openairinterface5G merge requestshttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests2024-03-29T08:55:42Zhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2652Provide additional troubleshooting documentation for QtScope2024-03-29T08:55:42ZRobert SchmidtProvide additional troubleshooting documentation for QtScopeOn Wayland (e.g., Gnome on Ubuntu 22), additional steps might be
necessary to make QtScope run. Describe them in the doc.
Closes: #696
[Direct link to document section](../dqt-troubleshoot-doc/openair1/PHY/TOOLS/readme.md#troubleshoot)On Wayland (e.g., Gnome on Ubuntu 22), additional steps might be
necessary to make QtScope run. Describe them in the doc.
Closes: #696
[Direct link to document section](../dqt-troubleshoot-doc/openair1/PHY/TOOLS/readme.md#troubleshoot)REVIEW_CAN_STARTArash SahbafardArash Sahbafardhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2651Nrppa procedures srs2024-03-29T07:37:48ZFlorian KaltenbergerNrppa procedures srshttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2650Fix Msg4 false feedback in UE2024-03-29T13:42:04ZSakthivel VelumaniFix Msg4 false feedback in UE## Issue
If the msg4 feedback is lost, the UE doesn't send feedback for retransmissions of msg4 and the gNB reaches max harq retransmissions. Hence, the UE thinks it is connected while the RA has failed at the gNB. The reason is that aft...## Issue
If the msg4 feedback is lost, the UE doesn't send feedback for retransmissions of msg4 and the gNB reaches max harq retransmissions. Hence, the UE thinks it is connected while the RA has failed at the gNB. The reason is that after successful reception of msg4, the UE goes into connected state and stops monitoring the random access search space. So the subsequent DCIs for msg4 retransmissions are not received.
## Fix
After successful decoding of msg4, the UE continues to monitor the random access search space for some time.REVIEW_CAN_STARThttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2649NR SRS RRC assrtion removal2024-03-28T22:39:42ZFrancesco ManiNR SRS RRC assrtion removalThis MR removes an assertion in SRS RRC config that was already been removed in 28fb7a37d7b45cf3dd6371709a5f2d3eb38b3fb6 but somehow has been put back in the codeThis MR removes an assertion in SRS RRC config that was already been removed in 28fb7a37d7b45cf3dd6371709a5f2d3eb38b3fb6 but somehow has been put back in the codeREVIEW_COMPLETED_AND_APPROVEDhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2648Minor CI fixes, stack protection, fix E1+F1 Quectel test, L2sim5G test2024-03-29T08:04:52ZRobert SchmidtMinor CI fixes, stack protection, fix E1+F1 Quectel test, L2sim5G test- install gdb in gnb-aw2s images
- fixes for debugging core dump script
- experimental: add stack protection, [source](https://best.openssf.org/Compiler-Hardening-Guides/Compiler-Options-Hardening-Guide-for-C-and-C++.html#enable-run-time...- install gdb in gnb-aw2s images
- fixes for debugging core dump script
- experimental: add stack protection, [source](https://best.openssf.org/Compiler-Hardening-Guides/Compiler-Options-Hardening-Guide-for-C-and-C++.html#enable-run-time-checks-for-stack-based-buffer-overflows). *if it does not work automatically, will be removed*
- fix radio of E1+F1 Quectel test
- remove L2 sim 5G testREVIEW_IN_PROGRESSThomas Laurentlaurent.thomas@open-cells.comThomas Laurentlaurent.thomas@open-cells.comhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2647NR UE mixed fixes2024-03-29T13:55:30ZFrancesco ManiNR UE mixed fixesShould close #765Should close #765https://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2646NRUE fix asn_sequence_del to clear a list2024-03-28T21:06:59ZFrancesco ManiNRUE fix asn_sequence_del to clear a listThis MR addresses the issue of element deletion from the sequences with a backward iteration method to prevent skipping elements during deletion.This MR addresses the issue of element deletion from the sequences with a backward iteration method to prevent skipping elements during deletion.REVIEW_COMPLETED_AND_APPROVEDhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2645add error code for pdu session reject in the ue2024-03-28T07:53:40ZThomas Laurentlaurent.thomas@open-cells.comadd error code for pdu session reject in the uethis message was processed as unknown message, this MR adds explicit message when the 5GC refuses the pdu session
it should be continued later, but we don't know what to do currently when 5GC refuses the pdu sessionthis message was processed as unknown message, this MR adds explicit message when the 5GC refuses the pdu session
it should be continued later, but we don't know what to do currently when 5GC refuses the pdu sessionOK_TO_BE_MERGEDhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2644NR UE prevent segfault at detach2024-03-28T07:52:36ZFrancesco ManiNR UE prevent segfault at detachDCI formats 11 and 01 assume the presence of an active BWP, if not present return with an error to prevent segfault.
UE logs from CI:
```
610694.175746 [NAS] I [UE 0] Received NAS_DEREGISTRATION_REQ
610694.194338 [NAS] I [UE 0] Receive...DCI formats 11 and 01 assume the presence of an active BWP, if not present return with an error to prevent segfault.
UE logs from CI:
```
610694.175746 [NAS] I [UE 0] Received NAS_DEREGISTRATION_REQ
610694.194338 [NAS] I [UE 0] Received NAS_DOWNLINK_DATA_IND: length 3 , buffer 0x6020000302d0
610694.194348 [NAS] I received deregistration accept
610694.400593 [NR_RRC] I [UE 0] Received RRC Release (gNB 0)
610694.400606 [PDCP] I release DRB 1 of UE 0
../../../openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.c:3407:47: runtime error: member access within null pointer of type 'const struct NR_UE_DL_BWP_t'
```OK_TO_BE_MERGEDhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2643Enable AT command interface in NR UE2024-03-29T01:34:08ZSakthivel VelumaniEnable AT command interface in NR UEThis MR is a continuation of !2485.
## List of changes
1. Changes to NAS to send multiple PDU session request.
2. Enable AT command interface. The command parser and API already exists. Added new handler functions to set and activate PD...This MR is a continuation of !2485.
## List of changes
1. Changes to NAS to send multiple PDU session request.
2. Enable AT command interface. The command parser and API already exists. Added new handler functions to set and activate PDP contexts.
3. Extended parser function to parse NSSAI from AT command.
4. Added two command line parameters. One to enable usage of AT-command interface (`--AT`) and another to enter the virtual serial port to be used by the AT interface (`--AT-interface`).
## Usage
1. Run `socat` to connect two virtual serial interfaces. This outputs two interface names.
2. Run UE with `--AT --AT-interface "<interface-1>"`
3. Open interface-2 with `screen` or `minicom` and send the AT commands.
4. Supported commands: `at+cgdcont` and `at+cgact`.REVIEW_CAN_STARTSakthivel VelumaniSakthivel Velumanihttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2642NR UE demoting to LOG_D another log that cause flooding2024-03-27T07:12:34ZFrancesco ManiNR UE demoting to LOG_D another log that cause floodingdemoting to LOG_D another log that otherwise would be printed too many times at the UE because of periodic calling of to_nrarfcn (plus some code refactoring)demoting to LOG_D another log that otherwise would be printed too many times at the UE because of periodic calling of to_nrarfcn (plus some code refactoring)OK_TO_BE_MERGEDhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2641NR DL and UL channel estimation fix2024-03-28T10:26:55ZFrancesco ManiNR DL and UL channel estimation fixIn both DL and UL channel estimation, when filling `ul_ch_estimates` the DMRS antenna port was used. Instead when looking into the same structure `ul_ch_estimates` we would use the layer index. The two parameters may not be the same, the...In both DL and UL channel estimation, when filling `ul_ch_estimates` the DMRS antenna port was used. Instead when looking into the same structure `ul_ch_estimates` we would use the layer index. The two parameters may not be the same, the latter should be used.REVIEW_COMPLETED_AND_APPROVEDhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2639NR gNB PUCCH2 CQI workaround2024-03-28T07:54:32ZFrancesco ManiNR gNB PUCCH2 CQI workaroundSNR for PUCCH2 is not correctly computed. CQI cannot be used at MAC, otherwise it would lead to wrong assumptions for power handling.SNR for PUCCH2 is not correctly computed. CQI cannot be used at MAC, otherwise it would lead to wrong assumptions for power handling.OK_TO_BE_MERGEDhttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2636Fix various F1 problems2024-03-29T13:11:43ZRobert SchmidtFix various F1 problems- Fix a bug related to #706
- Fix a bug where receive a F1 message unknown to the CU makes it crash: closes #762
- Take time stamps of when last UE RRC activity has occurred
- ~Implement signal handler to do RRC "house keeping": inacti...- Fix a bug related to #706
- Fix a bug where receive a F1 message unknown to the CU makes it crash: closes #762
- Take time stamps of when last UE RRC activity has occurred
- ~Implement signal handler to do RRC "house keeping": inactive UEs will be released, as requested by @raymond.knopp~REVIEW_COMPLETED_AND_APPROVEDSagar Arorasagar.arora@openairinterface.orgSagar Arorasagar.arora@openairinterface.orghttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2635SCTP problems: allow to bind separate F1-C/F1-U interfaces; allow to use DNS2024-03-29T07:03:14ZRobert SchmidtSCTP problems: allow to bind separate F1-C/F1-U interfaces; allow to use DNS- revert commit to fix issue #727
- introduce new option to bind on a separate address for F1U; if not specified, will reuse F1C; fix in CI; add doc
- rewrite SCTP functions to use C's `getaddrinfo()` to look up DNS names. This also al...- revert commit to fix issue #727
- introduce new option to bind on a separate address for F1U; if not specified, will reuse F1C; fix in CI; add doc
- rewrite SCTP functions to use C's `getaddrinfo()` to look up DNS names. This also allows to IPv6, but more work in the stack is necessary to fully use IPv6 (C-plane works, U-plane doesn't)
- remove all interface name config options (are not used), remove `ipv6` config options (handled through `ipv4`)
Closes #727REVIEW_COMPLETED_AND_APPROVEDLuis Pereiralpereira@allbesmart.ptLuis Pereiralpereira@allbesmart.pthttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2633Handle F1 Reset at DU, and ensure state re-init on F1Setup2024-03-21T10:01:06ZXin Zhe KhooiHandle F1 Reset at DU, and ensure state re-init on F1SetupThis MR implements the following:
1. F1Reset handling at the DU -- DU_handle_RESET, and DU_send_RESET_ACKNOWLEDGE.
* This includes one patch to the RACH handling ( 721c16051d79f1c29b455e3b7653ec1842694945) when the UE states are abse...This MR implements the following:
1. F1Reset handling at the DU -- DU_handle_RESET, and DU_send_RESET_ACKNOWLEDGE.
* This includes one patch to the RACH handling ( 721c16051d79f1c29b455e3b7653ec1842694945) when the UE states are absent after RESET to prevent the DU from crashing
2. Ensure the UE-related states are always cleared on every F1SetupRequest (especially when DU re-establishes its SCTP connection with the CU).
* Quoted from TS38.473 8.2.3.2, "This procedure erases any existing application level configuration data in the two nodes and replaces it by the one received. This procedure also re-initialises the F1AP UE-related contexts (if any) and erases all related signalling connections in the two nodes like a Reset procedure would do."
* Impact: Now, restarting the CU midway will not result in a crash.
Both (1) and (2) use similar logic in re-init/initializing the UE-related states in the DU, i.e., clear MAC/RLC, F1, and GTP.
**HOWEVER**, I am not 100% sure whether my interpretation of the standards is correct for the above, i.e., Reset and F1Setup.
So, feedback and discussions are welcome.
Thanks!
Note:
* F1Reset handling at the DU is currently limited to the entire "f1-Interface", i.e., all states are cleared.
* F1Reset handling is not implemented at the CU; neither CU/DU currently sends out the F1Reset.
* F1Reset handling at the DU is tested using a transparent SCTP proxy in F1-C.https://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2631chore(ci): some docker improvements2024-03-22T07:54:51ZRaphael Defosseuxchore(ci): some docker improvements* to build `ran-base` only a few files are in fact required.
* It will help the cache usage
* Testing some target images improvements* to build `ran-base` only a few files are in fact required.
* It will help the cache usage
* Testing some target images improvementsREVIEW_IN_PROGRESSJaroslava FiedlerovaJaroslava Fiedlerovahttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2629Msg2 and Msg3 at gNB: check allocation before committing to Msg22024-03-29T14:34:49ZRobert SchmidtMsg2 and Msg3 at gNB: check allocation before committing to Msg2This MR improves Msg2/Msg3 handling in the following ways:
1. previously, on reception of the preamble, the scheduler calculated a single Msg2 "occasion" from the Msg1 received frame/slot (using slot_ahead, i.e., L1 internals). tihs is ...This MR improves Msg2/Msg3 handling in the following ways:
1. previously, on reception of the preamble, the scheduler calculated a single Msg2 "occasion" from the Msg1 received frame/slot (using slot_ahead, i.e., L1 internals). tihs is modified to only mark once we received a preamble, and then the TX scheduler will pick it up and try to allocate Msg2, possible until the end of RA window
2. Msg3 allocation was partially hardcoded, or used AssertFatal(). In this MR, we now check if Msg3 can be allocated before committing to Msg2
3. various refactorings, and log improvementsREVIEW_IN_PROGRESShttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2624NR UE MAC SR improvements2024-03-27T18:25:39ZFrancesco ManiNR UE MAC SR improvementsImprovements in handling scheduling request at NR MAC UE. This MR depends on !2604.Improvements in handling scheduling request at NR MAC UE. This MR depends on !2604.REVIEW_CAN_STARThttps://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/2620Fix cppcheck errors2024-03-27T09:46:27ZFrancesco ManiFix cppcheck errorsCurrently cppcheck in CI is not enforced. This MR fixes some of the reported errors and warnings.Currently cppcheck in CI is not enforced. This MR fixes some of the reported errors and warnings.REVIEW_CAN_START