[BUG] arecord fails with non-zero exit status with check-pause-resume-playback-100 #9191

Activity

1. kv2019i

   added

   bugSomething isn't working as expectedSomething isn't working as expected

   P1Blocker bugs or important featuresBlocker bugs or important features

   LNLApplies to Lunar Lake platformApplies to Lunar Lake platform

   on Jun 4, 2024
2. kv2019i

   added a commit that references this issue on Jun 4, 2024

   app: boards: ace20_lnl: set DMA_INTEL_ADSP_HDA_TIMING_L1_EXIT=y

   ...

   8319b67
3. kv2019i commented on Jun 5, 2024

   

   kv2019i

   on Jun 5, 2024

   CollaboratorAuthor

   More actions

   @abonislawski @serhiy-katsyuba-intel It seems just enabling the L1_EXIT is not sufficient, record failure still seen (see comment #9194 (comment) ).

   The logs don't have much, but this very much seems like a case of host DMA not moving, and arecord exiting with error as no data is received.
4. marc-hb commented on Jun 7, 2024

   

   marc-hb

   on Jun 7, 2024

   Collaborator

   More actions

   Is thesofproject/linux#5048 a duplicate? Then maybe close this one because 5048 has a better looking description :-)
5. lgirdwood

   added this to the v2.10 milestone on Jun 10, 2024
6. lgirdwood commented on Jun 10, 2024

   

   lgirdwood

   on Jun 10, 2024

   Member

   More actions

   Lets keep it open for the moment, just as a place holder on the FW side for v2.10
7. lgirdwood

   removed

   P1Blocker bugs or important featuresBlocker bugs or important features

   on Jun 13, 2024
8. lgirdwood

   modified the milestones: v2.10, v2.11 on Jun 13, 2024
9. kv2019i

   self-assigned this

   on Sep 4, 2024
10. kv2019i

    changed the title [-][BUG] arecord fails with non-zero exit status[/-] [+][BUG] arecord fails with non-zero exit status with check-pause-resume-playback-100[/+] on Sep 4, 2024
11. kv2019i commented on Sep 4, 2024

    

    kv2019i

    on Sep 4, 2024

    CollaboratorAuthor

    More actions

    Looking at this again. Narrowing the scope of this bug to specifically cover the failure with pause-resume case with a recent occurence in:
    https://sof-ci.01.org/linuxpr/PR4733/build4501/devicetest/index.html?model=LNLM_SDW_AIOC&testcase=check-pause-resume-playback-100

    Failure fingerprint in sof-test console:

    t=14110 ms: aplay: (29/100) Found   === PAUSE ===  ,  pausing for 166 ms
    t=14110 ms: aplay: ERROR: aplay: do_pause:1567: pause push error: File descriptor in bad state
    

    The errors happen when there are no clear errors in either kernel dmesg.txt or FW mtrace.txt. There does seem to be a timing anomaly seen in the logs, like here:

    [  758.664736] kernel: snd_sof_intel_hda_common:hda_dai_trigger: soundwire_intel soundwire_intel.link.3: cmd=0 dai SDW3 Pin2 direction 0
    [  759.996415] kernel: snd_sof_intel_hda_common:hda_dsp_stream_trigger: sof-audio-pci-intel-lnl 0000:00:1f.3: FW Poll Status: reg[0x1e0]=0x140000 successful
    [  759.996434] kernel: snd_sof:sof_ipc4_trigger_pipelines: sof-audio-pci-intel-lnl 0000:00:1f.3: trigger cmd: 0 state: 2
    

    This is clearly larger pause in activity than the ~0.1-0.2ms delays that are used in the test case.
12. kv2019i

    added

    P2Critical bugs or normal featuresCritical bugs or normal features

    on Sep 4, 2024
13. kv2019i

    removed their assignment

    on Sep 9, 2024
14. kv2019i

    modified the milestones: v2.11, v2.12 on Sep 9, 2024
15. lgirdwood

    removed

    P2Critical bugs or normal featuresCritical bugs or normal features

    on Sep 17, 2024
16. kv2019i commented on Dec 13, 2024

    

    kv2019i

    on Dec 13, 2024

    CollaboratorAuthor

    More actions

    Not seen in PR/daily tests for a week, closing.
17. kv2019i

    closed this as completedon Dec 13, 2024