Implementation of STOP exchange
This implements the STOP.request/indication for all 3 transport mechanisms, the
VNF will send the STOP.request to the PNF, indicating it will disconnect, and
await a STOP.indication from the PNF
The WLS VNF is reworked to be able to start before the PNF, by calling call
rte_eal_init and rte_dev_probe to determine if dpdk has been initialized
properly, before calling rte_eal_init in the intended process when it is
appropriate to do so
Note: after this commit, monolithic works but nFAPI operation is broken
and fixed in the next commit, to keep the changes small(er). All
simulators work.
This commit removes the use of msgDataTx's intermediate buffers for FAPI
messages and the call of the "Schedule response". The latter incurs an
additional copy of FAPI messages for the TX chain (for the RX chain, it
is and will remain there) which we can avoid.
Instead, it uses an NR_Sched_Rsp_t-typed variable sched_response to
store the results of the scheduler when called from tx_func(). Thus, the
scheduler remains unchanged (it gets a pointer to where to put FAPI
messages), but we don't call nr_schedule_response() to then copy FAPI
messages into the L1, but have them in a local variable sched_response.
sched_response on on the data segment, as it is big and would overflow
the stack; at the same time, this is ok because only one tx_func() runs
at a time.
Since we don't use allocate_sched_response() anymore, we don't need
deref_sched_response().
clear_slot_beamid() is moved to tx_func(), as it was called in
nr_schedule_response().
Since the "intermediate" NR_gNB_DLSCH_t structure groups both the
DL_tti_pdsch and TX_data.req structures, a pointer has been created to
point to pdsch_pdu. This affects a number of simulators, as they have to
put some messages on the stack:
- nr_dlschsim: put the corresponding variable on the stack
- nr_pbchsim: a new array for the SSB PDUs has been introduced
Further, these changes are now necessary:
- nr_dlsim: there were variables "rel15" and "pdsch_pdu_rel15" that
point to the same PDSCH PDU. At least "rel15" would not exist, as the
corresponding pointer is now populated in phy_procedures_gNB_TX(), and
we therefore refer to the single PDSCH PDU variable with
"pdsch_pdu_rel15", but at the place where "rel15" used to be
initialized.
- nr_ulsim: use nr_save_ul_tti_req() to load PDUs into the RX chain
instead of nr_schedule_response(), which does not exist anymore.
The next commit will remove msgDataTx. We can therefore not use
msgDataTx to retain the PDSCH/DLSCH variables, and move them to the gNB.
Note that the only reason for NR_gNB_DLSCH_t after this commit is to
hold various large buffers (c, b, f). Future work could be done to
remove this, in which case the array inside PHY_VARS_gNB could be
deleted.
On this occasion, clean up the use of init_DLSCH_struct() and
reset_DLSCH_struct() (renamed destroy_DLSCH_struct()) and centralize
their call in phy_init_nr_gNB()/phy_free_nr_gNB(), and correct
simulators accordingly.
DLSCH encoding is not concerned with HARQ processes directly. Instead,
it gets the redundancy version and encodes a specific redundancy version
for a given TB. The rest is up to the MAC, the L1 is stateless in that
regard. Hence, remove the HARQ type and put it into DLSCH, which also
simplifies the code elsewhere.
There is no point in tracking the "active" state: the SSB PDU will be
sent from MAC to L1, encoded, and then "discarded" (by marking the
number of SSB PDUs as 0). Hence, remove this intermediate type.
Refactor DLSCH scheduler
Refactor the DLSCH post processor similarly to what has been done in
!3521 with the goal of saving an additional iteration through all UEs,
hopefully making the DLSCH scheduler faster.
It further includes changes to make working with and debugging multi-UE
operation easier:
- add some measurements in L1 (for total TX/RX time)
- a new option MACRLCs.[0].stats_max_ue to limit the number of UEs in
the periodic output (default: 8, 0=disable)
- avoid stack overflow in the MAC
- fix the writing of stats to nrMAC_stats.log and nrL1_stats.log when
the written output reduces by truncating the files
- avoid to schedule retransmission with 0 RBs.
saves about 600MB of memory allocation
Move the definition of PRACH items and the PRACH list to the
defs_nr_commons.h file.
Do not re-write the entire structure in nr_schedule_rx_prach(), as it is
quite large now.
Call the post-processor when the allocation is "fixed" (i.e., nothing
changes anymore), instead of having a final loop across all UEs that
might potentially be costly.
In order to calculate the PF metric, note that the statistics "reset" is
moved into pf_dl() instead of the post-processing loop, as it has to be
done on each slot.
See also: dd98030202 ("Refactor to call post-processor in place")
Similarly to parent commit(s), assign sched_pdsch in a single place
instead of relying on sched_ctrl. In a follow-up commit, the
post-processor will be called in a single place.
The function rx_nr_prach become autonomous, it doesn't read anymore global variables
So, it can become (later) a threadpool work, that will send a message when it finishes
rx_nr_prach_ru() still use some global variables, but it is executed in sequence, inside the ru main loop, that limits the probability of race conditions
This change is needed due to the previous addition of the function nfapi_stop_l1() in gnb_config.c which in turn calls 2 functions that do not exist in the simulators, since they're not linked against the nfapi libraries
1. Refactor PDCCH PHY functions to take rxdataF of one OFDM symbol and
produce LLR for each search space. The modified functions take
rxdataF in a different format than that produced by nr_slot_fep(). So
for now a memcpy is used to change the array structure. Once
nr_slot_fep() is modified, this memcpy will no longer be needed and
will be removed.
2. Modify DCI functions to decode DCI in the last PDCCH symbol and
inidcate to MAC once.
3. Fix PDCCH monitoring for start symbol > 0.
4. Fix multiple PDCCH monitoring within slot.
Co-authored-by: Sakthivel Velumani <mail@sakthi.me>
1.Create new funtion nr_symbol_fep() to do OFDM demod of one symbol on
all antennas. This is called by nr_slot_fep() for now. In the
upcoming MR, this function will be called directly after reading
each OFDM symbol from the radio.
2.Removed redundant memory alignment. DFT function performs memory
alignment if the input memory is not aligned.
Optimization of PDSCH precoding operation
x86/aarch64 optimizations for precoding, more complete timing reporting in
nr_dlsim. Timing report in nr_dlsim now shows the total PDSCH generation time
and is broken down into Layer mapping, resource mapping and precoding.
The optimizations for x86 are basically reorganizing loops (for AVX512,AVX2,
NEON but not SSE2) to minimize repetitive instructions when formatting of the
precoder weights, and creation of inline functions for complex multiply-add.
AVX512 optimization was added. requiring AVX512BW. Detection of up to 4 PRBs
with common PMI for efficiency with AVX512 (was 2 in develop for AVX2).
For aarch64, native coding of the precoder to avoid inefficient SIMDe
translations for this operation.
nr_dlsim now uses default precoding indices depending on the scenario instead
of 0 which doesn't do precoding at all.
See merge request description for nr_dlsim test results.
CI: Enable Physim threshold check for tests on Caracal and GH
This MR adds support for configurable timing threshold files in PhySim
tests.
- Introduces a new CMake cache variable PHYSIM_CHECK_FILES to specify
one or more timing threshold files (semicolon-separated).
- Enables optional enforcement of timing thresholds check in PhySim
tests without changing existing test definitions.
- Defines timing thresholds for LDPC encoding/decoding in nr_ulsim,
nr_dlsim and ldpctest on Caracal and GH machines.
SRS noise power
This MR implements the calculation of the noise power based on SRS in a
different way, and check whether SNR is more accurate or not. This MR is
built on top of the MR: !3460
A graph was added in T_Tracer to observe UL-SNR calculated using SRS.
See merge request description for measurement results.
Replace the previous add_timed_physim_test() macro with add_physim_test()
to use a single, unified function for defining physim tests. This simplifies
the test configuration and removes redundant macro definitions while keeping
all existing functionality.
This change follows up on the earlier addition of support for including
custom timing threshold files (PHYSIM_CHECK_FILES).
Update existing tests accordingly.
For some physical simulators, we have multiple SNR iterations.
The previous version of the script correctly retained only the last
measurement, but if a previous measurement condition set RC=1, the
script would fail although the final "good" run might fulfil the
measurement condition (IOW, it would show SUCCESS, but still fail on the
check).
Instead, check if all measurement lines finish on SUCCESS, which handles
the above case.
The awk script in analyze-timing.sh used syntax that works with gawk but fails
on systems where mawk is the default implementation, producing errors such as:
awk: /dev/fd/3: line 6: syntax error at or near ,
-- test command finished with SIGPIPE
Concretely, the match(text, pattern, variable) function does not exist
in mawk, as it cannot capture matches in variables. Work around by
substituing strings, and force a numeric comparison.
Make all targets compile
This MR enables us to compile all targets, and does that in one CI test. This
closes#726. Now, this should work:
cd ~/openairinterface5g
mkdir build && cd build
cmake .. -GNinja && ninja
To achieve this, I did the following:
- some targets have been fixed (warnings, errors)
- some targets have been removed when they did not seem to serve a good purpose
(e.g., an almost empty test)
- some targets have been commented (e.g., LTE simulators, some unitary tests)
which we maybe want to revive in the future
- some targets built when just building with ninja, but not all. For instance,
telnetsrv_enb would typically build, but it was missing a dependency. Doing
cmake .. -GNinja && ninja telnetsrv_enb would show this (in develop)
To help with the last bullet point, there is a script compile_targets.sh that
can compile targets independently (and in parallel, to not wait for ages) to
show that targets compile properly compile-targets.patch.
In almost all cases, the problem is that a specific module uses the logging
module, which depends on T, which needs to dynamically create headers. The way
to do this is to have a dependency, and many targets use generate_T or {T_LIB}
({} intentional: T might be absent, so then it evaluates to nothing). However,
not all do this, and those that depend on T actually don't depend on T, but on
the logging module, which in turn depends on T. So in many places, a simple
dependency onto log_headers fixes this problem, and generate_T or {T_LIB}
simplifies the dependency (but not always). The commits should have more info.
Note that at least the targets in FlexRIC do not all compile, at least in the
configuration that we use in docker/Dockerfile.build.ubuntu. So for the latter,
we cannot build with a simple ninja, but FlexRIC seems to be the only one that
does not compile. Since FlexRIC is a separate repository, that needs to be
addressed in the corresponding repository.
Since there is no UCI on PUSCH implementation in gNB, it is good to test
UE's implementation with matlab.
This commit allows to test with matlab in two ways:
1. generate OFDM waveform from ulsim and decode it in matlab using
uci_on_pusch_decode.m
2. generate codeword in matlab using uci_on_pusch_encode.m read it in
ulsim and compare the bits. ulsim option -o is used
Code cleanup (SRS, config files and more)
- Remove deprecated elements from some configuration files (also closes#997)
- Improve SRS code with a better separation between common code and gNB/UE code
- Other minor compilation improvements
CI: Generalize PhySim tests, run PhySims on Gracehopper
This changeset generalizes/harmonizes the CI infrastructure for physim
testing, and adds a new pipeline for physim tests on Gracehopper
machines.
Specifically:
- Refactor the existing CI code for running and analyzing physical
simulators. The existing code assumed those tests are deployed using
OpenShift. Instead, OpenShift/Helm-specificities (e.g., running a
"release") have been removed, and instead of hardcoding the deployment
of the physical simulators using a script for Helm, the function takes
any script that can deploy using Helm/docker/from source. The overall
appearance is the same.
- Add a new pipeline RAN-PhySim-GraceHopper-5G that executes physical
simulators on gracehopper. Reporting is done in the same way as for
existing physical simulators (e.g., like in the RAN-PhySim-Cluster-5G
pipeline), but deployed in docker instead of helm.
-> some tests don't work reliably on ARM yet and are skipped
- the RAN-gNB-N300-Timing-Phytest-LDPC pipeline has been reworked to use
the same code for testing T2 offload, with ctest, and the same
reporting.
- the CUDA LDPC test XML has been removed, as these tests don't work
reliably.
- add a new macro add_timed_physim_test() that allows to define
thresholds to be checked, so that ctest will check for them. see the
documentation for more information
Prior to this change, each test would have, as label, it's description.
This leads to many labels, which are normally reserved for grouping
tests. Move the description into a custom property instead.
Since we rely on the JSON output for the generation of HTML reports, we
need to also extract the description from the generated JSON. However,
custom properties are only logged starting from cmake 3.30. For
back-wards compatibility, write an additional, separate environment
variable TEST_DESCRIPTION from which the description can be read as a
fallback.
For the add_timed_physim_test(), this has been already done in a parent
commit.
Jaroslava Fiedlerova