Skip to content

[IPC error] sof-test/multiple-pause-resume.sh reported error when dma trace was disabled on TGL #3673

Closed
Closed
[IPC error] sof-test/multiple-pause-resume.sh reported error when dma trace was disabled on TGL#3673
Labels
IPC errorIPC error is observedP1Blocker bugs or important featuresSDWSoundWireTGLApplies to Tiger LakebugSomething isn't working as expectedxrunXRUN is observed and firmware may not recover

Description

@RanderWang
RanderWang
opened on Dec 7, 2020

Describe the bug
[ 55.864574] sof-audio-pci 0000:00:1f.3: ipc tx: 0x60070000: GLB_STREAM_MSG: TRIG_RELEASE
[ 55.864773] sof-audio-pci 0000:00:1f.3: ipc tx succeeded: 0x60070000: GLB_STREAM_MSG: TRIG_RELEASE
[ 56.033318] sof-audio-pci 0000:00:1f.3: pcm: trigger stream 1 dir 1 cmd 3
[ 56.033332] sof-audio-pci 0000:00:1f.3: ipc tx: 0x60060000: GLB_STREAM_MSG: TRIG_PAUSE
[ 56.033556] sof-audio-pci 0000:00:1f.3: error: ipc error for 0x60060000 size 12, err -22
[ 56.033592] sof-audio-pci 0000:00:1f.3: FW Poll Status: reg=0x20140000 successful
[ 56.033610] sof-audio-pci 0000:00:1f.3: ASoC: error at snd_soc_pcm_component_trigger on 0000:00:1f.3: -22
[ 56.033628] Headset mic: ASoC: trigger FE cmd: 3 failed: -22
[ 56.033650] sof-audio-pci 0000:00:1f.3: pcm: trigger stream 1 dir 1 cmd 4
[ 56.033984] sof-audio-pci 0000:00:1f.3: FW Poll Status: reg=0x2014001e successful
[ 56.033999] sof-audio-pci 0000:00:1f.3: ipc tx: 0x60070000: GLB_STREAM_MSG: TRIG_RELEASE
[ 56.034232] sof-audio-pci 0000:00:1f.3: error: ipc error for 0x60070000 size 12, err -22
[ 56.034244] sof-audio-pci 0000:00:1f.3: ASoC: error at snd_soc_pcm_component_trigger on 0000:00:1f.3: -22
[ 56.034252] Headset mic: ASoC: trigger FE cmd: 4 failed: -22

DMA trace error:
TIMESTAMP DELTA C# COMPONENT LOCATION CONTENT
[ 1025566.458333] ( 1025566.437500) c0 dw-dma src/drivers/dw/dma.c:1095 ERROR dw_dma_get_data_size(): xrun detected
[ 1025575.052083] ( 8.593750) c0 dai 2.10 src/audio/dai.c:785 ERROR dai_report_xrun(): overrun due to no space available
[ 1025582.708333] ( 7.656250) c0 dai 2.10 src/audio/dai.c:684 ERROR comp_overrun(): sink->free = 0, copy_bytes = 0
[ 1025615.260417] ( 32.552082) c0 pipe 2.11 src/audio/pipeline.c:1029 ERROR pipeline_copy(): ret = -61, start->comp.id = 10, dir = 0
[ 1025625.260417] ( 10.000000) c0 pipe 2.11 src/audio/pipeline.c:1219 ERROR pipeline_task(): xrun recover failed! pipeline will be stopped!
[ 1893947.604167] ( 868322.375000) c0 dai 2.10 src/audio/component.c:209 ERROR comp_set_state(): wrong state = 1, COMP_TRIGGER_STOP
[ 1893958.177083] ( 10.572917) c0 pipe 2.11 src/audio/pipeline.c:884 ERROR pipeline_trigger(): ret = -22, host->comp.id = 6, cmd = 0
[ 1893968.958333] ( 10.781250) c0 ipc src/ipc/handler.c:441 ERROR ipc: comp 6 trigger 0x50000 failed -22

To Reproduce
on TGL device, disable dma trace in kernel driver and run sof-test/test-case/multiple-pause-resume.sh

Reproduction Rate
etrace.txt
dmesg.txt

100%

Environment

  1. Branch name and commit hash of the 2 repositories: sof (firmware/topology) and linux (kernel driver).

  2. Name of the platform(s) on which the bug is observed.

    • Platform: TGL-U RVP with ALC711 and ALC1308 in SDW mode by default (no HW rework is needed), TGL Chrome device Volteer with RT5682 and Max98373 in SDW mode.

Activity

  1. RanderWang
    added
    bugSomething isn't working as expected
    TGLApplies to Tiger Lake
    on Dec 7, 2020
  2. mengdonglin
    added
    P1Blocker bugs or important features
    on Dec 8, 2020

  3. mengdonglin commented on Dec 8, 2020

    @mengdonglin
    mengdonglin
    on Dec 8, 2020
    Collaborator

    @keqiaozhang will check if this issue can also be reproduced on TGL Chrome device with I2S audio

  4. slawblauciak
    self-assigned this
    on Dec 8, 2020

  5. keqiaozhang commented on Dec 8, 2020

    @keqiaozhang
    keqiaozhang
    on Dec 8, 2020
    Collaborator

    Tried sof-dev + tgl-011-drop-stable branch on TGL Chrome-I2S, no reproduction so far(50+ iterations passed), only observed some xrun errors in error trace when testing "multiple-pause-resume.sh". will perform more tests.

    [1763781595.156250] (     2012.031250) c0 hda-dma            ..../intel/hda/hda-dma.c:865  ERROR hda_dma_link_check_xrun(): underrun detected
[1763789650.208333] (     8055.052246) c0 hda-dma            ..../intel/hda/hda-dma.c:865  ERROR hda_dma_link_check_xrun(): underrun detected
[1763793618.489583] (     3968.281250) c0 hda-dma            ..../intel/hda/hda-dma.c:865  ERROR hda_dma_link_check_xrun(): underrun detected
[1764379567.500000] (   585949.000000) c0 hda-dma            ..../intel/hda/hda-dma.c:865  ERROR hda_dma_link_check_xrun(): underrun detected

    Edited: No reproductions on TGL-Chrome-I2S, tested with "./multiple-pause-resume.sh -r 1000" and passed.

  6. bardliao commented on Dec 24, 2020

    @bardliao
    bardliao
    on Dec 24, 2020
    Collaborator

    @slawblauciak @RanderWang I think the issue is that we don't handle xrun on firmware. ie. #define NO_XRUN_RECOVERY 1
    I can see below sof-logger if I add pipe_err(p, "No xrun_recover"); in pipeline_xrun_recover

    [104075289.843750] (  2812062.500000) c0 dw-dma                 src/drivers/dw/dma.c:1101 ERROR dw_dma_get_data_size(): xrun detected
[104075308.020833] (       18.177084) c0 dai          4.23           src/audio/dai.c:788  ERROR dai_report_xrun(): overrun due to no space available
[104075324.635417] (       16.614584) c0 dai          4.23           src/audio/dai.c:685  ERROR comp_overrun(): sink->free = 224, copy_bytes = 0
[104075415.312500] (       90.677086) c0 pipe         4.24      src/audio/pipeline.c:1029 ERROR pipeline_copy(): ret = -61, start->comp.id = 23, dir = 0
[104075437.291667] (       21.979166) c0 pipe         4.24      src/audio/pipeline.c:1143 ERROR No xrun_recover
[104075457.187500] (       19.895834) c0 pipe         4.24      src/audio/pipeline.c:1220 ERROR pipeline_task(): xrun recover failed! pipeline will be stopped!

    diff --git a/src/audio/pipeline.c b/src/audio/pipeline.c
index 345773b7d..99617baf8 100644
--- a/src/audio/pipeline.c
+++ b/src/audio/pipeline.c
@@ -1140,6 +1140,7 @@ void pipeline_xrun(struct pipeline *p, struct comp_dev *dev,
 /* recover the pipeline from a XRUN condition */
 static int pipeline_xrun_recover(struct pipeline *p)
 {
+       pipe_err(p, "No xrun_recover");
        return -EINVAL;
 }

  7. xiulipan commented on Dec 25, 2020

    @xiulipan
    xiulipan
    on Dec 25, 2020
    Contributor

    @aiChaoSONG Is #3714 a duplicate with this?
    @bardliao @RanderWang I think we do not have XRUN recovery at all now. All XRUN will break pipeline and XRUN is not tolerated now. But I remember we have some exception in WoV pipelines @keyonjie could you clarify here.

  8. aiChaoSONG commented on Dec 25, 2020

    @aiChaoSONG
    aiChaoSONG
    on Dec 25, 2020
    Collaborator

    @xiulipan #3714 is reproduced in our daily test, I don't think dma trace is disabled in the default kernel, but issue also reproduced.

  9. bardliao commented on Dec 28, 2020

    @bardliao
    bardliao
    on Dec 28, 2020
    Collaborator

    @xiulipan @aiChaoSONG IMHO, #3714 and #3673 are the same issue. Disabling DMA trace is just a way to reproduce/trigger the XRUN.

  10. mengdonglin
    added
    xrunXRUN is observed and firmware may not recover
    IPC errorIPC error is observed
    on Jan 5, 2021
  11. mengdonglin
    added
    SDWSoundWire
    on Jan 19, 2021

  12. ranj063 commented on Jan 19, 2021

    @ranj063
    ranj063
    on Jan 19, 2021
    Collaborator

    @bardliao the ipc error is because the FW stops the pipeline in case of an xrun and then when the application stops the PCM after the xrun, the state is incorrect. This needs to be fixed @keyonjie FYI.

    But why does the xrun happen in the first place?

  13. bardliao commented on Jan 20, 2021

    @bardliao
    bardliao
    on Jan 20, 2021
    Collaborator

    But why does the xrun happen in the first place?

    I have no idea so far.

  14. lgirdwood commented on Jan 22, 2021

    @lgirdwood
    lgirdwood
    on Jan 22, 2021
    Member

    @mwasko fyi

  15. slawblauciak commented on Feb 26, 2021

    @slawblauciak
    slawblauciak
    on Feb 26, 2021
    Collaborator

    @RanderWang is the issue still reproducible on the latest master branch?

  16. RanderWang commented on Mar 9, 2021

    @RanderWang
    RanderWang
    on Mar 9, 2021
    CollaboratorAuthor

    @slawblauciak very sorry, so late. I tested FW in today CI build. The XRUN error in FW appears after a few cycles of test instead of occurring immediately.

  17. plbossart commented on Jun 23, 2021

    @plbossart
    plbossart
    on Jun 23, 2021
    Member

    @RanderWang is this issue still relevant? It's not tracked in CI or daily tests.

  18. RanderWang commented on Jun 24, 2021

    @RanderWang
    RanderWang
    on Jun 24, 2021
    CollaboratorAuthor

    @RanderWang is this issue still relevant? It's not tracked in CI or daily tests.

    This is a hint or fast way to trigger multiple-pause-resume issue, not a root reason. In my opinion, it is caused by schedule in FW

  19. slawblauciak
    removed their assignment
    on Aug 2, 2021

  20. abonislawski commented on Jan 25, 2022

    @abonislawski
    abonislawski
    on Jan 25, 2022
    Member

    @RanderWang could you please check if this issue is still valid?

  21. lgirdwood commented on Jan 25, 2022

    @lgirdwood
    lgirdwood
    on Jan 25, 2022
    Member

    @XiaoyunWu6666 can you comment if you still see this issue ? if not we can close.

  22. RanderWang commented on Jan 26, 2022

    @RanderWang
    RanderWang
    on Jan 26, 2022
    CollaboratorAuthor

    @RanderWang could you please check if this issue is still valid?

    I will check it. too old

  23. RanderWang commented on Jan 28, 2022

    @RanderWang
    RanderWang
    on Jan 28, 2022
    CollaboratorAuthor

    I checked with latest kernel commit ddcea4bef6 and fw 0bf7b73 today. I can't reproduce the bug. Let's close it

  24. RanderWang
    closed this as completedon Jan 28, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    IPC errorIPC error is observedP1Blocker bugs or important featuresSDWSoundWireTGLApplies to Tiger LakebugSomething isn't working as expectedxrunXRUN is observed and firmware may not recover

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions