Repository navigation
[BUG] glitch in test 22_05_TestDsmMulticoreSsp48000Hz32b32b2ch #4052
Description
Describe the bug
Glitch was found in test 22_05_TestDsmMulticoreSsp48000Hz32b32b2ch
Topology
+----------------------------------------------+
| +------+ +---+ +- -+ +---+ +-------+ |
| | Host |-->|Buf|-->|Dsm|-->|Buf|-->|SSP Dai|-------+
| +------+ +---+ +---+ +---+ +-------+ | |
+------------------+ ^ +-------------------+ |
| | | |
| +---+ | |
| |Buf| | |
| +---+ | |
+---^---+ |
+----------------------|------------------------+ |
| +------+ +---+ +-|--+ +---+ +-------+ | |
| | Host |<--|Buf|<--|Dmux|<--|Buf|<--|SSP Dai|<-----+
| +------+ +---+ +----+ +---+ +-------+ |
+-----------------------------------------------+
To Reproduce
Run test with Diagnostic driver 22_05_TestDsmMulticoreSsp48000Hz32b32b2ch
Reproduction Rate
Sporadic.
Environment
- Branch name and commit hash of the 2 repositories: sof (firmware/topology) and linux (kernel driver).
- SOF: BRANCH: main HASH: be065e5 (SOF_FW_MASTER_01.01.00.2246)
- Name of the platform(s) on which the bug is observed.
- Platform: ICL RVP
- Test environment:
BRANCH: master HASH: b7ed915d911acda055bbb822800b76557e9fe7de
Screenshots or console output
Activity
lgirdwood commentedon Apr 19, 2021
@kkarask I'm assuming all is good when DSM is disabled ? It looks like we have missed a deadline and slipped some data. @slawblauciak fyi.
lgirdwood commentedon Apr 19, 2021
@kkarask I also assume that your topology is all core 0 and can you confirm if you are using low power clocks ?
keyonjie commentedon Apr 20, 2021
@kkarask what is "SOF: BRANCH: main HASH: b7ed915d911acda"? I can't find this commit in the main branch?
kkarask commentedon Apr 20, 2021
@keyonjie Sorry, my mistake. Correct HASH is: be065e5
@lgirdwood There are no glithes in other Multi core tests apart from only one issue: [(https://github.com//issues/4050)]
All components use core = 0 except DSM = 1 in test 22_05_TestDsmMulticoreSsp48000Hz32b32b2ch.
Regarding power clocks could you please check dma_trace_log.
slawblauciak commentedon Apr 27, 2021
@kkarask can you please re-check if the issue is reproducible with the current main branch? There was a relevant fix on a different platform, I wonder if it also applies to ICL.
kkarask commentedon Apr 27, 2021
Hi, @slawblauciak
Unfortunately the problem still exists on current main branch on platforms: TGL and ICL. There are glitches in whole wav files.
Logs:
TGL 22_06_TestDsmMulticoreSdw48000Hz24b32b2ch :
Logs 4052 TGL.zip
ICL 22_05_TestDsmMulticoreSsp48000Hz32b32b2ch:
Logs 4052 ICL.zip
lgirdwood commentedon Apr 27, 2021
DSM and multicore is showing up in a few of these failures recently. @kkarask what's your take here since you are doing the bulk of the test runs.
It would be good to know if these issues are more likely scheduling work on other cores or are DSM overruns that then impact the pipeline.
kkarask commentedon Apr 28, 2021
@lgirdwood I will run stress tests and give you answer tomorrow. We will see how fix #4088 works.
kkarask commentedon Apr 29, 2021
All results from stress tests will be available tomorrow. I will comment about reproduction rate.
lgirdwood commentedon Apr 29, 2021
@kkarask thanks, it occurred to me that the pcm converter/copier component is also in this pipeline and also the simple pipelines with similar glitches so could be getting stuck or some confusion between DMA and module R/W pointers.
Btw, I can't read the logs.
kkarask commentedon Apr 30, 2021
@lgirdwood Logs should be readable now.
I had some problems with reproduction due to the bug #4101 witch covers that issue.
Reproduction rate was 1/100 counting only cases when the other bug #4101 didn't occur.
1 remaining item
lgirdwood commentedon May 13, 2021
Duplicate of simpler #4076
juimonen commentedon May 18, 2021
@kkarask do you have any idea when these "glitch" bugs started to appear? So do you recall running these previously at some point without these errors? Just trying to little bit bisect the area... I can't reproduce the loopback glitches with linux driver (or it is quite difficult to do at least)...
kkarask commentedon May 19, 2021
@juimonen Unfortunately I don't have logs, but as long as I can see the test fail on this hash: 4c1f41a I'm not sure about reason because I don't have logs:( maybe this help you a bit...
mwasko commentedon Jun 7, 2021
This issue is specific to scenario with DSM component on Core 1 while the rest of the pipeline is on Core 0. The issue is not reproducing in case when entire pipeline is on Core 1.
mwasko commentedon Jun 7, 2021
@juimonen the issue was not reproducible on tgl-11-drop-stable branch. However when we merged scheduling refinements from tgl-11 to main branch then we started to catch quite a lot of multi-core problems (including this one). Most of the issues has already been addressed, but this one still remain - reproduced recently on tgl-13-drop-stable.
github-actions commentedon Jun 19, 2025
This issue has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this issue will automatically be closed in 14 days. Note, that you can always re-open a closed issue at any time.