The memory alignment is required to use c16multaddVectRealComplex
in SRS channel interpolation. With a change in SRS comb_offset, the
existing algorithm was giving segmentation fault.
The existing srs channel estimation required memory alignment at two
stages. 1. When k = 0 (the first srs subcarrier), 2. When subcarrier <
K_TC (wrapped around after ofdm_symbol_size). Thus two memory alignments
are required leading to complications.
The srs channel estimation is now simplified to do it linearly instead of
wrapping around. This simplified the memory alignment required only when
k = 0.
When doing analogue beamforming, the beam ID for a pre-defined beam
table is passed directly to RU and there is no need to store the beam
table in PHY.
1. Removed function pointers for functions get_samples_per_slot(),
get_samples_slot_timestamp() and get_slot_from_timestamp() as it does
not offer any advantage. The functions are now called directly from
everywhere.
2. Optimized these functions by removing loops where ever possible.
All nFAPI message have copy helpers, so use that for "loading" nFAPI
messages into the L1. This also fixes a bug/error message
2. 0 no corresponding tx_data.request for dl_tti.request index 11, dropping
because the loop over DL_tti.request assumed that all messages are PDSCH
PDUs, when we can have others such as PDCCH, etc. In other words, when
looking up indices ("11" in the example above), we were checking also
inside other messages for SSB, PDCCH, etc, which is of course wrong.
That loop was an attempt to validate matching PDSCH PDUs with
TX_data.requests. This check is already done inside
phy_procedures_gNB_TX() through an assert, but this commit replaces with
the equivalent check previously inside nr_pnf_p7_get_msgs(). This is ok,
as the L1 cannot trust the scheduler (in monolithic) to provide the
correct information, either.
Fixes: eefb9da69a ("Reimplement nFAPI message exchange after msgDataTx removal")
Remove schedule response and update L1 threading documentation
This merge requests removes the somewhat superfluous "schedule response"
mechanism. "schedule response" was run after running the scheduler
itself and copied various FAPI messages into a global msgDataTx
structure from which the various TX/RX jobs were triggered. It is not
necessary, though, given that the scheduler is run inside the tx_func()
(running the TX jobs of the L1), and so the scheduler output FAPI
messages can directly be read into a variable of the existing
NR_Sched_Rsp_t type. This allows to delete a couple of hundred lines of
code (while retaining functionality), and simplifies the msgDataTx type
(which is still there to trigger TX job (including the scheduler) from
the ru thread).
In short, prior, there was this call chain:
- tx_func() calls run_scheduler_monolithic() (through fptr
NR_slot_indication())
- get some memory through allocate_sched_response() which has enough
space for NR_Sched_Rsp_t type variable
- call the scheduler and fill the variable
- call nr_schedule_response() (through fptr NR_schedule_response())
which copies into L1: to msgDataTx for TX, to separate structures
for RX
this is changed to
- tx_func() runs one message at a time, hence we can use memory in the
data segment for NR_Sched_Rsp_t. It calls run_scheduler_monolithic()
through the fptr
- pointers into NR_Sched_Rsp_t are passed to the scheduler, which
fills as before
- we return, memory is in its right place and does not need an extra
copy.
Additionally, this merge request updates documentation on threading and
cleans up some unused code along the way.
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.
Similarly to the parent commit, simplify the array type to the FAPI PDU
towards further refactoring, and encapsulate the code in a function for
easier reuse in a later refactoring step.
Note on ZP-CSI: after this commit, we continue treating CSI PDUs. The
prior code was returning, which I think is wrong (as this information is
only for the currently treated CSI-RS PDU).
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.
Both (DL) PDCCH and UL PDCCH PDU types are fundamentally of the same
type, so correct the array's type to facilitate refactoring later (a
later commit will reuse a single "generate DCI" function for both
cases).
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
Synchronized Real-Time Data Recording Application Addition
The main Features of Synchronized Real-Time Data Recording Application:
- The Data Recording Application runs in parallel to the gNB or UE
server or in another server as Data Recording or Data Lake Server
- It makes use of (an extended version) of OAI’s T-tracer framework
- It communicates with base station, and user terminal
- It synchronizes and combines data traces and control information
(meta-data) from Base station and User terminal into consistent SigMF
data sets
Details on implementation addition:
- T_tracer Extension to Support up to 35 Parameters in the T() macro
- Add traces on gNB PHY UL and related requirement: FD PUSH IQ, FD DMRS, and UL Channel estimates wi/wo interpolation, and Rx Payload bits:
1. The required memories have been created and initialized in:
openair1/PHY/NR_TRANSPORT/nr_ulsch_demodulation.c
2. Those memories: memory for DMRS signals, memory for channel
estimates based on DMRS positions, memory to store slot grid with
channel coefficients based on DMRS positions after interpolation,
and memory to store extracted data including PUSCH + DMRS.
3. The input parameters to nr_pusch_channel_estimation() has been
updated to get the pointer to the added memories of
pusch_dmrs_slot_mem and pusch_ch_est_dmrs_pos_slot_mem
(openair1/PHY/NR_ESTIMATION/nr_ul_estimation.h)
4. The DMRS grid of NR slot and the estimated channel coefficients at
DMRS locations are logged to created memories in:
openair1/PHY/NR_ESTIMATION/nr_ul_channel_estimation.c
5. The function inner_rx in
openair1/PHY/NR_TRANSPORT/nr_ulsch_demodulation.c is updated to log
the extracted ULSCH PRBs and the interpolated channel coefficients.
6. The T-macro() has been called in
(openair1/PHY/NR_TRANSPORT/nr_ulsch_demodulation.c) to trace added
traces: T_GNB_PHY_UL_FD_DMRS, T_GNB_PHY_UL_FD_CHAN_EST_DMRS_POS,
T_GNB_PHY_UL_FD_PUSCH_IQ, T_GNB_PHY_UL_FD_CHAN_EST_DMRS_INTERPL
- Add traces on UE PHY UL and related requirement: Tx Scrambled bits and Tx payload bits
1. Update input parameters to nr_ulsch_encoding () in
nr_transport_proto_ue.h to get number_dmrs_symbols to be added to
meta-data.
2. In nr_ulsch_coding.c: Trace UE PHY UL Payload TX bits (T_message:
T_UE_PHY_UL_PAYLOAD_TX_BITS) including related meta-data and UTC
time.
3. In nr_ulsch_ue.c: Trace UE PHY UL Scrambled TX bits (T_message:
T_UE_PHY_UL_SCRAMBLED_TX_BITS) including related meta-data and UTC
time.
4. Fix the input prameters in nr_ulsch_encoding () called in
(openair1/SIMULATION/NR_PHY/ulschsim.c) since a new input parameter
has been added.
- Add Data Collection Services (T-Tracers for gNB and UE)
1. Add _tracer_app_gnb and t_tracer_app_ue apps to common/utils/T/tracer/Makefile
2. Add Data collection Service for gNB: common/utils/T/tracer/t_tracer_app_gnb.c
3. Add Data Collection Service for UE: common/utils/T/tracer/t_tracer_app_ue.c
- Add Shared memory configuration
(common/utils/T/tracer/shared_memory_config.h), since we will have too
much memory consumption on the stack if the T-tracer is not fast
enough to process the traces or if we have too much traces. As a
result, tracers start to capture data, write data in shared memory,
then close the T-Tracer after getting the recording of N slots. The
Data recording App will read data from the shared memory.
- Add Main app of Data Recording Application v1.0 including: Config
files, SigMF Interface, Config Interface, Sync Service, and def
messages: The data recording App includes mainly the following files:
1. The main app: data_recording_app_v1.0.py that has the data control
service, interact with tracers, data sync and conversion service.
2. The configuration file:
common/utils/data_recording/config/config_data_recording.json
3. The wireless dictionary file:
common/utils/data_recording/config/wireless_link_parameter_map.yaml
4. Config Interface
(common/utils/data_recording/lib/config_interface.py) that reads
the configuration for both gNB and UE.
5. Sync interface (common/utils/data_recording/lib/sync_service.py)
that sync captured data for gNB and UE before saving them in SigMF
format
6. The SigMF Interface
(common/utils/data_recording/lib/sigmf_interface.py) to store
recorded data and meta-data in SigMF format
7. A simple script
(common/utils/data_recording/sync_validation_demo.py) to validate
that the recorded bits from gNB and UE are in Sync
- For more info look to doc/data_recording.md
Rakesh Mundlamuri