Cellular Network Infrastructure » Radio Access Network » OsmoPCU

11bit RACH is actually being worked on, assigning this to aravind. It might even be complete for some BTS models by now.

Currently it is supported and integration tested for NuRAN LC1.5 and LC1.0. However there is no support for osmo-trx.

Encoding/decoding support for 11-bit RACH for libosmocoding is available in gerrit 5062.

Gerrit 5062 has been merged. Have to verify whether 11-bit RACH is supported on sysmoBTS.

Test with "gprs 11bit_rach_support_for_egprs" failed - seems like smth wasn't merged/missing. See related BTS bug.

The following patch should implement 11-bit RACH support of osmo-bts-trx:

https://gerrit.osmocom.org/#/c/osmo-bts/+/6315/

but should be properly tested.

Both osmo-trx and osmo-bts-trx do support 11-bit encoded Access Bursts now. We also have a TTCN-3 test case for that (see BTS_Tests.TC_pcu_ext_rach_content) and it passes. However, all my attempts to test the network with the real phones have been unsuccessful. Unfortunately, I have no time to investigate further now.

I'm trying to add some TTCN3 tests with 11bit RACH support for osmo-pcu in PCU_Tests_RAW as part of #4338. I'd say that's all what's missing here.

I'd say that's all what's missing here.

I don't think so. OsmoPCU does not seem to handle 11-bit RACH correctly when running with osmo-trx and osmo-bts-trx.

OsmoPCU does not seem to handle 11-bit RACH correctly when running with osmo-trx and osmo-bts-trx.

https://gerrit.osmocom.org/c/libosmocore/+/17692/ tests/coding: add 11-bit Access Burst samples from a real phone
https://gerrit.osmocom.org/c/libosmocore/+/17693/ coding: fix bit ordering in 11-bit RACH coding functions

With this fix applied, OsmoPCU works much better when running with osmo-bts-trx.
I also have a (draft) TTCN-3 test case for OsmoPCU.

I also have a (draft) TTCN-3 test case for OsmoPCU.

https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/17704/ library/RLCMAC_CSN1_Types: add EGPRS Packet Channel Request definition
https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/17706/ PCU: add test cases for EGPRS Packet Channel Request

For some reason, TITAN decodes different extended RA value in IA Rest Octets...
Any help would be appreciated. Wireshark shows that everything is ok (as expected).

TTCN-3 test cases have been merged and, as can be seen, they're passing in Jenkins.

Though I am still not satisfied by some code parts handling 11 bit EGPRS Packet Channel Request in osmo-pcu, for example:

static inline uint16_t mslot_class_from_ra(uint16_t ra, bool is_11bit)
{
        if (is_11bit)
                return (ra & 0x3e0) >> 5;

        /* set multislot class to 0 for 8-bit RACH, since we don't know it yet */
        return 0;
}

The way we parse the multislot class does not seem to be correct, because it's optional and its presence depends on type of EGPRS Packet Channel Request. I believe we should implement proper parsing, and here is a draft change for that: https://gerrit.osmocom.org/c/osmo-pcu/+/17718.

I just realized that RR Immediate Assignment Reject (IAR Rest Octets) should also contain Extended RA, when sent in response to EGPRS Packet Access Request. More details can be found in 3GPP TS 44.018, section 10.5.2.17.

Actually, Encoding::write_immediate_assignment_reject() does support writing IAR Rest Octets with Extended RA. We just need a TTCN-3 test case.

Actually, Encoding::write_immediate_assignment_reject() does support writing IAR Rest Octets with Extended RA. We just need a TTCN-3 test case.

https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/18420/ library/GSM_RR_Types: fix definition of FeatureIndicator
https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/18421/ library/GSM_RR_Types: add IAR Rest Octets definition
https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/18422/ library/GSM_RR_Types: add receive templates for IMM ASS Reject
https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/18423/ PCU: introduce TC_egprs_pkt_chan_req_reject_content
https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/18424/ PCU: introduce TC_egprs_pkt_chan_req_reject_emergency

https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/18425 PCU: introduce TC_egprs_pkt_chan_req_reject_exhaustion

While working on EGPRS Packet Channel Request coding support, I discovered and fixed some problems in CSN.1 codec:

https://gerrit.osmocom.org/c/osmo-pcu/+/18430/ csn1: fix M_CHOICE: restirct maximum length of the choice list
https://gerrit.osmocom.org/c/osmo-pcu/+/18431/ csn1: fix csnStreamEncoder(): also check length of the choice list
https://gerrit.osmocom.org/c/osmo-pcu/+/18432/ csn1: fix csnStreamEncoder(): always check the choice index
https://gerrit.osmocom.org/c/osmo-pcu/+/18433/ csn1: fix: never use enumerated types in codec structures

All previously submitted changes have been merged. These two new changes are waiting for review:

https://gerrit.osmocom.org/c/osmo-pcu/+/18386/ bts: refactor handling and parsing of RACH.ind
https://gerrit.osmocom.org/c/osmo-pcu/+/18387/ BTS::parse_rach_ind(): properly handle EGPRS Packet Channel Request

as soon as they're merged, we can finally close this ticket.

All patches have been merged.