This was only present on non x86 systems.
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Integration: 2026.w28

- #235 RRC fixes
- #247 Enforce N_PRB=2^x3^y5^z in SC-FDMA uplink scheduler
- #274 Fix PSS correlation peak value and average value on the OAI UE logs
- #261 ci: Speedup fhi72 dockerfiles
- #273 Simplifications for xran builds
- #270 xran: fixes for FDD operation
- #281 Handle invalid input in nr_get_E
- #149 fix(zmq): speedup zmq radio
- #260 O-RU first fix pass
- #202 doc: add a Git guide (commit signing, branch management, rerere)
- #249 SDAP header handling fixes
- #199 UE: implement TRS processing in PHY
- #197 Bump SIMDE commit to the most up-to-date (1c68d9a)

Closes: #118
Closes: #127
Closes: #128
Closes: #148
Closes: #262

Bump SIMDE commit to the most up-to-date (1c68d9a) (#197)
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>

UE: implement TRS processing in PHY (#199)

This PR implements PHY layer processing for CSI-RS tracking signal that
estimates the CFO and sends a frequency correction command to the radio
when CFO goes above a threshold of 500Hz.

A CI test is added with a GNU radio script between OCUDU gNB and the UE
that simulates a continuous frequency offset. The test passes if the UE
sends a frequency correction command.
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>
Reviewed-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>
Reviewed-by: Francesco Mani <email@francescomani.it>

Signed-off-by: Luis Pereira <lpereira@allbesmart.pt>

SDAP header handling fixes (#249)

OAI SDAP treated SDAP header presence as a single PDU-session flag
(enable_sdap), so a valid asymmetric RRC config (UL header present, DL
header absent) caused the UE to parse downlink payload as an SDAP QFI
and drop traffic (e.g. IPv4 0x45 -> bogus QFI 5).

SDAP now follows RRC SDAP-Config per DRB and per direction from (TS
38.331 / TS 37.324). Header presence, QFI-to-DRB mapping, and
default-DRB fallback all derive from what RRC configured for each DRB.

In the new implementation, RRC (SDAP-Config IE) fills sdap_config_t per
DRB (role, defaultDRB, mapped QFIs), SDAP entity stores it in
qfi2drb_table + default_drb row to be used in the TX/RX data path.

Also, simplify mapping logic, improves E1 QoS-flow remap, and completes
UE reflective QoS mapping with the correct entity_role on new QFI rows.
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>
Reviewed-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Signed-off-by: Sakthivel Velumani <s.velumani@northeastern.edu>

measurement bitmap for CSI-RS is 0 means CSI-RS for tracking. So don't
return and proceed with processing the CSI-RS.
Signed-off-by: Sakthivel Velumani <s.velumani@northeastern.edu>

TRS resources if configured by gnb, is present in two consequtive slots
and two symbols per slot. LS estimates from the first symbol is saved
and in second symbol frequency offset and time offset is estimated. The
estimated FO is saved in UE global structure which is used in the last
TRS slot. If the FO is above a threshold, the radio's center frequency
is adjusted.

Time offset is estimated but not used because time is already tracked
using PBCH DMRS.
Signed-off-by: Sakthivel Velumani <s.velumani@northeastern.edu>

doc: add a Git guide (commit signing, branch management, rerere) (#202)

Adds doc/git-guide.md, a single entry point for the practical Git
knowledge needed to contribute to OAI: commit signing setup (DCO +
verified commits), branch management (including fixup commits and
--autosquash), submodules, recovering from mistakes, and reusing
conflict resolutions with git rerere.

This started as a rerere-only guide. Review discussion (@luispereira106,
@rorsc, @sgarg00) converged on a broader scope: the Git how-tos are
scattered across the documentation (signing in CONTRIBUTING.md, branch
management in code-style-contrib.md), and configuring commit signing in
particular was reported as hard to get right. This PR consolidates the
how-tos in one guide while the policy documents stay authoritative for
the rules.

Single commit (review rounds squashed per the linear-history policy):

- New doc/git-guide.md with sections on commit signing (moved from
  CONTRIBUTING.md, led by a copy-pasteable SSH recipe), branch
  management (moved from code-style-contrib.md; git switch,
  fixup/--autosquash, --fixup=amend:), submodules (unintended
  pointer-update pitfall), recovering from mistakes
  (restore/reset/reflog), and git rerere.
- CONTRIBUTING.md keeps the normative DCO/Verified requirements and
  links to the guide; code-style-contrib.md keeps the workflow policy
  and links to the guide; doc/README.md lists the guide under Developer
  tools.
- Cross-references (not moved, since they are coupled to their own
  docs): clang-format.md pre-commit hook, GET_SOURCES.md branch/tag
  model, doc_best_practices.md docs: commit prefix.
- Small typo fix in the T tracer addconsoletrace.md documentation.
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>
Reviewed-by: Francesco Mani <email@francescomani.it>
Reviewed-by: Shubhika Garg <shubhika.garg@openairinterface.org>

O-RU first fix pass (#260)

Added several fixes requested in the original PR

- simplify code
- add example config files & test procedure
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>

fix(zmq): speedup zmq radio (#149)

- Utilize ZMQ "zero-copy" inteface on TX
- Skip one memcpy in RX
- Use AVX-512 optimized cf_t <-> c16_t conversion loops in tx and rx
- Parallelize TX and RX polling threads for faster execution

Closes: #118
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>

Handle invalid input in nr_get_E (#281)

This PR prevents the function nr_get_E to crash with an assertion for
invalid input by handling the scenario with an error code
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>

xran: fixes for FDD operation (#270)

Two fixes in the xran fronthaul found while bringing up an FDD cell
(Band n1, 10 MHz).  Verified on a Band n1 FDD 10 MHz cell using Keysight
RUsim/UEsim, End-to-end UE attach and traffic work.

- Derive TTI from the configured numerology instead of a hardcoded 20
  slots per frame: xran_fh_tx_send_slot() assumed 30 kHz SCS. Compute
  slots_per_frame as 10 << mu from the fronthaul config (mu_number[0]
  for K release, frame_conf.nNumerology for F release), so the TTI is
  correct for any numerology. With the original code, PRACH was not
  working properly.
- Fix DL transmission in FDD mode: the TDD DL/guard slot check skipped
  every DL slot in FDD, where no TDD pattern exists and all slots carry
  DL. The check is now only applied when the frame duplex type is not
  FDD. Without this, no DL traffic was sent.
Reviewed-by: Teodora Vladić <teodora.vladic@openairinterface.org>
Reviewed-by: Robert Schmidt <robert.schmidt@openairinterface.org>

Simplifications for xran builds (#273)

- make xran_DOWNLOAD=ON default - skips an extra argument in cmake
  command
- Automatically detect aarch64 and select correct target for external
  xran build
Reviewed-by: Jaroslava Fiedlerova <jaroslava.fiedlerova@openairinterface.org>
Reviewed-by: Teodora Vladić <teodora.vladic@openairinterface.org>

Add doc/git-guide.md, a single entry point for the practical Git
knowledge needed to contribute to OAI, consolidating the how-tos
that were scattered across the documentation:

- Setting up commit signing: moved from CONTRIBUTING.md, which keeps
  the normative DCO/Verified requirements and links to the guide; led
  by a compact SSH recipe, with the allowed_signers troubleshooting
  and a note on repository-level configuration for shared servers.
- Managing your own branch: moved from code-style-contrib.md, which
  keeps the workflow policy and links to the guide; mentions git
  switch and covers fixup commits with git rebase --autosquash,
  including the --fixup=amend:<commit> variant.
- Working with submodules: the update command and the unintended
  submodule-pointer-update pitfall seen in past PRs.
- Recovering from mistakes: git restore --staged, git restore,
  git reset --soft/--hard, and git reflog.
- Reusing conflict resolutions with git rerere: enabling, typical
  flow, inspecting, seeding from existing history, sharing the cache,
  and caveats, clarifying that contribution branches must keep a
  linear history while forks may carry merge commits.

The guide is referenced from doc/README.md (Developer tools),
code-style-contrib.md, and CONTRIBUTING.md. Guidance coupled to its
own document is cross-linked instead of moved: clang-format.md
(pre-commit hook), GET_SOURCES.md (branch and tag model), and
doc_best_practices.md (docs: commit prefix).

Also fix a typo in the T tracer addconsoletrace documentation.
Signed-off-by: Andrea Lacava <thecave003@gmail.com>

Log, param help, and rrc-usage.md described enable_sdap as toggling the
SDAP layer, while it only sets sdap-HeaderUL/DL in RRC SDAP-Config.
Signed-off-by: Guido Casati <guido.casati@openairinterface.org>

Signed-off-by: Karim Boutiba <karim.boutiba@openairinterface.org>

Previously, is_tdd_ul_symbol() and is_tdd_dl_symbol() returned true
unconditionally in FDD. This is correct for UL/DL presence (in FDD,
every symbol carries both UL and DL), but it breaks logic that relies
on the TDD property that a UL symbol is not a DL symbol (and vice
versa), e.g. the computation of the last DL symbol in mixed slots. In
FDD, this assumption does not hold, leading to incorrect behavior.

Remove the FDD shortcut from the TDD symbol helpers so they strictly
answer the TDD question, and make the FDD case explicit at each
function call (nFrameDuplexType != XRAN_FDD && ...). This harmonizes
the TDD and FDD paths: in FDD, no slot/symbol is skipped on RX or TX,
and guard-slot handling only applies to TDD.
Signed-off-by: Karim Boutiba <karim.boutiba@openairinterface.org>

The TTI computation in xran_fh_tx_send_slot() assumed 20 slots per
frame (30 kHz SCS only). Compute slots_per_frame as 10 << mu from the
fronthaul configuration instead, so the TTI is correct for any numerology.
Signed-off-by: Karim Boutiba <karim.boutiba@openairinterface.org>

In nr_sdap_rx_entity, the RX callback receives the DRB identity from PDCP
(drb-Identity). Rename pdcp_entity to drb_id to be aligned with TS 37.324 terminology.

Changes:
- Rename the RX callback argument through the full SDAP RX surface:
sdap_data_ind(), the rx_entity function pointer table, and nr_sdap_rx_entity().
Signed-off-by: Guido Casati <guido.casati@openairinterface.org>

After new per-DRB SDAP roles changes, UE downlink reflective mapping
(TS 37.324 §5.3.2) still stored only drb_id on the new QFI row.
entity_role stayed NO_SDAP_HEADER, so later uplink TX and §5.3.1
end-marker logic read the wrong header config from that row.

When storing the reflective UL rule, copy the receiving DRB entity_role
(from nr_sdap_drb_lookup) together with drb_id, matching qfi2drb_map_add.

Changes:
- nr_sdap_rx_entity(): set qfi2drb_table[qfi].entity_role on RDI store

Refs:
- TS 37.324 §5.3.1, §5.3.2
- TS 38.331 SDAP-Config (sdap-HeaderUL / sdap-HeaderDL)
Signed-off-by: Guido Casati <guido.casati@openairinterface.org>

In fill_e1_drb_to_setup(), derive E1 SDAP-Config header-present booleans
with logical negation instead of redundant ternary expressions.

Changes:
- NGAP: sDAP_Header_UL/DL assignment in fill_e1_drb_to_setup()
Signed-off-by: Guido Casati <guido.casati@openairinterface.org>

Parent commits route QFI-to-DRB lookup through qfi2drb_map (including
default-DRB fallback per TS 37.324 clause 5.2.1), but gNB UL RX still
checked the raw mapping table and accepted any mapped QFI on whatever
DRB received the PDU.

On gNB UL SDAP PDUs with a header, nr_sdap_rx_entity resolves the QFI
through entity->qfi2drb_map, drops when no rule and no default DRB apply,
and drops when the resolved DRB does not match the receiving DRB before
GTP-U forwarding.

Changes:
- parse UL header with nr_sdap_ul_hdr_t
- call qfi2drb_map for QFI validation
- verify receiving-DRB match
- simplify gNB UL block

Refs:
- TS 37.324 clause 5.2.1 (Uplink), clause 6.2.2.3 (UL data PDU)
Signed-off-by: Guido Casati <guido.casati@openairinterface.org>

E1 DRB-To-Modify already carries the new QFI list. The old
nr_sdap_entity_update_qos_flows() duplicated that as a bare qfi[] plus other params
while map_update expects a full sdap_config_t (add/release lists and per-DRB
role). Route the E1 path through one struct instead.

Changes:
- nr_sdap_entity_update_qos_flows(): take sdap_config_t *, caller fills the struct,
  function sets role and defaultDRB from qfi2drb_table, builds mappedQFIs2Release
  for QFIs dropped from the DRB, then calls qfi2drb_map_update()
- cucp_cuup_handler.c: build stack sdap_config_t from E1 qosFlows[] and call
  the new API
Signed-off-by: Guido Casati <guido.casati@openairinterface.org>

A DRB may have an uplink SDAP header but no downlink header (valid TS 38.331).
The old entity->enable_sdap flag was true if either direction had a header, so UE
downlink wrongly read byte 0 as QFI. IPv4 payloads start with 0x45 -> bogus QFI 5
and dropped traffic before the TUN (PR !158).

Header presence must follow TS 37.324 §5.2.1/§5.2.2 per DRB and direction, not a
single PDU-session-wide enable-sdap.

Also, store the default DRB as a `qfi2drb_t` mapping in the SDAP entity and use
it directly for fallback paths.

Changes:
- Add nr_sdap_drb_lookup() to return first qfi2drb_table row for a DRB
- nr_sdap_rx_entity(): parse QFI only when the receiving direction has a header,
  drop if the DRB has no table row
- nr_sdap_tx_entity(): UL/DL TX header bits from qfi2drb_map() (drb_id + entity_role),
  including default-DRB fallback
- nr_sdap_qfi2drb(): const qfi2drb_t *, remove SDAP_MAP_RULE_EMPTY
- nr_sdap_map_ctrl_pdu(): handle NULL from qfi2drb_map
- nr_sdap_qfi2drb_map_update(): sync entity_role on
  all QFIs for the DRB. At most one DRB when both headers absent.
- nr_sdap_entity_update_qos_flows(): keep per-DRB role from nr_sdap_drb_lookup(),
  skip update when DRB has no row
- Cleanup entity->enable_sdap
- nr_sdap.c: remove TUN-thread QFI drop, tx_entity owns unmappable-QFI handling
- replace default_drb with qfi2drb_t default_drb in SDAP entity and init
  default DRB id+role at SDAP entity creation

Refs:
- PR #158
- TS 37.324 §5.2.1, §5.2.2, §6.2.2.1
- TS 38.331 SDAP-Config (sdap-HeaderUL / sdap-HeaderDL)
Signed-off-by: Guido Casati <guido.casati@openairinterface.org>

To run the test with OAI UE and vrtsim, use the following commands, assuming
the executables are in the build directory.

OAI O-RU:
  ./nr-oru -O ../targets/PROJECTS/GENERIC-NR-5GC/CONF/ru.band77.mu1.106rb.1x1.conf --device.name vrtsim --vrtsim.role server

OAI 7.2 gNB:
  ./nr-softmodem -O ../targets/PROJECTS/GENERIC-NR-5GC/CONF/gnb.band77.mu1.106rb.fhi.1x1.conf

OAI NR UE:
  ./nr-uesoftmodem -C 4049760000 -r 106 --numerology 1 --ssb 516 --device.name vrtsim
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Replace the single polling thread that handled all channels
with separate, dedicated polling threads per TX and RX channel.
This avoids calling `zmq_poll` across all sockets simultaneously,
reduces lock contention, and enables concurrent socket operations
on multiple antennas.
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Add `TxRxSamplesSIMD` to validate the correctness of the new ZMQ
device transmit and receive paths with int16 complex samples.
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Optimize the TX path by converting int16 complex samples to float
complex format directly into ZMQ message buffers using SIMD. Queue
these ZMQ messages (`zmq_msg_t`) directly, avoiding temporary internal
copying and enabling ZMQ's zero-copy interface during actual transmit
in the polling thread.
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Optimize the RX path by converting float complex samples received on ZMQ
to int16 complex format using SIMD right inside the polling thread. This
allows storing int16 complex (c16_t) samples directly inside the overflow_buffer,
skipping a temporary float buffer allocation and a memcpy in the receive function.
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Support float and int16 complex types
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Add AVX-512 and AVX2 optimized conversion loops between float complex (cf_t)
and int16 complex (c16_t) sample formats in radio/zmq/zmq_simd.h.
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Add a performance benchmark `BenchmarkThroughput` to measure the sample rate capacity
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Signed-off-by: Francesco Mani <email@francescomani.it>

This update allows automatic discovery of target platform when building xran.

Supproted platforms: armv8 and x86_64 with avx512
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>

Make xran_DOWNLOAD ON by default - this simplifies work in the most common
usecase and still allows to overwrite xran directory with xran_LOCATION.
Signed-off-by: Bartosz Podrygajlo <bartosz.podrygajlo@openairinterface.org>