[srslte-users] Issue when receiving long SIB by UE

Taillefer Simon simon.taillefer at telespazio.com
Mon Mar 15 15:13:14 UTC 2021


Hello Francisco,

It does not change the behaviour of the UE even if it changes the waiting TBS.
I can still receive the SIBs with a size of 200 and 212 bytes but not more.

Thank you for your help,

Simon

De : Francisco Paisana <francisco.paisana at softwareradiosystems.com>
Envoyé : samedi 13 mars 2021 19:51
À : Taillefer Simon <simon.taillefer at telespazio.com>
Cc : srslte-users at lists.softwareradiosystems.com; Arzel Laurent <laurent.arzel at telespazio.com>; Ayachi Soufian <soufian.ayachi at telespazio.com>
Objet : Re: [srslte-users] Issue when receiving long SIB by UE

Hey,

It is possible that there is a bug in the scheduler. Could you, in this line https://github.com/srsLTE/srsLTE/blob/master/srsenb/src/stack/mac/scheduler_grid.cc#L921, swap "bc_alloc.req_bytes" for "tbs / 8"? Let's see if that's the cause for the mismatch in the TBS encoding/decoding.

Kind Regards,
Francisco


On Wed, Mar 3, 2021 at 12:38 PM Taillefer Simon <simon.taillefer at telespazio.com<mailto:simon.taillefer at telespazio.com>> wrote:
Hello Francisco,

We want to transmit assistance data information from an E-SMLC server to multiple UEs.

In order to do this we receive LPPa messages through the LCS-AP and S1-AP protocols.
Then, these assistance data are encapsulated in posSIBs within an SI messages using the function generate_sibs in srsenb/src/stack/rrc/rrc.cc file exactly like the SIBs. And finally, they are received by the UE and decoded.

What you describe is exactly what is happening but we never touched the tbs neither in the ENB nor in the UE.
I think the timeout error is caused because the UE cannot decode the message or do not find it where it is waiting for as you said (shift in tbs).

There is no mismatch between the UE and ENB tbs when it works with a size of 121 bytes (see logs below).
“16:08:01.968978 [MAC ] [D] [   25] SCHED: SIB3, rbgs=(0,1), dci=(2,0), rv=0, len=121, period=8, mcs=16”
“16:08:06.313944 [PHY0] [I] [ 3869] PDSCH: cc=0, rnti=0xffff, nof_prb=4, nof_re=600, tbs={121}, mod={2}, rv={0}, crc={OK}, it=0.5, snr=37.1 dB”

Also, it works sometimes with 212 bytes even if it does not work for 205 bytes for example. But it works with 200 bytes even if there is a tbs shift. Please find attached the logs.

Thank you for your help,
Best regards,

Simon Taillefer

De : Francisco Paisana <francisco.paisana at softwareradiosystems.com<mailto:francisco.paisana at softwareradiosystems.com>>
Envoyé : lundi 1 mars 2021 17:11
À : Taillefer Simon <simon.taillefer at telespazio.com<mailto:simon.taillefer at telespazio.com>>
Cc : srslte-users at lists.softwareradiosystems.com<mailto:srslte-users at lists.softwareradiosystems.com>; Arzel Laurent <laurent.arzel at telespazio.com<mailto:laurent.arzel at telespazio.com>>; Ayachi Soufian <soufian.ayachi at telespazio.com<mailto:soufian.ayachi at telespazio.com>>
Objet : Re: [srslte-users] Issue when receiving long SIB by UE

Hi again,

I am not really sure what you are trying to achieve, but looking at the logs I was able to make the following observations:
- the UE sets the receive window for the SIB3 at win_start=1140 (ue logs):
14:29:13.531544 [RRC ] [I] Proc "SI Acquire" - Starting SI Acquisition procedure for posSIBs
14:29:13.531548 [MAC ] [I] [  970] SCHED: Searching for DL dci for SI-RNTI window_st=1140, window_len=20
14:29:13.531551 [RRC ] [I] Proc "SI Acquire" - Instructed MAC to search for SIB3, win_start=1140, win_len=20, period=16, sched_index=1

- the eNB sends the SIB of interest at tti=1155 (enb logs):
14:29:13.708292 [MAC ] [D] [ 1155] SCHED: SIB3, rbgs=(0,1), dci=(2,0), rv=2, len=223, period=8, mcs=24

- the UE detects the SIB and tries to decode it. Unfortunately, the decoding fails - see ack=KO (ue logs):
14:29:13.714318 [PHY0] [D] [ 1159] PDCCH looking for rnti=0xffff
14:29:13.714375 [PHY0] [I] [ 1159] PDCCH: cc=0, f=1A, cce= 0, L=3, riv=300, pid=0, mcs={24}, ndi={0}, tpc_pucch=0, snr=37.7 dB
14:29:13.714388 [MAC ] [D] [ 1159] Set HARQ for retransmission
14:29:13.714605 [PHY0] [I] [ 1159] PDSCH: cc=0, rnti=0xffff, nof_prb=4, nof_re=600, tbs={225}, mod={2}, rv={2}, crc={KO}, it=4.0, snr=37.7 dB
14:29:13.714617 [MAC ] [I] [ 1159] DL 0 (TB 0):  reTX  tbs=225, rv=2, ack=KO, ndi=0

The eNB tbs=223 (Transport Block Size) and the UE decoded tbs=225 do not seem to match.

It looks to me that the set tbs=223 in the eNB is not valid and is causing the mismatch. Can you try to pad it with 2 zeros, to see if it fixes the problem?

Kind Regards,
Francisco


On Mon, Mar 1, 2021 at 3:12 PM Taillefer Simon <simon.taillefer at telespazio.com<mailto:simon.taillefer at telespazio.com>> wrote:
Hello Francisco Paisana,

Please find attached the logs of the ENB and UE. I removed some lines in the UE logs because the files were too heavy.

Please also note that the message causing the issue is the message containing the posSIBs and the assistance data. It is also called SIB3 because it is the third SIB message sent to the UE (after the SIB1 and the message that contains all the other SIBs like SIB2, SIB3 and etc.).

Thank you for your help,

Simon Taillefer

De : Francisco Paisana <francisco.paisana at softwareradiosystems.com<mailto:francisco.paisana at softwareradiosystems.com>>
Envoyé : lundi 1 mars 2021 13:44
À : Taillefer Simon <simon.taillefer at telespazio.com<mailto:simon.taillefer at telespazio.com>>
Cc : srslte-users at lists.softwareradiosystems.com<mailto:srslte-users at lists.softwareradiosystems.com>
Objet : Re: [srslte-users] Issue when receiving long SIB by UE

Hello Taillefer Simon,

Can you share the logs of the eNB and UE?

Kind Regards,
Francisco


On Sun, Feb 28, 2021 at 2:00 PM Taillefer Simon <simon.taillefer at telespazio.com<mailto:simon.taillefer at telespazio.com>> wrote:
Hello srsLTE team,

We are having trouble with the reception of a long SIB message by the UE.

More precisely, the UE is not able to connect to the ENB because it timeouts if the SIB message with all other SIBs (SIB2, SIB3, …) created on the generate_sibs function in the srsenb/src/stack/rrc/rrc.cc file has a length of more than 173 bytes.
The UE loops on the si_acquire_proc::init function in the srsLTE/srsue/src/stack/rrc/rrc_procedures.cc.
The raised error are the following:

-          [RRC ] [I] Proc "SI Acquire" - SI Acquire Retry Timeout for SIB2 from srsLTE/srsue/src/stack/rrc/rrc_procedures.cc file line 369.

-          [RRC ] [E] Proc "SI Acquire" - Timeout while acquiring SIB2 from srsLTE/srsue/src/stack/rrc/rrc_procedures.cc file line 376

-          [RRC ] [E] Proc "SI Acquire" - Failed to acquire SIB2 from srsLTE/srsue/src/stack/rrc/rrc_procedures.cc file line 320

Do you know if there is a solution to receive longer messages to allow the UE to connect with the ENB?

Thank you in advance for your help,

Simon Taillefer




-WARNING: This message contains confidential and/or proprietary information which may be subject to privilege or immunity and which is intended for the use of its addressee only.
Should you receive this message in error, you are kindly requested to inform the sender and to definitively remove it from any paper or electronic format.
Any other use of this e-mail is strictly forbidden. Thank you in advance for your cooperation.

Please consider the environment before printing this e-mail.
_______________________________________________
srslte-users mailing list
srslte-users at lists.softwareradiosystems.com<mailto:srslte-users at lists.softwareradiosystems.com>
http://www.softwareradiosystems.com/mailman/listinfo/srslte-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.srsran.com/pipermail/srsran-users/attachments/20210315/caf66620/attachment.htm>


More information about the srsran-users mailing list