Testing RA with OAI gNB and UE both using X300 via PCIe link indicated a value of 15 was way too small.
The correct value for this setup seems to be 65.
Tests with our USRP X300, connected via PCIe, show that tx_sample_advance of 45 is slightly too much.
Unfortunately, transmitting the PRACH preamble too early is very bad, slightly too late is no problem.
So change this value from 45 to 40. This helps receiving the PRACH preamble reliably.
Update in SIB1 implementation to solve REG to CCE mapping issues and also correction in pdcch RE extraction when BWP_start>0.
* MR 1031: LTE/NR X2AP: fix content of SgNBAdditionRequest
Wireshark (version 3.3.2) was not happy. https://asn1.io/asn1playground/ was not happy.
* MR 1036: LTE ULSCH bugfix
Minor fix in the UL scheduler
* MR 1008: NR SA ITTI sim update for core connectivity
Updates for connections to 5G core node using ITTI sim. Usage documentation have also been added.
* MR 1025: Cleanup NR RA procedure - mostly UE
MR1006 : itti enhancement
-new feature: creates itti queue during runtime
-modified itti message instance
MR1012 : 5G NR: UL transform precoding (SC-FDMA)
MR1022 : Wireshark t hack ueid
MR997 : LTE uplink improvement
-LTE update about uplink performance improvement and some bug fixes from Fujitsu.
-Fix uplink calculation error in SCD and PHY layer.
-Add interference calculation for PUSCH.
-Update TPC control.
-Fix nFAPI hang-up
-Fix DRB is non active after Re-establishment.
-Msg3 PHICH problem.
-Add statistical logs.
MR1030 : CI RHEL container update
-CI adaptation to build the shared and target images on asterix and obelix Python framework
MR1023 : ci ue parallel
-Enables OAI UE parallelization in CI tests (where applicable)
- bugfix in nr_ue_dci_configuration
- cleanup of irrelevant RA-related data structure members in UE
- moved msg1 init condition to MAC
- minor fixes concerning init_ra function
Recently we stared using the BWP_indicator signalled in DCI, before we always assumed BWP_Id == 1.
Now "fake-DCI" with bandwidth_indicator == 0 lead to segmentation faults, so we discard these.
Note: There are still several placed where we assume BWP_Id == 1!
Thomas Schlichter