prefix_type" -#define GNB_CONFIG_STRING_EUTRA_BAND "eutra_band" +#define GNB_CONFIG_STRING_NR_BAND "nr_band" #define GNB_CONFIG_STRING_DOWNLINK_FREQUENCY "downlink_frequency" #define GNB_CONFIG_STRING_UPLINK_FREQUENCY_OFFSET "uplink_frequency_offset" #define GNB_CONFIG_STRING_NID_CELL "Nid_cell" diff --git a/openair2/GNB_APP/gnb_config.c b/openair2/GNB_APP/gnb_config.c index d695833ed3..4484cc18a8 100644 --- a/openair2/GNB_APP/gnb_config.c +++ b/openair2/GNB_APP/gnb_config.c @@ -118,9 +118,9 @@ void RCconfig_nr_flexran() *SearchSpace_searchSpaceType, *ue_Specific__dci_Formats, *RateMatchPatternLTE_CRS_subframeAllocation_choice; - long long int downlink_frequency; + uint64_t downlink_frequency; - int32_t eutra_band, uplink_frequency_offset, N_RB_DL, nb_antenna_ports, + int32_t nr_band, uplink_frequency_offset, N_RB_DL, nb_antenna_ports, MIB_subCarrierSpacingCommon, MIB_ssb_SubcarrierOffset, MIB_dmrs_TypeA_Position, pdcch_ConfigSIB1, SIB1_ssb_PeriodicityServingCell, SIB1_ss_PBCH_BlockPower, absoluteFrequencySSB, DL_FreqBandIndicatorNR, @@ -431,8 +431,8 @@ void RCconfig_NRRRC(MessageDef *msg_p, uint32_t i, gNB_RRC_INST *rrc) { char* DL_prefix_type = NULL; char* UL_prefix_type = NULL; - int32_t eutra_band = 0; - long long int downlink_frequency = 0; + int32_t nr_band = 0; + uint64_t downlink_frequency = 0; int32_t uplink_frequency_offset = 0; int32_t Nid_cell = 0; int32_t N_RB_DL = 0; @@ -629,7 +629,7 @@ void RCconfig_NRRRC(MessageDef *msg_p, uint32_t i, gNB_RRC_INST *rrc) { paramdef_t GNBParams[] = GNBPARAMS_DESC; paramlist_def_t GNBParamList = {GNB_CONFIG_STRING_GNB_LIST,NULL,0}; ////////// Physical parameters - //checkedparam_t config_check_CCparams[] = NRCCPARAMS_CHECK; + checkedparam_t config_check_CCparams[] = NRCCPARAMS_CHECK; paramdef_t CCsParams[] = NRCCPARAMS_DESC; paramlist_def_t CCsParamList = {GNB_CONFIG_STRING_COMPONENT_CARRIERS,NULL,0}; @@ -788,8 +788,8 @@ void RCconfig_NRRRC(MessageDef *msg_p, uint32_t i, gNB_RRC_INST *rrc) { RC.config_file_name, i, UL_prefix_type); } - NRRRC_CONFIGURATION_REQ (msg_p).eutra_band[j] = eutra_band; - NRRRC_CONFIGURATION_REQ (msg_p).downlink_frequency[j] = (uint32_t) downlink_frequency; + NRRRC_CONFIGURATION_REQ (msg_p).nr_band[j] = nr_band; + NRRRC_CONFIGURATION_REQ (msg_p).downlink_frequency[j] = (uint64_t) downlink_frequency; NRRRC_CONFIGURATION_REQ (msg_p).uplink_frequency_offset[j] = (unsigned int) uplink_frequency_offset; NRRRC_CONFIGURATION_REQ (msg_p).Nid_cell[j]= Nid_cell; @@ -844,12 +844,12 @@ void RCconfig_NRRRC(MessageDef *msg_p, uint32_t i, gNB_RRC_INST *rrc) { RC.config_file_name, i, UL_prefix_type); } - NRRRC_CONFIGURATION_REQ (msg_p).eutra_band[j] = eutra_band; - NRRRC_CONFIGURATION_REQ (msg_p).downlink_frequency[j] = (uint32_t) downlink_frequency; + NRRRC_CONFIGURATION_REQ (msg_p).nr_band[j] = nr_band; + NRRRC_CONFIGURATION_REQ (msg_p).downlink_frequency[j] = (uint64_t) downlink_frequency; NRRRC_CONFIGURATION_REQ (msg_p).uplink_frequency_offset[j] = (unsigned int) uplink_frequency_offset; if (config_check_band_frequencies(j, - NRRRC_CONFIGURATION_REQ (msg_p).eutra_band[j], + NRRRC_CONFIGURATION_REQ (msg_p).nr_band[j], NRRRC_CONFIGURATION_REQ (msg_p).downlink_frequency[j], NRRRC_CONFIGURATION_REQ (msg_p).uplink_frequency_offset[j], NRRRC_CONFIGURATION_REQ (msg_p).frame_type[j])) { diff --git a/openair2/GNB_APP/gnb_paramdef.h b/openair2/GNB_APP/gnb_paramdef.h index a457a3b1a4..15a2e97239 100644 --- a/openair2/GNB_APP/gnb_paramdef.h +++ b/openair2/GNB_APP/gnb_paramdef.h @@ -602,12 +602,12 @@ typedef enum { /* optname helpstr paramflags XXXptr defXXXval type numelt checked_param */ /*---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------*/ #define NRCCPARAMS_DESC { \ -{GNB_CONFIG_STRING_FRAME_TYPE, NULL, 0, strptr:&frame_type, defstrval:"FDD", TYPE_STRING, 0}, \ +{GNB_CONFIG_STRING_FRAME_TYPE, NULL, 0, strptr:&frame_type, defstrval:"TDD", TYPE_STRING, 0}, \ {GNB_CONFIG_STRING_DL_PREFIX_TYPE, NULL, 0, strptr:&DL_prefix_type, defstrval:"NORMAL", TYPE_STRING, 0}, \ {GNB_CONFIG_STRING_UL_PREFIX_TYPE, NULL, 0, strptr:&UL_prefix_type, defstrval:"NORMAL", TYPE_STRING, 0}, \ -{GNB_CONFIG_STRING_EUTRA_BAND, NULL, 0, iptr:&eutra_band, defintval:7, TYPE_UINT, 0}, \ -{GNB_CONFIG_STRING_DOWNLINK_FREQUENCY, NULL, 0, i64ptr:(int64_t *)&downlink_frequency, defint64val:2680000000, TYPE_UINT64, 0}, \ -{GNB_CONFIG_STRING_UPLINK_FREQUENCY_OFFSET, NULL, 0, iptr:&uplink_frequency_offset, defintval:-120000000, TYPE_INT, 0}, \ +{GNB_CONFIG_STRING_NR_BAND, NULL, 0, iptr:&nr_band, defintval:78, TYPE_UINT, 0}, \ +{GNB_CONFIG_STRING_DOWNLINK_FREQUENCY, NULL, 0, i64ptr:(int64_t *)&downlink_frequency, defint64val:3600000000, TYPE_UINT64, 0}, \ +{GNB_CONFIG_STRING_UPLINK_FREQUENCY_OFFSET, NULL, 0, iptr:&uplink_frequency_offset, defintval:0, TYPE_INT, 0}, \ {GNB_CONFIG_STRING_NID_CELL, NULL, 0, iptr:&Nid_cell, defintval:0, TYPE_UINT, 0}, \ {GNB_CONFIG_STRING_N_RB_DL, NULL, 0, iptr:&N_RB_DL, defintval:25, TYPE_UINT, 0}, \ {GNB_CONFIG_STRING_NB_ANT_PORTS, NULL, 0, iptr:&nb_antenna_ports, defintval:15, TYPE_UINT, 0}, \ @@ -751,7 +751,7 @@ typedef enum { #define GNB_CONFIG_FRAME_TYPE_IDX 0 #define GNB_CONFIG_DL_PREFIX_TYPE_IDX 1 #define GNB_CONFIG_UL_PREFIX_TYPE_IDX 2 -#define GNB_CONFIG_EUTRA_BAND_IDX 3 +#define GNB_CONFIG_NR_BAND_IDX 3 #define GNB_CONFIG_DOWNLINK_FREQUENCY_IDX 4 #define GNB_CONFIG_UPLINK_FREQUENCY_OFFSET_IDX 5 #define GNB_CONFIG_NID_CELL_IDX 6 diff --git a/openair2/LAYER2/NR_MAC_UE/mac_proto.h b/openair2/LAYER2/NR_MAC_UE/mac_proto.h index 24662aaa7c..f34000bc98 100755 --- a/openair2/LAYER2/NR_MAC_UE/mac_proto.h +++ b/openair2/LAYER2/NR_MAC_UE/mac_proto.h @@ -127,5 +127,8 @@ void nr_ue_process_mac_pdu( uint8_t *pduP, uint16_t mac_pdu_len, uint8_t eNB_index); + +int8_t nr_ue_process_dlsch(module_id_t module_id, int cc_id, uint8_t gNB_index, fapi_nr_dci_indication_t *dci_ind, void *pduP, uint32_t pdu_len); + #endif -/** @}*/ \ No newline at end of file +/** @}*/ diff --git a/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c b/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c index 157ce9d84d..44f1cf05ea 100644 --- a/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c +++ b/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c @@ -95,7 +95,7 @@ int8_t nr_ue_process_dlsch(module_id_t module_id, int cc_id, uint8_t gNB_index, ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.n_layers = 0; ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.tpmi = 0; ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.n_dmrs_cdm_groups = 0; - ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.dmrs_ports[4] = 0; + for (int i=0;i<4;i++) ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.dmrs_ports[i]=0; ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.n_front_load_symb = 0; //ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.srs_config = 0; ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15.csi_reportTriggerSize = 0; @@ -125,7 +125,7 @@ int8_t nr_ue_process_dlsch(module_id_t module_id, int cc_id, uint8_t gNB_index, ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.pucch_GroupHopping = phy_config->config_req.ul_bwp_common.pucch_config_common.pucch_group_hopping; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.hoppingId = phy_config->config_req.ul_bwp_common.pucch_config_common.hopping_id; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.p0_nominal = phy_config->config_req.ul_bwp_common.pucch_config_common.p0_nominal; - ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.deltaF_PUCCH_f[NUMBER_PUCCH_FORMAT_NR] = 0; + for (int i=0;iul_config_list[ul_config->number_pdus].pucch_config_pdu.deltaF_PUCCH_f[i] = 0; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.p0_PUCCH_Id = 0; /* INTEGER (1..8) */ ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.p0_PUCCH_Value = 0; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.twoPUCCH_PC_AdjustmentStates = 0; @@ -180,14 +180,14 @@ int8_t nr_ue_process_dlsch(module_id_t module_id, int cc_id, uint8_t gNB_index, ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.pucch_GroupHopping = phy_config->config_req.ul_bwp_common.pucch_config_common.pucch_group_hopping; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.hoppingId = phy_config->config_req.ul_bwp_common.pucch_config_common.hopping_id; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.p0_nominal = phy_config->config_req.ul_bwp_common.pucch_config_common.p0_nominal; - ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.deltaF_PUCCH_f[NUMBER_PUCCH_FORMAT_NR] = 0; + for (int i=0;iul_config_list[ul_config->number_pdus].pucch_config_pdu.deltaF_PUCCH_f[i] = 0; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.p0_PUCCH_Id = 0; /* INTEGER (1..8) */ ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.p0_PUCCH_Value = 0; ul_config->ul_config_list[ul_config->number_pdus].pucch_config_pdu.twoPUCCH_PC_AdjustmentStates = 0; } if(mac->if_module != NULL && mac->if_module->ul_indication != NULL){ - mac->if_module->ul_indication(&mac->scheduled_response); + mac->if_module->dl_indication(&mac->scheduled_response); } return 0; } @@ -659,9 +659,15 @@ NR_UE_L2_STATE_t nr_ue_scheduler( dl_config->dl_config_list[dl_config->number_pdus].dci_config_pdu.dci_config_rel15 = mac->type0_pdcch_dci_config; dl_config->dl_config_list[dl_config->number_pdus].pdu_type = FAPI_NR_DL_CONFIG_TYPE_DCI; - dl_config->number_pdus = dl_config->number_pdus + 1; dl_config->dl_config_list[dl_config->number_pdus].dci_config_pdu.dci_config_rel15.rnti = 0xaaaa; // to be set + dl_config->dl_config_list[dl_config->number_pdus].dci_config_pdu.dci_config_rel15.N_RB_BWP = 106; // to be set + /* + printf("nr_ue_scheduler Type0 PDCCH with rnti %x, BWP %d\n", + dl_config->dl_config_list[dl_config->number_pdus].dci_config_pdu.dci_config_rel15.rnti, + dl_config->dl_config_list[dl_config->number_pdus].dci_config_pdu.dci_config_rel15.N_RB_BWP); + */ + dl_config->number_pdus = dl_config->number_pdus + 1; } if(search_space_mask & type0a_pdcch){ @@ -1301,9 +1307,9 @@ printf("\n>>> nr_ue_process_dci at MAC layer with dci_format=%d\n",dci_format); fapi_nr_ul_config_pusch_pdu_rel15_t *ulsch_config_pdu_0_0 = &ul_config->ul_config_list[ul_config->number_pdus].ulsch_config_pdu.ulsch_pdu_rel15; /* IDENTIFIER_DCI_FORMATS */ /* FREQ_DOM_RESOURCE_ASSIGNMENT_UL */ - nr_ue_process_dci_freq_dom_resource_assignment(&ulsch_config_pdu_0_0,NULL,n_RB_ULBWP,0,dci->freq_dom_resource_assignment_UL); + nr_ue_process_dci_freq_dom_resource_assignment(ulsch_config_pdu_0_0,NULL,n_RB_ULBWP,0,dci->freq_dom_resource_assignment_UL); /* TIME_DOM_RESOURCE_ASSIGNMENT */ - nr_ue_process_dci_time_dom_resource_assignment(&ulsch_config_pdu_0_0,NULL,dci->time_dom_resource_assignment,mac->mib->dmrs_TypeA_Position); + nr_ue_process_dci_time_dom_resource_assignment(ulsch_config_pdu_0_0,NULL,dci->time_dom_resource_assignment,mac->mib->dmrs_TypeA_Position); /* FREQ_HOPPING_FLAG */ if ((mac->phy_config.config_req.ul_bwp_dedicated.pusch_config_dedicated.resource_allocation != 0) && (mac->phy_config.config_req.ul_bwp_dedicated.pusch_config_dedicated.frequency_hopping !=0)) @@ -1377,9 +1383,9 @@ printf("\n>>> nr_ue_process_dci at MAC layer with dci_format=%d\n",dci_format); /* BANDWIDTH_PART_IND */ ulsch_config_pdu_0_1->bandwidth_part_ind = dci->bandwidth_part_ind; /* FREQ_DOM_RESOURCE_ASSIGNMENT_UL */ - nr_ue_process_dci_freq_dom_resource_assignment(&ulsch_config_pdu_0_1,NULL,n_RB_ULBWP,0,dci->freq_dom_resource_assignment_UL); + nr_ue_process_dci_freq_dom_resource_assignment(ulsch_config_pdu_0_1,NULL,n_RB_ULBWP,0,dci->freq_dom_resource_assignment_UL); /* TIME_DOM_RESOURCE_ASSIGNMENT */ - nr_ue_process_dci_time_dom_resource_assignment(&ulsch_config_pdu_0_1,NULL,dci->time_dom_resource_assignment,mac->mib->dmrs_TypeA_Position); + nr_ue_process_dci_time_dom_resource_assignment(ulsch_config_pdu_0_1,NULL,dci->time_dom_resource_assignment,mac->mib->dmrs_TypeA_Position); /* FREQ_HOPPING_FLAG */ if ((mac->phy_config.config_req.ul_bwp_dedicated.pusch_config_dedicated.resource_allocation != 0) && (mac->phy_config.config_req.ul_bwp_dedicated.pusch_config_dedicated.frequency_hopping !=0)) @@ -1769,8 +1775,7 @@ printf("\n>>> nr_ue_process_dci at MAC layer with dci_format=%d\n",dci_format); //for (int k=0;k<1000;k++) printf(">>> %d ",k); //mac->if_module->dl_indication(&mac->dl_info); //mac->if_module->dl_indication(&mac->dl_config_request); - mac->if_module->dl_indication(&mac->phy_config); - + mac->if_module->dl_indication(&mac->phy_config); break; diff --git a/openair2/LAYER2/NR_MAC_gNB/config.c b/openair2/LAYER2/NR_MAC_gNB/config.c index 98bd01e71b..30917fd543 100644 --- a/openair2/LAYER2/NR_MAC_gNB/config.c +++ b/openair2/LAYER2/NR_MAC_gNB/config.c @@ -52,99 +52,108 @@ extern uint8_t nfapi_mode; int32_t **rxdata; int32_t **txdata; -typedef struct eutra_bandentry_s { +typedef struct nr_bandentry_s { int16_t band; - uint32_t ul_min; - uint32_t ul_max; - uint32_t dl_min; - uint32_t dl_max; - uint32_t N_OFFs_DL; -} eutra_bandentry_t; + uint64_t ul_min; + uint64_t ul_max; + uint64_t dl_min; + uint64_t dl_max; + uint64_t N_OFFs_DL; + uint64_t step_size; +} nr_bandentry_t; typedef struct band_info_s { int nbands; - eutra_bandentry_t band_info[100]; -} band_info_t; + nr_bandentry_t band_info[100]; +} nr_band_info_t; -static const eutra_bandentry_t eutra_bandtable[] = { - {1, 19200, 19800, 21100, 21700, 0}, - {2, 18500, 19100, 19300, 19900, 6000}, - {3, 17100, 17850, 18050, 18800, 12000}, - {4, 17100, 17550, 21100, 21550, 19500}, - {5, 8240, 8490, 8690, 8940, 24000}, - {6, 8300, 8400, 8750, 8850, 26500}, - {7, 25000, 25700, 26200, 26900, 27500}, - {8, 8800, 9150, 9250, 9600, 34500}, - {9, 17499, 17849, 18449, 18799, 38000}, - {10, 17100, 17700, 21100, 21700, 41500}, - {11, 14279, 14529, 14759, 15009, 47500}, - {12, 6980, 7160, 7280, 7460, 50100}, - {13, 7770, 7870, 7460, 7560, 51800}, - {14, 7880, 7980, 7580, 7680, 52800}, - {17, 7040, 7160, 7340, 7460, 57300}, - {18, 8150, 9650, 8600, 10100, 58500}, - {19, 8300, 8450, 8750, 8900, 60000}, - {20, 8320, 8620, 7910, 8210, 61500}, - {21, 14479, 14629, 14959, 15109, 64500}, - {22, 34100, 34900, 35100, 35900, 66000}, - {23, 20000, 20200, 21800, 22000, 75000}, - {24, 16126, 16605, 15250, 15590, 77000}, - {25, 18500, 19150, 19300, 19950, 80400}, - {26, 8140, 8490, 8590, 8940, 86900}, - {27, 8070, 8240, 8520, 8690, 90400}, - {28, 7030, 7580, 7580, 8130, 92100}, - {29, 0, 0, 7170, 7280, 96600}, - {30, 23050, 23250, 23500, 23600, 97700}, - {31, 45250, 34900, 46250, 35900, 98700}, - {32, 0, 0, 14520, 14960, 99200}, - {33, 19000, 19200, 19000, 19200, 36000}, - {34, 20100, 20250, 20100, 20250, 36200}, - {35, 18500, 19100, 18500, 19100, 36350}, - {36, 19300, 19900, 19300, 19900, 36950}, - {37, 19100, 19300, 19100, 19300, 37550}, - {38, 25700, 26200, 25700, 26300, 37750}, - {39, 18800, 19200, 18800, 19200, 38250}, - {40, 23000, 24000, 23000, 24000, 38650}, - {41, 24960, 26900, 24960, 26900, 39650}, - {42, 34000, 36000, 34000, 36000, 41590}, - {43, 36000, 38000, 36000, 38000, 43590}, - {44, 7030, 8030, 7030, 8030, 45590}, - {45, 14470, 14670, 14470, 14670, 46590}, - {46, 51500, 59250, 51500, 59250, 46790}, - {65, 19200, 20100, 21100, 22000, 65536}, - {66, 17100, 18000, 21100, 22000, 66436}, - {67, 0, 0, 7380, 7580, 67336}, - {68, 6980, 7280, 7530, 7830, 67536} +static const nr_bandentry_t nr_bandtable[] = { + {1, 1920000, 1980000, 2110000, 2170000, 20, 422000}, + {2, 1850000, 1910000, 1930000, 1990000, 20, 386000}, + {3, 1710000, 1785000, 1805000, 1880000, 20, 361000}, + {5, 824000, 849000, 869000, 894000, 20, 173800}, + {7, 2500000, 2570000, 2620000, 2690000, 20, 524000}, + {8, 880000, 915000, 925000, 960000, 20, 185000}, + {12, 698000, 716000, 728000, 746000, 20, 145800}, + {20, 832000, 862000, 791000, 821000, 20, 158200}, + {25, 1850000, 1915000, 1930000, 1995000, 20, 386000}, + {28, 703000, 758000, 758000, 813000, 20, 151600}, + {34, 2010000, 2025000, 2010000, 2025000, 20, 402000}, + {38, 2570000, 2620000, 2570000, 2630000, 20, 514000}, + {39, 1880000, 1920000, 1880000, 1920000, 20, 376000}, + {40, 2300000, 2400000, 2300000, 2400000, 20, 460000}, + {41, 2496000, 2690000, 2496000, 2690000, 3, 499200}, + {50, 1432000, 1517000, 1432000, 1517000, 20, 286400}, + {51, 1427000, 1432000, 1427000, 1432000, 20, 285400}, + {66, 1710000, 1780000, 2110000, 2200000, 20, 422000}, + {70, 1695000, 1710000, 1995000, 2020000, 20, 399000}, + {71, 663000, 698000, 617000, 652000, 20, 123400}, + {74, 1427000, 1470000, 1475000, 1518000, 20, 295000}, + {75, 000, 000, 1432000, 1517000, 20, 286400}, + {76, 000, 000, 1427000, 1432000, 20, 285400}, + {77, 3300000, 4200000, 3300000, 4200000, 1, 620000}, + {78, 3300000, 3800000, 3300000, 3800000, 1, 620000}, + {79, 4400000, 5000000, 4400000, 5000000, 2, 693334}, + {80, 1710000, 1785000, 000, 000, 20, 342000}, + {81, 860000, 915000, 000, 000, 20, 176000}, + {82, 832000, 862000, 000, 000, 20, 166400}, + {83, 703000, 748000, 000, 000, 20, 140600}, + {84, 1920000, 1980000, 000, 000, 20, 384000}, + {86, 1710000, 1785000, 000, 000, 20, 342000} }; -uint32_t nr_to_earfcn(int eutra_bandP, uint32_t dl_CarrierFreq, uint32_t bw) +uint32_t to_nrarfcn(int nr_bandP, uint64_t dl_CarrierFreq, uint32_t bw) { - uint32_t dl_CarrierFreq_by_100k = dl_CarrierFreq / 100000; - int bw_by_100 = bw / 100; + uint64_t dl_CarrierFreq_by_1k = dl_CarrierFreq / 1000; + int bw_kHz = bw / 1000; int i; - AssertFatal(eutra_bandP < 69, "eutra_band %d > 68\n", eutra_bandP); - for (i = 0; i < 69 && eutra_bandtable[i].band != eutra_bandP; i++); + LOG_I(MAC,"Searching for nr band %d DL Carrier frequency %llu bw %u\n",nr_bandP,(long long unsigned int)dl_CarrierFreq,bw); + AssertFatal(nr_bandP < 86, "nr_band %d > 86\n", nr_bandP); + for (i = 0; i < 30 && nr_bandtable[i].band != nr_bandP; i++); - AssertFatal(dl_CarrierFreq_by_100k >= eutra_bandtable[i].dl_min, - "Band %d, bw %u : DL carrier frequency %u Hz < %u\n", - eutra_bandP, bw, dl_CarrierFreq, - eutra_bandtable[i].dl_min); - AssertFatal(dl_CarrierFreq_by_100k <= - (eutra_bandtable[i].dl_max - bw_by_100), - "Band %d, bw %u: DL carrier frequency %u Hz > %d\n", - eutra_bandP, bw, dl_CarrierFreq, - eutra_bandtable[i].dl_max - bw_by_100); + AssertFatal(dl_CarrierFreq_by_1k >= nr_bandtable[i].dl_min, + "Band %d, bw %u : DL carrier frequency %llu kHz < %llu\n", + nr_bandP, bw, (long long unsigned int)dl_CarrierFreq_by_1k, + (long long unsigned int)nr_bandtable[i].dl_min); + AssertFatal(dl_CarrierFreq_by_1k <= + (nr_bandtable[i].dl_max - bw_kHz), + "Band %d, dl_CarrierFreq %llu bw %u: DL carrier frequency %llu kHz > %llu\n", + nr_bandP, (long long unsigned int)dl_CarrierFreq,bw, (long long unsigned int)dl_CarrierFreq_by_1k, + (long long unsigned int)(nr_bandtable[i].dl_max - bw_kHz)); + + int deltaFglobal; + if (dl_CarrierFreq < 3e9) deltaFglobal = 5; + else deltaFglobal = 15; - return (dl_CarrierFreq_by_100k - eutra_bandtable[i].dl_min + - (eutra_bandtable[i].N_OFFs_DL / 10)); + // This is equation before Table 5.4.2.1-1 in 38101-1-f30 + // F_REF=F_REF_Offs + deltaF_Global(N_REF-NREF_REF_Offs) + return (((dl_CarrierFreq_by_1k - nr_bandtable[i].dl_min)/deltaFglobal) + + nr_bandtable[i].N_OFFs_DL); } +uint64_t from_nrarfcn(int nr_bandP, uint32_t dl_nrarfcn) +{ + + int i; + int deltaFglobal; + + if (nr_bandP < 77 || nr_bandP > 79) deltaFglobal = 5; + else deltaFglobal = 15; + + AssertFatal(nr_bandP < 87, "nr_band %d > 86\n", nr_bandP); + for (i = 0; i < 31 && nr_bandtable[i].band != nr_bandP; i++); + AssertFatal(dl_nrarfcn>=nr_bandtable[i].N_OFFs_DL,"dl_nrarfcn %u < N_OFFs_DL %llu\n",dl_nrarfcn, (long long unsigned int)nr_bandtable[i].N_OFFs_DL); + + return 1000*(nr_bandtable[i].dl_min + (dl_nrarfcn - nr_bandtable[i].N_OFFs_DL) * deltaFglobal); + +} + void config_nr_mib(int Mod_idP, int CC_idP, int p_gNBP, @@ -159,24 +168,31 @@ void config_nr_mib(int Mod_idP, cfg->num_tlv=0; - cfg->rf_config.tx_antenna_ports.value = p_gNBP; - cfg->rf_config.tx_antenna_ports.tl.tag = NFAPI_RF_CONFIG_TX_ANTENNA_PORTS_TAG; + cfg->rf_config.dl_subcarrierspacing.value = subCarrierSpacingCommon; + + cfg->rf_config.dl_subcarrierspacing.tl.tag = NFAPI_NR_RF_CONFIG_DL_SUBCARRIERSPACING_TAG; cfg->num_tlv++; - cfg->sch_config.ssb_subcarrier_offset.value = ssb_SubcarrierOffset; + cfg->rf_config.ul_subcarrierspacing.value = subCarrierSpacingCommon; + cfg->rf_config.ul_subcarrierspacing.tl.tag = NFAPI_NR_RF_CONFIG_UL_SUBCARRIERSPACING_TAG; + cfg->num_tlv++; - + cfg->sch_config.ssb_subcarrier_offset.value = ssb_SubcarrierOffset; + cfg->sch_config.ssb_subcarrier_offset.tl.tag = NFAPI_NR_SCH_CONFIG_SSB_SUBCARRIER_OFFSET_TAG; + cfg->num_tlv++; } void config_common(int Mod_idP, int CC_idP, - int eutra_bandP, - int dl_CarrierFreqP, - int dl_BandwidthP + int nr_bandP, + uint64_t dl_CarrierFreqP, + uint32_t dl_BandwidthP ){ nfapi_nr_config_request_t *cfg = &RC.nrmac[Mod_idP]->config[CC_idP]; + int mu = 1; + // FDD cfg->subframe_config.duplex_mode.value = 1; cfg->subframe_config.duplex_mode.tl.tag = NFAPI_SUBFRAME_CONFIG_DUPLEX_MODE_TAG; @@ -184,15 +200,15 @@ void config_common(int Mod_idP, /// In NR DL and UL will be different band cfg->nfapi_config.rf_bands.number_rf_bands = 1; - cfg->nfapi_config.rf_bands.rf_band[0] = eutra_bandP; + cfg->nfapi_config.rf_bands.rf_band[0] = nr_bandP; cfg->nfapi_config.rf_bands.tl.tag = NFAPI_PHY_RF_BANDS_TAG; cfg->num_tlv++; - cfg->nfapi_config.earfcn.value = nr_to_earfcn(eutra_bandP,dl_CarrierFreqP,dl_BandwidthP*180/100); - cfg->nfapi_config.earfcn.tl.tag = NFAPI_NFAPI_EARFCN_TAG; + cfg->nfapi_config.nrarfcn.value = to_nrarfcn(nr_bandP,dl_CarrierFreqP,dl_BandwidthP*180000*(1+mu)); + cfg->nfapi_config.nrarfcn.tl.tag = NFAPI_NR_NFAPI_NRARFCN_TAG; cfg->num_tlv++; - cfg->subframe_config.numerology_index_mu.value = 1; + cfg->subframe_config.numerology_index_mu.value = mu; //cfg->subframe_config.tl.tag = //cfg->num_tlv++; @@ -217,8 +233,8 @@ void config_common(int Mod_idP, int rrc_mac_config_req_gNB(module_id_t Mod_idP, int CC_idP, int p_gNB, - int eutra_bandP, - int dl_CarrierFreqP, + int nr_bandP, + uint64_t dl_CarrierFreqP, int dl_BandwidthP, NR_BCCH_BCH_Message_t *mib, NR_ServingCellConfigCommon_t *servingcellconfigcommon @@ -245,7 +261,7 @@ int rrc_mac_config_req_gNB(module_id_t Mod_idP, if( servingcellconfigcommon != NULL ){ config_common(Mod_idP, CC_idP, - eutra_bandP, + nr_bandP, dl_CarrierFreqP, dl_BandwidthP ); diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c index c78f52280b..d29a9fb908 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c @@ -71,7 +71,7 @@ uint16_t nr_pdcch_order_table[6] = { 31, 31, 511, 2047, 2047, 8191 }; void clear_nr_nfapi_information(gNB_MAC_INST * gNB, int CC_idP, frame_t frameP, - sub_frame_t subframeP){ + sub_frame_t slotP){ nfapi_nr_dl_config_request_t *DL_req = &gNB->DL_req[0]; nfapi_ul_config_request_t *UL_req = &gNB->UL_req[0]; @@ -87,7 +87,7 @@ void clear_nr_nfapi_information(gNB_MAC_INST * gNB, DL_req[CC_idP].dl_config_request_body.number_pdsch_rnti = 0; //DL_req[CC_idP].dl_config_request_body.transmission_power_pcfich = 6000; - HI_DCI0_req[CC_idP].hi_dci0_request_body.sfnsf = subframeP + (frameP<<4); + HI_DCI0_req[CC_idP].hi_dci0_request_body.sfnsf = slotP + (frameP<<7); HI_DCI0_req[CC_idP].hi_dci0_request_body.number_of_dci = 0; @@ -104,7 +104,7 @@ void check_nr_ul_failure(module_id_t module_idP, int CC_id, int UE_id, frame_t frameP, - sub_frame_t subframeP) { + sub_frame_t slotP) { UE_list_t *UE_list = &RC.nrmac[module_idP]->UE_list; nfapi_nr_dl_config_request_t *DL_req = &RC.nrmac[module_idP]->DL_req[0]; @@ -259,14 +259,14 @@ void schedule_nr_SRS(module_id_t module_idP, frame_t frameP, sub_frame_t subfram } } */ -void copy_nr_ulreq(module_id_t module_idP, frame_t frameP, sub_frame_t subframeP) +void copy_nr_ulreq(module_id_t module_idP, frame_t frameP, sub_frame_t slotP) { int CC_id; gNB_MAC_INST *mac = RC.nrmac[module_idP]; for (CC_id = 0; CC_id < MAX_NUM_CCs; CC_id++) { - nfapi_ul_config_request_t *ul_req_tmp = &mac->UL_req_tmp[CC_id][subframeP]; + nfapi_ul_config_request_t *ul_req_tmp = &mac->UL_req_tmp[CC_id][slotP]; nfapi_ul_config_request_t *ul_req = &mac->UL_req[CC_id]; nfapi_ul_config_request_pdu_t *ul_req_pdu = ul_req->ul_config_request_body.ul_config_pdu_list; @@ -274,12 +274,12 @@ void copy_nr_ulreq(module_id_t module_idP, frame_t frameP, sub_frame_t subframeP // Restore the pointer ul_req->ul_config_request_body.ul_config_pdu_list = ul_req_pdu; - ul_req->sfn_sf = (frameP<<4) + subframeP; + ul_req->sfn_sf = (frameP<<7) + slotP; ul_req_tmp->ul_config_request_body.number_of_pdus = 0; if (ul_req->ul_config_request_body.number_of_pdus>0) { - LOG_D(PHY, "%s() active NOW (frameP:%d subframeP:%d) pdus:%d\n", __FUNCTION__, frameP, subframeP, ul_req->ul_config_request_body.number_of_pdus); + LOG_D(PHY, "%s() active NOW (frameP:%d slotP:%d) pdus:%d\n", __FUNCTION__, frameP, slotP, ul_req->ul_config_request_body.number_of_pdus); } memcpy((void*)ul_req->ul_config_request_body.ul_config_pdu_list, @@ -290,7 +290,7 @@ void copy_nr_ulreq(module_id_t module_idP, frame_t frameP, sub_frame_t subframeP void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, frame_t frameP, - sub_frame_t subframeP){ + sub_frame_t slotP){ protocol_ctxt_t ctxt; int CC_id, i = -1; @@ -303,7 +303,8 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_ENB_DLSCH_ULSCH_SCHEDULER,VCD_FUNCTION_IN); RC.nrmac[module_idP]->frame = frameP; - RC.nrmac[module_idP]->subframe = subframeP; + RC.nrmac[module_idP]->slot = slotP; + for (CC_id = 0; CC_id < MAX_NUM_CCs; CC_id++) { //mbsfn_status[CC_id] = 0; @@ -312,17 +313,20 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, memset(cc[CC_id].vrb_map, 0, 100); memset(cc[CC_id].vrb_map_UL, 0, 100); - clear_nr_nfapi_information(RC.nrmac[module_idP], CC_id, frameP, subframeP); + clear_nr_nfapi_information(RC.nrmac[module_idP], CC_id, frameP, slotP); } // refresh UE list based on UEs dropped by PHY in previous subframe for (i = 0; i < MAX_MOBILES_PER_GNB; i++) { if (UE_list->active[i]) { + nfapi_nr_config_request_t *cfg = &RC.nrmac[module_idP]->config[CC_id]; + rnti = UE_RNTI(module_idP, i); CC_id = UE_PCCID(module_idP, i); - - if (((frameP&127) == 0) && (subframeP == 0)) { + int spf = get_spf(cfg); + + if (((frameP&127) == 0) && (slotP == 0)) { LOG_I(MAC, "UE rnti %x : %s, PHR %d dB DL CQI %d PUSCH SNR %d PUCCH SNR %d\n", rnti, @@ -334,11 +338,11 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, (UE_list->UE_sched_ctrl[i].pucch1_snr[CC_id] - 128) / 2); } - RC.gNB[module_idP][CC_id]->pusch_stats_bsr[i][(frameP * 10) + subframeP] = -63; + RC.gNB[module_idP][CC_id]->pusch_stats_bsr[i][to_absslot(cfg,frameP,slotP)] = -63; if (i == UE_list->head) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_UE0_BSR,RC.gNB[module_idP][CC_id]-> - pusch_stats_bsr[i][(frameP * 10) + subframeP]); + pusch_stats_bsr[i][to_absslot(cfg,frameP,slotP)]); // increment this, it is cleared when we receive an sdu RC.nrmac[module_idP]->UE_list.UE_sched_ctrl[i].ul_inactivity_timer++; @@ -407,19 +411,21 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, } //END if (UE_list->active[i]) } //END for (i = 0; i < MAX_MOBILES_PER_GNB; i++) - PROTOCOL_CTXT_SET_BY_MODULE_ID(&ctxt, module_idP, ENB_FLAG_YES,NOT_A_RNTI, frameP, subframeP,module_idP); + PROTOCOL_CTXT_SET_BY_MODULE_ID(&ctxt, module_idP, ENB_FLAG_YES,NOT_A_RNTI, frameP, slotP,module_idP); pdcp_run(&ctxt); //rrc_rx_tx(&ctxt, CC_id); // This schedules MIB - if((subframeP == 0) && (frameP & 7) == 0){ - schedule_nr_mib(module_idP, frameP, subframeP); + if((slotP == 0) && (frameP & 7) == 0){ + schedule_nr_mib(module_idP, frameP, slotP); } // Phytest scheduling/ option not activated because of pending bug - nr_schedule_css_dlsch_phytest(module_idP, frameP, subframeP); + + if (slotP==2) + nr_schedule_css_dlsch_phytest(module_idP, frameP, slotP); /* diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c index 95f685a536..8dbdbc93fa 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c @@ -57,7 +57,7 @@ extern RAN_CONTEXT_t RC; -void schedule_nr_mib(module_id_t module_idP, frame_t frameP, sub_frame_t subframeP){ +void schedule_nr_mib(module_id_t module_idP, frame_t frameP, sub_frame_t slotP){ gNB_MAC_INST *gNB = RC.nrmac[module_idP]; NR_COMMON_channels_t *cc; @@ -70,9 +70,9 @@ void schedule_nr_mib(module_id_t module_idP, frame_t frameP, sub_frame_t subfram int mib_sdu_length; int CC_id; - uint16_t sfn_sf = frameP << 4 | subframeP; + uint16_t sfn_sf = frameP << 7 | slotP; - AssertFatal(subframeP == 0, "Subframe must be 0\n"); + AssertFatal(slotP == 0, "Subframe must be 0\n"); AssertFatal((frameP & 7) == 0, "Frame must be a multiple of 8\n"); for (CC_id = 0; CC_id < MAX_NUM_CCs; CC_id++) { @@ -83,11 +83,11 @@ void schedule_nr_mib(module_id_t module_idP, frame_t frameP, sub_frame_t subfram mib_sdu_length = mac_rrc_nr_data_req(module_idP, CC_id, frameP, MIBCH, 1, &cc->MIB_pdu.payload[0]); // not used in this case - LOG_D(MAC, "Frame %d, subframe %d: BCH PDU length %d\n", frameP, subframeP, mib_sdu_length); + LOG_D(MAC, "Frame %d, slot %d: BCH PDU length %d\n", frameP, slotP, mib_sdu_length); if (mib_sdu_length > 0) { - LOG_D(MAC, "Frame %d, subframe %d: Adding BCH PDU in position %d (length %d)\n", frameP, subframeP, dl_req->number_pdu, mib_sdu_length); + LOG_D(MAC, "Frame %d, slot %d: Adding BCH PDU in position %d (length %d)\n", frameP, slotP, dl_req->number_pdu, mib_sdu_length); if ((frameP & 1023) < 80){ LOG_D(MAC,"[gNB %d] Frame %d : MIB->BCH CC_id %d, Received %d bytes\n",module_idP, frameP, CC_id, mib_sdu_length); diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c index 2e73775f29..131e64a93b 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c @@ -30,60 +30,101 @@ #include "nr_mac_gNB.h" #include "SCHED_NR/sched_nr.h" +#include "mac_proto.h" +#include "PHY/NR_TRANSPORT/nr_dlsch.h" +#include "PHY/NR_TRANSPORT/nr_dci.h" extern RAN_CONTEXT_t RC; /*Scheduling of DLSCH with associated DCI in common search space - * current version has only a DCI for type 1 PDCCH for RA-RNTI*/ + * current version has only a DCI for type 1 PDCCH for C_RNTI*/ void nr_schedule_css_dlsch_phytest(module_id_t module_idP, frame_t frameP, - sub_frame_t subframeP) + sub_frame_t slotP) { uint8_t CC_id; gNB_MAC_INST *nr_mac = RC.nrmac[module_idP]; //NR_COMMON_channels_t *cc = nr_mac->common_channels; nfapi_nr_dl_config_request_body_t *dl_req; - nfapi_nr_dl_config_request_pdu_t *dl_config_pdu; + nfapi_nr_dl_config_request_pdu_t *dl_config_dci_pdu; + nfapi_nr_dl_config_request_pdu_t *dl_config_dlsch_pdu; nfapi_tx_request_pdu_t *TX_req; - uint16_t sfn_sf = frameP << 4 | subframeP; + nfapi_nr_config_request_t *cfg = &nr_mac->config[0]; + uint16_t rnti = 0x1234; + + uint16_t sfn_sf = frameP << 7 | slotP; + int dl_carrier_bandwidth = cfg->rf_config.dl_carrier_bandwidth.value; + + // everything here is hard-coded to 30 kHz + int scs = get_dlscs(cfg); + int slots_per_frame = get_spf(cfg); for (CC_id=0; CC_idgNB_config; - NR_DL_FRAME_PARMS *fp = &gNB->frame_parms; dl_req = &nr_mac->DL_req[CC_id].dl_config_request_body; - dl_config_pdu = &dl_req->dl_config_pdu_list[dl_req->number_pdu]; - memset((void*)dl_config_pdu,0,sizeof(nfapi_nr_dl_config_request_pdu_t)); - dl_config_pdu->pdu_type = NFAPI_NR_DL_CONFIG_DCI_DL_PDU_TYPE; - dl_config_pdu->pdu_size = (uint8_t)(2+sizeof(nfapi_nr_dl_config_dci_dl_pdu)); + dl_config_dci_pdu = &dl_req->dl_config_pdu_list[dl_req->number_pdu]; + memset((void*)dl_config_dci_pdu,0,sizeof(nfapi_nr_dl_config_request_pdu_t)); + dl_config_dci_pdu->pdu_type = NFAPI_NR_DL_CONFIG_DCI_DL_PDU_TYPE; + dl_config_dci_pdu->pdu_size = (uint8_t)(2+sizeof(nfapi_nr_dl_config_dci_dl_pdu)); - nfapi_nr_dl_config_dci_dl_pdu_rel15_t *pdu_rel15 = &dl_config_pdu->dci_dl_pdu.dci_dl_pdu_rel15; - nfapi_nr_dl_config_pdcch_parameters_rel15_t *params_rel15 = &dl_config_pdu->dci_dl_pdu.pdcch_params_rel15; + dl_config_dlsch_pdu = &dl_req->dl_config_pdu_list[dl_req->number_pdu+1]; + memset((void*)dl_config_dlsch_pdu,0,sizeof(nfapi_nr_dl_config_request_pdu_t)); + dl_config_dlsch_pdu->pdu_type = NFAPI_NR_DL_CONFIG_DLSCH_PDU_TYPE; + dl_config_dlsch_pdu->pdu_size = (uint8_t)(2+sizeof(nfapi_nr_dl_config_dlsch_pdu)); - nr_configure_css_dci_from_mib(&gNB->pdcch_type0_params, - kHz30, kHz30, nr_FR1, 0, 0, - fp->slots_per_frame, - cfg->rf_config.dl_carrier_bandwidth.value); - memcpy((void*)params_rel15, (void*)&gNB->pdcch_type0_params, sizeof(nfapi_nr_dl_config_pdcch_parameters_rel15_t)); + nfapi_nr_dl_config_dci_dl_pdu_rel15_t *pdu_rel15 = &dl_config_dci_pdu->dci_dl_pdu.dci_dl_pdu_rel15; + nfapi_nr_dl_config_pdcch_parameters_rel15_t *params_rel15 = &dl_config_dci_pdu->dci_dl_pdu.pdcch_params_rel15; + nfapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_pdu_rel15 = &dl_config_dlsch_pdu->dlsch_pdu.dlsch_pdu_rel15; - pdu_rel15->frequency_domain_assignment = 5; - pdu_rel15->time_domain_assignment = 2; - pdu_rel15->vrb_to_prb_mapping = 0; - pdu_rel15->mcs = 12; + dlsch_pdu_rel15->start_prb = 0; + dlsch_pdu_rel15->n_prb = 50; + dlsch_pdu_rel15->start_symbol = 2; + dlsch_pdu_rel15->nb_symbols = 8; + dlsch_pdu_rel15->rnti = rnti; + dlsch_pdu_rel15->nb_layers =1; + dlsch_pdu_rel15->nb_codewords = 1; + dlsch_pdu_rel15->mcs_idx = 9; + dlsch_pdu_rel15->ndi = 1; + dlsch_pdu_rel15->redundancy_version = 0; + + + nr_configure_css_dci_initial(params_rel15, + scs, scs, nr_FR1, 0, 0, + slots_per_frame, + dl_carrier_bandwidth); + + params_rel15->first_slot = 0; + + pdu_rel15->frequency_domain_assignment = get_RIV(dlsch_pdu_rel15->start_prb, dlsch_pdu_rel15->n_prb, cfg->rf_config.dl_carrier_bandwidth.value); + pdu_rel15->time_domain_assignment = get_SLIV(dlsch_pdu_rel15->start_symbol, dlsch_pdu_rel15->nb_symbols); + pdu_rel15->vrb_to_prb_mapping = 1; + pdu_rel15->mcs = 9; pdu_rel15->tb_scaling = 1; - LOG_D(MAC, "[gNB scheduler phytest] DCI type 1 payload: freq_alloc %d, time_alloc %d, vrb to prb %d, mcs %d tb_scaling %d\n", + + pdu_rel15->ra_preamble_index = 25; + pdu_rel15->format_indicator = 1; + pdu_rel15->ndi = 1; + pdu_rel15->rv = 0; + pdu_rel15->harq_pid = 0; + pdu_rel15->dai = 2; + pdu_rel15->tpc = 2; + pdu_rel15->pucch_resource_indicator = 7; + pdu_rel15->pdsch_to_harq_feedback_timing_indicator = 7; + + LOG_I(MAC, "[gNB scheduler phytest] DCI type 1 payload: freq_alloc %d, time_alloc %d, vrb to prb %d, mcs %d tb_scaling %d ndi %d rv %d\n", pdu_rel15->frequency_domain_assignment, pdu_rel15->time_domain_assignment, pdu_rel15->vrb_to_prb_mapping, pdu_rel15->mcs, - pdu_rel15->tb_scaling); + pdu_rel15->tb_scaling, + pdu_rel15->ndi, + pdu_rel15->rv); - params_rel15->rnti = 0x03; - params_rel15->rnti_type = NFAPI_NR_RNTI_RA; + params_rel15->rnti = rnti; + params_rel15->rnti_type = NFAPI_NR_RNTI_C; params_rel15->dci_format = NFAPI_NR_DL_DCI_FORMAT_1_0; //params_rel15->aggregation_level = 1; LOG_D(MAC, "DCI type 1 params: rmsi_pdcch_config %d, rnti %d, rnti_type %d, dci_format %d\n \ @@ -102,15 +143,35 @@ void nr_schedule_css_dlsch_phytest(module_id_t module_idP, params_rel15->nb_slots, params_rel15->sfn_mod2, params_rel15->first_slot); + nr_get_tbs(&dl_config_dlsch_pdu->dlsch_pdu, dl_config_dci_pdu->dci_dl_pdu, *cfg); + LOG_I(MAC, "DLSCH PDU: start PRB %d n_PRB %d start symbol %d nb_symbols %d nb_layers %d nb_codewords %d mcs %d\n", + dlsch_pdu_rel15->start_prb, + dlsch_pdu_rel15->n_prb, + dlsch_pdu_rel15->start_symbol, + dlsch_pdu_rel15->nb_symbols, + dlsch_pdu_rel15->nb_layers, + dlsch_pdu_rel15->nb_codewords, + dlsch_pdu_rel15->mcs_idx); + dl_req->number_dci++; - dl_req->number_pdu++; + dl_req->number_pdsch_rnti++; + dl_req->number_pdu+=2; TX_req = &nr_mac->TX_req[CC_id].tx_request_body.tx_pdu_list[nr_mac->TX_req[CC_id].tx_request_body.number_of_pdus]; TX_req->pdu_length = 6; TX_req->pdu_index = nr_mac->pdu_index[CC_id]++; TX_req->num_segments = 1; TX_req->segments[0].segment_length = 8; - //TX_req->segments[0].segment_data = (uint8_t*)pdu_rel15; + nr_mac->TX_req[CC_id].tx_request_body.number_of_pdus++; + nr_mac->TX_req[CC_id].sfn_sf = sfn_sf; + nr_mac->TX_req[CC_id].tx_request_body.tl.tag = NFAPI_TX_REQUEST_BODY_TAG; + nr_mac->TX_req[CC_id].header.message_id = NFAPI_TX_REQUEST; + + TX_req = &nr_mac->TX_req[CC_id].tx_request_body.tx_pdu_list[nr_mac->TX_req[CC_id].tx_request_body.number_of_pdus+1]; + TX_req->pdu_length = dlsch_pdu_rel15->transport_block_size; + TX_req->pdu_index = nr_mac->pdu_index[CC_id]++; + TX_req->num_segments = 1; + TX_req->segments[0].segment_length = 8; nr_mac->TX_req[CC_id].tx_request_body.number_of_pdus++; nr_mac->TX_req[CC_id].sfn_sf = sfn_sf; nr_mac->TX_req[CC_id].tx_request_body.tl.tag = NFAPI_TX_REQUEST_BODY_TAG; diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c index a2a96799fd..ca0f103a43 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c @@ -65,59 +65,175 @@ extern RAN_CONTEXT_t RC; extern int n_active_slices; -int is_nr_UL_sf(NR_COMMON_channels_t * ccP, sub_frame_t subframeP){ - // if FDD return dummy value - if (ccP->tdd_Config == NULL) +/// LUT for the number of symbols in the coreset indexed by coreset index (4 MSB rmsi_pdcch_config) +uint8_t nr_coreset_nsymb_pdcch_type_0_b40Mhz[16] = {2,2,2,2,2,3,3,3,3,3,1,1,1,2,2,2}; // below 40Mhz bw +uint8_t nr_coreset_nsymb_pdcch_type_0_a40Mhz[10] = {2,2,3,3,1,1,2,2,3,3}; // above 40Mhz bw +/// LUT for the number of RBs in the coreset indexed by coreset index +uint8_t nr_coreset_rb_offset_pdcch_type_0_b40Mhz[16] = {0,1,2,3,4,0,1,2,3,4,12,14,16,12,14,16}; +uint8_t nr_coreset_rb_offset_pdcch_type_0_a40Mhz[10] = {0,4,0,4,0,28,0,28,0,28}; +/// LUT for monitoring occasions param O indexed by ss index (4 LSB rmsi_pdcch_config) +uint8_t nr_ss_param_O_type_0_mux1_FR1[16] = {0,0,2,2,5,5,7,7,0,5,0,0,2,2,5,5}; +uint8_t nr_ss_param_O_type_0_mux1_FR2[14] = {0,0,2.5,2.5,5,5,0,2.5,5,7.5,7.5,7.5,0,5}; +/// LUT for number of SS sets per slot indexed by ss index +uint8_t nr_ss_sets_per_slot_type_0_FR1[16] = {1,2,1,2,1,2,1,2,1,1,1,1,1,1,1,1}; +uint8_t nr_ss_sets_per_slot_type_0_FR2[14] = {1,2,1,2,1,2,2,2,2,1,2,2,1,1}; +/// LUT for monitoring occasions param M indexed by ss index +uint8_t nr_ss_param_M_type_0_mux1_FR1[16] = {1,0.5,1,0.5,1,0.5,1,0.5,2,2,1,1,1,1,1,1}; +uint8_t nr_ss_param_M_type_0_mux1_FR2[14] = {1,0.5,1,0.5,1,0.5,0.5,0.5,0.5,1,0.5,0.5,2,2}; +/// LUT for SS first symbol index indexed by ss index +uint8_t nr_ss_first_symb_idx_type_0_mux1_FR1[8] = {0,0,1,2,1,2,1,2}; + +int is_nr_UL_slot(NR_COMMON_channels_t * ccP, int slot){ + return (0); +} + +void nr_configure_css_dci_initial(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, + nr_scs_e scs_common, + nr_scs_e pdcch_scs, + nr_frequency_range_e freq_range, + uint8_t rmsi_pdcch_config, + uint8_t ssb_idx, + uint16_t nb_slots_per_frame, + uint16_t N_RB) +{ + uint8_t O, M; + uint8_t ss_idx = rmsi_pdcch_config&0xf; + uint8_t cset_idx = (rmsi_pdcch_config>>4)&0xf; + uint8_t mu; + + /// Coreset params + switch(scs_common) { + + case kHz15: + mu = 0; + break; + + case kHz30: + mu = 1; + + if (N_RB < 106) { // Minimum 40Mhz bandwidth not satisfied + switch(pdcch_scs) { + case kHz15: + AssertFatal(1==0,"kHz15 not supported yet\n"); + break; + + case kHz30: + pdcch_params->mux_pattern = NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1; + pdcch_params->n_rb = (cset_idx < 10)? 24 : 48; + pdcch_params->n_symb = nr_coreset_nsymb_pdcch_type_0_b40Mhz[cset_idx]; + pdcch_params->rb_offset = nr_coreset_rb_offset_pdcch_type_0_b40Mhz[cset_idx]; + break; + + default: + AssertFatal(1==0,"Invalid scs_common/pdcch_scs combination %d/%d \n", scs_common, pdcch_scs); + } + } + + else { + AssertFatal(ss_idx<10 ,"Invalid scs_common/pdcch_scs combination %d/%d \n", scs_common, pdcch_scs); + switch(pdcch_scs) { + case kHz15: + AssertFatal(1==0,"15 kHz SCS not supported yet\n"); + break; + + case kHz30: + pdcch_params->mux_pattern = NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1; + pdcch_params->n_rb = (cset_idx < 4)? 24 : 48; + pdcch_params->n_symb = nr_coreset_nsymb_pdcch_type_0_b40Mhz[cset_idx]; + pdcch_params->rb_offset = nr_coreset_rb_offset_pdcch_type_0_b40Mhz[cset_idx]; + break; + + default: + AssertFatal(1==0,"Invalid scs_common/pdcch_scs combination %d/%d \n", scs_common, pdcch_scs); + } + } + + case kHz60: + mu = 2; + break; + + case kHz120: + mu = 3; + break; + + default: + AssertFatal(1==0,"Invalid common subcarrier spacing %d\n", scs_common); + + } + + /// Search space params + switch(pdcch_params->mux_pattern) { + + case NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE1: + if (freq_range == nr_FR1) { + O = nr_ss_param_O_type_0_mux1_FR1[ss_idx]; + pdcch_params->nb_ss_sets_per_slot = nr_ss_sets_per_slot_type_0_FR1[ss_idx]; + M = nr_ss_param_M_type_0_mux1_FR1[ss_idx]; + pdcch_params->first_symbol = (ss_idx < 8)? ( (ss_idx&1)? pdcch_params->n_symb : 0 ) : nr_ss_first_symb_idx_type_0_mux1_FR1[ss_idx - 8]; + } + + else { + AssertFatal(ss_idx<14 ,"Invalid search space index for multiplexing type 1 and FR2 %d\n", ss_idx); + O = nr_ss_param_O_type_0_mux1_FR2[ss_idx]; + pdcch_params->nb_ss_sets_per_slot = nr_ss_sets_per_slot_type_0_FR2[ss_idx]; + M = nr_ss_param_M_type_0_mux1_FR2[ss_idx]; + pdcch_params->first_symbol = (ss_idx < 12)? ( (ss_idx&1)? 7 : 0 ) : 0; + } + pdcch_params->nb_slots = 2; + pdcch_params->sfn_mod2 = ((uint8_t)(floor( (O*pow(2, mu) + floor(ssb_idx*M)) / nb_slots_per_frame )) & 1)? 1 : 0; + pdcch_params->first_slot = (uint8_t)(O*pow(2, mu) + floor(ssb_idx*M)) % nb_slots_per_frame; + + break; + + case NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE2: + break; + + case NFAPI_NR_SSB_AND_CSET_MUX_PATTERN_TYPE3: + break; + + default: + AssertFatal(1==0, "Invalid SSB and coreset multiplexing pattern %d\n", pdcch_params->mux_pattern); + } + pdcch_params->config_type = NFAPI_NR_CSET_CONFIG_MIB_SIB1; + pdcch_params->cr_mapping_type = NFAPI_NR_CCE_REG_MAPPING_INTERLEAVED; + pdcch_params->precoder_granularity = NFAPI_NR_CSET_SAME_AS_REG_BUNDLE; + pdcch_params->reg_bundle_size = 6; + pdcch_params->interleaver_size = 2; + // set initial banwidth part to full bandwidth + pdcch_params->n_RB_BWP = N_RB; + + +} + +void nr_configure_css_dci_from_pdcch_config(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, + nfapi_nr_coreset_t* coreset, + nfapi_nr_search_space_t* search_space) { + + +} + +int get_dlscs(nfapi_nr_config_request_t *cfg) { + + return(cfg->rf_config.dl_subcarrierspacing.value); +} + + +int get_ulscs(nfapi_nr_config_request_t *cfg) { + + return(cfg->rf_config.ul_subcarrierspacing.value); +} + +int get_spf(nfapi_nr_config_request_t *cfg) { + + int mu = cfg->rf_config.dl_subcarrierspacing.value; + AssertFatal(mu>=0&&mu<4,"Illegal scs %d\n",mu); + + return(10 * (1<tdd_Config->subframeAssignment) { - case 1: - switch (subframeP) { - case 0: - case 4: - case 5: - case 9: - return (0); - break; - case 2: - case 3: - case 7: - case 8: - return (1); - break; - default: - return (0); - break; - } - break; - case 3: - if ((subframeP <= 1) || (subframeP >= 5)) - return (0); - else if ((subframeP > 1) && (subframeP < 5)) - return (1); - else - AssertFatal(1 == 0, "Unknown subframe number\n"); - break; - case 4: - if ((subframeP <= 1) || (subframeP >= 4)) - return (0); - else if ((subframeP > 1) && (subframeP < 4)) - return (1); - else - AssertFatal(1 == 0, "Unknown subframe number\n"); - break; - case 5: - if ((subframeP <= 1) || (subframeP >= 3)) - return (0); - else if ((subframeP > 1) && (subframeP < 3)) - return (1); - else - AssertFatal(1 == 0, "Unknown subframe number\n"); - break; - default: - AssertFatal(1 == 0, - "subframe %d Unsupported TDD configuration %d\n", - subframeP, (int) ccP->tdd_Config->subframeAssignment); - break; - } } diff --git a/openair2/LAYER2/NR_MAC_gNB/mac_proto.h b/openair2/LAYER2/NR_MAC_gNB/mac_proto.h index c205fa5834..784496ff9e 100644 --- a/openair2/LAYER2/NR_MAC_gNB/mac_proto.h +++ b/openair2/LAYER2/NR_MAC_gNB/mac_proto.h @@ -39,14 +39,14 @@ void mac_top_init_gNB(void); int rrc_mac_config_req_gNB(module_id_t Mod_idP, int CC_id, int p_gNB, - int eutra_bandP, - int dl_CarrierFreqP, + int nr_bandP, + uint64_t dl_CarrierFreqP, int dl_BandwidthP, NR_BCCH_BCH_Message_t *mib, NR_ServingCellConfigCommon_t *servingcellconfigcommon ); -int is_nr_UL_sf(NR_COMMON_channels_t * ccP, sub_frame_t subframeP); +int is_nr_UL_slot(NR_COMMON_channels_t * ccP, int slotP); void clear_nr_nfapi_information(gNB_MAC_INST * gNB, int CC_idP, @@ -59,4 +59,33 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, void schedule_nr_mib(module_id_t module_idP, frame_t frameP, sub_frame_t subframeP); + + +void nr_schedule_css_dlsch_phytest(module_id_t module_idP, + frame_t frameP, + sub_frame_t subframeP); + + +void nr_configure_css_dci_initial(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, + nr_scs_e scs_common, + nr_scs_e pdcch_scs, + nr_frequency_range_e freq_range, + uint8_t rmsi_pdcch_config, + uint8_t ssb_idx, + uint16_t nb_slots_per_frame, + uint16_t N_RB); + + +void nr_configure_css_dci_from_pdcch_config(nfapi_nr_dl_config_pdcch_parameters_rel15_t* pdcch_params, + nfapi_nr_coreset_t* coreset, + nfapi_nr_search_space_t* search_space); + +int get_dlscs(nfapi_nr_config_request_t *cfg); + +int get_ulscs(nfapi_nr_config_request_t *cfg); + +int get_spf(nfapi_nr_config_request_t *cfg); + +int to_absslot(nfapi_nr_config_request_t *cfg,int frame,int slot); + #endif /*__LAYER2_NR_MAC_PROTO_H__*/ diff --git a/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h b/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h index 1799721325..3af2a4a7c9 100644 --- a/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h +++ b/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h @@ -67,7 +67,7 @@ typedef struct { int physCellId; int p_gNB; int Ncp; - int eutra_band; + int nr_band; uint32_t dl_CarrierFreq; NR_BCCH_BCH_Message_t *mib; TDD_Config_t *tdd_Config; @@ -107,8 +107,8 @@ typedef struct gNB_MAC_INST_s { module_id_t Mod_id; /// frame counter frame_t frame; - /// subframe counter - sub_frame_t subframe; + /// slot counter + int slot; /// Pointer to IF module instance for PHY NR_IF_Module_t *if_inst; /// Common cell resources diff --git a/openair2/NR_PHY_INTERFACE/NR_IF_Module.c b/openair2/NR_PHY_INTERFACE/NR_IF_Module.c index 05adde0fe4..1bde86873b 100644 --- a/openair2/NR_PHY_INTERFACE/NR_IF_Module.c +++ b/openair2/NR_PHY_INTERFACE/NR_IF_Module.c @@ -59,7 +59,7 @@ void handle_nr_rach(NR_UL_IND_t *UL_info) { AssertFatal(UL_info->rach_ind.rach_indication_body.number_of_preambles==1,"More than 1 preamble not supported\n"); UL_info->rach_ind.rach_indication_body.number_of_preambles=0; - LOG_D(MAC,"UL_info[Frame %d, Subframe %d] Calling initiate_ra_proc RACH:SFN/SF:%d\n",UL_info->frame,UL_info->subframe, NFAPI_SFNSF2DEC(UL_info->rach_ind.sfn_sf)); + LOG_D(MAC,"UL_info[Frame %d, Slot %d] Calling initiate_ra_proc RACH:SFN/SF:%d\n",UL_info->frame,UL_info->slot, NFAPI_SFNSF2DEC(UL_info->rach_ind.sfn_sf)); initiate_ra_proc(UL_info->module_id, UL_info->CC_id, NFAPI_SFNSF2SFN(UL_info->rach_ind.sfn_sf), @@ -80,12 +80,12 @@ void handle_nr_rach(NR_UL_IND_t *UL_info) { for (i=0;irach_ind_br.rach_indication_body.number_of_preambles;i++) { AssertFatal(UL_info->rach_ind_br.rach_indication_body.preamble_list[i].preamble_rel13.rach_resource_type>0, "Got regular PRACH preamble, not BL/CE\n"); - LOG_D(MAC,"Frame %d, Subframe %d Calling initiate_ra_proc (CE_level %d)\n",UL_info->frame,UL_info->subframe, + LOG_D(MAC,"Frame %d, Slot %d Calling initiate_ra_proc (CE_level %d)\n",UL_info->frame,UL_info->slot, UL_info->rach_ind_br.rach_indication_body.preamble_list[i].preamble_rel13.rach_resource_type-1); initiate_ra_proc(UL_info->module_id, UL_info->CC_id, UL_info->frame, - UL_info->subframe, + UL_info->slot, UL_info->rach_ind_br.rach_indication_body.preamble_list[i].preamble_rel8.preamble, UL_info->rach_ind_br.rach_indication_body.preamble_list[i].preamble_rel8.timing_advance, UL_info->rach_ind_br.rach_indication_body.preamble_list[i].preamble_rel8.rnti, @@ -113,7 +113,7 @@ void handle_nr_sr(NR_UL_IND_t *UL_info) { SR_indication(UL_info->module_id, UL_info->CC_id, UL_info->frame, - UL_info->subframe, + UL_info->slot, UL_info->sr_ind.sr_indication_body.sr_pdu_list[i].rx_ue_information.rnti, UL_info->sr_ind.sr_indication_body.sr_pdu_list[i].ul_cqi_information.ul_cqi); } @@ -133,7 +133,7 @@ void handle_nr_cqi(NR_UL_IND_t *UL_info) { nfapi_cqi_indication_t ind; ind.header.message_id = NFAPI_RX_CQI_INDICATION; - ind.sfn_sf = UL_info->frame<<4 | UL_info->subframe; + ind.sfn_sf = UL_info->frame<<4 | UL_info->slot; ind.cqi_indication_body = UL_info->cqi_ind; oai_nfapi_cqi_indication(&ind); @@ -147,7 +147,7 @@ void handle_nr_cqi(NR_UL_IND_t *UL_info) { cqi_indication(UL_info->module_id, UL_info->CC_id, UL_info->frame, - UL_info->subframe, + UL_info->slot, UL_info->cqi_ind.cqi_pdu_list[i].rx_ue_information.rnti, &UL_info->cqi_ind.cqi_pdu_list[i].cqi_indication_rel9, UL_info->cqi_ind.cqi_raw_pdu_list[i].pdu, @@ -220,11 +220,11 @@ void handle_nr_ulsch(NR_UL_IND_t *UL_info) { UL_info->rx_ind.rx_indication_body.rx_pdu_list[i].rx_ue_information.rnti) { LOG_D(PHY, "UL_info->crc_ind.crc_indication_body.crc_pdu_list[%d].crc_indication_rel8.crc_flag:%d\n", j, UL_info->crc_ind.crc_indication_body.crc_pdu_list[j].crc_indication_rel8.crc_flag); if (UL_info->crc_ind.crc_indication_body.crc_pdu_list[j].crc_indication_rel8.crc_flag == 1) { // CRC error indication - LOG_D(MAC,"Frame %d, Subframe %d Calling rx_sdu (CRC error) \n",UL_info->frame,UL_info->subframe); + LOG_D(MAC,"Frame %d, Slot %d Calling rx_sdu (CRC error) \n",UL_info->frame,UL_info->slot); rx_sdu(UL_info->module_id, UL_info->CC_id, NFAPI_SFNSF2SFN(UL_info->rx_ind.sfn_sf), //UL_info->frame, - NFAPI_SFNSF2SF(UL_info->rx_ind.sfn_sf), //UL_info->subframe, + NFAPI_SFNSF2SF(UL_info->rx_ind.sfn_sf), //UL_info->slot, UL_info->rx_ind.rx_indication_body.rx_pdu_list[i].rx_ue_information.rnti, (uint8_t *)NULL, UL_info->rx_ind.rx_indication_body.rx_pdu_list[i].rx_indication_rel8.length, @@ -232,11 +232,11 @@ void handle_nr_ulsch(NR_UL_IND_t *UL_info) { UL_info->rx_ind.rx_indication_body.rx_pdu_list[i].rx_indication_rel8.ul_cqi); } else { - LOG_D(MAC,"Frame %d, Subframe %d Calling rx_sdu (CRC ok) \n",UL_info->frame,UL_info->subframe); + LOG_D(MAC,"Frame %d, Slot %d Calling rx_sdu (CRC ok) \n",UL_info->frame,UL_info->slot); rx_sdu(UL_info->module_id, UL_info->CC_id, NFAPI_SFNSF2SFN(UL_info->rx_ind.sfn_sf), //UL_info->frame, - NFAPI_SFNSF2SF(UL_info->rx_ind.sfn_sf), //UL_info->subframe, + NFAPI_SFNSF2SF(UL_info->rx_ind.sfn_sf), //UL_info->slot, UL_info->rx_ind.rx_indication_body.rx_pdu_list[i].rx_ue_information.rnti, UL_info->rx_ind.rx_indication_body.rx_pdu_list[i].data, UL_info->rx_ind.rx_indication_body.rx_pdu_list[i].rx_indication_rel8.length, @@ -250,12 +250,12 @@ void handle_nr_ulsch(NR_UL_IND_t *UL_info) { } // for (i=0;irx_ind.number_of_pdus;i++) UL_info->crc_ind.crc_indication_body.number_of_crcs=0; UL_info->rx_ind.rx_indication_body.number_of_pdus = 0; - } // UL_info->rx_ind.rx_indication_body.number_of_pdus>0 && UL_info->subframe && UL_info->crc_ind.crc_indication_body.number_of_crcs>0 + } // UL_info->rx_ind.rx_indication_body.number_of_pdus>0 && UL_info->slot && UL_info->crc_ind.crc_indication_body.number_of_crcs>0 else if (UL_info->rx_ind.rx_indication_body.number_of_pdus!=0 || UL_info->crc_ind.crc_indication_body.number_of_crcs!=0) { LOG_E(PHY,"hoping not to have mis-match between CRC ind and RX ind - hopefully the missing message is coming shortly rx_ind:%d(SFN/SF:%05d) crc_ind:%d(SFN/SF:%05d) UL_info(SFN/SF):%04d%d\n", UL_info->rx_ind.rx_indication_body.number_of_pdus, NFAPI_SFNSF2DEC(UL_info->rx_ind.sfn_sf), UL_info->crc_ind.crc_indication_body.number_of_crcs, NFAPI_SFNSF2DEC(UL_info->crc_ind.sfn_sf), - UL_info->frame, UL_info->subframe); + UL_info->frame, UL_info->slot); } } } @@ -276,7 +276,7 @@ void NR_UL_indication(NR_UL_IND_t *UL_info) gNB_MAC_INST *mac = RC.nrmac[module_id]; LOG_D(PHY,"SFN/SF:%d%d module_id:%d CC_id:%d UL_info[rx_ind:%d harqs:%d crcs:%d cqis:%d preambles:%d sr_ind:%d]\n", - UL_info->frame,UL_info->subframe, + UL_info->frame,UL_info->slot, module_id,CC_id, UL_info->rx_ind.rx_indication_body.number_of_pdus, UL_info->harq_ind.harq_indication_body.number_of_harqs, UL_info->crc_ind.crc_indication_body.number_of_crcs, UL_info->cqi_ind.number_of_cqis, UL_info->rach_ind.rach_indication_body.number_of_preambles, UL_info->sr_ind.sr_indication_body.number_of_srs); @@ -284,18 +284,18 @@ void NR_UL_indication(NR_UL_IND_t *UL_info) { if (ifi->CC_mask==0) { ifi->current_frame = UL_info->frame; - ifi->current_subframe = UL_info->subframe; + ifi->current_slot = UL_info->slot; } else { AssertFatal(UL_info->frame != ifi->current_frame,"CC_mask %x is not full and frame has changed\n",ifi->CC_mask); - AssertFatal(UL_info->subframe != ifi->current_subframe,"CC_mask %x is not full and subframe has changed\n",ifi->CC_mask); + AssertFatal(UL_info->slot != ifi->current_slot,"CC_mask %x is not full and slot has changed\n",ifi->CC_mask); } ifi->CC_mask |= (1<frame,UL_info->subframe); + clear_nr_nfapi_information(mac,CC_id,UL_info->frame,UL_info->slot); handle_nr_rach(UL_info); @@ -315,24 +315,26 @@ void NR_UL_indication(NR_UL_IND_t *UL_info) if (ifi->CC_mask == ((1<frame+((UL_info->subframe>(9-sf_ahead))?1:0)) % 1024, - (UL_info->subframe+sf_ahead)%10); + (UL_info->frame+((UL_info->slot>(9-sf_ahead))?1:0)) % 1024, + (UL_info->slot+sf_ahead)%10); */ - + nfapi_nr_config_request_t *cfg = &mac->config[CC_id]; + int spf = get_spf(cfg); + gNB_dlsch_ulsch_scheduler(module_id, - (UL_info->frame+((UL_info->subframe>(9-sf_ahead))?1:0)) % 1024, - (UL_info->subframe+sf_ahead)%10); + (UL_info->frame+((UL_info->slot>(spf-1-sf_ahead))?1:0)) % 1024, + (UL_info->slot+sf_ahead)%spf); ifi->CC_mask = 0; sched_info->module_id = module_id; sched_info->CC_id = CC_id; - sched_info->frame = (UL_info->frame + ((UL_info->subframe>(9-sf_ahead)) ? 1 : 0)) % 1024; - sched_info->subframe = (UL_info->subframe+sf_ahead)%10; + sched_info->frame = (UL_info->frame + ((UL_info->slot>(spf-1-sf_ahead)) ? 1 : 0)) % 1024; + sched_info->slot = (UL_info->slot+sf_ahead)%spf; sched_info->DL_req = &mac->DL_req[CC_id]; sched_info->HI_DCI0_req = &mac->HI_DCI0_req[CC_id]; if ((mac->common_channels[CC_id].tdd_Config==NULL) || - (is_nr_UL_sf(&mac->common_channels[CC_id],(sched_info->subframe+sf_ahead)%10)>0)) + (is_nr_UL_slot(&mac->common_channels[CC_id],(sched_info->slot+sf_ahead)%spf)>0)) sched_info->UL_req = &mac->UL_req[CC_id]; else sched_info->UL_req = NULL; @@ -352,7 +354,7 @@ void NR_UL_indication(NR_UL_IND_t *UL_info) ifi->NR_Schedule_response(sched_info); } - LOG_D(PHY,"NR_Schedule_response: SFN_SF:%d%d dl_pdus:%d\n",sched_info->frame,sched_info->subframe,sched_info->DL_req->dl_config_request_body.number_pdu); + LOG_D(PHY,"NR_Schedule_response: SFN_SF:%d%d dl_pdus:%d\n",sched_info->frame,sched_info->slot,sched_info->DL_req->dl_config_request_body.number_pdu); } } } @@ -374,4 +376,4 @@ NR_IF_Module_t *NR_IF_Module_init(int Mod_id){ "allocation of if_inst[%d]->if_mutex fails\n",Mod_id); } return if_inst[Mod_id]; -} \ No newline at end of file +} diff --git a/openair2/NR_PHY_INTERFACE/NR_IF_Module.h b/openair2/NR_PHY_INTERFACE/NR_IF_Module.h index 7849c034fe..63cbd68f22 100644 --- a/openair2/NR_PHY_INTERFACE/NR_IF_Module.h +++ b/openair2/NR_PHY_INTERFACE/NR_IF_Module.h @@ -57,8 +57,8 @@ typedef struct{ int CC_id; /// frame frame_t frame; - /// subframe - sub_frame_t subframe; + /// slot + slot_t slot; /// harq indication list nfapi_harq_indication_t harq_ind; @@ -88,7 +88,7 @@ typedef struct{ } NR_UL_IND_t; -// Downlink subframe P7 +// Downlink slot P7 typedef struct{ @@ -98,8 +98,8 @@ typedef struct{ uint8_t CC_id; /// frame frame_t frame; - /// subframe - sub_frame_t subframe; + /// slot + slot_t slot; /// nFAPI DL Config Request nfapi_nr_dl_config_request_t *DL_req; /// nFAPI UL Config Request @@ -123,7 +123,7 @@ typedef struct NR_IF_Module_s{ void (*NR_PHY_config_req)(NR_PHY_Config_t* config_INFO); uint32_t CC_mask; uint16_t current_frame; - uint8_t current_subframe; + uint8_t current_slot; pthread_mutex_t if_mutex; } NR_IF_Module_t; @@ -137,4 +137,4 @@ void NR_UL_indication(NR_UL_IND_t *UL_INFO); /*Interface for Downlink, transmitting the DLSCH SDU, DCI SDU*/ void NR_Schedule_Response(NR_Sched_Rsp_t *Sched_INFO); -#endif /*_NFAPI_INTERFACE_NR_H_*/ \ No newline at end of file +#endif /*_NFAPI_INTERFACE_NR_H_*/ diff --git a/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c b/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c index 282a15ca47..4e8b72614d 100644 --- a/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c +++ b/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c @@ -33,6 +33,7 @@ #include "NR_IF_Module.h" #include "mac_proto.h" #include "assertions.h" +#include "LAYER2/NR_MAC_UE/mac_extern.h" #include @@ -41,217 +42,270 @@ static nr_ue_if_module_t *nr_ue_if_module_inst[MAX_IF_MODULES]; // L2 Abstraction Layer -int8_t handle_bcch_bch(module_id_t module_id, int cc_id, uint8_t gNB_index, uint8_t *pduP, uint8_t additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id){ +int handle_bcch_bch(module_id_t module_id, int cc_id, unsigned int gNB_index, uint8_t *pduP, unsigned int additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id){ - return nr_ue_decode_mib( module_id, - cc_id, - gNB_index, - additional_bits, - ssb_length, // Lssb = 64 is not support - ssb_index, - pduP, - cell_id); + return nr_ue_decode_mib( module_id, + cc_id, + gNB_index, + additional_bits, + ssb_length, // Lssb = 64 is not support + ssb_index, + pduP, + cell_id); } // L2 Abstraction Layer -int8_t handle_bcch_dlsch(module_id_t module_id, int cc_id, uint8_t gNB_index, uint32_t sibs_mask, uint8_t *pduP, uint32_t pdu_len){ +int handle_bcch_dlsch(module_id_t module_id, int cc_id, unsigned int gNB_index, uint32_t sibs_mask, uint8_t *pduP, uint32_t pdu_len){ - return 0; + return 0; } // L2 Abstraction Layer -int8_t handle_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fapi_nr_dci_pdu_rel15_t *dci, uint16_t rnti, uint32_t dci_type){ +int handle_dci(module_id_t module_id, int cc_id, unsigned int gNB_index, fapi_nr_dci_pdu_rel15_t *dci, uint16_t rnti, uint32_t dci_type){ - return nr_ue_process_dci(module_id, cc_id, gNB_index, dci, rnti, dci_type); + printf("handle_dci: rnti %x,dci_type %d\n",rnti,dci_type); + return nr_ue_process_dci(module_id, cc_id, gNB_index, dci, rnti, dci_type); } // L2 Abstraction Layer int8_t handle_dlsch (module_id_t module_id, int cc_id, uint8_t gNB_index, fapi_nr_dci_indication_t *dci_ind, uint8_t *pduP, uint32_t pdu_len){ -// return 0; - return nr_ue_process_dlsch( module_id, - cc_id, - gNB_index, - &dci_ind, - pduP, - pdu_len); + // return 0; + return nr_ue_process_dlsch( module_id, + cc_id, + gNB_index, + &dci_ind, + pduP, + pdu_len); } -int8_t nr_ue_ul_indication(nr_uplink_indication_t *ul_info){ +int nr_ue_ul_indication(nr_uplink_indication_t *ul_info){ - NR_UE_L2_STATE_t ret; - module_id_t module_id = ul_info->module_id; - NR_UE_MAC_INST_t *mac = get_mac_inst(module_id); + NR_UE_L2_STATE_t ret; + module_id_t module_id = ul_info->module_id; + NR_UE_MAC_INST_t *mac = get_mac_inst(module_id); - // clean previous FAPI messages - mac->tx_request.number_of_pdus = 0; - mac->ul_config_request.number_pdus = 0; - mac->dl_config_request.number_pdus = 0; - // clean previous FAPI messages + // clean previous FAPI messages + mac->tx_request.number_of_pdus = 0; + mac->ul_config_request.number_pdus = 0; + mac->dl_config_request.number_pdus = 0; + // clean previous FAPI messages - ret = nr_ue_scheduler( - ul_info->module_id, - ul_info->gNB_index, - ul_info->cc_id, - ul_info->frame, - ul_info->slot, - ul_info->ssb_index, - 0, 0); // TODO check tx/rx frame/slot is need for NR version + ret = nr_ue_scheduler( + ul_info->module_id, + ul_info->gNB_index, + ul_info->cc_id, + ul_info->frame, + ul_info->slot, + ul_info->ssb_index, + 0, 0); // TODO check tx/rx frame/slot is need for NR version - switch(ret){ - case CONNECTION_OK: - break; - case CONNECTION_LOST: - break; - case PHY_RESYNCH: - break; - case PHY_HO_PRACH: - break; - default: - break; - } + switch(ret){ + case CONNECTION_OK: + break; + case CONNECTION_LOST: + break; + case PHY_RESYNCH: + break; + case PHY_HO_PRACH: + break; + default: + break; + } - mac->if_module->scheduled_response(&mac->scheduled_response); + mac->if_module->scheduled_response(&mac->scheduled_response); - return 0; + return 0; } -int8_t nr_ue_dl_indication(nr_downlink_indication_t *dl_info){ +int nr_ue_dl_indication(nr_downlink_indication_t *dl_info){ - int32_t i; - uint32_t ret_mask = 0x0; - module_id_t module_id = dl_info->module_id; - NR_UE_MAC_INST_t *mac = get_mac_inst(module_id); + int32_t i; + uint32_t ret_mask = 0x0; + module_id_t module_id = dl_info->module_id; + NR_UE_MAC_INST_t *mac = get_mac_inst(module_id); - // clean up scheduled_response structure + // clean up scheduled_response structure - if(dl_info->dci_ind != NULL){ - printf("[L2][IF MODULE][DL INDICATION][DCI_IND]\n"); - for(i=0; idci_ind->number_of_dcis; ++i){ - //printf(">>>NR_IF_Module i=%d, dl_info->dci_ind->number_of_dcis=%d\n",i,dl_info->dci_ind->number_of_dcis); - fapi_nr_dci_pdu_rel15_t *dci = &dl_info->dci_ind->dci_list[i].dci; + if(dl_info->dci_ind != NULL){ + printf("[L2][IF MODULE][DL INDICATION][DCI_IND]\n"); + for(i=0; idci_ind->number_of_dcis; ++i){ + printf(">>>NR_IF_Module i=%d, dl_info->dci_ind->number_of_dcis=%d\n",i,dl_info->dci_ind->number_of_dcis); + fapi_nr_dci_pdu_rel15_t *dci = &dl_info->dci_ind->dci_list[i].dci; + /* + ret_mask |= (handle_dci( + dl_info->module_id, + dl_info->cc_id, + dl_info->gNB_index, + dci, + (dl_info->dci_ind->dci_list+i)->rnti, + (dl_info->dci_ind->dci_list+i)->dci_format)) << FAPI_NR_DCI_IND; + */ - ret_mask |= (handle_dci( - dl_info->module_id, - dl_info->cc_id, - dl_info->gNB_index, - dci, - (dl_info->dci_ind->dci_list+i)->rnti, - (dl_info->dci_ind->dci_list+i)->dci_format)) << FAPI_NR_DCI_IND; - - - /*switch((dl_info->dci_ind->dci_list+i)->dci_type){ - case FAPI_NR_DCI_TYPE_0_0: - case FAPI_NR_DCI_TYPE_0_1: - case FAPI_NR_DCI_TYPE_1_1: - case FAPI_NR_DCI_TYPE_2_0: - case FAPI_NR_DCI_TYPE_2_1: - case FAPI_NR_DCI_TYPE_2_2: - case FAPI_NR_DCI_TYPE_2_3: - AssertFatal(1==0, "Not yet support at this moment!\n"); - break; + /*switch((dl_info->dci_ind->dci_list+i)->dci_type){ + case FAPI_NR_DCI_TYPE_0_0: + case FAPI_NR_DCI_TYPE_0_1: + case FAPI_NR_DCI_TYPE_1_1: + case FAPI_NR_DCI_TYPE_2_0: + case FAPI_NR_DCI_TYPE_2_1: + case FAPI_NR_DCI_TYPE_2_2: + case FAPI_NR_DCI_TYPE_2_3: + AssertFatal(1==0, "Not yet support at this moment!\n"); + break; - case FAPI_NR_DCI_TYPE_1_0: + case FAPI_NR_DCI_TYPE_1_0: - dl_config->dl_config_list[dl_config->number_pdus].pdu_type = FAPI_NR_DL_CONFIG_TYPE_DLSCH; + dl_config->dl_config_list[dl_config->number_pdus].pdu_type = FAPI_NR_DL_CONFIG_TYPE_DLSCH; - // mapping into DL_CONFIG_REQ for DL-SCH - fapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_config_pdu = &dl_config->dl_config_list[dl_config->number_pdus].dlsch_config_pdu.dlsch_config_rel15; + // mapping into DL_CONFIG_REQ for DL-SCH + fapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_config_pdu = &dl_config->dl_config_list[dl_config->number_pdus].dlsch_config_pdu.dlsch_config_rel15; - dl_config->dl_config_list[dl_config->number_pdus].dlsch_config_pdu.rnti = 0x0000; // TX RNTI: UE-spec - memcpy(dlsch_config_pdu, dci, sizeof(fapi_nr_dci_pdu_rel15_t)); + dl_config->dl_config_list[dl_config->number_pdus].dlsch_config_pdu.rnti = 0x0000; // TX RNTI: UE-spec + memcpy(dlsch_config_pdu, dci, sizeof(fapi_nr_dci_pdu_rel15_t)); - dl_config->number_pdus = dl_config->number_pdus + 1; + dl_config->number_pdus = dl_config->number_pdus + 1; - ret_mask |= (handle_dci( - dl_info->module_id, - dl_info->cc_id, - dl_info->gNB_index, - dci, - (dl_info->dci_ind->dci_list+i)->rnti, - (dl_info->dci_ind->dci_list+i)->dci_type)) << FAPI_NR_DCI_IND; + ret_mask |= (handle_dci( + dl_info->module_id, + dl_info->cc_id, + dl_info->gNB_index, + dci, + (dl_info->dci_ind->dci_list+i)->rnti, + (dl_info->dci_ind->dci_list+i)->dci_type)) << FAPI_NR_DCI_IND; - break; + break; - default: - break; - }*/ + default: + break; + }*/ - //(dl_info->dci_list+i)->rnti + //(dl_info->dci_list+i)->rnti - } } + } - if(dl_info->rx_ind != NULL){ - printf("[L2][IF MODULE][DL INDICATION][RX_IND]\n"); - for(i=0; irx_ind->number_pdus; ++i){ - switch(dl_info->rx_ind->rx_indication_body[i].pdu_type){ - case FAPI_NR_RX_PDU_TYPE_MIB: - ret_mask |= (handle_bcch_bch(dl_info->module_id, dl_info->cc_id, dl_info->gNB_index, - (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.pdu, - (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.additional_bits, - (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.ssb_index, - (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.ssb_length, - (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.cell_id )) << FAPI_NR_RX_PDU_TYPE_MIB; - break; - case FAPI_NR_RX_PDU_TYPE_SIB: - ret_mask |= (handle_bcch_dlsch(dl_info->module_id, dl_info->cc_id, dl_info->gNB_index, - (dl_info->rx_ind->rx_indication_body+i)->sib_pdu.sibs_mask, - (dl_info->rx_ind->rx_indication_body+i)->sib_pdu.pdu, - (dl_info->rx_ind->rx_indication_body+i)->sib_pdu.pdu_length )) << FAPI_NR_RX_PDU_TYPE_SIB; - break; - case FAPI_NR_RX_PDU_TYPE_DLSCH: -// ret_mask |= (0) << FAPI_NR_RX_PDU_TYPE_DLSCH; - ret_mask |= (handle_dlsch(dl_info->module_id, dl_info->cc_id, dl_info->gNB_index, &dl_info->dci_ind, - (dl_info->rx_ind->rx_indication_body+i)->pdsch_pdu.pdu, - (dl_info->rx_ind->rx_indication_body+i)->pdsch_pdu.pdu_length)) << FAPI_NR_RX_PDU_TYPE_DLSCH; + if(dl_info->rx_ind != NULL){ + printf("[L2][IF MODULE][DL INDICATION][RX_IND]\n"); + for(i=0; irx_ind->number_pdus; ++i){ + switch(dl_info->rx_ind->rx_indication_body[i].pdu_type){ + case FAPI_NR_RX_PDU_TYPE_MIB: + ret_mask |= (handle_bcch_bch(dl_info->module_id, dl_info->cc_id, dl_info->gNB_index, + (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.pdu, + (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.additional_bits, + (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.ssb_index, + (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.ssb_length, + (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.cell_id )) << FAPI_NR_RX_PDU_TYPE_MIB; + break; + case FAPI_NR_RX_PDU_TYPE_SIB: + ret_mask |= (handle_bcch_dlsch(dl_info->module_id, dl_info->cc_id, dl_info->gNB_index, + (dl_info->rx_ind->rx_indication_body+i)->sib_pdu.sibs_mask, + (dl_info->rx_ind->rx_indication_body+i)->sib_pdu.pdu, + (dl_info->rx_ind->rx_indication_body+i)->sib_pdu.pdu_length )) << FAPI_NR_RX_PDU_TYPE_SIB; + break; + case FAPI_NR_RX_PDU_TYPE_DLSCH: + // ret_mask |= (0) << FAPI_NR_RX_PDU_TYPE_DLSCH; + ret_mask |= (handle_dlsch(dl_info->module_id, dl_info->cc_id, dl_info->gNB_index, dl_info->dci_ind, + (dl_info->rx_ind->rx_indication_body+i)->pdsch_pdu.pdu, + (dl_info->rx_ind->rx_indication_body+i)->pdsch_pdu.pdu_length)) << FAPI_NR_RX_PDU_TYPE_DLSCH; - break; - default: + break; + default: - break; + break; - } - } + } } + } + //clean up nr_downlink_indication_t *dl_info + dl_info->rx_ind = NULL; + dl_info->dci_ind = NULL; - AssertFatal( nr_ue_if_module_inst[module_id] != NULL, "IF module is void!\n" ); - nr_ue_if_module_inst[module_id]->scheduled_response(&mac->scheduled_response); + AssertFatal( nr_ue_if_module_inst[module_id] != NULL, "IF module is void!\n" ); + nr_ue_if_module_inst[module_id]->scheduled_response(&mac->scheduled_response); - return 0; + return 0; } nr_ue_if_module_t *nr_ue_if_module_init(uint32_t module_id){ - if (nr_ue_if_module_inst[module_id] == NULL) { - nr_ue_if_module_inst[module_id] = (nr_ue_if_module_t *)malloc(sizeof(nr_ue_if_module_t)); - memset((void*)nr_ue_if_module_inst[module_id],0,sizeof(nr_ue_if_module_t)); + if (nr_ue_if_module_inst[module_id] == NULL) { + nr_ue_if_module_inst[module_id] = (nr_ue_if_module_t *)malloc(sizeof(nr_ue_if_module_t)); + memset((void*)nr_ue_if_module_inst[module_id],0,sizeof(nr_ue_if_module_t)); - nr_ue_if_module_inst[module_id]->cc_mask=0; - nr_ue_if_module_inst[module_id]->current_frame = 0; - nr_ue_if_module_inst[module_id]->current_slot = 0; - nr_ue_if_module_inst[module_id]->phy_config_request = NULL; - nr_ue_if_module_inst[module_id]->scheduled_response = NULL; - nr_ue_if_module_inst[module_id]->dl_indication = nr_ue_dl_indication; - nr_ue_if_module_inst[module_id]->ul_indication = nr_ue_ul_indication; - } + nr_ue_if_module_inst[module_id]->cc_mask=0; + nr_ue_if_module_inst[module_id]->current_frame = 0; + nr_ue_if_module_inst[module_id]->current_slot = 0; + nr_ue_if_module_inst[module_id]->phy_config_request = NULL; + nr_ue_if_module_inst[module_id]->scheduled_response = NULL; + nr_ue_if_module_inst[module_id]->dl_indication = nr_ue_dl_indication; + nr_ue_if_module_inst[module_id]->ul_indication = nr_ue_ul_indication; + } - return nr_ue_if_module_inst[module_id]; + return nr_ue_if_module_inst[module_id]; } -int8_t nr_ue_if_module_kill(uint32_t module_id) { +int nr_ue_if_module_kill(uint32_t module_id) { - if (nr_ue_if_module_inst[module_id] != NULL){ - free(nr_ue_if_module_inst[module_id]); - } - return 0; + if (nr_ue_if_module_inst[module_id] != NULL){ + free(nr_ue_if_module_inst[module_id]); + } + return 0; +} + +int nr_ue_dcireq(nr_dcireq_t *dcireq) { + + fapi_nr_dl_config_request_t *dl_config=&dcireq->dl_config_req; + + // Type0 PDCCH search space + dl_config->number_pdus = 1; + dl_config->dl_config_list[0].pdu_type = FAPI_NR_DL_CONFIG_TYPE_DCI; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.rnti = 0x1234; // to be set + + uint64_t mask = 0x0; + uint16_t num_rbs=24; + uint16_t rb_offset=0; + uint16_t cell_id=0; + uint16_t num_symbols=2; + for(int i=0; i<(num_rbs/6); ++i){ // 38.331 Each bit corresponds a group of 6 RBs + mask = mask >> 1; + mask = mask | 0x100000000000; + } + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.frequency_domain_resource = mask; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.rb_offset = rb_offset; // additional parameter other than coreset + + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.duration = num_symbols; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.cce_reg_mapping_type =CCE_REG_MAPPING_TYPE_INTERLEAVED; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.cce_reg_interleaved_reg_bundle_size = 6; // L 38.211 7.3.2.2 + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.cce_reg_interleaved_interleaver_size = 2; // R 38.211 7.3.2.2 + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.cce_reg_interleaved_shift_index = cell_id; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.precoder_granularity = PRECODER_GRANULARITY_SAME_AS_REG_BUNDLE; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.coreset.pdcch_dmrs_scrambling_id = cell_id; + + uint32_t number_of_search_space_per_slot=1; + uint32_t first_symbol_index=0; + uint32_t search_space_duration=0; // element of search space + uint32_t coreset_duration; // element of coreset + + coreset_duration = num_symbols * number_of_search_space_per_slot; + + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.number_of_candidates[0] = table_38213_10_1_1_c2[0]; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.number_of_candidates[1] = table_38213_10_1_1_c2[1]; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.number_of_candidates[2] = table_38213_10_1_1_c2[2]; // CCE aggregation level = 4 + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.number_of_candidates[3] = table_38213_10_1_1_c2[3]; // CCE aggregation level = 8 + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.number_of_candidates[4] = table_38213_10_1_1_c2[4]; // CCE aggregation level = 16 + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.duration = search_space_duration; + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.monitoring_symbols_within_slot = (0x3fff << first_symbol_index) & (0x3fff >> (14-coreset_duration-first_symbol_index)) & 0x3fff; + + dl_config->dl_config_list[0].dci_config_pdu.dci_config_rel15.N_RB_BWP = 106; + + } diff --git a/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.h b/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.h index a9b385d878..0ecba9736d 100755 --- a/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.h +++ b/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.h @@ -38,6 +38,21 @@ +typedef struct { + /// module id + module_id_t module_id; + /// gNB index + uint32_t gNB_index; + /// component carrier id + int cc_id; + /// frame + frame_t frame; + /// slot + int slot; + + fapi_nr_dl_config_request_t dl_config_req; +} nr_dcireq_t; + typedef struct { /// module id module_id_t module_id; @@ -48,7 +63,7 @@ typedef struct { /// frame frame_t frame; /// slot - uint8_t slot; + int slot; /// NR UE FAPI-like P7 message, direction: L1 to L2 /// data reception indication structure @@ -72,7 +87,7 @@ typedef struct { /// slot uint32_t slot; /// ssb_index, if ssb is not present in current TTI, thie value set to -1 - int8_t ssb_index; + int ssb_index; } nr_uplink_indication_t; // Downlink subframe P7 @@ -82,13 +97,13 @@ typedef struct { /// module id module_id_t module_id; /// component carrier id - uint8_t CC_id; + int CC_id; /// frame frame_t frame; /// subframe sub_frame_t subframe; /// slot - uint8_t slot; + int slot; /// NR UE FAPI-like P7 message, direction: L2 to L1 /// downlink transmission configuration request structure @@ -142,7 +157,7 @@ typedef int8_t (nr_ue_phy_config_request_f)(nr_phy_config_t *phy_config); * -1: Failed to consume bytes. Abort the mission. * Non-negative return values indicate success, and ignored. */ -typedef int8_t (nr_ue_dl_indication_f)(nr_downlink_indication_t *dl_info); +typedef int (nr_ue_dl_indication_f)(nr_downlink_indication_t *dl_info); /* * Generic type of an application-defined callback to return various @@ -151,19 +166,21 @@ typedef int8_t (nr_ue_dl_indication_f)(nr_downlink_indication_t *dl_info); * -1: Failed to consume bytes. Abort the mission. * Non-negative return values indicate success, and ignored. */ -typedef int8_t (nr_ue_ul_indication_f)(nr_uplink_indication_t *ul_info); +typedef int (nr_ue_ul_indication_f)(nr_uplink_indication_t *ul_info); + +typedef int (nr_ue_dcireq_f)(nr_dcireq_t *ul_info); // TODO check this stuff can be reuse of need modification typedef struct nr_ue_if_module_s { - nr_ue_scheduled_response_f *scheduled_response; - nr_ue_phy_config_request_f *phy_config_request; - nr_ue_dl_indication_f *dl_indication; - nr_ue_ul_indication_f *ul_indication; - - uint32_t cc_mask; - uint32_t current_frame; - uint32_t current_slot; - //pthread_mutex_t nr_if_mutex; + nr_ue_scheduled_response_f *scheduled_response; + nr_ue_phy_config_request_f *phy_config_request; + nr_ue_dl_indication_f *dl_indication; + nr_ue_ul_indication_f *ul_indication; + //nr_ue_dcireq_f *dcireq; + uint32_t cc_mask; + uint32_t current_frame; + uint32_t current_slot; + //pthread_mutex_t nr_if_mutex; } nr_ue_if_module_t; @@ -174,14 +191,16 @@ nr_ue_if_module_t *nr_ue_if_module_init(uint32_t module_id); /**\brief done free of memory allocation by module_id and release to pointer pool. \param module_id module id*/ -int8_t nr_ue_if_module_kill(uint32_t module_id); +int nr_ue_if_module_kill(uint32_t module_id); /**\brief interface between L1/L2, indicating the downlink related information, like dci_ind and rx_req \param dl_info including dci_ind and rx_request messages*/ -int8_t nr_ue_dl_indication(nr_downlink_indication_t *dl_info); +int nr_ue_dl_indication(nr_downlink_indication_t *dl_info); -int8_t nr_ue_ul_indication(nr_uplink_indication_t *ul_info); +int nr_ue_ul_indication(nr_uplink_indication_t *ul_info); + +int nr_ue_dcireq(nr_dcireq_t *dcireq); // TODO check /**\brief handle BCCH-BCH message from dl_indication @@ -190,15 +209,15 @@ int8_t nr_ue_ul_indication(nr_uplink_indication_t *ul_info); \param ssb_index SSB index within 0 - (L_ssb-1) corresponding to 38.331 ch.13 parameter i \param ssb_length corresponding to L1 parameter L_ssb \param cell_id cell id */ -int8_t handle_bcch_bch(module_id_t module_id, int cc_id, uint8_t gNB_index, uint8_t *pduP, uint8_t additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id); +int handle_bcch_bch(module_id_t module_id, int cc_id, unsigned int gNB_index, uint8_t *pduP, unsigned int additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id); // TODO check /**\brief handle BCCH-DL-SCH message from dl_indication \param pdu_len length(bytes) of pdu \param pduP pointer to pdu*/ -int8_t handle_bcch_dlsch(module_id_t module_id, int cc_id, uint8_t gNB_index, uint32_t sibs_mask, uint8_t *pduP, uint32_t pdu_len); +int handle_bcch_dlsch(module_id_t module_id, int cc_id, unsigned int gNB_index, uint32_t sibs_mask, uint8_t *pduP, uint32_t pdu_len); -int8_t handle_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fapi_nr_dci_pdu_rel15_t *dci, uint16_t rnti, uint32_t dci_type); +int handle_dci(module_id_t module_id, int cc_id, unsigned int gNB_index, fapi_nr_dci_pdu_rel15_t *dci, uint16_t rnti, uint32_t dci_type); #endif diff --git a/openair2/RRC/LTE/MESSAGES/asn1_msg.c b/openair2/RRC/LTE/MESSAGES/asn1_msg.c index a39746bc2c..c8e1e31571 100644 --- a/openair2/RRC/LTE/MESSAGES/asn1_msg.c +++ b/openair2/RRC/LTE/MESSAGES/asn1_msg.c @@ -39,7 +39,7 @@ #include #include /* for _ASN_DEFAULT_STACK_MAX */ #include - +//#include "COMMON/messages_types.h" #include "assertions.h" #include "RRCConnectionRequest.h" #include "UL-CCCH-Message.h" diff --git a/openair2/RRC/NR/MESSAGES/asn1_msg.c b/openair2/RRC/NR/MESSAGES/asn1_msg.c index ca7206744c..df45ae5e05 100644 --- a/openair2/RRC/NR/MESSAGES/asn1_msg.c +++ b/openair2/RRC/NR/MESSAGES/asn1_msg.c @@ -42,7 +42,6 @@ #include "asn1_msg.h" #include "RRC/NR/nr_rrc_extern.h" - #if defined(NR_Rel15) #include "NR_SCS-SpecificCarrier.h" #include "NR_TDD-UL-DL-ConfigCommon.h" diff --git a/openair2/RRC/NR/nr_rrc_defs.h b/openair2/RRC/NR/nr_rrc_defs.h index 3c72250fa9..2290a409e1 100644 --- a/openair2/RRC/NR/nr_rrc_defs.h +++ b/openair2/RRC/NR/nr_rrc_defs.h @@ -385,8 +385,8 @@ typedef struct { int Ncp_UL; //cyclic prefix for UL int p_gNB; //number of tx antenna port int p_rx_gNB; //number of receiving antenna ports - uint32_t dl_CarrierFreq; //detected by the UE - uint32_t ul_CarrierFreq; //detected by the UE + uint64_t dl_CarrierFreq; //detected by the UE + uint64_t ul_CarrierFreq; //detected by the UE //are the only static one (memory has been already allocated) NR_BCCH_BCH_Message_t mib; diff --git a/openair2/RRC/NR/rrc_gNB.c b/openair2/RRC/NR/rrc_gNB.c index 7a9c16f699..10fc6b6a4e 100644 --- a/openair2/RRC/NR/rrc_gNB.c +++ b/openair2/RRC/NR/rrc_gNB.c @@ -265,7 +265,7 @@ static void init_NR_SI(const protocol_ctxt_t* const ctxt_pP, rrc_mac_config_req_gNB(ctxt_pP->module_id, CC_id, RC.nrrrc[ctxt_pP->module_id]->carrier[CC_id].p_gNB, - configuration->eutra_band[CC_id], + configuration->nr_band[CC_id], RC.nrrrc[ctxt_pP->module_id]->carrier[CC_id].dl_CarrierFreq, configuration->N_RB_DL[CC_id], (NR_BCCH_BCH_Message_t *)&RC.nrrrc[ctxt_pP->module_id]->carrier[CC_id].mib, diff --git a/targets/ARCH/ADRV9371_ZC706/SYRTEM_IQ_QuickStart.readme b/targets/ARCH/ADRV9371_ZC706/SYRTEM_IQ_QuickStart.readme index d7f6585a75..cd74fdb397 100755 --- a/targets/ARCH/ADRV9371_ZC706/SYRTEM_IQ_QuickStart.readme +++ b/targets/ARCH/ADRV9371_ZC706/SYRTEM_IQ_QuickStart.readme @@ -115,8 +115,16 @@ sudo cset shield --reset cd eur_oai_develop-nr/openairinterface5g/ source oaienv -sudo -E ./cmake_targets/build_oai -c --gNB -w ADRV9371_ZC706 -I -sudo -E ./cmake_targets/build_oai -c --gNB -w ADRV9371_ZC706 +sudo -E ./cmake_targets/build_oai -c --gNB -w ADRV9371_ZC706 -I --enable-cpu-affinity + +/* CPU available shall be mapped on CPU2 to get_nprocs() number with hyperthreading disabled */ +/* because : */ +/* - cset shield leave CPU0 for system */ +/* - CPU1 is reserved for HwRxTx thread for I/Q acquisition ! */ +/* Note: "/usr/local/etc/syriq/cpu-irq.sh" shall be changed if more than 8 cores are used */ +/* ie. "cset shield --force --kthread on -c 1-7" where "-c 1-7" reserves CPU1 to CPU7 for application (CPU0 is for system) */ + +sudo -E ./cmake_targets/build_oai -c --gNB -w ADRV9371_ZC706 --enable-cpu-affinity diff --git a/targets/ARCH/COMMON/common_lib.h b/targets/ARCH/COMMON/common_lib.h index ae5f4d91d9..685a628b26 100644 --- a/targets/ARCH/COMMON/common_lib.h +++ b/targets/ARCH/COMMON/common_lib.h @@ -204,6 +204,7 @@ typedef struct { //! clock source clock_source_t clock_source; //! Manual SDR IP address +//#if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR) char *sdr_addrs; //! Auto calibration flag int autocal[4]; diff --git a/targets/COMMON/threads_t.h b/targets/COMMON/threads_t.h index 99fb0d8cf4..0671f40988 100644 --- a/targets/COMMON/threads_t.h +++ b/targets/COMMON/threads_t.h @@ -2,19 +2,20 @@ #define _THREADS_T_H_ typedef struct threads_s { - int iq; - int one; - int two; - int three; - int slot1_proc_one; - int slot1_proc_two; - int slot1_proc_three; - int dlsch_td_one; - int dlsch_td_two; - int dlsch_td_three; - int dlsch_td1_one; - int dlsch_td1_two; - int dlsch_td1_three; + int main; + int sync; + int one; + int two; + int three; + int slot1_proc_one; + int slot1_proc_two; + int slot1_proc_three; + //int dlsch_td_one; + //int dlsch_td_two; + //int dlsch_td_three; + //int dlsch_td1_one; + //int dlsch_td1_two; + //int dlsch_td1_three; } threads_t; #endif /* _THREADS_T_H_ */ diff --git a/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpn300.conf b/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpn300.conf index b197ea7576..3aaf0c5506 100644 --- a/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpn300.conf +++ b/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpn300.conf @@ -262,9 +262,9 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_SINGLE_THREAD"; + parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" - worker_config = "WORKER_DISABLE"; + worker_config = "WORKER_ENABLE"; } ); diff --git a/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpx300.conf b/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpx300.conf index 721c307f71..d62b06aaa5 100644 --- a/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpx300.conf +++ b/targets/PROJECTS/GENERIC-LTE-EPC/CONF/gnb.band78.tm1.106PRB.usrpx300.conf @@ -262,7 +262,7 @@ RUs = ( THREAD_STRUCT = ( { #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" - parallel_config = "PARALLEL_RU_L1_TRX_SPLIT"; + parallel_config = "PARALLEL_SINGLE_THREAD"; #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" worker_config = "WORKER_DISABLE"; } diff --git a/targets/RT/USER/gNB_usrp.gtkw b/targets/RT/USER/gNB_usrp.gtkw index f7ca2c291a..b40f946bc1 100644 --- a/targets/RT/USER/gNB_usrp.gtkw +++ b/targets/RT/USER/gNB_usrp.gtkw @@ -24,9 +24,9 @@ variables.trx_tst[63:0] functions.eNB_thread_rxtx0 @24 variables.frame_number_RX0_RU[63:0] -variables.subframe_number_RX0_RU[63:0] +variables.tti_number_RX0_RU[63:0] variables.frame_number_TX0_RU[63:0] -variables.subframe_number_TX0_RU[63:0] +variables.tti_number_TX0_RU[63:0] @28 functions.mac_schedule_dlsch functions.macxface_eNB_dlsch_ulsch_scheduler @@ -34,16 +34,16 @@ functions.macxface_ue_scheduler functions.phy_eNB_ofdm_mod_l @24 variables.frame_number_RX0_gNB[63:0] -variables.subframe_number_RX0_gNB[63:0] +variables.slot_number_RX0_gNB[63:0] variables.frame_number_TX0_gNB[63:0] -variables.subframe_number_TX0_gNB[63:0] +variables.slot_number_TX0_gNB[63:0] @28 functions.gNB_thread_rxtx0 @24 variables.frame_number_RX1_gNB[63:0] -variables.subframe_number_RX1_gNB[63:0] +variables.slot_number_RX1_gNB[63:0] variables.frame_number_TX1_gNB[63:0] -variables.subframe_number_TX1_gNB[63:0] +variables.slot_number_TX1_gNB[63:0] @28 functions.gNB_thread_rxtx1 functions.phy_procedures_ru_feprx0 diff --git a/targets/RT/USER/lte-enb.c b/targets/RT/USER/lte-enb.c index 79f531df0c..7145974a4a 100644 --- a/targets/RT/USER/lte-enb.c +++ b/targets/RT/USER/lte-enb.c @@ -453,13 +453,13 @@ void eNB_top(PHY_VARS_eNB *eNB, int frame_rx, int subframe_rx, char *string,RU_t proc_rxtx->timestamp_tx = ru_proc->timestamp_rx + (sf_ahead*fp->samples_per_tti); proc_rxtx->frame_rx = ru_proc->frame_rx; - proc_rxtx->subframe_rx = ru_proc->subframe_rx; + proc_rxtx->subframe_rx = ru_proc->tti_rx; proc_rxtx->frame_tx = (proc_rxtx->subframe_rx > (9-sf_ahead)) ? (proc_rxtx->frame_rx+1)&1023 : proc_rxtx->frame_rx; proc_rxtx->subframe_tx = (proc_rxtx->subframe_rx + sf_ahead)%10; if (rxtx(eNB,proc_rxtx,string) < 0) LOG_E(PHY,"eNB %d CC_id %d failed during execution\n",eNB->Mod_id,eNB->CC_id); ru_proc->timestamp_tx = proc_rxtx->timestamp_tx; - ru_proc->subframe_tx = proc_rxtx->subframe_tx; + ru_proc->tti_tx = proc_rxtx->subframe_tx; ru_proc->frame_tx = proc_rxtx->frame_tx; } } @@ -474,24 +474,24 @@ int wakeup_txfh(eNB_rxtx_proc_t *proc,RU_proc_t *ru_proc) { if(wait_on_condition(&ru_proc->mutex_eNBs,&ru_proc->cond_eNBs,&ru_proc->ru_tx_ready,"wakeup_txfh")<0) { - LOG_E(PHY,"Frame %d, subframe %d: TX FH not ready\n", ru_proc->frame_tx, ru_proc->subframe_tx); + LOG_E(PHY,"Frame %d, subframe %d: TX FH not ready\n", ru_proc->frame_tx, ru_proc->tti_tx); return(-1); } if (release_thread(&ru_proc->mutex_eNBs,&ru_proc->ru_tx_ready,"wakeup_txfh")<0) return(-1); if (ru_proc->instance_cnt_eNBs == 0) { - LOG_E(PHY,"Frame %d, subframe %d: TX FH thread busy, dropping Frame %d, subframe %d\n", ru_proc->frame_tx, ru_proc->subframe_tx, proc->frame_rx, proc->subframe_rx); + LOG_E(PHY,"Frame %d, subframe %d: TX FH thread busy, dropping Frame %d, subframe %d\n", ru_proc->frame_tx, ru_proc->tti_tx, proc->frame_rx, proc->subframe_rx); return(-1); } if (pthread_mutex_timedlock(&ru_proc->mutex_eNBs,&wait) != 0) { - LOG_E( PHY, "[eNB] ERROR pthread_mutex_lock for eNB TX1 thread %d (IC %d)\n", ru_proc->subframe_rx&1,ru_proc->instance_cnt_eNBs ); + LOG_E( PHY, "[eNB] ERROR pthread_mutex_lock for eNB TX1 thread %d (IC %d)\n", ru_proc->tti_rx&1,ru_proc->instance_cnt_eNBs ); exit_fun( "error locking mutex_eNB" ); return(-1); } ++ru_proc->instance_cnt_eNBs; ru_proc->timestamp_tx = proc->timestamp_tx; - ru_proc->subframe_tx = proc->subframe_tx; + ru_proc->tti_tx = proc->subframe_tx; ru_proc->frame_tx = proc->frame_tx; // the thread can now be woken up @@ -622,7 +622,7 @@ int wakeup_rxtx(PHY_VARS_eNB *eNB,RU_t *ru) { // and proc->subframe_tx = proc->subframe_rx+sf_ahead proc_rxtx0->timestamp_tx = ru_proc->timestamp_rx + (sf_ahead*fp->samples_per_tti); proc_rxtx0->frame_rx = ru_proc->frame_rx; - proc_rxtx0->subframe_rx = ru_proc->subframe_rx; + proc_rxtx0->subframe_rx = ru_proc->tti_rx; proc_rxtx0->frame_tx = (proc_rxtx0->subframe_rx > (9-sf_ahead)) ? (proc_rxtx0->frame_rx+1)&1023 : proc_rxtx0->frame_rx; proc_rxtx0->subframe_tx = (proc_rxtx0->subframe_rx + sf_ahead)%10; diff --git a/targets/RT/USER/lte-ru.c b/targets/RT/USER/lte-ru.c index 8ecd0783d5..a35c22f4eb 100644 --- a/targets/RT/USER/lte-ru.c +++ b/targets/RT/USER/lte-ru.c @@ -311,21 +311,21 @@ int connect_rau(RU_t *ru) { // southbound IF5 fronthaul for 16-bit OAI format static inline void fh_if5_south_out(RU_t *ru) { if (ru == RC.ru[0]) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, ru->proc.timestamp_tx&0xffffffff ); - send_IF5(ru, ru->proc.timestamp_tx, ru->proc.subframe_tx, &ru->seqno, IF5_RRH_GW_DL); + send_IF5(ru, ru->proc.timestamp_tx, ru->proc.tti_tx, &ru->seqno, IF5_RRH_GW_DL); } // southbound IF5 fronthaul for Mobipass packet format static inline void fh_if5_mobipass_south_out(RU_t *ru) { if (ru == RC.ru[0]) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, ru->proc.timestamp_tx&0xffffffff ); - send_IF5(ru, ru->proc.timestamp_tx, ru->proc.subframe_tx, &ru->seqno, IF5_MOBIPASS); + send_IF5(ru, ru->proc.timestamp_tx, ru->proc.tti_tx, &ru->seqno, IF5_MOBIPASS); } // southbound IF4p5 fronthaul static inline void fh_if4p5_south_out(RU_t *ru) { if (ru == RC.ru[0]) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, ru->proc.timestamp_tx&0xffffffff ); - LOG_D(PHY,"Sending IF4p5 for frame %d subframe %d\n",ru->proc.frame_tx,ru->proc.subframe_tx); - if (subframe_select(ru->frame_parms,ru->proc.subframe_tx)!=SF_UL) - send_IF4p5(ru,ru->proc.frame_tx, ru->proc.subframe_tx, IF4p5_PDLFFT); + LOG_D(PHY,"Sending IF4p5 for frame %d subframe %d\n",ru->proc.frame_tx,ru->proc.tti_tx); + if (subframe_select(ru->frame_parms,ru->proc.tti_tx)!=SF_UL) + send_IF4p5(ru,ru->proc.frame_tx, ru->proc.tti_tx, IF4p5_PDLFFT); } /*************************************************************/ @@ -340,11 +340,11 @@ void fh_if5_south_in(RU_t *ru,int *frame, int *subframe) { recv_IF5(ru, &proc->timestamp_rx, *subframe, IF5_RRH_GW_UL); proc->frame_rx = (proc->timestamp_rx / (fp->samples_per_tti*10))&1023; - proc->subframe_rx = (proc->timestamp_rx / fp->samples_per_tti)%10; + proc->tti_rx = (proc->timestamp_rx / fp->samples_per_tti)%10; if (proc->first_rx == 0) { - if (proc->subframe_rx != *subframe){ - LOG_E(PHY,"Received Timestamp doesn't correspond to the time we think it is (proc->subframe_rx %d, subframe %d)\n",proc->subframe_rx,*subframe); + if (proc->tti_rx != *subframe){ + LOG_E(PHY,"Received Timestamp doesn't correspond to the time we think it is (proc->tti_rx %d, subframe %d)\n",proc->tti_rx,*subframe); exit_fun("Exiting"); } @@ -355,7 +355,7 @@ void fh_if5_south_in(RU_t *ru,int *frame, int *subframe) { } else { proc->first_rx = 0; *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *subframe = proc->tti_rx; } VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TS, proc->timestamp_rx&0xffffffff ); @@ -396,16 +396,16 @@ void fh_if4p5_south_in(RU_t *ru,int *frame,int *subframe) { } while(proc->symbol_mask[*subframe] != symbol_mask_full); //caculate timestamp_rx, timestamp_tx based on frame and subframe - proc->subframe_rx = sf; + proc->tti_rx = sf; proc->frame_rx = f; - proc->timestamp_rx = ((proc->frame_rx * 10) + proc->subframe_rx ) * fp->samples_per_tti ; + proc->timestamp_rx = ((proc->frame_rx * 10) + proc->tti_rx ) * fp->samples_per_tti ; // proc->timestamp_tx = proc->timestamp_rx + (4*fp->samples_per_tti); - proc->subframe_tx = (sf+sf_ahead)%10; + proc->tti_tx = (sf+sf_ahead)%10; proc->frame_tx = (sf>(9-sf_ahead)) ? (f+1)&1023 : f; if (proc->first_rx == 0) { - if (proc->subframe_rx != *subframe){ - LOG_E(PHY,"Received Timestamp (IF4p5) doesn't correspond to the time we think it is (proc->subframe_rx %d, subframe %d)\n",proc->subframe_rx,*subframe); + if (proc->tti_rx != *subframe){ + LOG_E(PHY,"Received Timestamp (IF4p5) doesn't correspond to the time we think it is (proc->tti_rx %d, subframe %d)\n",proc->tti_rx,*subframe); exit_fun("Exiting"); } if (proc->frame_rx != *frame) { @@ -415,14 +415,14 @@ void fh_if4p5_south_in(RU_t *ru,int *frame,int *subframe) { } else { proc->first_rx = 0; *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *subframe = proc->tti_rx; } if (ru == RC.ru[0]) { VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX0_RU, f ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_RU, sf ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX0_RU, sf ); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, proc->frame_tx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, proc->subframe_tx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, proc->tti_tx ); } proc->symbol_mask[sf] = 0; @@ -455,22 +455,22 @@ void fh_if5_south_asynch_in_mobipass(RU_t *ru,int *frame,int *subframe) { pthread_mutex_lock(&proc->mutex_asynch_rxtx); int offset_mobipass = 40120; pthread_mutex_lock(&proc->mutex_asynch_rxtx); - proc->subframe_rx = ((proc->timestamp_rx-offset_mobipass)/fp->samples_per_tti)%10; + proc->tti_rx = ((proc->timestamp_rx-offset_mobipass)/fp->samples_per_tti)%10; proc->frame_rx = ((proc->timestamp_rx-offset_mobipass)/(fp->samples_per_tti*10))&1023; - proc->subframe_rx = (proc->timestamp_rx/fp->samples_per_tti)%10; + proc->tti_rx = (proc->timestamp_rx/fp->samples_per_tti)%10; proc->frame_rx = (proc->timestamp_rx/(10*fp->samples_per_tti))&1023; if (proc->first_rx == 1) { proc->first_rx =2; - *subframe = proc->subframe_rx; + *subframe = proc->tti_rx; *frame = proc->frame_rx; - LOG_E(PHY,"[Mobipass]timestamp_rx:%llu, frame_rx %d, subframe: %d\n",(unsigned long long int)proc->timestamp_rx,proc->frame_rx,proc->subframe_rx); + LOG_E(PHY,"[Mobipass]timestamp_rx:%llu, frame_rx %d, subframe: %d\n",(unsigned long long int)proc->timestamp_rx,proc->frame_rx,proc->tti_rx); } else { - if (proc->subframe_rx != *subframe) { + if (proc->tti_rx != *subframe) { proc->first_rx++; - LOG_E(PHY,"[Mobipass]timestamp:%llu, subframe_rx %d is not what we expect %d, first_rx:%d\n",(unsigned long long int)proc->timestamp_rx, proc->subframe_rx,*subframe, proc->first_rx); + LOG_E(PHY,"[Mobipass]timestamp:%llu, tti_rx %d is not what we expect %d, first_rx:%d\n",(unsigned long long int)proc->timestamp_rx, proc->tti_rx,*subframe, proc->first_rx); //exit_fun("Exiting"); } if (proc->frame_rx != *frame) { @@ -479,7 +479,7 @@ void fh_if5_south_asynch_in_mobipass(RU_t *ru,int *frame,int *subframe) { // exit_fun("Exiting"); } // temporary solution - *subframe = proc->subframe_rx; + *subframe = proc->tti_rx; *frame = proc->frame_rx; } @@ -503,15 +503,15 @@ void fh_if4p5_south_asynch_in(RU_t *ru,int *frame,int *subframe) { prach_rx = 0; do { // Blocking, we need a timeout on this !!!!!!!!!!!!!!!!!!!!!!! - recv_IF4p5(ru, &proc->frame_rx, &proc->subframe_rx, &packet_type, &symbol_number); + recv_IF4p5(ru, &proc->frame_rx, &proc->tti_rx, &packet_type, &symbol_number); // grab first prach information for this new subframe if (got_prach_info==0) { - prach_rx = is_prach_subframe(fp, proc->frame_rx, proc->subframe_rx); + prach_rx = is_prach_subframe(fp, proc->frame_rx, proc->tti_rx); got_prach_info = 1; } if (proc->first_rx != 0) { *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *subframe = proc->tti_rx; proc->first_rx = 0; } else { @@ -519,8 +519,8 @@ void fh_if4p5_south_asynch_in(RU_t *ru,int *frame,int *subframe) { LOG_E(PHY,"frame_rx %d is not what we expect %d\n",proc->frame_rx,*frame); exit_fun("Exiting"); } - if (proc->subframe_rx != *subframe) { - LOG_E(PHY,"subframe_rx %d is not what we expect %d\n",proc->subframe_rx,*subframe); + if (proc->tti_rx != *subframe) { + LOG_E(PHY,"tti_rx %d is not what we expect %d\n",proc->tti_rx,*subframe); exit_fun("Exiting"); } } @@ -564,7 +564,7 @@ void fh_if4p5_north_in(RU_t *ru,int *frame,int *subframe) { // dump VCD output for first RU in list if (ru == RC.ru[0]) { VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, *frame ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, *subframe ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, *subframe ); } } @@ -572,23 +572,23 @@ void fh_if5_north_asynch_in(RU_t *ru,int *frame,int *subframe) { LTE_DL_FRAME_PARMS *fp = ru->frame_parms; RU_proc_t *proc = &ru->proc; - int subframe_tx,frame_tx; + int tti_tx,frame_tx; openair0_timestamp timestamp_tx; recv_IF5(ru, ×tamp_tx, *subframe, IF5_RRH_GW_DL); - // printf("Received subframe %d (TS %llu) from RCC\n",subframe_tx,timestamp_tx); + // printf("Received subframe %d (TS %llu) from RCC\n",tti_tx,timestamp_tx); - subframe_tx = (timestamp_tx/fp->samples_per_tti)%10; + tti_tx = (timestamp_tx/fp->samples_per_tti)%10; frame_tx = (timestamp_tx/(fp->samples_per_tti*10))&1023; if (proc->first_tx != 0) { - *subframe = subframe_tx; + *subframe = tti_tx; *frame = frame_tx; proc->first_tx = 0; } else { - AssertFatal(subframe_tx == *subframe, - "subframe_tx %d is not what we expect %d\n",subframe_tx,*subframe); + AssertFatal(tti_tx == *subframe, + "tti_tx %d is not what we expect %d\n",tti_tx,*subframe); AssertFatal(frame_tx == *frame, "frame_tx %d is not what we expect %d\n",frame_tx,*frame); } @@ -601,28 +601,28 @@ void fh_if4p5_north_asynch_in(RU_t *ru,int *frame,int *subframe) { uint16_t packet_type; uint32_t symbol_number,symbol_mask,symbol_mask_full; - int subframe_tx,frame_tx; + int tti_tx,frame_tx; LOG_D(PHY, "%s(ru:%p frame, subframe)\n", __FUNCTION__, ru); symbol_number = 0; symbol_mask = 0; symbol_mask_full = ((subframe_select(fp,*subframe) == SF_S) ? (1<dl_symbols_in_S_subframe) : (1<symbols_per_tti))-1; do { - recv_IF4p5(ru, &frame_tx, &subframe_tx, &packet_type, &symbol_number); - if ((subframe_select(fp,subframe_tx) == SF_DL) && (symbol_number == 0)) start_meas(&ru->rx_fhaul); + recv_IF4p5(ru, &frame_tx, &tti_tx, &packet_type, &symbol_number); + if ((subframe_select(fp,tti_tx) == SF_DL) && (symbol_number == 0)) start_meas(&ru->rx_fhaul); LOG_D(PHY,"subframe %d (%d): frame %d, subframe %d, symbol %d\n", - *subframe,subframe_select(fp,*subframe),frame_tx,subframe_tx,symbol_number); + *subframe,subframe_select(fp,*subframe),frame_tx,tti_tx,symbol_number); if (proc->first_tx != 0) { *frame = frame_tx; - *subframe = subframe_tx; + *subframe = tti_tx; proc->first_tx = 0; symbol_mask_full = ((subframe_select(fp,*subframe) == SF_S) ? (1<dl_symbols_in_S_subframe) : (1<symbols_per_tti))-1; } else { AssertFatal(frame_tx == *frame, "frame_tx %d is not what we expect %d\n",frame_tx,*frame); - AssertFatal(subframe_tx == *subframe, - "subframe_tx %d is not what we expect %d\n",subframe_tx,*subframe); + AssertFatal(tti_tx == *subframe, + "tti_tx %d is not what we expect %d\n",tti_tx,*subframe); } if (packet_type == IF4p5_PDLFFT) { symbol_mask = symbol_mask | (1<rx_fhaul); + if (subframe_select(fp,tti_tx) == SF_DL) stop_meas(&ru->rx_fhaul); - proc->subframe_tx = subframe_tx; + proc->tti_tx = tti_tx; proc->frame_tx = frame_tx; - if ((frame_tx == 0)&&(subframe_tx == 0)) proc->frame_tx_unwrap += 1024; + if ((frame_tx == 0)&&(tti_tx == 0)) proc->frame_tx_unwrap += 1024; - proc->timestamp_tx = ((((uint64_t)frame_tx + (uint64_t)proc->frame_tx_unwrap) * 10) + (uint64_t)subframe_tx) * (uint64_t)fp->samples_per_tti; + proc->timestamp_tx = ((((uint64_t)frame_tx + (uint64_t)proc->frame_tx_unwrap) * 10) + (uint64_t)tti_tx) * (uint64_t)fp->samples_per_tti; - LOG_D(PHY,"RU %d/%d TST %llu, frame %d, subframe %d\n",ru->idx,0,(long long unsigned int)proc->timestamp_tx,frame_tx,subframe_tx); + LOG_D(PHY,"RU %d/%d TST %llu, frame %d, subframe %d\n",ru->idx,0,(long long unsigned int)proc->timestamp_tx,frame_tx,tti_tx); // dump VCD output for first RU in list if (ru == RC.ru[0]) { VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, frame_tx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, subframe_tx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, tti_tx ); } if (ru->feptx_ofdm) ru->feptx_ofdm(ru); @@ -657,7 +657,7 @@ void fh_if5_north_out(RU_t *ru) { /// **** send_IF5 of rxdata to BBU **** /// VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_SEND_IF5, 1 ); - send_IF5(ru, proc->timestamp_rx, proc->subframe_rx, &seqno, IF5_RRH_GW_UL); + send_IF5(ru, proc->timestamp_rx, proc->tti_rx, &seqno, IF5_RRH_GW_UL); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_SEND_IF5, 0 ); } @@ -667,17 +667,17 @@ void fh_if4p5_north_out(RU_t *ru) { RU_proc_t *proc=&ru->proc; LTE_DL_FRAME_PARMS *fp = ru->frame_parms; - const int subframe = proc->subframe_rx; - if (ru->idx==0) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_RU, proc->subframe_rx ); + const int subframe = proc->tti_rx; + if (ru->idx==0) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX0_RU, proc->tti_rx ); if ((fp->frame_type == TDD) && (subframe_select(fp,subframe)!=SF_UL)) { /// **** in TDD during DL send_IF4 of ULTICK to RCC **** /// - send_IF4p5(ru, proc->frame_rx, proc->subframe_rx, IF4p5_PULTICK); + send_IF4p5(ru, proc->frame_rx, proc->tti_rx, IF4p5_PULTICK); return; } start_meas(&ru->tx_fhaul); - send_IF4p5(ru, proc->frame_rx, proc->subframe_rx, IF4p5_PULFFT); + send_IF4p5(ru, proc->frame_rx, proc->tti_rx, IF4p5_PULFFT); stop_meas(&ru->tx_fhaul); } @@ -780,38 +780,38 @@ void rx_rf(RU_t *ru,int *frame,int *subframe) { } proc->frame_rx = (proc->timestamp_rx / (fp->samples_per_tti*10))&1023; - proc->subframe_rx = (proc->timestamp_rx / fp->samples_per_tti)%10; + proc->tti_rx = (proc->timestamp_rx / fp->samples_per_tti)%10; // synchronize first reception to frame 0 subframe 0 #ifdef PHY_TX_THREAD proc->timestamp_phy_tx = proc->timestamp_rx+((sf_ahead-1)*fp->samples_per_tti); - proc->subframe_phy_tx = (proc->subframe_rx+(sf_ahead-1))%10; - proc->frame_phy_tx = (proc->subframe_rx>(9-(sf_ahead-1))) ? (proc->frame_rx+1)&1023 : proc->frame_rx; + proc->subframe_phy_tx = (proc->tti_rx+(sf_ahead-1))%10; + proc->frame_phy_tx = (proc->tti_rx>(9-(sf_ahead-1))) ? (proc->frame_rx+1)&1023 : proc->frame_rx; #else proc->timestamp_tx = proc->timestamp_rx+(sf_ahead*fp->samples_per_tti); - proc->subframe_tx = (proc->subframe_rx+sf_ahead)%10; - proc->frame_tx = (proc->subframe_rx>(9-sf_ahead)) ? (proc->frame_rx+1)&1023 : proc->frame_rx; + proc->tti_tx = (proc->tti_rx+sf_ahead)%10; + proc->frame_tx = (proc->tti_rx>(9-sf_ahead)) ? (proc->frame_rx+1)&1023 : proc->frame_rx; #endif //proc->timestamp_tx = proc->timestamp_rx+(sf_ahead*fp->samples_per_tti); - //proc->subframe_tx = (proc->subframe_rx+sf_ahead)%10; - //proc->frame_tx = (proc->subframe_rx>(9-sf_ahead)) ? (proc->frame_rx+1)&1023 : proc->frame_rx; + //proc->subframe_tx = (proc->tti_rx+sf_ahead)%10; + //proc->frame_tx = (proc->tti_rx>(9-sf_ahead)) ? (proc->frame_rx+1)&1023 : proc->frame_rx; LOG_D(PHY,"RU %d/%d TS %llu (off %d), frame %d, subframe %d\n", ru->idx, 0, (unsigned long long int)proc->timestamp_rx, - (int)ru->ts_offset,proc->frame_rx,proc->subframe_rx); + (int)ru->ts_offset,proc->frame_rx,proc->tti_rx); // dump VCD output for first RU in list if (ru == RC.ru[0]) { VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX0_RU, proc->frame_rx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_RU, proc->subframe_rx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX0_RU, proc->tti_rx ); } if (proc->first_rx == 0) { - if (proc->subframe_rx != *subframe){ - LOG_E(PHY,"Received Timestamp (%llu) doesn't correspond to the time we think it is (proc->subframe_rx %d, subframe %d)\n",(long long unsigned int)proc->timestamp_rx,proc->subframe_rx,*subframe); + if (proc->tti_rx != *subframe){ + LOG_E(PHY,"Received Timestamp (%llu) doesn't correspond to the time we think it is (proc->tti_rx %d, subframe %d)\n",(long long unsigned int)proc->timestamp_rx,proc->tti_rx,*subframe); exit_fun("Exiting"); } @@ -822,10 +822,10 @@ void rx_rf(RU_t *ru,int *frame,int *subframe) { } else { proc->first_rx = 0; *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *subframe = proc->tti_rx; } - //printf("timestamp_rx %lu, frame %d(%d), subframe %d(%d)\n",ru->timestamp_rx,proc->frame_rx,frame,proc->subframe_rx,subframe); + //printf("timestamp_rx %lu, frame %d(%d), subframe %d(%d)\n",ru->timestamp_rx,proc->frame_rx,frame,proc->tti_rx,subframe); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TS, proc->timestamp_rx&0xffffffff ); @@ -849,12 +849,12 @@ void tx_rf(RU_t *ru) { unsigned int txs; int i; - T(T_ENB_PHY_OUTPUT_SIGNAL, T_INT(0), T_INT(0), T_INT(proc->frame_tx), T_INT(proc->subframe_tx), - T_INT(0), T_BUFFER(&ru->common.txdata[0][proc->subframe_tx * fp->samples_per_tti], fp->samples_per_tti * 4)); + T(T_ENB_PHY_OUTPUT_SIGNAL, T_INT(0), T_INT(0), T_INT(proc->frame_tx), T_INT(proc->tti_tx), + T_INT(0), T_BUFFER(&ru->common.txdata[0][proc->tti_tx * fp->samples_per_tti], fp->samples_per_tti * 4)); - lte_subframe_t SF_type = subframe_select(fp,proc->subframe_tx%10); - lte_subframe_t prevSF_type = subframe_select(fp,(proc->subframe_tx+9)%10); - lte_subframe_t nextSF_type = subframe_select(fp,(proc->subframe_tx+1)%10); + lte_subframe_t SF_type = subframe_select(fp,proc->tti_tx%10); + lte_subframe_t prevSF_type = subframe_select(fp,(proc->tti_tx+9)%10); + lte_subframe_t nextSF_type = subframe_select(fp,(proc->tti_tx+1)%10); int sf_extension = 0; if ((SF_type == SF_DL) || @@ -892,7 +892,7 @@ void tx_rf(RU_t *ru) { #endif for (i=0; inb_tx; i++) - txp[i] = (void*)&ru->common.txdata[i][(proc->subframe_tx*fp->samples_per_tti)-sf_extension]; + txp[i] = (void*)&ru->common.txdata[i][(proc->tti_tx*fp->samples_per_tti)-sf_extension]; /* add fail safe for late command */ if(late_control!=STATE_BURST_NORMAL){//stop burst @@ -926,7 +926,7 @@ void tx_rf(RU_t *ru) { /* add fail safe for late command end */ VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, proc->frame_tx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, proc->subframe_tx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, proc->tti_tx ); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, (proc->timestamp_tx-ru->openair0_cfg.tx_sample_advance)&0xffffffff ); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_TRX_WRITE, 1 ); @@ -940,7 +940,7 @@ void tx_rf(RU_t *ru) { flags); LOG_D(PHY,"[TXPATH] RU %d tx_rf, writing to TS %llu, frame %d, unwrapped_frame %d, subframe %d\n",ru->idx, - (long long unsigned int)proc->timestamp_tx,proc->frame_tx,proc->frame_tx_unwrap,proc->subframe_tx); + (long long unsigned int)proc->timestamp_tx,proc->frame_tx,proc->frame_tx_unwrap,proc->tti_tx); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_TRX_WRITE, 0 ); @@ -1033,7 +1033,7 @@ void wakeup_slaves(RU_proc_t *proc) { int cnt_slave = ++slave_proc->instance_cnt_FH; slave_proc->frame_rx = proc->frame_rx; - slave_proc->subframe_rx = proc->subframe_rx; + slave_proc->tti_rx = proc->tti_rx; slave_proc->timestamp_rx = proc->timestamp_rx; slave_proc->timestamp_tx = proc->timestamp_tx; @@ -1263,7 +1263,7 @@ void wakeup_eNBs(RU_t *ru) { char string[20]; sprintf(string,"Incoming RU %d",ru->idx); LOG_D(PHY,"RU %d Call eNB_top\n",ru->idx); - ru->eNB_top(eNB_list[0],ru->proc.frame_rx,ru->proc.subframe_rx,string,ru); + ru->eNB_top(eNB_list[0],ru->proc.frame_rx,ru->proc.tti_rx,string,ru); ru->proc.emulate_rf_busy = 0; } else { @@ -1276,7 +1276,7 @@ void wakeup_eNBs(RU_t *ru) { eNB_list[i]->proc.ru_proc = &ru->proc; if (ru->wakeup_rxtx!=0 && ru->wakeup_rxtx(eNB_list[i],ru) < 0) { - LOG_E(PHY,"could not wakeup eNB rxtx process for subframe %d\n", ru->proc.subframe_rx); + LOG_E(PHY,"could not wakeup eNB rxtx process for subframe %d\n", ru->proc.tti_rx); } ru->proc.emulate_rf_busy = 0; } @@ -1298,12 +1298,12 @@ static inline int wakeup_prach_ru(RU_t *ru) { if (ru->proc.instance_cnt_prach==-1) { ++ru->proc.instance_cnt_prach; ru->proc.frame_prach = ru->proc.frame_rx; - ru->proc.subframe_prach = ru->proc.subframe_rx; + ru->proc.subframe_prach = ru->proc.tti_rx; // DJP - think prach_procedures() is looking at eNB frame_prach if (ru->eNB_list[0]) { ru->eNB_list[0]->proc.frame_prach = ru->proc.frame_rx; - ru->eNB_list[0]->proc.subframe_prach = ru->proc.subframe_rx; + ru->eNB_list[0]->proc.subframe_prach = ru->proc.tti_rx; } LOG_D(PHY,"RU %d: waking up PRACH thread\n",ru->idx); // the thread can now be woken up @@ -1331,7 +1331,7 @@ static inline int wakeup_prach_ru_br(RU_t *ru) { if (ru->proc.instance_cnt_prach_br==-1) { ++ru->proc.instance_cnt_prach_br; ru->proc.frame_prach_br = ru->proc.frame_rx; - ru->proc.subframe_prach_br = ru->proc.subframe_rx; + ru->proc.subframe_prach_br = ru->proc.tti_rx; LOG_D(PHY,"RU %d: waking up PRACH thread\n",ru->idx); // the thread can now be woken up @@ -1571,10 +1571,10 @@ static void* ru_thread_tx( void* param ) { else { for (int i=0; inb_tx; i++) { if(proc->frame_tx == 2) { - sprintf(filename,"txdataF%d_frame%d_sf%d.m",i,proc->frame_tx,proc->subframe_tx); + sprintf(filename,"txdataF%d_frame%d_sf%d.m",i,proc->frame_tx,proc->tti_tx); LOG_M(filename,"txdataF_frame",ru->common.txdataF_BF[i],fp->symbols_per_tti*fp->ofdm_symbol_size, 1, 1); } - if(proc->frame_tx == 2 && proc->subframe_tx==0){ + if(proc->frame_tx == 2 && proc->tti_tx==0){ sprintf(filename,"txdata%d_frame%d.m",i,proc->frame_tx); LOG_M(filename,"txdata_frame",ru->common.txdata[i],fp->samples_per_tti*10, 1, 1); } @@ -1748,13 +1748,13 @@ static void* ru_thread( void* param ) { LOG_D(PHY,"RU thread (do_prach %d, is_prach_subframe %d), received frame %d, subframe %d\n", ru->do_prach, - is_prach_subframe(fp, proc->frame_rx, proc->subframe_rx), - proc->frame_rx,proc->subframe_rx); - if ((ru->do_prach>0) && (is_prach_subframe(fp, proc->frame_rx, proc->subframe_rx)==1)) { + is_prach_subframe(fp, proc->frame_rx, proc->tti_rx), + proc->frame_rx,proc->tti_rx); + if ((ru->do_prach>0) && (is_prach_subframe(fp, proc->frame_rx, proc->tti_rx)==1)) { wakeup_prach_ru(ru); } #if (RRC_VERSION >= MAKE_VERSION(14, 0, 0)) - else if ((ru->do_prach>0) && (is_prach_subframe(fp, proc->frame_rx, proc->subframe_rx)>1)) { + else if ((ru->do_prach>0) && (is_prach_subframe(fp, proc->frame_rx, proc->tti_rx)>1)) { wakeup_prach_ru_br(ru); } #endif @@ -1817,10 +1817,10 @@ static void* ru_thread( void* param ) { else { for (int i=0; inb_tx; i++) { if(proc->frame_tx == 2) { - sprintf(filename,"txdataF%d_frame%d_sf%d.m",i,proc->frame_tx,proc->subframe_tx); + sprintf(filename,"txdataF%d_frame%d_sf%d.m",i,proc->frame_tx,proc->tti_tx); LOG_M(filename,"txdataF_frame",ru->common.txdataF_BF[i],fp->symbols_per_tti*fp->ofdm_symbol_size, 1, 1); } - if(proc->frame_tx == 2 && proc->subframe_tx==0){ + if(proc->frame_tx == 2 && proc->tti_tx==0){ sprintf(filename,"txdata%d_frame%d.m",i,proc->frame_tx); LOG_M(filename,"txdata_frame",ru->common.txdata[i],fp->samples_per_tti*10, 1, 1); } @@ -2018,7 +2018,7 @@ static void* eNB_thread_phy_tx( void* param ) { LOG_D(PHY,"Running eNB phy tx procedures\n"); if(ru->num_eNB == 1){ - proc_rxtx.subframe_tx = proc->subframe_phy_tx; + proc_rxtx.tti_tx = proc->subframe_phy_tx; proc_rxtx.frame_tx = proc->frame_phy_tx; phy_procedures_eNB_TX(eNB_list[0], &proc_rxtx, 1); phy_tx_txdataF_end = 1; @@ -2029,7 +2029,7 @@ static void* eNB_thread_phy_tx( void* param ) { if (ru->proc.instance_cnt_rf_tx==-1) { ++ru->proc.instance_cnt_rf_tx; ru->proc.frame_tx = proc->frame_phy_tx; - ru->proc.subframe_tx = proc->subframe_phy_tx; + ru->proc.tti_tx = proc->subframe_phy_tx; ru->proc.timestamp_tx = proc->timestamp_phy_tx; // the thread can now be woken up diff --git a/targets/RT/USER/lte-softmodem.c b/targets/RT/USER/lte-softmodem.c index 03a461454f..8dbd1d51c9 100644 --- a/targets/RT/USER/lte-softmodem.c +++ b/targets/RT/USER/lte-softmodem.c @@ -253,7 +253,7 @@ extern char uecap_xer[1024]; char uecap_xer_in=0; int oaisim_flag=0; -threads_t threads= {-1,-1,-1,-1,-1,-1,-1}; +threads_t threads= {-1,-1,-1,-1,-1,-1,-1,-1}; /* see file openair2/LAYER2/MAC/main.c for why abstraction_flag is needed * this is very hackish - find a proper solution @@ -906,14 +906,16 @@ int main( int argc, char **argv ) char cpu_affinity[1024]; CPU_ZERO(&cpuset); #ifdef CPU_AFFINITY + int j; if (get_nprocs() > 2) { - CPU_SET(0, &cpuset); + for (j = 2; j < get_nprocs(); j++) + CPU_SET(j, &cpuset); s = pthread_setaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); if (s != 0) { perror( "pthread_setaffinity_np"); exit_fun("Error setting processor affinity"); } - LOG_I(HW, "Setting the affinity of main function to CPU 0, for device library to use CPU 0 only!\n"); + LOG_I(HW, "Setting the affinity of main function to all CPUs, for device library to use CPU 0 only!\n"); } #endif diff --git a/targets/RT/USER/lte-softmodem.h b/targets/RT/USER/lte-softmodem.h index 96ab799c3d..0870405fac 100644 --- a/targets/RT/USER/lte-softmodem.h +++ b/targets/RT/USER/lte-softmodem.h @@ -169,7 +169,8 @@ {"clock", CONFIG_HLP_CLK, 0, uptr:&clock_source, defintval:0, TYPE_UINT, 0}, \ {"wait-for-sync", NULL, PARAMFLAG_BOOL, iptr:&wait_for_sync, defintval:0, TYPE_INT, 0}, \ {"single-thread-enable", CONFIG_HLP_NOSNGLT, PARAMFLAG_BOOL, iptr:&single_thread_flag, defintval:0, TYPE_INT, 0}, \ -{"threadIQ", NULL, 0, iptr:&(threads.iq), defintval:1, TYPE_INT, 0}, \ +{"threadMain", NULL, 0, iptr:&(threads.main), defintval:1, TYPE_INT, 0}, \ +{"threadSync", NULL, 0, iptr:&(threads.sync), defintval:1, TYPE_INT, 0}, \ {"threadOneSubframe", NULL, 0, iptr:&(threads.one), defintval:1, TYPE_INT, 0}, \ {"threadTwoSubframe", NULL, 0, iptr:&(threads.two), defintval:1, TYPE_INT, 0}, \ {"threadThreeSubframe", NULL, 0, iptr:&(threads.three), defintval:1, TYPE_INT, 0}, \ diff --git a/targets/RT/USER/lte-ue.c b/targets/RT/USER/lte-ue.c index 965203f47e..a9a880dde7 100644 --- a/targets/RT/USER/lte-ue.c +++ b/targets/RT/USER/lte-ue.c @@ -456,8 +456,8 @@ static void *UE_thread_synch(void *arg) cpu_set_t cpuset; CPU_ZERO(&cpuset); - if ( threads.iq != -1 ) - CPU_SET(threads.iq, &cpuset); + if ( threads.sync != -1 ) + CPU_SET(threads.sync, &cpuset); // this thread priority must be lower that the main acquisition thread sprintf(threadname, "sync UE %d\n", UE->Mod_id); init_thread(100000, 500000, FIFO_PRIORITY-1, &cpuset, threadname); @@ -1435,8 +1435,8 @@ void *UE_thread(void *arg) { cpu_set_t cpuset; CPU_ZERO(&cpuset); - if ( threads.iq != -1 ) - CPU_SET(threads.iq, &cpuset); + if ( threads.main != -1 ) + CPU_SET(threads.main, &cpuset); init_thread(100000, 500000, FIFO_PRIORITY, &cpuset, "UHD Threads"); diff --git a/targets/RT/USER/lte-uesoftmodem.c b/targets/RT/USER/lte-uesoftmodem.c index 3d2995e5d1..617756e722 100644 --- a/targets/RT/USER/lte-uesoftmodem.c +++ b/targets/RT/USER/lte-uesoftmodem.c @@ -260,7 +260,7 @@ extern char uecap_xer[1024]; char uecap_xer_in=0; int oaisim_flag=0; -threads_t threads= {-1,-1,-1,-1,-1,-1,-1}; +threads_t threads = {-1,-1,-1,-1,-1,-1,-1}; /* see file openair2/LAYER2/MAC/main.c for why abstraction_flag is needed * this is very hackish - find a proper solution @@ -1014,14 +1014,17 @@ printf("~~~~~~~~~~~~~~~~~~~~successfully get the parallel config[%d], worker con char cpu_affinity[1024]; CPU_ZERO(&cpuset); #ifdef CPU_AFFINITY + int j; if (get_nprocs() > 2) { - CPU_SET(0, &cpuset); + for (j = 2; j < get_nprocs(); j++) + CPU_SET(j, &cpuset); + s = pthread_setaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); if (s != 0) { perror( "pthread_setaffinity_np"); exit_fun("Error setting processor affinity"); } - LOG_I(HW, "Setting the affinity of main function to CPU 0, for device library to use CPU 0 only!\n"); + LOG_I(HW, "Setting the affinity of main function to all CPUs, for device library to use CPU 0 only!\n"); } #endif diff --git a/targets/RT/USER/nr-gnb.c b/targets/RT/USER/nr-gnb.c index eae095b936..5d8c98d645 100644 --- a/targets/RT/USER/nr-gnb.c +++ b/targets/RT/USER/nr-gnb.c @@ -118,8 +118,8 @@ extern openair0_config_t openair0_cfg[MAX_CARDS]; extern int transmission_mode; +uint16_t sl_ahead=4; uint16_t sf_ahead=4; - //pthread_t main_gNB_thread; time_stats_t softmodem_stats_mt; // main thread @@ -168,14 +168,14 @@ static inline int rxtx(PHY_VARS_gNB *gNB,gNB_L1_rxtx_proc_t *proc, char *thread_ // I am a PNF and I need to let nFAPI know that we have a (sub)frame tick uint16_t frame = proc->frame_rx; - uint16_t subframe = proc->subframe_rx; + uint16_t slot = proc->slot_rx; //add_subframe(&frame, &subframe, 4); //oai_subframe_ind(proc->frame_tx, proc->subframe_tx); //LOG_D(PHY, "oai_subframe_ind(frame:%u, subframe:%d) - NOT CALLED ********\n", frame, subframe); start_meas(&nfapi_meas); - oai_subframe_ind(frame, subframe); + oai_subframe_ind(frame, slot); stop_meas(&nfapi_meas); if (gNB->UL_INFO.rx_ind.rx_indication_body.number_of_pdus|| @@ -190,8 +190,8 @@ static inline int rxtx(PHY_VARS_gNB *gNB,gNB_L1_rxtx_proc_t *proc, char *thread_ NFAPI_SFNSF2DEC(gNB->UL_INFO.crc_ind.sfn_sf), gNB->UL_INFO.crc_ind.crc_indication_body.number_of_crcs, NFAPI_SFNSF2DEC(gNB->UL_INFO.rach_ind.sfn_sf), gNB->UL_INFO.rach_ind.rach_indication_body.number_of_preambles, gNB->UL_INFO.cqi_ind.number_of_cqis, - proc->frame_rx, proc->subframe_rx, - proc->frame_tx, proc->subframe_tx); + proc->frame_rx, proc->slot_rx, + proc->frame_tx, proc->slot_tx); } } @@ -200,11 +200,11 @@ static inline int rxtx(PHY_VARS_gNB *gNB,gNB_L1_rxtx_proc_t *proc, char *thread_ // **************************************** // Common RX procedures subframe n - T(T_GNB_PHY_DL_TICK, T_INT(gNB->Mod_id), T_INT(proc->frame_tx), T_INT(proc->subframe_tx)); + T(T_GNB_PHY_DL_TICK, T_INT(gNB->Mod_id), T_INT(proc->frame_tx), T_INT(proc->slot_tx)); /* // if this is IF5 or 3GPP_gNB if (gNB && gNB->RU_list && gNB->RU_list[0] && gNB->RU_list[0]->function < NGFI_RAU_IF4p5) { - wakeup_prach_gNB(gNB,NULL,proc->frame_rx,proc->subframe_rx); + wakeup_prach_gNB(gNB,NULL,proc->frame_rx,proc->slot_rx); } // UE-specific RX processing for subframe n @@ -215,7 +215,7 @@ static inline int rxtx(PHY_VARS_gNB *gNB,gNB_L1_rxtx_proc_t *proc, char *thread_ pthread_mutex_lock(&gNB->UL_INFO_mutex); gNB->UL_INFO.frame = proc->frame_rx; - gNB->UL_INFO.subframe = proc->subframe_rx; + gNB->UL_INFO.slot = proc->slot_rx; gNB->UL_INFO.module_id = gNB->Mod_id; gNB->UL_INFO.CC_id = gNB->CC_id; @@ -225,7 +225,7 @@ static inline int rxtx(PHY_VARS_gNB *gNB,gNB_L1_rxtx_proc_t *proc, char *thread_ /// end // ***************************************** - // TX processing for subframe n+sf_ahead + // TX processing for subframe n+sl_ahead // run PHY TX procedures the one after the other for all CCs to avoid race conditions // (may be relaxed in the future for performance reasons) // ***************************************** @@ -238,7 +238,7 @@ static inline int rxtx(PHY_VARS_gNB *gNB,gNB_L1_rxtx_proc_t *proc, char *thread_ stop_meas( &softmodem_stats_rxtx_sf ); - LOG_D(PHY,"%s() Exit proc[rx:%d%d tx:%d%d]\n", __FUNCTION__, proc->frame_rx, proc->subframe_rx, proc->frame_tx, proc->subframe_tx); + LOG_D(PHY,"%s() Exit proc[rx:%d%d tx:%d%d]\n", __FUNCTION__, proc->frame_rx, proc->slot_rx, proc->frame_tx, proc->slot_tx); #if 0 LOG_D(PHY, "rxtx:%lld nfapi:%lld phy:%lld tx:%lld rx:%lld prach:%lld ofdm:%lld ", @@ -311,8 +311,8 @@ static void* gNB_L1_thread_tx(void* param) { // (may be relaxed in the future for performance reasons) // ***************************************** - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX1_GNB,proc->subframe_tx); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX1_GNB,proc->subframe_rx); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_TX1_GNB,proc->slot_tx); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_RX1_GNB,proc->slot_rx); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX1_GNB,proc->frame_tx); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX1_GNB,proc->frame_rx); @@ -364,8 +364,8 @@ static void* gNB_L1_thread( void* param ) { if (wait_on_condition(&proc->mutex,&proc->cond,&proc->instance_cnt,thread_name)<0) break; VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_gNB_PROC_RXTX0, 1 ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_GNB,proc->subframe_tx); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_GNB,proc->subframe_rx); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_TX0_GNB,proc->slot_tx); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_SLOT_NUMBER_RX0_GNB,proc->slot_rx); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_GNB,proc->frame_tx); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME(VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX0_GNB,proc->frame_rx); @@ -420,7 +420,7 @@ static void wait_system_ready (char *message, volatile int *start_flag) { -void gNB_top(PHY_VARS_gNB *gNB, int frame_rx, int subframe_rx, char *string, struct RU_t_s *ru) +void gNB_top(PHY_VARS_gNB *gNB, int frame_rx, int slot_rx, char *string, struct RU_t_s *ru) { gNB_L1_proc_t *proc = &gNB->proc; gNB_L1_rxtx_proc_t *L1_proc = &proc->L1_proc; @@ -428,20 +428,20 @@ void gNB_top(PHY_VARS_gNB *gNB, int frame_rx, int subframe_rx, char *string, str RU_proc_t *ru_proc=&ru->proc; proc->frame_rx = frame_rx; - proc->subframe_rx = subframe_rx; + proc->slot_rx = slot_rx; if (!oai_exit) { - T(T_ENB_MASTER_TICK, T_INT(0), T_INT(proc->frame_rx), T_INT(proc->subframe_rx)); + T(T_ENB_MASTER_TICK, T_INT(0), T_INT(proc->frame_rx), T_INT(proc->slot_rx)); - L1_proc->timestamp_tx = ru_proc->timestamp_rx + (sf_ahead*fp->samples_per_subframe); + L1_proc->timestamp_tx = ru_proc->timestamp_rx + (sl_ahead*fp->samples_per_slot); L1_proc->frame_rx = ru_proc->frame_rx; - L1_proc->subframe_rx = ru_proc->subframe_rx; - L1_proc->frame_tx = (L1_proc->subframe_rx > (9-sf_ahead)) ? (L1_proc->frame_rx+1)&1023 : L1_proc->frame_rx; - L1_proc->subframe_tx = (L1_proc->subframe_rx + sf_ahead)%10; + L1_proc->slot_rx = ru_proc->tti_rx; + L1_proc->frame_tx = (L1_proc->slot_rx > (fp->slots_per_frame-1-sl_ahead)) ? (L1_proc->frame_rx+1)&1023 : L1_proc->frame_rx; + L1_proc->slot_tx = (L1_proc->slot_rx + sl_ahead)%fp->slots_per_frame; if (rxtx(gNB,L1_proc,string) < 0) LOG_E(PHY,"gNB %d CC_id %d failed during execution\n",gNB->Mod_id,gNB->CC_id); ru_proc->timestamp_tx = L1_proc->timestamp_tx; - ru_proc->subframe_tx = L1_proc->subframe_tx; + ru_proc->tti_tx = L1_proc->slot_tx; ru_proc->frame_tx = L1_proc->frame_tx; } } @@ -455,10 +455,10 @@ int wakeup_txfh(gNB_L1_rxtx_proc_t *proc,PHY_VARS_gNB *gNB) { struct timespec wait; wait.tv_sec=0; wait.tv_nsec=5000000L; -//printf("~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~inside wakeup_txfh %d.%d IC_RU = %d\n", proc->frame_tx, proc->subframe_tx, proc->instance_cnt_RUs); +//printf("~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~inside wakeup_txfh %d.%d IC_RU = %d\n", proc->frame_tx, proc->slot_tx, proc->instance_cnt_RUs); if(wait_on_condition(&proc->mutex_RUs,&proc->cond_RUs,&proc->instance_cnt_RUs,"wakeup_txfh")<0) { - LOG_E(PHY,"Frame %d, subframe %d: TX FH not ready\n", proc->frame_tx, proc->subframe_tx); + LOG_E(PHY,"Frame %d, subframe %d: TX FH not ready\n", proc->frame_tx, proc->slot_tx); return(-1); } pthread_mutex_lock(&gNB->proc.mutex_RU_tx); @@ -471,21 +471,21 @@ int wakeup_txfh(gNB_L1_rxtx_proc_t *proc,PHY_VARS_gNB *gNB) { ru = gNB->RU_list[i]; ru_proc = &ru->proc; if (ru_proc->instance_cnt_gNBs == 0) { - LOG_E(PHY,"Frame %d, subframe %d: TX FH thread busy, dropping Frame %d, subframe %d\n", ru_proc->frame_tx, ru_proc->subframe_tx, proc->frame_rx, proc->subframe_rx); + LOG_E(PHY,"Frame %d, subframe %d: TX FH thread busy, dropping Frame %d, subframe %d\n", ru_proc->frame_tx, ru_proc->tti_tx, proc->frame_rx, proc->slot_rx); return(-1); } if (pthread_mutex_timedlock(&ru_proc->mutex_gNBs,&wait) != 0) { - LOG_E( PHY, "[eNB] ERROR pthread_mutex_lock for eNB TX1 thread %d (IC %d)\n", ru_proc->subframe_rx&1,ru_proc->instance_cnt_gNBs ); + LOG_E( PHY, "[eNB] ERROR pthread_mutex_lock for eNB TX1 thread %d (IC %d)\n", ru_proc->tti_rx&1,ru_proc->instance_cnt_gNBs ); exit_fun( "error locking mutex_gNB" ); return(-1); } ru_proc->instance_cnt_gNBs = 0; ru_proc->timestamp_tx = proc->timestamp_tx; - ru_proc->subframe_tx = proc->subframe_tx; + ru_proc->tti_tx = proc->slot_tx; ru_proc->frame_tx = proc->frame_tx; - + LOG_I(PHY,"Signaling tx_thread_fh for %d.%d\n",ru_proc->frame_tx,ru_proc->tti_tx); // the thread can now be woken up if (pthread_cond_signal(&ru_proc->cond_gNBs) != 0) { LOG_E( PHY, "[gNB] ERROR pthread_cond_signal for gNB TXnp4 thread\n"); @@ -524,9 +524,9 @@ int wakeup_tx(PHY_VARS_gNB *gNB) { L1_proc_tx->instance_cnt = 0; - L1_proc_tx->subframe_rx = L1_proc->subframe_rx; + L1_proc_tx->slot_rx = L1_proc->slot_rx; L1_proc_tx->frame_rx = L1_proc->frame_rx; - L1_proc_tx->subframe_tx = L1_proc->subframe_tx; + L1_proc_tx->slot_tx = L1_proc->slot_tx; L1_proc_tx->frame_tx = L1_proc->frame_tx; L1_proc_tx->timestamp_tx = L1_proc->timestamp_tx; @@ -557,7 +557,7 @@ int wakeup_rxtx(PHY_VARS_gNB *gNB,RU_t *ru) { if (ru == gNB->RU_list[i]) { if ((proc->RU_mask&(1< 0) LOG_E(PHY,"gNB %d frame %d, subframe %d : previous information from RU %d (num_RU %d,mask %x) has not been served yet!\n", - gNB->Mod_id,proc->frame_rx,proc->subframe_rx,ru->idx,gNB->num_RU,proc->RU_mask); + gNB->Mod_id,proc->frame_rx,proc->slot_rx,ru->idx,gNB->num_RU,proc->RU_mask); proc->RU_mask |= (1<instance_cnt == 0; i++) { - LOG_W( PHY,"[gNB] Frame %d Subframe %d, gNB RXn-TXnp4 thread busy!! (cnt %i)\n", L1_proc->frame_tx, L1_proc->subframe_tx, L1_proc->instance_cnt); - usleep(500); + for (i = 0; i < fp->slots_per_frame && L1_proc->instance_cnt == 0; i++) { + LOG_W( PHY,"[gNB] SFN.SL %d.%d, gNB RXn-TXnp4 thread busy!! (i %d, cnt %i)\n", L1_proc->frame_tx, L1_proc->slot_tx, i,L1_proc->instance_cnt); + usleep(100); } if (L1_proc->instance_cnt == 0) { exit_fun( "TX thread busy" ); return(-1); } - // wake up TX for subframe n+sf_ahead + // wake up TX for subframe n+sl_ahead // lock the TX mutex and make sure the thread is ready if (pthread_mutex_timedlock(&L1_proc->mutex,&wait) != 0) { - LOG_E( PHY, "[gNB] ERROR pthread_mutex_lock for gNB RXTX thread %d (IC %d)\n", L1_proc->subframe_rx&1,L1_proc->instance_cnt ); + LOG_E( PHY, "[gNB] ERROR pthread_mutex_lock for gNB RXTX thread %d (IC %d)\n", L1_proc->slot_rx&1,L1_proc->instance_cnt ); exit_fun( "error locking mutex" ); return(-1); } @@ -599,15 +599,15 @@ int wakeup_rxtx(PHY_VARS_gNB *gNB,RU_t *ru) { // TS_rx is the last received timestamp (start of 1st slot), TS_tx is the desired // transmitted timestamp of the next TX slot (first). // The last (TS_rx mod samples_per_frame) was n*samples_per_tti, - // we want to generate subframe (n+sf_ahead), so TS_tx = TX_rx+sf_ahead*samples_per_tti, - // and proc->subframe_tx = proc->subframe_rx+sf_ahead - L1_proc->timestamp_tx = ru_proc->timestamp_rx + (sf_ahead*fp->samples_per_subframe); + // we want to generate subframe (n+sl_ahead), so TS_tx = TX_rx+sl_ahead*samples_per_tti, + // and proc->slot_tx = proc->slot_rx+sl_ahead + L1_proc->timestamp_tx = ru_proc->timestamp_rx + (sl_ahead*fp->samples_per_slot); L1_proc->frame_rx = ru_proc->frame_rx; - L1_proc->subframe_rx = ru_proc->subframe_rx; - L1_proc->frame_tx = (L1_proc->subframe_rx > (9-sf_ahead)) ? (L1_proc->frame_rx+1)&1023 : L1_proc->frame_rx; - L1_proc->subframe_tx = (L1_proc->subframe_rx + sf_ahead)%10; + L1_proc->slot_rx = ru_proc->tti_rx; + L1_proc->frame_tx = (L1_proc->slot_rx > (fp->slots_per_frame-1-sl_ahead)) ? (L1_proc->frame_rx+1)&1023 : L1_proc->frame_rx; + L1_proc->slot_tx = (L1_proc->slot_rx + sl_ahead)%fp->slots_per_frame; -//printf("~~~~~~~~~~~~~~~~~~~~~~passing parameter IC = %d, RX: %d.%d, TX: %d.%d to L1 sf_ahead = %d\n", L1_proc->instance_cnt, L1_proc->frame_rx, L1_proc->subframe_rx, L1_proc->frame_tx, L1_proc->subframe_tx, sf_ahead); + LOG_I(PHY,"wakeupL1: passing parameter IC = %d, RX: %d.%d, TX: %d.%d to L1 sl_ahead = %d\n", L1_proc->instance_cnt, L1_proc->frame_rx, L1_proc->slot_rx, L1_proc->frame_tx, L1_proc->slot_tx, sl_ahead); // the thread can now be woken up if (pthread_cond_signal(&L1_proc->cond) != 0) { @@ -904,87 +904,7 @@ void print_opp_meas(void) { print_meas(&softmodem_stats_rx_sf,"[gNB][total_phy_proc_rx]",NULL,NULL); } } -/* -void free_transport(PHY_VARS_gNB *gNB) -{ - int i; - int j; - for (i=0; idlsch[i][j]); - - LOG_I(PHY, "Freeing Transport Channel Buffer for ULSCH, UE %d\n",i); - free_gNB_ulsch(gNB->ulsch[1+i]); - } - free_gNB_ulsch(gNB->ulsch[0]); -}*/ - -/* -void init_transport(PHY_VARS_gNB *gNB) { - - int i; - int j; - NR_DL_FRAME_PARMS *fp = &gNB->frame_parms; - - LOG_I(PHY, "Initialise transport\n"); - - for (i=0; idlsch[i][j] = new_gNB_dlsch(1,8,NSOFT,fp->N_RB_DL,0,fp); - if (!gNB->dlsch[i][j]) { - LOG_E(PHY,"Can't get gNB dlsch structures for UE %d \n", i); - exit(-1); - } else { - gNB->dlsch[i][j]->rnti=0; - LOG_D(PHY,"dlsch[%d][%d] => %p rnti:%d\n",i,j,gNB->dlsch[i][j], gNB->dlsch[i][j]->rnti); - } - } - - LOG_I(PHY,"Allocating Transport Channel Buffer for ULSCH, UE %d\n",i); - gNB->ulsch[1+i] = new_gNB_ulsch(MAX_TURBO_ITERATIONS,fp->N_RB_UL, 0); - - if (!gNB->ulsch[1+i]) { - LOG_E(PHY,"Can't get gNB ulsch structures\n"); - exit(-1); - } - - // this is the transmission mode for the signalling channels - // this will be overwritten with the real transmission mode by the RRC once the UE is connected - gNB->transmission_mode[i] = fp->nb_antenna_ports_gNB==1 ? 1 : 2; - } - // ULSCH for RA - gNB->ulsch[0] = new_gNB_ulsch(MAX_TURBO_ITERATIONS, fp->N_RB_UL, 0); - - if (!gNB->ulsch[0]) { - LOG_E(PHY,"Can't get gNB ulsch structures\n"); - exit(-1); - } - gNB->dlsch_SI = new_gNB_dlsch(1,8,NSOFT,fp->N_RB_DL, 0, fp); - LOG_D(PHY,"gNB %d.%d : SI %p\n",gNB->Mod_id,gNB->CC_id,gNB->dlsch_SI); - gNB->dlsch_ra = new_gNB_dlsch(1,8,NSOFT,fp->N_RB_DL, 0, fp); - LOG_D(PHY,"gNB %d.%d : RA %p\n",gNB->Mod_id,gNB->CC_id,gNB->dlsch_ra); - gNB->dlsch_MCH = new_gNB_dlsch(1,8,NSOFT,fp->N_RB_DL, 0, fp); - LOG_D(PHY,"gNB %d.%d : MCH %p\n",gNB->Mod_id,gNB->CC_id,gNB->dlsch_MCH); - - - gNB->rx_total_gain_dB=130; - - for(i=0; imu_mimo_mode[i].dl_pow_off = 2; - - gNB->check_for_total_transmissions = 0; - - gNB->check_for_MUMIMO_transmissions = 0; - - gNB->FULL_MUMIMO_transmissions = 0; - - gNB->check_for_SUMIMO_transmissions = 0; - - fp->pucch_config_common.deltaPUCCH_Shift = 1; - -} */ /// eNB kept in function name for nffapi calls, TO FIX void init_eNB_afterRU(void) { @@ -1005,7 +925,6 @@ void init_eNB_afterRU(void) { // map antennas and PRACH signals to gNB RX if (0) AssertFatal(gNB->num_RU>0,"Number of RU attached to gNB %d is zero\n",gNB->Mod_id); LOG_I(PHY,"Mapping RX ports from %d RUs to gNB %d\n",gNB->num_RU,gNB->Mod_id); - gNB->gNB_config.rf_config.tx_antenna_ports.value = 0; //LOG_I(PHY,"Overwriting gNB->prach_vars.rxsigF[0]:%p\n", gNB->prach_vars.rxsigF[0]); @@ -1014,7 +933,6 @@ void init_eNB_afterRU(void) { LOG_I(PHY,"gNB->num_RU:%d\n", gNB->num_RU); for (ru_id=0,aa=0;ru_idnum_RU;ru_id++) { - gNB->gNB_config.rf_config.tx_antenna_ports.value += gNB->RU_list[ru_id]->nb_rx; AssertFatal(gNB->RU_list[ru_id]->common.rxdataF!=NULL, "RU %d : common.rxdataF is NULL\n", @@ -1037,31 +955,7 @@ void init_eNB_afterRU(void) { * In monolithic mode, we come here with nb_antennas_rx == 0 * (not tested in other modes). */ - if (gNB->gNB_config.rf_config.tx_antenna_ports.value < 1) - { - LOG_I(PHY, "%s() ************* DJP ***** gNB->gNB_config.rf_config.tx_antenna_ports:%d - GOING TO HARD CODE TO 1", __FUNCTION__, gNB->gNB_config.rf_config.tx_antenna_ports.value); - gNB->gNB_config.rf_config.tx_antenna_ports.value = 1; - } - else - { - //LOG_I(PHY," Delete code\n"); - } - if (gNB->gNB_config.rf_config.tx_antenna_ports.value < 1) - { - LOG_I(PHY, "%s() ************* DJP ***** gNB->gNB_config.rf_config.tx_antenna_ports:%d - GOING TO HARD CODE TO 1", __FUNCTION__, gNB->gNB_config.rf_config.tx_antenna_ports.value); - gNB->gNB_config.rf_config.tx_antenna_ports.value = 1; - } - else - { - //LOG_I(PHY," Delete code\n"); - } - - AssertFatal(gNB->gNB_config.rf_config.tx_antenna_ports.value >0, - "inst %d, CC_id %d : nb_antennas_rx %d\n",inst,CC_id,gNB->gNB_config.rf_config.tx_antenna_ports.value); - LOG_I(PHY,"inst %d, CC_id %d : nb_antennas_rx %d\n",inst,CC_id,gNB->gNB_config.rf_config.tx_antenna_ports.value); -/// Transport init necessary for NR synchro - //init_transport(gNB); //init_precoding_weights(RC.gNB[inst][CC_id]); } init_gNB_proc(inst); diff --git a/targets/RT/USER/nr-ru.c b/targets/RT/USER/nr-ru.c index 3b22f901cc..af3916604e 100644 --- a/targets/RT/USER/nr-ru.c +++ b/targets/RT/USER/nr-ru.c @@ -7,7 +7,6 @@ the Free Software Foundation, either version 3 of the License, or (at your option) any later version. - OpenAirInterface is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the @@ -135,7 +134,7 @@ int attach_rru(RU_t *ru); int connect_rau(RU_t *ru); -extern uint16_t sf_ahead; +extern uint16_t sl_ahead; extern int emulate_rf; extern int numerology; @@ -309,40 +308,40 @@ int connect_rau(RU_t *ru) { // southbound IF5 fronthaul for 16-bit OAI format static inline void fh_if5_south_out(RU_t *ru) { if (ru == RC.ru[0]) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, ru->proc.timestamp_tx&0xffffffff ); - send_IF5(ru, ru->proc.timestamp_tx, ru->proc.subframe_tx, &ru->seqno, IF5_RRH_GW_DL); + send_IF5(ru, ru->proc.timestamp_tx, ru->proc.tti_tx, &ru->seqno, IF5_RRH_GW_DL); } // southbound IF5 fronthaul for Mobipass packet format static inline void fh_if5_mobipass_south_out(RU_t *ru) { if (ru == RC.ru[0]) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, ru->proc.timestamp_tx&0xffffffff ); - send_IF5(ru, ru->proc.timestamp_tx, ru->proc.subframe_tx, &ru->seqno, IF5_MOBIPASS); + send_IF5(ru, ru->proc.timestamp_tx, ru->proc.tti_tx, &ru->seqno, IF5_MOBIPASS); } // southbound IF4p5 fronthaul static inline void fh_if4p5_south_out(RU_t *ru) { if (ru == RC.ru[0]) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, ru->proc.timestamp_tx&0xffffffff ); - LOG_D(PHY,"Sending IF4p5 for frame %d subframe %d\n",ru->proc.frame_tx,ru->proc.subframe_tx); - if (nr_subframe_select(&ru->gNB_list[0]->gNB_config,ru->proc.subframe_tx)!=SF_UL) - send_IF4p5(ru,ru->proc.frame_tx, ru->proc.subframe_tx, IF4p5_PDLFFT); + LOG_D(PHY,"Sending IF4p5 for frame %d subframe %d\n",ru->proc.frame_tx,ru->proc.tti_tx); + if (nr_slot_select(&ru->gNB_list[0]->gNB_config,ru->proc.tti_tx)!=SF_UL) + send_IF4p5(ru,ru->proc.frame_tx, ru->proc.tti_tx, IF4p5_PDLFFT); } /*************************************************************/ /* Input Fronthaul from south RCC/RAU */ // Synchronous if5 from south -void fh_if5_south_in(RU_t *ru,int *frame, int *subframe) { +void fh_if5_south_in(RU_t *ru,int *frame, int *tti) { NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; RU_proc_t *proc = &ru->proc; - recv_IF5(ru, &proc->timestamp_rx, *subframe, IF5_RRH_GW_UL); + recv_IF5(ru, &proc->timestamp_rx, *tti, IF5_RRH_GW_UL); - proc->frame_rx = (proc->timestamp_rx / (fp->samples_per_subframe*10))&1023; - proc->subframe_rx = (proc->timestamp_rx / fp->samples_per_subframe)%10; + proc->frame_rx = (proc->timestamp_rx / (fp->samples_per_slot*20))&1023; + proc->tti_rx = (proc->timestamp_rx / fp->samples_per_slot)%20; if (proc->first_rx == 0) { - if (proc->subframe_rx != *subframe){ - LOG_E(PHY,"Received Timestamp doesn't correspond to the time we think it is (proc->subframe_rx %d, subframe %d)\n",proc->subframe_rx,*subframe); + if (proc->tti_rx != *tti){ + LOG_E(PHY,"Received Timestamp doesn't correspond to the time we think it is (proc->tti_rx %d, subframe %d)\n",proc->tti_rx,*tti); exit_fun("Exiting"); } @@ -353,7 +352,7 @@ void fh_if5_south_in(RU_t *ru,int *frame, int *subframe) { } else { proc->first_rx = 0; *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *tti = proc->tti_rx; } VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TS, proc->timestamp_rx&0xffffffff ); @@ -361,49 +360,49 @@ void fh_if5_south_in(RU_t *ru,int *frame, int *subframe) { } // Synchronous if4p5 from south -void fh_if4p5_south_in(RU_t *ru,int *frame,int *subframe) { +void fh_if4p5_south_in(RU_t *ru,int *frame,int *slot) { NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; RU_proc_t *proc = &ru->proc; - int f,sf; + int f,sl; uint16_t packet_type; uint32_t symbol_number=0; uint32_t symbol_mask_full=0; /* - if ((fp->frame_type == TDD) && (subframe_select(fp,*subframe)==SF_S)) + if ((fp->frame_type == TDD) && (subframe_select(fp,*slot)==SF_S)) symbol_mask_full = (1<ul_symbols_in_S_subframe)-1; else - symbol_mask_full = (1<symbols_per_tti)-1; + symbol_mask_full = (1<symbols_per_slot)-1; - AssertFatal(proc->symbol_mask[*subframe]==0,"rx_fh_if4p5: proc->symbol_mask[%d] = %x\n",*subframe,proc->symbol_mask[*subframe]);*/ + AssertFatal(proc->symbol_mask[*slot]==0,"rx_fh_if4p5: proc->symbol_mask[%d] = %x\n",*slot,proc->symbol_mask[*slot]);*/ do { // Blocking, we need a timeout on this !!!!!!!!!!!!!!!!!!!!!!! - recv_IF4p5(ru, &f, &sf, &packet_type, &symbol_number); + recv_IF4p5(ru, &f, &sl, &packet_type, &symbol_number); - if (packet_type == IF4p5_PULFFT) proc->symbol_mask[sf] = proc->symbol_mask[sf] | (1<symbol_mask[sl] = proc->symbol_mask[sl] | (1<first_rx==0) && (f!=*frame)) LOG_E(PHY,"rx_fh_if4p5: PULTICK received frame %d != expected %d\n",f,*frame); - if ((proc->first_rx==0) && (sf!=*subframe)) LOG_E(PHY,"rx_fh_if4p5: PULTICK received subframe %d != expected %d (first_rx %d)\n",sf,*subframe,proc->first_rx); + if ((proc->first_rx==0) && (sl!=*slot)) LOG_E(PHY,"rx_fh_if4p5: PULTICK received subframe %d != expected %d (first_rx %d)\n",sl,*slot,proc->first_rx); break; } else if (packet_type == IF4p5_PRACH) { // nothing in RU for RAU } - LOG_D(PHY,"rx_fh_if4p5: subframe %d symbol mask %x\n",*subframe,proc->symbol_mask[*subframe]); - } while(proc->symbol_mask[*subframe] != symbol_mask_full); + LOG_D(PHY,"rx_fh_if4p5: subframe %d symbol mask %x\n",*slot,proc->symbol_mask[sl]); + } while(proc->symbol_mask[sl] != symbol_mask_full); //caculate timestamp_rx, timestamp_tx based on frame and subframe - proc->subframe_rx = sf; + proc->tti_rx = sl; proc->frame_rx = f; - proc->timestamp_rx = ((proc->frame_rx * 10) + proc->subframe_rx ) * fp->samples_per_subframe ; + proc->timestamp_rx = ((proc->frame_rx * fp->slots_per_frame) + proc->tti_rx ) * fp->samples_per_slot ; // proc->timestamp_tx = proc->timestamp_rx + (4*fp->samples_per_subframe); - proc->subframe_tx = (sf+sf_ahead)%10; - proc->frame_tx = (sf>(9-sf_ahead)) ? (f+1)&1023 : f; + proc->tti_tx = (sl+sl_ahead)%fp->slots_per_frame; + proc->frame_tx = (sl>(fp->slots_per_frame-sl_ahead)) ? (f+1)&1023 : f; if (proc->first_rx == 0) { - if (proc->subframe_rx != *subframe){ - LOG_E(PHY,"Received Timestamp (IF4p5) doesn't correspond to the time we think it is (proc->subframe_rx %d, subframe %d)\n",proc->subframe_rx,*subframe); + if (proc->tti_rx != *slot){ + LOG_E(PHY,"Received Timestamp (IF4p5) doesn't correspond to the time we think it is (proc->tti_rx %d, subframe %d)\n",proc->tti_rx,*slot); exit_fun("Exiting"); } if (proc->frame_rx != *frame) { @@ -413,81 +412,23 @@ void fh_if4p5_south_in(RU_t *ru,int *frame,int *subframe) { } else { proc->first_rx = 0; *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *slot = proc->tti_rx; } if (ru == RC.ru[0]) { VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX0_RU, f ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_RU, sf ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX0_RU, sl); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, proc->frame_tx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, proc->subframe_tx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, proc->tti_tx ); } - proc->symbol_mask[sf] = 0; + proc->symbol_mask[proc->tti_rx] = 0; VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TS, proc->timestamp_rx&0xffffffff ); LOG_D(PHY,"RU %d: fh_if4p5_south_in sleeping ...\n",ru->idx); - usleep(100); } -// Dummy FH from south for getting synchronization from master RU -void fh_slave_south_in(RU_t *ru,int *frame,int *subframe) { - // This case is for synchronization to another thread - // it just waits for an external event. The actual rx_fh is handle by the asynchronous RX thread - RU_proc_t *proc=&ru->proc; - - if (wait_on_condition(&proc->mutex_FH,&proc->cond_FH,&proc->instance_cnt_FH,"fh_slave_south_in") < 0) - return; - - release_thread(&proc->mutex_FH,&proc->instance_cnt_FH,"rx_fh_slave_south_in"); - - -} - -// asynchronous inbound if5 fronthaul from south (Mobipass) -void fh_if5_south_asynch_in_mobipass(RU_t *ru,int *frame,int *subframe) { - - RU_proc_t *proc = &ru->proc; - NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; - - recv_IF5(ru, &proc->timestamp_rx, *subframe, IF5_MOBIPASS); - pthread_mutex_lock(&proc->mutex_asynch_rxtx); - int offset_mobipass = 40120; - pthread_mutex_lock(&proc->mutex_asynch_rxtx); - proc->subframe_rx = ((proc->timestamp_rx-offset_mobipass)/fp->samples_per_subframe)%10; - proc->frame_rx = ((proc->timestamp_rx-offset_mobipass)/(fp->samples_per_subframe*10))&1023; - - proc->subframe_rx = (proc->timestamp_rx/fp->samples_per_subframe)%10; - proc->frame_rx = (proc->timestamp_rx/(10*fp->samples_per_subframe))&1023; - - if (proc->first_rx == 1) { - proc->first_rx =2; - *subframe = proc->subframe_rx; - *frame = proc->frame_rx; - LOG_E(PHY,"[Mobipass]timestamp_rx:%llu, frame_rx %d, subframe: %d\n",(unsigned long long int)proc->timestamp_rx,proc->frame_rx,proc->subframe_rx); - } - else { - if (proc->subframe_rx != *subframe) { - proc->first_rx++; - LOG_E(PHY,"[Mobipass]timestamp:%llu, subframe_rx %d is not what we expect %d, first_rx:%d\n",(unsigned long long int)proc->timestamp_rx, proc->subframe_rx,*subframe, proc->first_rx); - //exit_fun("Exiting"); - } - if (proc->frame_rx != *frame) { - proc->first_rx++; - LOG_E(PHY,"[Mobipass]timestamp:%llu, frame_rx %d is not what we expect %d, first_rx:%d\n",(unsigned long long int)proc->timestamp_rx,proc->frame_rx,*frame, proc->first_rx); - // exit_fun("Exiting"); - } - // temporary solution - *subframe = proc->subframe_rx; - *frame = proc->frame_rx; - } - - pthread_mutex_unlock(&proc->mutex_asynch_rxtx); - - -} // gNodeB_3GPP_BBU - // asynchronous inbound if4p5 fronthaul from south -void fh_if4p5_south_asynch_in(RU_t *ru,int *frame,int *subframe) { +void fh_if4p5_south_asynch_in(RU_t *ru,int *frame,int *slot) { NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; RU_proc_t *proc = &ru->proc; @@ -497,19 +438,19 @@ void fh_if4p5_south_asynch_in(RU_t *ru,int *frame,int *subframe) { // uint32_t got_prach_info=0; symbol_number = 0; - symbol_mask = (1<<(fp->symbols_per_slot * fp->slots_per_subframe))-1; + symbol_mask = (1<<(fp->symbols_per_slot))-1; prach_rx = 0; do { // Blocking, we need a timeout on this !!!!!!!!!!!!!!!!!!!!!!! - recv_IF4p5(ru, &proc->frame_rx, &proc->subframe_rx, &packet_type, &symbol_number); + recv_IF4p5(ru, &proc->frame_rx, &proc->tti_rx, &packet_type, &symbol_number); // grab first prach information for this new subframe /*if (got_prach_info==0) { - prach_rx = is_prach_subframe(fp, proc->frame_rx, proc->subframe_rx); + prach_rx = is_prach_subframe(fp, proc->frame_rx, proc->tti_rx); got_prach_info = 1; }*/ if (proc->first_rx != 0) { *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *slot = proc->tti_rx; proc->first_rx = 0; } else { @@ -517,8 +458,8 @@ void fh_if4p5_south_asynch_in(RU_t *ru,int *frame,int *subframe) { LOG_E(PHY,"frame_rx %d is not what we expect %d\n",proc->frame_rx,*frame); exit_fun("Exiting"); } - if (proc->subframe_rx != *subframe) { - LOG_E(PHY,"subframe_rx %d is not what we expect %d\n",proc->subframe_rx,*subframe); + if (proc->tti_rx != *slot) { + LOG_E(PHY,"tti_rx %d is not what we expect %d\n",proc->tti_rx,*slot); exit_fun("Exiting"); } } @@ -536,7 +477,7 @@ void fh_if4p5_south_asynch_in(RU_t *ru,int *frame,int *subframe) { // RRU IF4p5 TX fronthaul receiver. Assumes an if_device on input and if or rf device on output // receives one subframe's worth of IF4p5 OFDM symbols and OFDM modulates -void fh_if4p5_north_in(RU_t *ru,int *frame,int *subframe) { +void fh_if4p5_north_in(RU_t *ru,int *frame,int *slot) { uint32_t symbol_number=0; uint32_t symbol_mask, symbol_mask_full; @@ -546,47 +487,47 @@ void fh_if4p5_north_in(RU_t *ru,int *frame,int *subframe) { /// **** incoming IF4p5 from remote RCC/RAU **** /// symbol_number = 0; symbol_mask = 0; - symbol_mask_full = (1<<(ru->nr_frame_parms->symbols_per_slot* ru->nr_frame_parms->slots_per_subframe))-1; + symbol_mask_full = (1<<(ru->nr_frame_parms->symbols_per_slot))-1; do { - recv_IF4p5(ru, frame, subframe, &packet_type, &symbol_number); + recv_IF4p5(ru, frame, slot, &packet_type, &symbol_number); symbol_mask = symbol_mask | (1<nr_frame_parms; RU_proc_t *proc = &ru->proc; - int subframe_tx,frame_tx; + int tti_tx,frame_tx; openair0_timestamp timestamp_tx; - recv_IF5(ru, ×tamp_tx, *subframe, IF5_RRH_GW_DL); - // printf("Received subframe %d (TS %llu) from RCC\n",subframe_tx,timestamp_tx); + recv_IF5(ru, ×tamp_tx, *slot, IF5_RRH_GW_DL); + // printf("Received subframe %d (TS %llu) from RCC\n",tti_tx,timestamp_tx); - subframe_tx = (timestamp_tx/fp->samples_per_subframe)%10; - frame_tx = (timestamp_tx/(fp->samples_per_subframe*10))&1023; + tti_tx = (timestamp_tx/fp->samples_per_slot)%fp->slots_per_frame; + frame_tx = (timestamp_tx/(fp->samples_per_slot*fp->slots_per_frame))&1023; if (proc->first_tx != 0) { - *subframe = subframe_tx; + *slot = tti_tx; *frame = frame_tx; proc->first_tx = 0; } else { - AssertFatal(subframe_tx == *subframe, - "subframe_tx %d is not what we expect %d\n",subframe_tx,*subframe); + AssertFatal(tti_tx == *slot, + "tti_tx %d is not what we expect %d\n",tti_tx,*slot); AssertFatal(frame_tx == *frame, "frame_tx %d is not what we expect %d\n",frame_tx,*frame); } } -void fh_if4p5_north_asynch_in(RU_t *ru,int *frame,int *subframe) { +void fh_if4p5_north_asynch_in(RU_t *ru,int *frame,int *slot) { NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; nfapi_nr_config_request_t *cfg = &ru->gNB_list[0]->gNB_config; @@ -594,28 +535,28 @@ void fh_if4p5_north_asynch_in(RU_t *ru,int *frame,int *subframe) { uint16_t packet_type; uint32_t symbol_number,symbol_mask,symbol_mask_full=0; - int subframe_tx,frame_tx; + int slot_tx,frame_tx; LOG_D(PHY, "%s(ru:%p frame, subframe)\n", __FUNCTION__, ru); symbol_number = 0; symbol_mask = 0; -// symbol_mask_full = ((subframe_select(fp,*subframe) == SF_S) ? (1<dl_symbols_in_S_subframe) : (1<symbols_per_tti))-1; +// symbol_mask_full = ((subframe_select(fp,*slot) == SF_S) ? (1<dl_symbols_in_S_subframe) : (1<symbols_per_slot))-1; do { - recv_IF4p5(ru, &frame_tx, &subframe_tx, &packet_type, &symbol_number); - if ((nr_subframe_select(cfg,subframe_tx) == SF_DL) && (symbol_number == 0)) start_meas(&ru->rx_fhaul); + recv_IF4p5(ru, &frame_tx, &slot_tx, &packet_type, &symbol_number); + if ((nr_slot_select(cfg,slot_tx) == SF_DL) && (symbol_number == 0)) start_meas(&ru->rx_fhaul); LOG_D(PHY,"subframe %d (%d): frame %d, subframe %d, symbol %d\n", - *subframe,nr_subframe_select(cfg,*subframe),frame_tx,subframe_tx,symbol_number); + *slot,nr_slot_select(cfg,*slot),frame_tx,slot_tx,symbol_number); if (proc->first_tx != 0) { - *frame = frame_tx; - *subframe = subframe_tx; + *frame = frame_tx; + *slot = slot_tx; proc->first_tx = 0; - //symbol_mask_full = ((subframe_select(fp,*subframe) == SF_S) ? (1<dl_symbols_in_S_subframe) : (1<symbols_per_tti))-1; + //symbol_mask_full = ((subframe_select(fp,*slot) == SF_S) ? (1<dl_symbols_in_S_subframe) : (1<symbols_per_slot))-1; } else { AssertFatal(frame_tx == *frame, "frame_tx %d is not what we expect %d\n",frame_tx,*frame); - AssertFatal(subframe_tx == *subframe, - "subframe_tx %d is not what we expect %d\n",subframe_tx,*subframe); + AssertFatal(slot_tx == *slot, + "slot_tx %d is not what we expect %d\n",slot_tx,*slot); } if (packet_type == IF4p5_PDLFFT) { symbol_mask = symbol_mask | (1<rx_fhaul); + if (nr_slot_select(cfg,slot_tx) == SF_DL) stop_meas(&ru->rx_fhaul); - proc->subframe_tx = subframe_tx; + proc->tti_tx = slot_tx; proc->frame_tx = frame_tx; - if ((frame_tx == 0)&&(subframe_tx == 0)) proc->frame_tx_unwrap += 1024; + if ((frame_tx == 0)&&(slot_tx == 0)) proc->frame_tx_unwrap += 1024; - proc->timestamp_tx = ((((uint64_t)frame_tx + (uint64_t)proc->frame_tx_unwrap) * 10) + (uint64_t)subframe_tx) * (uint64_t)fp->samples_per_subframe; + proc->timestamp_tx = ((((uint64_t)frame_tx + (uint64_t)proc->frame_tx_unwrap) * fp->slots_per_frame) + (uint64_t)slot_tx) * (uint64_t)fp->samples_per_slot; - LOG_D(PHY,"RU %d/%d TST %llu, frame %d, subframe %d\n",ru->idx,0,(long long unsigned int)proc->timestamp_tx,frame_tx,subframe_tx); + LOG_D(PHY,"RU %d/%d TST %llu, frame %d, subframe %d\n",ru->idx,0,(long long unsigned int)proc->timestamp_tx,frame_tx,slot_tx); // dump VCD output for first RU in list if (ru == RC.ru[0]) { VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, frame_tx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, subframe_tx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, slot_tx ); } if (ru->feptx_ofdm) ru->feptx_ofdm(ru); @@ -650,7 +591,7 @@ void fh_if5_north_out(RU_t *ru) { /// **** send_IF5 of rxdata to BBU **** /// VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_SEND_IF5, 1 ); - send_IF5(ru, proc->timestamp_rx, proc->subframe_rx, &seqno, IF5_RRH_GW_UL); + send_IF5(ru, proc->timestamp_rx, proc->tti_rx, &seqno, IF5_RRH_GW_UL); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_SEND_IF5, 0 ); } @@ -660,17 +601,17 @@ void fh_if4p5_north_out(RU_t *ru) { RU_proc_t *proc=&ru->proc; //NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; - //const int subframe = proc->subframe_rx; - if (ru->idx==0) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_RU, proc->subframe_rx ); + //const int subframe = proc->tti_rx; + if (ru->idx==0) VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX0_RU, proc->tti_rx ); /* if ((fp->frame_type == TDD) && (subframe_select(fp,subframe)!=SF_UL)) { /// **** in TDD during DL send_IF4 of ULTICK to RCC **** /// - send_IF4p5(ru, proc->frame_rx, proc->subframe_rx, IF4p5_PULTICK); + send_IF4p5(ru, proc->frame_rx, proc->tti_rx, IF4p5_PULTICK); return; }*/ start_meas(&ru->tx_fhaul); - send_IF4p5(ru, proc->frame_rx, proc->subframe_rx, IF4p5_PULFFT); + send_IF4p5(ru, proc->frame_rx, proc->tti_rx, IF4p5_PULFFT); stop_meas(&ru->tx_fhaul); } @@ -703,7 +644,7 @@ static void* emulatedRF_thread(void* param) { return 0; } -void rx_rf(RU_t *ru,int *frame,int *subframe) { +void rx_rf(RU_t *ru,int *frame,int *slot) { RU_proc_t *proc = &ru->proc; NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; @@ -711,25 +652,28 @@ void rx_rf(RU_t *ru,int *frame,int *subframe) { unsigned int rxs; int i; openair0_timestamp ts,old_ts; - + + AssertFatal(*slotslots_per_frame && *slot>=0, "slot %d is illegal (%d)\n",*slot,fp->slots_per_frame); for (i=0; inb_rx; i++) - rxp[i] = (void*)&ru->common.rxdata[i][*subframe*fp->samples_per_subframe]; + rxp[i] = (void*)&ru->common.rxdata[i][*slot*fp->samples_per_slot]; VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_TRX_READ, 1 ); old_ts = proc->timestamp_rx; + LOG_I(PHY,"Reading %d samples for slot %d (%p)\n",fp->samples_per_slot,*slot,rxp[0]); + if(emulate_rf){ wait_on_condition(&proc->mutex_emulateRF,&proc->cond_emulateRF,&proc->instance_cnt_emulateRF,"emulatedRF_thread"); release_thread(&proc->mutex_emulateRF,&proc->instance_cnt_emulateRF,"emulatedRF_thread"); - rxs = fp->samples_per_subframe; + rxs = fp->samples_per_slot; ts = old_ts + rxs; } else{ rxs = ru->rfdevice.trx_read_func(&ru->rfdevice, &ts, rxp, - fp->samples_per_subframe, + fp->samples_per_slot, ru->nb_rx); } @@ -739,43 +683,43 @@ void rx_rf(RU_t *ru,int *frame,int *subframe) { //AssertFatal(rxs == fp->samples_per_subframe, //"rx_rf: Asked for %d samples, got %d from USRP\n",fp->samples_per_subframe,rxs); - if (rxs != fp->samples_per_subframe) LOG_E(PHY, "rx_rf: Asked for %d samples, got %d from USRP\n",fp->samples_per_subframe,rxs); + if (rxs != fp->samples_per_slot) LOG_E(PHY, "rx_rf: Asked for %d samples, got %d from USRP\n",fp->samples_per_slot,rxs); if (proc->first_rx == 1) { ru->ts_offset = proc->timestamp_rx; proc->timestamp_rx = 0; } else { - if (proc->timestamp_rx - old_ts != fp->samples_per_subframe) { - LOG_I(PHY,"rx_rf: rfdevice timing drift of %"PRId64" samples (ts_off %"PRId64")\n",proc->timestamp_rx - old_ts - fp->samples_per_subframe,ru->ts_offset); - ru->ts_offset += (proc->timestamp_rx - old_ts - fp->samples_per_subframe); + if (proc->timestamp_rx - old_ts != fp->samples_per_slot) { + LOG_I(PHY,"rx_rf: rfdevice timing drift of %"PRId64" samples (ts_off %"PRId64")\n",proc->timestamp_rx - old_ts - fp->samples_per_slot,ru->ts_offset); + ru->ts_offset += (proc->timestamp_rx - old_ts - fp->samples_per_slot); proc->timestamp_rx = ts-ru->ts_offset; } } - proc->frame_rx = (proc->timestamp_rx / (fp->samples_per_subframe*10))&1023; - proc->subframe_rx = (proc->timestamp_rx / fp->samples_per_subframe)%10; + proc->frame_rx = (proc->timestamp_rx / (fp->samples_per_slot*fp->slots_per_frame))&1023; + proc->tti_rx = (proc->timestamp_rx / fp->samples_per_slot)%fp->slots_per_frame; // synchronize first reception to frame 0 subframe 0 - proc->timestamp_tx = proc->timestamp_rx+(sf_ahead*fp->samples_per_subframe); - proc->subframe_tx = (proc->subframe_rx+sf_ahead)%10; - proc->frame_tx = (proc->subframe_rx>(9-sf_ahead)) ? (proc->frame_rx+1)&1023 : proc->frame_rx; + proc->timestamp_tx = proc->timestamp_rx+(sl_ahead*fp->samples_per_slot); + proc->tti_tx = (proc->tti_rx+sl_ahead)%fp->slots_per_frame; + proc->frame_tx = (proc->tti_rx>(fp->slots_per_frame-1-sl_ahead)) ? (proc->frame_rx+1)&1023 : proc->frame_rx; - LOG_D(PHY,"RU %d/%d TS %llu (off %d), frame %d, subframe %d\n", + LOG_I(PHY,"RU %d/%d TS %llu (off %d), frame %d, slot %d.%d / %d\n", ru->idx, 0, (unsigned long long int)proc->timestamp_rx, - (int)ru->ts_offset,proc->frame_rx,proc->subframe_rx); + (int)ru->ts_offset,proc->frame_rx,proc->tti_rx,proc->tti_tx,fp->slots_per_frame); // dump VCD output for first RU in list if (ru == RC.ru[0]) { VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_RX0_RU, proc->frame_rx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_RX0_RU, proc->subframe_rx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_RX0_RU, proc->tti_rx ); } if (proc->first_rx == 0) { - if (proc->subframe_rx != *subframe){ - LOG_E(PHY,"Received Timestamp (%llu) doesn't correspond to the time we think it is (proc->subframe_rx %d, subframe %d)\n",(long long unsigned int)proc->timestamp_rx,proc->subframe_rx,*subframe); + if (proc->tti_rx != *slot){ + LOG_E(PHY,"Received Timestamp (%llu) doesn't correspond to the time we think it is (proc->tti_rx %d, subframe %d)\n",(long long unsigned int)proc->timestamp_rx,proc->tti_rx,*slot); exit_fun("Exiting"); } @@ -786,14 +730,14 @@ void rx_rf(RU_t *ru,int *frame,int *subframe) { } else { proc->first_rx = 0; *frame = proc->frame_rx; - *subframe = proc->subframe_rx; + *slot = proc->tti_rx; } - //printf("timestamp_rx %lu, frame %d(%d), subframe %d(%d)\n",ru->timestamp_rx,proc->frame_rx,frame,proc->subframe_rx,subframe); + //printf("timestamp_rx %lu, frame %d(%d), subframe %d(%d)\n",ru->timestamp_rx,proc->frame_rx,frame,proc->tti_rx,subframe); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TS, proc->timestamp_rx&0xffffffff ); - if (rxs != fp->samples_per_subframe) + if (rxs != fp->samples_per_slot) { //exit_fun( "problem receiving samples" ); LOG_E(PHY, "problem receiving samples\n"); @@ -810,18 +754,17 @@ void tx_rf(RU_t *ru) { unsigned int txs; int i; - T(T_ENB_PHY_OUTPUT_SIGNAL, T_INT(0), T_INT(0), T_INT(proc->frame_tx), T_INT(proc->subframe_tx), - T_INT(0), T_BUFFER(&ru->common.txdata[0][proc->subframe_tx * fp->samples_per_subframe], fp->samples_per_subframe * 4)); + T(T_ENB_PHY_OUTPUT_SIGNAL, T_INT(0), T_INT(0), T_INT(proc->frame_tx), T_INT(proc->tti_tx), + T_INT(0), T_BUFFER(&ru->common.txdata[0][proc->tti_tx * fp->samples_per_slot], fp->samples_per_slot * 4)); - nr_subframe_t SF_type = nr_subframe_select(cfg,proc->subframe_tx%10); - /*nr_subframe_t prevSF_type = nr_subframe_select(cfg,(proc->subframe_tx+9)%10); - nr_subframe_t nextSF_type = nr_subframe_select(cfg,(proc->subframe_tx+1)%10);*/ + nr_subframe_t SF_type = nr_slot_select(cfg,proc->tti_tx%fp->slots_per_frame); int sf_extension = 0; if ((SF_type == SF_DL) || (SF_type == SF_S)) { - int siglen=fp->samples_per_subframe,flags=1; + int siglen=fp->samples_per_slot,flags=1; + /* if (SF_type == SF_S) { siglen = fp->dl_symbols_in_S_subframe*(fp->ofdm_symbol_size+fp->nb_prefix_samples0); @@ -843,16 +786,12 @@ void tx_rf(RU_t *ru) { sf_extension = ru->N_TA_offset<<1; } */ VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, proc->frame_tx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, proc->subframe_tx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, proc->tti_tx ); for (i=0; inb_tx; i++) - txp[i] = (void*)&ru->common.txdata[i][(proc->subframe_tx*fp->samples_per_subframe)-sf_extension]; -/* - if (proc->subframe_tx == 0){ - write_output("txdataF_frame.m","txdataF_frame",&ru->common.txdataF_BF[i],fp->samples_per_subframe_wCP, 1, 1); - write_output("txdata_frame.m","txdata_frame",&ru->common.txdata[i],fp->samples_per_subframe, 1, 1); - } -*/ + + txp[i] = (void*)&ru->common.txdata[i][(proc->tti_tx*fp->samples_per_slot)-sf_extension]; + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TRX_TST, (proc->timestamp_tx-ru->openair0_cfg.tx_sample_advance)&0xffffffff ); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_TRX_WRITE, 1 ); // prepare tx buffer pointers @@ -864,7 +803,7 @@ void tx_rf(RU_t *ru) { flags); LOG_D(PHY,"[TXPATH] RU %d tx_rf, writing to TS %llu, frame %d, unwrapped_frame %d, subframe %d\n",ru->idx, - (long long unsigned int)proc->timestamp_tx,proc->frame_tx,proc->frame_tx_unwrap,proc->subframe_tx); + (long long unsigned int)proc->timestamp_tx,proc->frame_tx,proc->frame_tx_unwrap,proc->tti_tx); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME( VCD_SIGNAL_DUMPER_FUNCTIONS_TRX_WRITE, 0 ); @@ -921,7 +860,7 @@ static void* ru_thread_asynch_rxtx( void* param ) { if (ru->fh_south_asynch_in) ru->fh_south_asynch_in(ru,&frame,&subframe); // asynchronous receive from north (RRU IF4/IF5) else if (ru->fh_north_asynch_in) { - if (nr_subframe_select(&ru->gNB_list[0]->gNB_config,subframe)!=SF_UL) + if (nr_slot_select(&ru->gNB_list[0]->gNB_config,subframe)!=SF_UL) ru->fh_north_asynch_in(ru,&frame,&subframe); } else AssertFatal(1==0,"Unknown function in ru_thread_asynch_rxtx\n"); @@ -934,47 +873,6 @@ static void* ru_thread_asynch_rxtx( void* param ) { -void wakeup_slaves(RU_proc_t *proc) { - - int i; - struct timespec wait; - - wait.tv_sec=0; - wait.tv_nsec=5000000L; - - for (i=0;inum_slaves;i++) { - RU_proc_t *slave_proc = proc->slave_proc[i]; - // wake up slave FH thread - // lock the FH mutex and make sure the thread is ready - if (pthread_mutex_timedlock(&slave_proc->mutex_FH,&wait) != 0) { - LOG_E( PHY, "ERROR pthread_mutex_lock for RU %d slave %d (IC %d)\n",proc->ru->idx,slave_proc->ru->idx,slave_proc->instance_cnt_FH); - exit_fun( "error locking mutex_rxtx" ); - break; - } - - int cnt_slave = ++slave_proc->instance_cnt_FH; - slave_proc->frame_rx = proc->frame_rx; - slave_proc->subframe_rx = proc->subframe_rx; - slave_proc->timestamp_rx = proc->timestamp_rx; - slave_proc->timestamp_tx = proc->timestamp_tx; - - pthread_mutex_unlock( &slave_proc->mutex_FH ); - - if (cnt_slave == 0) { - // the thread was presumably waiting where it should and can now be woken up - if (pthread_cond_signal(&slave_proc->cond_FH) != 0) { - LOG_E( PHY, "ERROR pthread_cond_signal for RU %d, slave RU %d\n",proc->ru->idx,slave_proc->ru->idx); - exit_fun( "ERROR pthread_cond_signal" ); - break; - } - } else { - LOG_W( PHY,"[RU] Frame %d, slave %d thread busy!! (cnt_FH %i)\n",slave_proc->frame_rx,slave_proc->ru->idx, cnt_slave); - exit_fun( "FH thread busy" ); - break; - } - } -} - /*! * \brief The prach receive thread of RU. * \param param is a \ref RU_proc_t structure which contains the info what to process. @@ -1147,7 +1045,7 @@ void wakeup_gNB_L1s(RU_t *ru) { char string[20]; sprintf(string,"Incoming RU %d",ru->idx); LOG_D(PHY,"RU %d Call gNB_top\n",ru->idx); - ru->gNB_top(gNB_list[0],ru->proc.frame_rx,ru->proc.subframe_rx,string,ru); + ru->gNB_top(gNB_list[0],ru->proc.frame_rx,ru->proc.tti_rx,string,ru); } else { @@ -1158,7 +1056,7 @@ void wakeup_gNB_L1s(RU_t *ru) { LOG_D(PHY,"ru->wakeup_rxtx:%p\n", ru->nr_wakeup_rxtx); if (ru->nr_wakeup_rxtx!=0 && ru->nr_wakeup_rxtx(gNB_list[i],ru) < 0) { - LOG_E(PHY,"could not wakeup gNB rxtx process for subframe %d\n", ru->proc.subframe_rx); + LOG_E(PHY,"could not wakeup gNB rxtx process for subframe %d\n", ru->proc.tti_rx); } } } @@ -1179,12 +1077,12 @@ static inline int wakeup_prach_ru(RU_t *ru) { if (ru->proc.instance_cnt_prach==-1) { ++ru->proc.instance_cnt_prach; ru->proc.frame_prach = ru->proc.frame_rx; - ru->proc.subframe_prach = ru->proc.subframe_rx; + ru->proc.subframe_prach = ru->proc.tti_rx; // DJP - think prach_procedures() is looking at gNB frame_prach if (ru->gNB_list[0]) { ru->gNB_list[0]->proc.frame_prach = ru->proc.frame_rx; - ru->gNB_list[0]->proc.subframe_prach = ru->proc.subframe_rx; + ru->gNB_list[0]->proc.slot_prach = ru->proc.tti_rx; } LOG_I(PHY,"RU %d: waking up PRACH thread\n",ru->idx); // the thread can now be woken up @@ -1388,7 +1286,7 @@ static void* ru_thread_tx( void* param ) { gNB_L1_rxtx_proc_t *L1_proc; NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; char filename[40]; - int print_frame = 2; + int print_frame = 8; int i = 0; cpu_set_t cpuset; @@ -1415,9 +1313,9 @@ static void* ru_thread_tx( void* param ) { wait_on_condition(&proc->mutex_gNBs,&proc->cond_gNBs,&proc->instance_cnt_gNBs,"ru_thread_tx"); if (oai_exit) break; -//printf("~~~~~~~~~~~~~~~~start process for ru_thread_tx %d.%d\n", proc->frame_tx, proc->subframe_tx); +//printf("~~~~~~~~~~~~~~~~start process for ru_thread_tx %d.%d\n", proc->frame_tx, proc->tti_tx); VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_FRAME_NUMBER_TX0_RU, proc->frame_tx ); - VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_SUBFRAME_NUMBER_TX0_RU, proc->subframe_tx ); + VCD_SIGNAL_DUMPER_DUMP_VARIABLE_BY_NAME( VCD_SIGNAL_DUMPER_VARIABLES_TTI_NUMBER_TX0_RU, proc->tti_tx ); // do TX front-end processing if needed (precoding and/or IDFTs) if (ru->feptx_prec) ru->feptx_prec(ru); @@ -1435,9 +1333,9 @@ static void* ru_thread_tx( void* param ) { { for (i=0; inb_tx; i++) { - sprintf(filename,"tx%ddataF_frame%d_sf%d.m", i, print_frame, proc->subframe_tx); + sprintf(filename,"tx%ddataF_frame%d_sf%d.m", i, print_frame, proc->tti_tx); LOG_M(filename,"txdataF_frame",&ru->common.txdataF_BF[i][0],fp->samples_per_subframe_wCP, 1, 1); - if(proc->subframe_tx == 9) + if(proc->tti_tx == 9) { sprintf(filename,"tx%ddata_frame%d.m", i, print_frame); LOG_M(filename,"txdata_frame",&ru->common.txdata[i][0],fp->samples_per_frame, 1, 1); @@ -1453,7 +1351,7 @@ static void* ru_thread_tx( void* param ) { else { LOG_E(PHY,"Cannot write to file %s\n",filename); } - }//if(proc->subframe_tx == 9) + }//if(proc->tti_tx == 9) }//for (i=0; inb_tx; i++) }//if(proc->frame_tx == print_frame) }//else emulate_rf @@ -1468,7 +1366,7 @@ static void* ru_thread_tx( void* param ) { if (ru == gNB->RU_list[j]) { if ((gNB_proc->RU_mask_tx&(1< 0) LOG_E(PHY,"eNB %d frame %d, subframe %d : previous information from RU tx %d (num_RU %d,mask %x) has not been served yet!\n", - gNB->Mod_id,gNB_proc->frame_rx,gNB_proc->subframe_rx,ru->idx,gNB->num_RU,gNB_proc->RU_mask_tx); + gNB->Mod_id,gNB_proc->frame_rx,gNB_proc->slot_rx,ru->idx,gNB->num_RU,gNB_proc->RU_mask_tx); gNB_proc->RU_mask_tx |= (1<proc; NR_DL_FRAME_PARMS *fp = ru->nr_frame_parms; int ret; - int subframe =9; + int slot = fp->slots_per_frame-1; int frame =1023; - char filename[40]; - int print_frame = 2; + char filename[40],threadname[40]; + int print_frame = 8; int i = 0; // set default return value @@ -1513,7 +1411,8 @@ static void* ru_thread( void* param ) { // set default return value - thread_top_init("ru_thread",0,870000,1000000,1000000); + sprintf(threadname,"ru_thread %d",ru->idx); + thread_top_init(threadname,0,870000,1000000,1000000); LOG_I(PHY,"Starting RU %d (%s,%s),\n",ru->idx,NB_functions[ru->function],NB_timing[ru->if_timing]); @@ -1594,36 +1493,36 @@ static void* ru_thread( void* param ) { // these are local subframe/frame counters to check that we are in synch with the fronthaul timing. // They are set on the first rx/tx in the underly FH routines. - if (subframe==9) { - subframe=0; + if (slot==(fp->slots_per_frame-1)) { + slot=0; frame++; frame&=1023; } else { - subframe++; + slot++; } // synchronization on input FH interface, acquire signals/data and block - if (ru->fh_south_in) ru->fh_south_in(ru,&frame,&subframe); + if (ru->fh_south_in) ru->fh_south_in(ru,&frame,&slot); else AssertFatal(1==0, "No fronthaul interface at south port"); - LOG_D(PHY,"AFTER fh_south_in - SFN/SF:%d%d RU->proc[RX:%d%d TX:%d%d] RC.gNB[0][0]:[RX:%d%d TX(SFN):%d]\n", - frame,subframe, - proc->frame_rx,proc->subframe_rx, - proc->frame_tx,proc->subframe_tx, - RC.gNB[0][0]->proc.frame_rx,RC.gNB[0][0]->proc.subframe_rx, + LOG_I(PHY,"AFTER fh_south_in - SFN/SL:%d%d RU->proc[RX:%d.%d TX:%d.%d] RC.gNB[0][0]:[RX:%d%d TX(SFN):%d]\n", + frame,slot, + proc->frame_rx,proc->tti_rx, + proc->frame_tx,proc->tti_tx, + RC.gNB[0][0]->proc.frame_rx,RC.gNB[0][0]->proc.slot_rx, RC.gNB[0][0]->proc.frame_tx); /* LOG_D(PHY,"RU thread (do_prach %d, is_prach_subframe %d), received frame %d, subframe %d\n", ru->do_prach, - is_prach_subframe(fp, proc->frame_rx, proc->subframe_rx), - proc->frame_rx,proc->subframe_rx); + is_prach_subframe(fp, proc->frame_rx, proc->tti_rx), + proc->frame_rx,proc->tti_rx); - if ((ru->do_prach>0) && (is_prach_subframe(fp, proc->frame_rx, proc->subframe_rx)==1)) { + if ((ru->do_prach>0) && (is_prach_subframe(fp, proc->frame_rx, proc->tti_rx)==1)) { wakeup_prach_ru(ru); }*/ // adjust for timing offset between RU -//printf("~~~~~~~~~~~~~~~~~~~~~~~~~~%d.%d in ru_thread is in process\n", proc->frame_rx, proc->subframe_rx); +//printf("~~~~~~~~~~~~~~~~~~~~~~~~~~%d.%d in ru_thread is in process\n", proc->frame_rx, proc->tti_rx); if (ru->idx!=0) proc->frame_tx = (proc->frame_tx+proc->frame_offset)&1023; @@ -1631,8 +1530,6 @@ static void* ru_thread( void* param ) { if (ru->feprx) ru->feprx(ru); // At this point, all information for subframe has been received on FH interface - // If this proc is to provide synchronization, do so - wakeup_slaves(proc); // wakeup all gNB processes waiting for this RU if (ru->num_gNB>0) wakeup_gNB_L1s(ru); @@ -1657,9 +1554,9 @@ static void* ru_thread( void* param ) { { for (i=0; inb_tx; i++) { - sprintf(filename,"tx%ddataF_frame%d_sf%d.m", i, print_frame, proc->subframe_tx); - LOG_M(filename,"txdataF_frame",&ru->common.txdataF_BF[i][0],fp->samples_per_subframe_wCP, 1, 1); - if(proc->subframe_tx == 9) + sprintf(filename,"tx%ddataF_frame%d_sf%d.m", i, print_frame, proc->tti_tx); + LOG_M(filename,"txdataF_frame",&ru->common.txdataF_BF[i][0],fp->samples_per_slot_wCP, 1, 1); + if(proc->tti_tx == 9) { sprintf(filename,"tx%ddata_frame%d.m", i, print_frame); LOG_M(filename,"txdata_frame",&ru->common.txdata[i][0],fp->samples_per_frame, 1, 1); @@ -1675,7 +1572,7 @@ static void* ru_thread( void* param ) { else { LOG_E(PHY,"Cannot write to file %s\n",filename); } - }//if(proc->subframe_tx == 9) + }//if(proc->tti_tx == 9) }//for (i=0; inb_tx; i++) }//if(proc->frame_tx == print_frame) }//else emulate_rf @@ -2247,16 +2144,9 @@ void set_function_spec_param(RU_t *ru) ru->feprx = (get_nprocs()<=2) ? fep_full : fep_full; // this is frequency-shift + DFTs ru->feptx_prec = feptx_prec; // need to do transmit Precoding + IDFTs ru->feptx_ofdm = (get_nprocs()<=2) ? nr_feptx_ofdm : nr_feptx_ofdm_2thread; // need to do transmit Precoding + IDFTs - if (ru->if_timing == synch_to_other) { - ru->fh_south_in = fh_slave_south_in; // synchronize to master - ru->fh_south_out = fh_if5_mobipass_south_out; // use send_IF5 for mobipass - ru->fh_south_asynch_in = fh_if5_south_asynch_in_mobipass; // UL is asynchronous - } - else { - ru->fh_south_in = fh_if5_south_in; // synchronous IF5 reception - ru->fh_south_out = fh_if5_south_out; // synchronous IF5 transmission - ru->fh_south_asynch_in = NULL; // no asynchronous UL - } + ru->fh_south_in = fh_if5_south_in; // synchronous IF5 reception + ru->fh_south_out = fh_if5_south_out; // synchronous IF5 transmission + ru->fh_south_asynch_in = NULL; // no asynchronous UL ru->start_rf = NULL; // no local RF ru->stop_rf = NULL; ru->start_if = start_if; // need to start if interface for IF5 diff --git a/targets/RT/USER/nr-softmodem.c b/targets/RT/USER/nr-softmodem.c index cb8c916a6d..3bc9bf49d9 100644 --- a/targets/RT/USER/nr-softmodem.c +++ b/targets/RT/USER/nr-softmodem.c @@ -103,6 +103,9 @@ static pthread_t forms_thread; //xforms #endif //XFORMS +short nr_mod_table[NR_MOD_TABLE_SIZE_SHORT] = {0,0,16384,16384,-16384,-16384,16384,16384,16384,-16384,-16384,16384,-16384,-16384,7327,7327,7327,21981,21981,7327,21981,21981,7327,-7327,7327,-21981,21981,-7327,21981,-21981,-7327,7327,-7327,21981,-21981,7327,-21981,21981,-7327,-7327,-7327,-21981,-21981,-7327,-21981,-21981,10726,10726,10726,3576,3576,10726,3576,3576,10726,17876,10726,25027,3576,17876,3576,25027,17876,10726,17876,3576,25027,10726,25027,3576,17876,17876,17876,25027,25027,17876,25027,25027,10726,-10726,10726,-3576,3576,-10726,3576,-3576,10726,-17876,10726,-25027,3576,-17876,3576,-25027,17876,-10726,17876,-3576,25027,-10726,25027,-3576,17876,-17876,17876,-25027,25027,-17876,25027,-25027,-10726,10726,-10726,3576,-3576,10726,-3576,3576,-10726,17876,-10726,25027,-3576,17876,-3576,25027,-17876,10726,-17876,3576,-25027,10726,-25027,3576,-17876,17876,-17876,25027,-25027,17876,-25027,25027,-10726,-10726,-10726,-3576,-3576,-10726,-3576,-3576,-10726,-17876,-10726,-25027,-3576,-17876,-3576,-25027,-17876,-10726,-17876,-3576,-25027,-10726,-25027,-3576,-17876,-17876,-17876,-25027,-25027,-17876,-25027,-25027,8886,8886,8886,12439,12439,8886,12439,12439,8886,5332,8886,1778,12439,5332,12439,1778,5332,8886,5332,12439,1778,8886,1778,12439,5332,5332,5332,1778,1778,5332,1778,1778,8886,19547,8886,15993,12439,19547,12439,15993,8886,23101,8886,26655,12439,23101,12439,26655,5332,19547,5332,15993,1778,19547,1778,15993,5332,23101,5332,26655,1778,23101,1778,26655,19547,8886,19547,12439,15993,8886,15993,12439,19547,5332,19547,1778,15993,5332,15993,1778,23101,8886,23101,12439,26655,8886,26655,12439,23101,5332,23101,1778,26655,5332,26655,1778,19547,19547,19547,15993,15993,19547,15993,15993,19547,23101,19547,26655,15993,23101,15993,26655,23101,19547,23101,15993,26655,19547,26655,15993,23101,23101,23101,26655,26655,23101,26655,26655,8886,-8886,8886,-12439,12439,-8886,12439,-12439,8886,-5332,8886,-1778,12439,-5332,12439,-1778,5332,-8886,5332,-12439,1778,-8886,1778,-12439,5332,-5332,5332,-1778,1778,-5332,1778,-1778,8886,-19547,8886,-15993,12439,-19547,12439,-15993,8886,-23101,8886,-26655,12439,-23101,12439,-26655,5332,-19547,5332,-15993,1778,-19547,1778,-15993,5332,-23101,5332,-26655,1778,-23101,1778,-26655,19547,-8886,19547,-12439,15993,-8886,15993,-12439,19547,-5332,19547,-1778,15993,-5332,15993,-1778,23101,-8886,23101,-12439,26655,-8886,26655,-12439,23101,-5332,23101,-1778,26655,-5332,26655,-1778,19547,-19547,19547,-15993,15993,-19547,15993,-15993,19547,-23101,19547,-26655,15993,-23101,15993,-26655,23101,-19547,23101,-15993,26655,-19547,26655,-15993,23101,-23101,23101,-26655,26655,-23101,26655,-26655,-8886,8886,-8886,12439,-12439,8886,-12439,12439,-8886,5332,-8886,1778,-12439,5332,-12439,1778,-5332,8886,-5332,12439,-1778,8886,-1778,12439,-5332,5332,-5332,1778,-1778,5332,-1778,1778,-8886,19547,-8886,15993,-12439,19547,-12439,15993,-8886,23101,-8886,26655,-12439,23101,-12439,26655,-5332,19547,-5332,15993,-1778,19547,-1778,15993,-5332,23101,-5332,26655,-1778,23101,-1778,26655,-19547,8886,-19547,12439,-15993,8886,-15993,12439,-19547,5332,-19547,1778,-15993,5332,-15993,1778,-23101,8886,-23101,12439,-26655,8886,-26655,12439,-23101,5332,-23101,1778,-26655,5332,-26655,1778,-19547,19547,-19547,15993,-15993,19547,-15993,15993,-19547,23101,-19547,26655,-15993,23101,-15993,26655,-23101,19547,-23101,15993,-26655,19547,-26655,15993,-23101,23101,-23101,26655,-26655,23101,-26655,26655,-8886,-8886,-8886,-12439,-12439,-8886,-12439,-12439,-8886,-5332,-8886,-1778,-12439,-5332,-12439,-1778,-5332,-8886,-5332,-12439,-1778,-8886,-1778,-12439,-5332,-5332,-5332,-1778,-1778,-5332,-1778,-1778,-8886,-19547,-8886,-15993,-12439,-19547,-12439,-15993,-8886,-23101,-8886,-26655,-12439,-23101,-12439,-26655,-5332,-19547,-5332,-15993,-1778,-19547,-1778,-15993,-5332,-23101,-5332,-26655,-1778,-23101,-1778,-26655,-19547,-8886,-19547,-12439,-15993,-8886,-15993,-12439,-19547,-5332,-19547,-1778,-15993,-5332,-15993,-1778,-23101,-8886,-23101,-12439,-26655,-8886,-26655,-12439,-23101,-5332,-23101,-1778,-26655,-5332,-26655,-1778,-19547,-19547,-19547,-15993,-15993,-19547,-15993,-15993,-19547,-23101,-19547,-26655,-15993,-23101,-15993,-26655,-23101,-19547,-23101,-15993,-26655,-19547,-26655,-15993,-23101,-23101,-23101,-26655,-26655,-23101,-26655,-26655}; + + pthread_cond_t nfapi_sync_cond; pthread_mutex_t nfapi_sync_mutex; int nfapi_sync_var=-1; //!< protected by mutex \ref nfapi_sync_mutex @@ -244,8 +247,6 @@ double cpuf; extern char uecap_xer[1024]; char uecap_xer_in=0; -threads_t threads= {-1,-1,-1,-1,-1,-1,-1}; - /* see file openair2/LAYER2/MAC/main.c for why abstraction_flag is needed * this is very hackish - find a proper solution */ @@ -384,11 +385,11 @@ void reset_stats(FL_OBJECT *button, long arg) for (i=0; idlsch[i][0]->Mlimit; j++) { +/* for (j=0; jdlsch[i][0]->Mlimit; j++) { phy_vars_gNB->UE_stats[i].dlsch_NAK[k][j]=0; phy_vars_gNB->UE_stats[i].dlsch_ACK[k][j]=0; phy_vars_gNB->UE_stats[i].dlsch_trials[k][j]=0; - } + }*/ phy_vars_gNB->UE_stats[i].dlsch_l2_errors[k]=0; phy_vars_gNB->UE_stats[i].ulsch_errors[k]=0; @@ -609,8 +610,6 @@ void set_default_frame_parms(nfapi_nr_config_request_t *config[MAX_NUM_CCs], NR_ config[CC_id]->subframe_config.dl_cyclic_prefix_type.value = 0; //NORMAL config[CC_id]->rf_config.dl_carrier_bandwidth.value = 106; config[CC_id]->rf_config.ul_carrier_bandwidth.value = 106; - config[CC_id]->rf_config.tx_antenna_ports.value = 1; - config[CC_id]->rf_config.rx_antenna_ports.value = 1; config[CC_id]->sch_config.physical_cell_id.value = 0; ///dl frequency to be filled in @@ -1033,14 +1032,19 @@ int main( int argc, char **argv ) char cpu_affinity[1024]; CPU_ZERO(&cpuset); #ifdef CPU_AFFINITY + int j; if (get_nprocs() > 2) { - CPU_SET(0, &cpuset); + // CPU_SET(1, &cpuset); + for (j = 2; j < get_nprocs(); j++) + { + CPU_SET(j, &cpuset); + } s = pthread_setaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); if (s != 0) { perror( "pthread_setaffinity_np"); exit_fun("Error setting processor affinity"); } - LOG_I(HW, "Setting the affinity of main function to CPU 0, for device library to use CPU 0 only!\n"); + LOG_I(HW, "Setting the affinity of main function to all CPUs, for device library to use CPU 0 only!\n"); } #endif diff --git a/targets/RT/USER/nr-softmodem.h b/targets/RT/USER/nr-softmodem.h index 9dce5f34bd..5ee2880f3d 100644 --- a/targets/RT/USER/nr-softmodem.h +++ b/targets/RT/USER/nr-softmodem.h @@ -152,12 +152,6 @@ {"external-clock", CONFIG_HLP_EXCCLK, PARAMFLAG_BOOL, uptr:&clock_source, defintval:0, TYPE_INT, 0}, \ {"wait-for-sync", NULL, PARAMFLAG_BOOL, iptr:&wait_for_sync, defintval:0, TYPE_INT, 0}, \ {"single-thread-disable", CONFIG_HLP_NOSNGLT, PARAMFLAG_BOOL, iptr:&single_thread_flag, defintval:1, TYPE_INT, 0}, \ -{"threadIQ", NULL, 0, iptr:&(threads.iq), defintval:1, TYPE_INT, 0}, \ -{"threadOneSubframe", NULL, 0, iptr:&(threads.one), defintval:1, TYPE_INT, 0}, \ -{"threadTwoSubframe", NULL, 0, iptr:&(threads.two), defintval:1, TYPE_INT, 0}, \ -{"threadThreeSubframe", NULL, 0, iptr:&(threads.three), defintval:1, TYPE_INT, 0}, \ -{"threadSlot1ProcOne", NULL, 0, iptr:&(threads.slot1_proc_one), defintval:1, TYPE_INT, 0}, \ -{"threadSlot1ProcTwo", NULL, 0, iptr:&(threads.slot1_proc_two), defintval:1, TYPE_INT, 0}, \ {"A" , CONFIG_HLP_TADV, 0, uptr:&timing_advance, defintval:0, TYPE_UINT, 0}, \ {"C" , CONFIG_HLP_DLF, 0, uptr:&(downlink_frequency[0][0]), defuintval:DEFAULT_DLF, TYPE_UINT, 0}, \ {"a" , CONFIG_HLP_CHOFF, 0, iptr:&chain_offset, defintval:0, TYPE_INT, 0}, \ diff --git a/targets/RT/USER/nr-ue.c b/targets/RT/USER/nr-ue.c index 41b20710e9..b289811b01 100644 --- a/targets/RT/USER/nr-ue.c +++ b/targets/RT/USER/nr-ue.c @@ -226,7 +226,7 @@ PHY_VARS_NR_UE* init_nr_ue_vars(NR_DL_FRAME_PARMS *frame_parms, // initialize all signal buffers init_nr_ue_signal(ue,1,abstraction_flag); // intialize transport - //init_nr_ue_transport(ue,abstraction_flag); + init_nr_ue_transport(ue,abstraction_flag); return(ue); } @@ -341,8 +341,8 @@ static void *UE_thread_synch(void *arg) { cpu_set_t cpuset; CPU_ZERO(&cpuset); - if ( threads.iq != -1 ) - CPU_SET(threads.iq, &cpuset); + if ( threads.sync != -1 ) + CPU_SET(threads.sync, &cpuset); // this thread priority must be lower that the main acquisition thread sprintf(threadname, "sync UE %d", UE->Mod_id); init_thread(100000, 500000, FIFO_PRIORITY-1, &cpuset, threadname); @@ -442,6 +442,9 @@ static void *UE_thread_synch(void *arg) { #endif if (nr_initial_sync( UE, UE->mode ) == 0) { + //write_output("txdata_sym.m", "txdata_sym", UE->common_vars.rxdata[0], (10*UE->frame_parms.samples_per_subframe), 1, 1); + + freq_offset = UE->common_vars.freq_offset; // frequency offset computed with pss in initial sync hw_slot_offset = (UE->rx_offset<<1) / UE->frame_parms.samples_per_subframe; printf("Got synch: hw_slot_offset %d, carrier off %d Hz, rxgain %d (DL %u, UL %u), UE_scan_carrier %d\n", hw_slot_offset, @@ -455,16 +458,13 @@ static void *UE_thread_synch(void *arg) { // rerun with new cell parameters and frequency-offset for (i=0; irf_map.card].rx_num_channels; i++) { openair0_cfg[UE->rf_map.card].rx_gain[UE->rf_map.chain+i] = UE->rx_total_gain_dB;//-USRP_GAIN_OFFSET; - if (UE->UE_scan_carrier == 1) { if (freq_offset >= 0) - openair0_cfg[UE->rf_map.card].rx_freq[UE->rf_map.chain+i] += abs(UE->common_vars.freq_offset); + openair0_cfg[UE->rf_map.card].rx_freq[UE->rf_map.chain+i] += abs(freq_offset); else - openair0_cfg[UE->rf_map.card].rx_freq[UE->rf_map.chain+i] -= abs(UE->common_vars.freq_offset); + openair0_cfg[UE->rf_map.card].rx_freq[UE->rf_map.chain+i] -= abs(freq_offset); openair0_cfg[UE->rf_map.card].tx_freq[UE->rf_map.chain+i] = openair0_cfg[UE->rf_map.card].rx_freq[UE->rf_map.chain+i]+uplink_frequency_offset[CC_id][i]; downlink_frequency[CC_id][i] = openair0_cfg[CC_id].rx_freq[i]; - freq_offset=0; - } } // reconfigure for potentially different bandwidth @@ -594,16 +594,16 @@ static void *UE_thread_synch(void *arg) { break; } -#ifdef XFORMS +#if 0 //defined XFORMS if (do_forms) { extern FD_lte_phy_scope_ue *form_ue[NUMBER_OF_UE_MAX]; phy_scope_UE(form_ue[0], PHY_vars_UE_g[0][0], - 0,0,7); + 0,0,1); } #endif - + AssertFatal ( 0== pthread_mutex_lock(&UE->proc.mutex_synch), ""); // indicate readiness UE->proc.instance_cnt_synch--; @@ -634,7 +634,7 @@ static void *UE_thread_rxn_txnp4(void *arg) { proc->instance_cnt_rxtx=-1; proc->subframe_rx=proc->sub_frame_start; - proc->dci_err_cnt=0; + proc->dci_err_cnt=0; char threadname[256]; sprintf(threadname,"UE_%d_proc_%d", UE->Mod_id, proc->sub_frame_start); cpu_set_t cpuset; @@ -671,6 +671,8 @@ static void *UE_thread_rxn_txnp4(void *arg) { // pickTime(current); // updateTimes(proc->gotIQs, &t2, 10000, "Delay to wake up UE_Thread_Rx (case 2)"); + + // Process Rx data for one sub-frame if (slot_select_nr(&UE->frame_parms, proc->frame_tx, proc->nr_tti_tx) & NR_DOWNLINK_SLOT) { @@ -679,11 +681,23 @@ static void *UE_thread_rxn_txnp4(void *arg) { UE->dci_ind.number_of_dcis = 0; //clean previous FAPI MESSAGE + // call L2 for DL_CONFIG (DCI) + UE->dcireq.module_id = UE->Mod_id; + UE->dcireq.gNB_index = 0; + UE->dcireq.cc_id = 0; + UE->dcireq.frame = proc->frame_rx; + UE->dcireq.slot = proc->nr_tti_rx; + nr_ue_dcireq(&UE->dcireq); //to be replaced with function pointer later + + NR_UE_MAC_INST_t *UE_mac = get_mac_inst(0); + UE_mac->scheduled_response.dl_config = &UE->dcireq.dl_config_req; + nr_ue_scheduled_response(&UE_mac->scheduled_response); + #ifdef UE_SLOT_PARALLELISATION - phy_procedures_slot_parallelization_UE_RX( UE, proc, 0, 0, 1, UE->mode, no_relay, NULL ); + phy_procedures_slot_parallelization_nrUE_RX( UE, proc, 0, 0, 1, UE->mode, no_relay, NULL ); #else - phy_procedures_nrUE_RX( UE, proc, 0, 0, 1, UE->mode, no_relay); - printf(">>> nr_ue_pdcch_procedures ended\n"); + phy_procedures_nrUE_RX( UE, proc, 0, 1, UE->mode); + // printf(">>> nr_ue_pdcch_procedures ended\n"); #endif } @@ -691,78 +705,27 @@ static void *UE_thread_rxn_txnp4(void *arg) { #if UE_TIMING_TRACE start_meas(&UE->generic_stat); #endif -printf(">>> mac init\n"); + //printf(">>> mac init\n"); if (UE->mac_enabled==1) { // trigger L2 to run ue_scheduler thru IF module // [TODO] mapping right after NR initial sync - if(1) if(UE->if_inst != NULL && UE->if_inst->ul_indication != NULL){ UE->ul_indication.module_id = 0; UE->ul_indication.gNB_index = 0; UE->ul_indication.cc_id = 0; - UE->ul_indication.slot = 0; // to be fill - UE->ul_indication.frame = 0; // to be fill - // [TODO] mapping right after NR initial sync UE->ul_indication.frame = proc->frame_rx; UE->ul_indication.slot = proc->nr_tti_rx; UE->if_inst->ul_indication(&UE->ul_indication); } - - - -#ifdef NEW_MAC - ret = mac_xface->ue_scheduler(UE->Mod_id, - proc->frame_rx, - proc->subframe_rx, - proc->nr_tti_rx, - proc->frame_tx, - proc->subframe_tx, - proc->nr_tti_tx%(UE->frame_parms.ttis_per_subframe), -#ifndef NO_RAT_NR - slot_select_nr(&UE->frame_parms, proc->frame_tx, proc->nr_tti_tx), -#else - subframe_select(&UE->frame_parms,proc->subframe_tx), -#endif - 0, - 0/*FIXME CC_id*/); -#endif - -/*#else - ret = mac_xface->ue_scheduler(UE->Mod_id, - proc->frame_rx, - proc->subframe_rx, - proc->frame_tx, - proc->subframe_tx, - subframe_select(&UE->frame_parms,proc->subframe_tx), - 0, */ -// 0/*FIXME CC_id*/); -//#endif - if ( ret != CONNECTION_OK) { - char *txt; - switch (ret) { - case CONNECTION_LOST: - txt="RRC Connection lost, returning to PRACH"; - break; - case PHY_RESYNCH: - txt="RRC Connection lost, trying to resynch"; - break; - case RESYNCH: - txt="return to PRACH and perform a contention-free access"; - break; - default: - txt="UNKNOWN RETURN CODE"; - }; - LOG_E( PHY, "[UE %"PRIu8"] Frame %"PRIu32", subframe %u %s\n", - UE->Mod_id, proc->frame_rx, proc->subframe_tx,txt ); - } - } + } + #if UE_TIMING_TRACE stop_meas(&UE->generic_stat); #endif -printf(">>> mac ended\n"); + //printf(">>> mac ended\n"); // Prepare the future Tx data #if 0 @@ -832,15 +795,15 @@ void *UE_thread(void *arg) { cpu_set_t cpuset; CPU_ZERO(&cpuset); - if ( threads.iq != -1 ) - CPU_SET(threads.iq, &cpuset); - init_thread(100000, 500000, FIFO_PRIORITY, &cpuset, - "UHD Threads"); + if ( threads.main != -1 ) + CPU_SET(threads.main, &cpuset); + sprintf(threadname, "Main UE %d", UE->Mod_id); + init_thread(100000, 500000, FIFO_PRIORITY, &cpuset,threadname); + if ((oaisim_flag == 0) && (UE->mode !=loop_through_memory)) AssertFatal(0== openair0_device_load(&(UE->rfdevice), &openair0_cfg[0]), ""); UE->rfdevice.host_type = RAU_HOST; - sprintf(threadname, "Main UE %d", UE->Mod_id); - pthread_setname_np(pthread_self(), threadname); + init_UE_threads(UE); #ifdef NAS_UE @@ -866,13 +829,14 @@ void *UE_thread(void *arg) { usleep(500*1000); } #endif + if (instance_cnt_synch < 0) { // we can invoke the synch // grab 10 ms of signal and wakeup synch thread for (int i=0; iframe_parms.nb_antennas_rx; i++) rxp[i] = (void*)&UE->common_vars.rxdata[i][0]; if (UE->mode != loop_through_memory) - AssertFatal( UE->frame_parms.samples_per_subframe*10 == + AssertFatal( UE->frame_parms.samples_per_subframe*10== UE->rfdevice.trx_read_func(&UE->rfdevice, ×tamp, rxp, @@ -925,10 +889,10 @@ void *UE_thread(void *arg) { } UE->rx_offset=0; UE->time_sync_cell=0; - //UE->proc.proc_rxtx[0].frame_rx++; + UE->proc.proc_rxtx[0].frame_rx++; //UE->proc.proc_rxtx[1].frame_rx++; - for (th_id=0; th_id < RX_NB_TH; th_id++) { - UE->proc.proc_rxtx[th_id].frame_rx++; + for (th_id=1; th_id < RX_NB_TH; th_id++) { + UE->proc.proc_rxtx[th_id].frame_rx = UE->proc.proc_rxtx[0].frame_rx; } //printf("first stream frame rx %d\n",UE->proc.proc_rxtx[0].frame_rx); @@ -947,6 +911,10 @@ void *UE_thread(void *arg) { rt_sleep_ns(1000*1000); } else { + thread_idx++; + if(thread_idx>=RX_NB_TH) + thread_idx = 0; + subframe_nr++; subframe_nr %= NR_NUMBER_OF_SUBFRAMES_PER_FRAME; UE_nr_rxtx_proc_t *proc = &UE->proc.proc_rxtx[thread_idx]; @@ -967,9 +935,6 @@ void *UE_thread(void *arg) { LOG_D(PHY,"Process subframe %d thread Idx %d \n", subframe_nr, UE->current_thread_id[subframe_nr]); - thread_idx++; - if(thread_idx>=RX_NB_TH) - thread_idx = 0; if (UE->mode != loop_through_memory) { for (i=0; iframe_parms.nb_antennas_rx; i++) @@ -1038,10 +1003,10 @@ void *UE_thread(void *arg) { // operate on thread sf mod 2 AssertFatal(pthread_mutex_lock(&proc->mutex_rxtx) ==0,""); if(subframe_nr == 0) { - //UE->proc.proc_rxtx[0].frame_rx++; + UE->proc.proc_rxtx[0].frame_rx++; //UE->proc.proc_rxtx[1].frame_rx++; - for (th_id=0; th_id < RX_NB_TH; th_id++) { - UE->proc.proc_rxtx[th_id].frame_rx++; + for (th_id=1; th_id < RX_NB_TH; th_id++) { + UE->proc.proc_rxtx[th_id].frame_rx = UE->proc.proc_rxtx[0].frame_rx; } #ifdef SAIF_ENABLED if (!(proc->frame_rx%4000)) @@ -1061,7 +1026,7 @@ void *UE_thread(void *arg) { proc->nr_tti_rx=subframe_nr; proc->subframe_rx=subframe_nr; - + proc->frame_tx = proc->frame_rx; proc->nr_tti_tx= subframe_nr + DURATION_RX_TO_TX; if (proc->nr_tti_tx > NR_NUMBER_OF_SUBFRAMES_PER_FRAME) { @@ -1106,42 +1071,60 @@ void *UE_thread(void *arg) { // pickStaticTime(lastTime); } //UE->mode != loop_through_memory else { - proc->nr_tti_rx=subframe_nr; - proc->subframe_rx=subframe_nr; - if(subframe_nr == 0) { - for (th_id=0; th_id < RX_NB_TH; th_id++) { - UE->proc.proc_rxtx[th_id].frame_rx++; - } + proc->nr_tti_rx=subframe_nr; + proc->subframe_rx=subframe_nr; + if(subframe_nr == 0) { + for (th_id=0; th_id < RX_NB_TH; th_id++) { + UE->proc.proc_rxtx[th_id].frame_rx++; } - proc->frame_tx = proc->frame_rx; - proc->nr_tti_tx= subframe_nr + DURATION_RX_TO_TX; - if (proc->nr_tti_tx > NR_NUMBER_OF_SUBFRAMES_PER_FRAME) { - proc->frame_tx = (proc->frame_tx + 1)%MAX_FRAME_NUMBER; - proc->nr_tti_tx %= NR_NUMBER_OF_SUBFRAMES_PER_FRAME; - } - proc->subframe_tx=proc->nr_tti_tx; + } + proc->frame_tx = proc->frame_rx; + proc->nr_tti_tx= subframe_nr + DURATION_RX_TO_TX; + if (proc->nr_tti_tx > NR_NUMBER_OF_SUBFRAMES_PER_FRAME) { + proc->frame_tx = (proc->frame_tx + 1)%MAX_FRAME_NUMBER; + proc->nr_tti_tx %= NR_NUMBER_OF_SUBFRAMES_PER_FRAME; + } + proc->subframe_tx=proc->nr_tti_tx; - printf("Processing subframe %d\n",proc->subframe_rx); + if (slot_select_nr(&UE->frame_parms, proc->frame_tx, proc->nr_tti_tx) & NR_DOWNLINK_SLOT) { + + //clean previous FAPI MESSAGE + UE->rx_ind.number_pdus = 0; + UE->dci_ind.number_of_dcis = 0; + //clean previous FAPI MESSAGE + + // call L2 for DL_CONFIG (DCI) + UE->dcireq.module_id = UE->Mod_id; + UE->dcireq.gNB_index = 0; + UE->dcireq.cc_id = 0; + UE->dcireq.frame = proc->frame_rx; + UE->dcireq.slot = proc->nr_tti_rx; + nr_ue_dcireq(&UE->dcireq); //to be replaced with function pointer later - if(UE->if_inst != NULL && UE->if_inst->ul_indication != NULL){ - UE->ul_indication.module_id = 0; - UE->ul_indication.gNB_index = 0; - UE->ul_indication.cc_id = 0; - UE->ul_indication.slot = 0; // to be fill - UE->ul_indication.frame = 0; // to be fill - // [TODO] mapping right after NR initial sync - UE->ul_indication.frame = proc->frame_rx; - UE->ul_indication.slot = proc->nr_tti_rx; - - UE->if_inst->ul_indication(&UE->ul_indication); - } + NR_UE_MAC_INST_t *UE_mac = get_mac_inst(0); + UE_mac->scheduled_response.dl_config = &UE->dcireq.dl_config_req; + nr_ue_scheduled_response(&UE_mac->scheduled_response); - phy_procedures_nrUE_RX( UE, proc, 0, 0, 1, UE->mode, no_relay); - getchar(); + + printf("Processing subframe %d\n",proc->subframe_rx); + phy_procedures_nrUE_RX( UE, proc, 0, 1, UE->mode); + } + + if(UE->if_inst != NULL && UE->if_inst->ul_indication != NULL){ + UE->ul_indication.module_id = 0; + UE->ul_indication.gNB_index = 0; + UE->ul_indication.cc_id = 0; + UE->ul_indication.frame = proc->frame_rx; + UE->ul_indication.slot = proc->nr_tti_rx; + + UE->if_inst->ul_indication(&UE->ul_indication); + } + + getchar(); } // else loop_through_memory } // start_rx_stream==1 } // UE->is_synchronized==1 - + } // while !oai_exit return NULL; } @@ -1186,11 +1169,11 @@ void init_UE_threads(PHY_VARS_NR_UE *UE) { #ifdef UE_DLSCH_PARALLELISATION pthread_mutex_init(&UE->proc.proc_rxtx[i].mutex_dlsch_td,NULL); pthread_cond_init(&UE->proc.proc_rxtx[i].cond_dlsch_td,NULL); - pthread_create(&UE->proc.proc_rxtx[i].pthread_dlsch_td,NULL,dlsch_decoding_2thread0, rtd); + pthread_create(&UE->proc.proc_rxtx[i].pthread_dlsch_td,NULL,nr_dlsch_decoding_2thread0, rtd); //thread 2 pthread_mutex_init(&UE->proc.proc_rxtx[i].mutex_dlsch_td1,NULL); pthread_cond_init(&UE->proc.proc_rxtx[i].cond_dlsch_td1,NULL); - pthread_create(&UE->proc.proc_rxtx[i].pthread_dlsch_td1,NULL,dlsch_decoding_2thread1, rtd); + pthread_create(&UE->proc.proc_rxtx[i].pthread_dlsch_td1,NULL,nr_dlsch_decoding_2thread1, rtd); #endif #ifdef UE_SLOT_PARALLELISATION diff --git a/targets/RT/USER/nr-uesoftmodem.c b/targets/RT/USER/nr-uesoftmodem.c index 02db7d1a02..d3d660e7f5 100644 --- a/targets/RT/USER/nr-uesoftmodem.c +++ b/targets/RT/USER/nr-uesoftmodem.c @@ -85,10 +85,10 @@ unsigned short config_frames[4] = {2,9,11,13}; // current status is that every UE has a DL scope for a SINGLE eNB (eNB_id=0) // at eNB 0, an UL scope for every UE FD_lte_phy_scope_ue *form_ue[NUMBER_OF_UE_MAX]; -FD_lte_phy_scope_enb *form_enb[MAX_NUM_CCs][NUMBER_OF_UE_MAX]; -FD_stats_form *form_stats=NULL,*form_stats_l2=NULL; +//FD_lte_phy_scope_enb *form_enb[MAX_NUM_CCs][NUMBER_OF_UE_MAX]; +//FD_stats_form *form_stats=NULL,*form_stats_l2=NULL; char title[255]; -unsigned char scope_enb_num_ue = 2; +//unsigned char scope_enb_num_ue = 2; static pthread_t forms_thread; //xforms #endif //XFORMS #include "nr-uesoftmodem.h" @@ -119,6 +119,8 @@ int threequarter_fs=0; uint32_t downlink_frequency[MAX_NUM_CCs][4]; int32_t uplink_frequency_offset[MAX_NUM_CCs][4]; +extern int16_t nr_dlsch_demod_shift; + //static char *conf_config_file_name = NULL; @@ -128,6 +130,7 @@ static char *itti_dump_file = NULL; int UE_scan = 0; int UE_scan_carrier = 0; +int UE_fo_compensation = 0; runmode_t mode = normal_txrx; FILE *input_fd=NULL; @@ -266,7 +269,8 @@ char uecap_xer[1024],uecap_xer_in=0; int oaisim_flag=0; int emulate_rf = 0; -threads_t threads= {-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1}; +threads_t threads = {-1,-1,-1,-1,-1,-1,-1,-1}; +int threads_offset = 0; char* usrp_args=NULL; char* usrp_clksrc=NULL; @@ -435,21 +439,21 @@ static void *scope_thread(void *arg) { #endif - while (!oai_exit) { + while (!oai_exit) { //len = dump_ue_stats (PHY_vars_UE_g[0][0], &PHY_vars_UE_g[0][0]->proc.proc_rxtx[0],stats_buffer, 0, mode,rx_input_level_dBm); //fl_set_object_label(form_stats->stats_text, stats_buffer); - fl_clear_browser(form_stats->stats_text); - fl_add_browser_line(form_stats->stats_text, stats_buffer); + //fl_clear_browser(form_stats->stats_text); + //fl_add_browser_line(form_stats->stats_text, stats_buffer); - if (PHY_vars_UE_g[0][0]->is_synchronized == 1) + //if (PHY_vars_UE_g[0][0]->is_synchronized == 1) phy_scope_UE(form_ue[0], PHY_vars_UE_g[0][0], - 0,0,7); + 0,0,1); //else it is done in the synch thread //printf("doing forms\n"); //usleep(100000); // 100 ms - sleep(1); + sleep(0.1); } // printf("%s",stats_buffer); @@ -547,7 +551,7 @@ static void get_options(void) { set_glog_onlinelog(online_log_messages); } if(config_isparamset(cmdline_logparams,CMDLINE_GLOGLEVEL_IDX)) { - set_glog(glog_level); + set_glog(glog_level); } if (start_telnetsrv) { @@ -656,8 +660,6 @@ void set_default_frame_parms(NR_DL_FRAME_PARMS *frame_parms[MAX_NUM_CCs]) { config[CC_id]->subframe_config.dl_cyclic_prefix_type.value = 0; //NORMAL config[CC_id]->rf_config.dl_carrier_bandwidth.value = 106; config[CC_id]->rf_config.ul_carrier_bandwidth.value = 106; - config[CC_id]->rf_config.tx_antenna_ports.value = 1; - config[CC_id]->rf_config.rx_antenna_ports.value = 1; config[CC_id]->sch_config.physical_cell_id.value = 0; frame_parms[CC_id]->frame_type = FDD; @@ -872,23 +874,6 @@ int main( int argc, char **argv ) { T_Config_Init(); #endif - // initialize the log (see log.h for details) - set_glog(OAILOG_DEBUG); - - set_log(HW, OAILOG_DEBUG); - set_log(PHY, OAILOG_DEBUG); - set_log(MAC, OAILOG_INFO); - set_log(RLC, OAILOG_INFO); - set_log(PDCP, OAILOG_INFO); - set_log(OTG, OAILOG_INFO); - set_log(RRC, OAILOG_INFO); -#if defined(ENABLE_ITTI) - set_log(SIM, OAILOG_INFO); -# if defined(ENABLE_USE_MME) - set_log(NAS, OAILOG_INFO); -# endif -#endif - //randominit (0); set_taus_seed (0); @@ -994,6 +979,7 @@ int main( int argc, char **argv ) { UE[CC_id]->UE_scan = UE_scan; UE[CC_id]->UE_scan_carrier = UE_scan_carrier; + UE[CC_id]->UE_fo_compensation = UE_fo_compensation; UE[CC_id]->mode = mode; printf("UE[%d]->mode = %d\n",CC_id,mode); @@ -1044,8 +1030,25 @@ int main( int argc, char **argv ) { //dump_frame_parms(frame_parms[0]); init_openair0(); - + // default threads.main = 2 + // if there are enough processors, all others on subsequent CPUs + if (threads_offset>0) { + if (threads_offset+8<=get_nprocs()) { + if (threads.main<0) threads.main=threads_offset+1; + if (threads.sync<0) threads.sync=threads_offset+2; + if (threads.one<0) threads.one=threads_offset+3; + if (threads.two<0) threads.two=threads_offset+4; + if (threads.three<0) threads.three=threads_offset+5; + if (threads.slot1_proc_one<0) threads.slot1_proc_one=threads_offset+6; + if (threads.slot1_proc_two<0) threads.slot1_proc_two=threads_offset+7; + if (threads.slot1_proc_three<0) threads.slot1_proc_three=threads_offset+8; + } + else { + LOG_E(HW,"Not enough CPUs available (nprocs (=%d) >= threadmain (=%d) + 10)\n",get_nprocs(),threads_offset); + exit_fun("Error setting thread affinity\n"); + } + } #ifndef DEADLINE_SCHEDULER @@ -1056,14 +1059,17 @@ int main( int argc, char **argv ) { char cpu_affinity[1024]; CPU_ZERO(&cpuset); #ifdef CPU_AFFINITY - if (get_nprocs() > 2) { - CPU_SET(0, &cpuset); + int j; + if (get_nprocs()>=2) { + for (j = 2; j < get_nprocs(); j++) + CPU_SET(j, &cpuset); + s = pthread_setaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); if (s != 0) { perror( "pthread_setaffinity_np"); exit_fun("Error setting processor affinity"); } - LOG_I(HW, "Setting the affinity of main function to CPU 0, for device library to use CPU 0 only!\n"); + LOG_I(HW, "Setting the affinity of main function to all CPUs, for device library to use CPU 0 only!\n"); } #endif @@ -1131,6 +1137,8 @@ int main( int argc, char **argv ) { pthread_cond_init(&sync_cond,NULL); pthread_mutex_init(&sync_mutex, NULL); + init_UE(1); + #ifdef XFORMS int UE_id; int fl_argc=1; @@ -1140,8 +1148,8 @@ int main( int argc, char **argv ) { fl_initialize (&fl_argc, argv, NULL, 0, 0); // restore the original command line args // argv = fl_get_cmdline_args( &argc ); - form_stats = create_form_stats_form(); - fl_show_form (form_stats->stats_form, FL_PLACE_HOTSPOT, FL_FULLBORDER, "stats"); + //form_stats = create_form_stats_form(); + //fl_show_form (form_stats->stats_form, FL_PLACE_HOTSPOT, FL_FULLBORDER, "stats"); UE_id = 0; form_ue[UE_id] = create_lte_phy_scope_ue(); sprintf (title, "NR DL SCOPE UE"); @@ -1171,7 +1179,6 @@ int main( int argc, char **argv ) { rt_sleep_ns(10*100000000ULL); - init_UE(1); number_of_cards = 1; for(CC_id=0; CC_idstats_form); - fl_free_form(form_stats->stats_form); + //fl_hide_form(form_stats->stats_form); + //fl_free_form(form_stats->stats_form); fl_hide_form(form_ue[0]->lte_phy_scope_ue); fl_free_form(form_ue[0]->lte_phy_scope_ue); } diff --git a/targets/RT/USER/nr-uesoftmodem.h b/targets/RT/USER/nr-uesoftmodem.h index ff3ec107de..7f96480b2b 100644 --- a/targets/RT/USER/nr-uesoftmodem.h +++ b/targets/RT/USER/nr-uesoftmodem.h @@ -55,6 +55,7 @@ #define CONFIG_HLP_UENANTR "set UE number of rx antennas\n" #define CONFIG_HLP_UENANTT "set UE number of tx antennas\n" #define CONFIG_HLP_UESCAN "set UE to scan around carrier\n" +#define CONFIG_HLP_UEFO "set UE to enable estimation and compensation of frequency offset\n" #define CONFIG_HLP_DUMPFRAME "dump UE received frame to rxsig_frame0.dat and exit\n" #define CONFIG_HLP_DLSHIFT "dynamic shift for LLR compuation for TM3/4 (default 0)\n" #define CONFIG_HLP_UELOOP "get softmodem (UE) to loop through memory instead of acquiring from HW\n" @@ -135,6 +136,7 @@ {"ue-nb-ant-rx", CONFIG_HLP_UENANTR, 0, u8ptr:&nb_antenna_rx, defuintval:1, TYPE_UINT8, 0}, \ {"ue-nb-ant-tx", CONFIG_HLP_UENANTT, 0, u8ptr:&nb_antenna_tx, defuintval:1, TYPE_UINT8, 0}, \ {"ue-scan-carrier", CONFIG_HLP_UESCAN, PARAMFLAG_BOOL, iptr:&UE_scan_carrier, defintval:0, TYPE_INT, 0}, \ +{"ue-fo-compensation", CONFIG_HLP_UEFO, PARAMFLAG_BOOL, iptr:&UE_fo_compensation, defintval:0, TYPE_INT, 0}, \ {"ue-max-power", NULL, 0, iptr:&(tx_max_power[0]), defintval:90, TYPE_INT, 0}, \ {"r" , CONFIG_HLP_PRB, 0, iptr:&(frame_parms[0]->N_RB_DL), defintval:25, TYPE_UINT, 0}, \ {"dlsch-demod-shift", CONFIG_HLP_DLSHIFT, 0, iptr:(int32_t *)&dlsch_demod_shift, defintval:0, TYPE_INT, 0}, \ @@ -157,13 +159,16 @@ {"external-clock", CONFIG_HLP_EXCCLK, PARAMFLAG_BOOL, uptr:&clock_source, defintval:0, TYPE_INT, 0}, \ {"wait-for-sync", NULL, PARAMFLAG_BOOL, iptr:&wait_for_sync, defintval:0, TYPE_INT, 0}, \ {"single-thread-disable", CONFIG_HLP_NOSNGLT, PARAMFLAG_BOOL, iptr:&single_thread_flag, defintval:1, TYPE_INT, 0}, \ -{"threadIQ", NULL, 0, iptr:&(threads.iq), defintval:1, TYPE_INT, 0}, \ -{"threadOneSubframe", NULL, 0, iptr:&(threads.one), defintval:1, TYPE_INT, 0}, \ -{"threadTwoSubframe", NULL, 0, iptr:&(threads.two), defintval:1, TYPE_INT, 0}, \ -{"threadThreeSubframe", NULL, 0, iptr:&(threads.three), defintval:1, TYPE_INT, 0}, \ -{"threadSlot1ProcOne", NULL, 0, iptr:&(threads.slot1_proc_one), defintval:1, TYPE_INT, 0}, \ -{"threadSlot1ProcTwo", NULL, 0, iptr:&(threads.slot1_proc_two), defintval:1, TYPE_INT, 0}, \ -{"dlsch-demod-shift", CONFIG_HLP_DLSHIFT, 0, iptr:(int32_t *)&dlsch_demod_shift, defintval:0, TYPE_INT, 0}, \ +{"threadoffset", NULL, 0, iptr:&(threads_offset), defintval:0, TYPE_INT, 0}, \ +{"threadMain", NULL, 0, iptr:&(threads.main), defintval:-1, TYPE_INT, 0}, \ +{"threadSync", NULL, 0, iptr:&(threads.sync), defintval:-1, TYPE_INT, 0}, \ +{"threadOneSubframe", NULL, 0, iptr:&(threads.one), defintval:-1, TYPE_INT, 0}, \ +{"threadTwoSubframe", NULL, 0, iptr:&(threads.two), defintval:-1, TYPE_INT, 0}, \ +{"threadThreeSubframe", NULL, 0, iptr:&(threads.three), defintval:-1, TYPE_INT, 0}, \ +{"threadSlot1ProcOne", NULL, 0, iptr:&(threads.slot1_proc_one), defintval:-1, TYPE_INT, 0}, \ +{"threadSlot1ProcTwo", NULL, 0, iptr:&(threads.slot1_proc_two), defintval:-1, TYPE_INT, 0}, \ +{"threadSlot1ProcThree", NULL, 0, iptr:&(threads.slot1_proc_three), defintval:-1, TYPE_INT, 0}, \ +{"nr-dlsch-demod-shift", CONFIG_HLP_DLSHIFT, 0, iptr:(int32_t *)&nr_dlsch_demod_shift, defintval:0, TYPE_INT, 0}, \ {"A" , CONFIG_HLP_TADV, 0, uptr:&timing_advance, defintval:0, TYPE_UINT, 0}, \ {"C" , CONFIG_HLP_DLF, 0, uptr:&(downlink_frequency[0][0]), defuintval:2680000000, TYPE_UINT, 0}, \ {"a" , CONFIG_HLP_CHOFF, 0, iptr:&chain_offset, defintval:0, TYPE_INT, 0}, \ diff --git a/targets/RT/USER/rt_wrapper.c b/targets/RT/USER/rt_wrapper.c index 443d183068..e4dc54f78f 100644 --- a/targets/RT/USER/rt_wrapper.c +++ b/targets/RT/USER/rt_wrapper.c @@ -294,20 +294,20 @@ void thread_top_init(char *thread_name, char cpu_affinity[1024]; cpu_set_t cpuset; - /* Set affinity mask to include CPUs 1 to MAX_CPUS */ + /* Set affinity mask to include CPUs 2 to MAX_CPUS */ /* CPU 0 is reserved for UHD threads */ /* CPU 1 is reserved for all RX_TX threads */ - /* Enable CPU Affinity only if number of CPUs >2 */ + /* Enable CPU Affinity only if number of CPUs > 2 */ CPU_ZERO(&cpuset); #ifdef CPU_AFFINITY - if (get_nprocs() > 2) + if (affinity == 0) { + LOG_W(HW,"thread_top_init() called with affinity==0, but overruled by #ifdef CPU_AFFINITY\n"); + } + else if (get_nprocs() > 2) { - if (affinity == 0) - CPU_SET(0,&cpuset); - else - for (j = 1; j < get_nprocs(); j++) - CPU_SET(j, &cpuset); + for (j = 2; j < get_nprocs(); j++) + CPU_SET(j, &cpuset); s = pthread_setaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); if (s != 0) { @@ -315,21 +315,29 @@ void thread_top_init(char *thread_name, exit_fun("Error setting processor affinity"); } } +#else //CPU_AFFINITY + if (affinity) { + LOG_W(HW,"thread_top_init() called with affinity>0, but overruled by #ifndef CPU_AFFINITY.\n"); + } #endif //CPU_AFFINITY /* Check the actual affinity mask assigned to the thread */ s = pthread_getaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); - if (s != 0) { + if (s != 0) + { perror( "pthread_getaffinity_np"); exit_fun("Error getting processor affinity "); } memset(cpu_affinity,0,sizeof(cpu_affinity)); for (j = 0; j < 1024; j++) - if (CPU_ISSET(j, &cpuset)) { + { + if (CPU_ISSET(j, &cpuset)) + { char temp[1024]; sprintf (temp, " CPU_%d", j); strcat(cpu_affinity, temp); } + } memset(&sparam, 0, sizeof(sparam)); sparam.sched_priority = sched_get_priority_max(SCHED_FIFO);