You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pipeline or multiple-pause-resume on MTLP_RVP_NOCODEC #6734
Describe the bug
CI observed an IPC error "FW reported error: 7 - Unsupported operation requested" for 0x13000004|0x1 when testing multiple-pipeline or multiple-pause-resume on MTLP_RVP_NOCODEC.
changed the title [-][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pause-resume on MTLP_RVP_NOCODEC[/-][+][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pause-resume or multiple-pipeline on MTLP_RVP_NOCODEC[/+]on Dec 8, 2022
changed the title [-][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pause-resume or multiple-pipeline on MTLP_RVP_NOCODEC[/-][+][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pipeline or multiple-pause-resume on MTLP_RVP_NOCODEC[/+]on Dec 19, 2022
@miRoox can you share logs with ipc payload
please contact with @softwarecki and share remote access to your rvp, because we don't have reproduction on our side
No repro found using FAT framework.
Logs starts with clean up of old topology and building the problematic one. Maybe it's necessary to buil&run both topologies to reproduce the error.
Could we get log with both topologies?
@miRoox probably you missed my previous comment. Are You able to catch ipc logs of both topologies (topology run before the error and topology with the error)
Activity
[-][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pause-resume on MTLP_RVP_NOCODEC[/-][+][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pause-resume or multiple-pipeline on MTLP_RVP_NOCODEC[/+][-][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pause-resume or multiple-pipeline on MTLP_RVP_NOCODEC[/-][+][BUG] FW reported error: 7 for 0x13000004|0x1 when multiple-pipeline or multiple-pause-resume on MTLP_RVP_NOCODEC[/+]keqiaozhang commentedon Dec 19, 2022
The reproduction rate is almost 100% on main branch, change it to P1
wszypelt commentedon Feb 3, 2023
@miRoox can you share logs with ipc payload
please contact with @softwarecki and share remote access to your rvp, because we don't have reproduction on our side
miRoox commentedon Feb 7, 2023
dmesg.txt
mtrace.txt
mszleszy commentedon Feb 10, 2023
working on reproduction using FAT framework
lgirdwood commentedon Feb 13, 2023
@mszleszy can you try with #7052 now being merged, it's improved and hopefully fixed the pause issue. @lyakh fyi.
mszleszy commentedon Feb 16, 2023
No repro found using FAT framework.
Logs starts with clean up of old topology and building the problematic one. Maybe it's necessary to buil&run both topologies to reproduce the error.
Could we get log with both topologies?
wszypelt commentedon Feb 17, 2023
@miRoox any update?
miRoox commentedon Feb 17, 2023
We still meet this issue every day.
mszleszy commentedon Feb 24, 2023
@miRoox probably you missed my previous comment. Are You able to catch ipc logs of both topologies (topology run before the error and topology with the error)
keqiaozhang commentedon Feb 27, 2023
This issue has not appeared in CI recently, will continue to monitor this issue in CI for another 2 weeks.
marc-hb commentedon Mar 2, 2023
Note every pause-resume test on every platform (including of course MTL) has been "spamming" the firmware logs with thousands of errors like this:
Example in https://sof-ci.01.org/sofpr/PR7202/build4268/devicetest/index.html?model=ADLP_RVP_SDW_IPC4ZPH&testcase=multiple-pause-resume-5 + many, many others.
The test is PASSing anyway because sof-test has never cared about firmware errors.
This makes it practically impossible to find other errors.
Should this be a warning instead? EDIT: work in progress in #7207 and #7210
cc: @RanderWang, @softwarecki, @kv2019i , @lyakh , @ranj063 , #6723
keqiaozhang commentedon Mar 6, 2023
This issue disappeared in CI for weeks, Closing it.