Repository navigation
[BUG][CML]" ipc error for 0x60060000 size 12"when do Capture-pause-release test using headset(0,0) #2404
Description
Describe the bug
I/O error when do Capture-pause-release test using headset (0,0)
After error occurred arecord stop,but if arecord again,it can work normally
To Reproduce
1.sudo reboot
2. ./sof-test-pause-release.sh -t pcm0c (script as attachment)
Reproduction Rate
10/10
Expected behavior
There is no I/O error,after error occurred arecord stop,but if arecord again,it can work normally
Actual result
I/O error
dmesgerror.log
[ 7739.251053] sof-audio-pci 0000:00:1f.3: ipc rx: 0x90020000: GLB_TRACE_MSG
[ 7739.251076] sof-audio-pci 0000:00:1f.3: ipc rx done: 0x90020000: GLB_TRACE_MSG
[ 7739.732448] sof-audio-pci 0000:00:1f.3: pcm: trigger stream 0 dir 1 cmd 3
[ 7739.732453] sof-audio-pci 0000:00:1f.3: ipc tx: 0x60060000: GLB_STREAM_MSG: TRIG_PAUSE
[ 7739.732726] sof-audio-pci 0000:00:1f.3: error: ipc error for 0x60060000 size 12
[ 7739.733241] sof-audio-pci 0000:00:1f.3: FW Poll Status: reg=0x20240000 successful
[ 7739.733245] HDA Analog: ASoC: trigger FE cmd: 3 failed: -22
[ 7739.733251] sof-audio-pci 0000:00:1f.3: pcm: trigger stream 0 dir 1 cmd 4
[ 7739.733755] sof-audio-pci 0000:00:1f.3: FW Poll Status: reg=0x2024001e successful
[ 7739.733760] sof-audio-pci 0000:00:1f.3: ipc tx: 0x60070000: GLB_STREAM_MSG: TRIG_RELEASE
[ 7739.733975] sof-audio-pci 0000:00:1f.3: error: ipc error for 0x60070000 size 12
[ 7739.733977] HDA Analog: ASoC: trigger FE cmd: 4 failed: -22
[ 7739.750774] sof-audio-pci 0000:00:1f.3: ipc rx: 0x90020000: GLB_TRACE_MSG
[ 7739.750796] sof-audio-pci 0000:00:1f.3: ipc rx done: 0x90020000: GLB_TRACE_MSG
[ 7740.240197] sof-audio-pci 0000:00:1f.3: pcm: trigger stream 0 dir 1 cmd 0
[ 7740.240209] sof-audio-pci 0000:00:1f.3: ipc tx: 0x60050000: GLB_STREAM_MSG: TRIG_STOP
[ 7740.240448] sof-audio-pci 0000:00:1f.3: ipc tx succeeded: 0x60050000: GLB_STREAM_MSG: TRIG_STOP
[ 7740.240462] sof-audio-pci 0000:00:1f.3: FW Poll Status: reg=0x20240000 successful
[ 7740.240471] sof-audio-pci 0000:00:1f.3: ipc tx: 0x60030000: GLB_STREAM_MSG: PCM_FREE
[ 7740.240837] sof-audio-pci 0000:00:1f.3: ipc tx succeeded: 0x60030000: GLB_STREAM_MSG: PCM_FREE
[ 7740.240856] sof-audio-pci 0000:00:1f.3: In hda_link_pcm_trigger cmd=0
[ 7740.240876] sof-audio-pci 0000:00:1f.3: ipc tx: 0x80010000: GLB_DAI_MSG: CONFIG
Environment
Platform: CML laptop with onboard codec ALC3204 in HDA mode
Firmware: 53d35e4 https://github.com/thesofproject/sof/commits/master
Kernel: b2d9823 https://github.com/thesofproject/linux/commits/topic/sof-dev
Topology: file: tools/topology/sof-hda-generic-2ch.tplg same as firmware
Activity
Liviali155 commentedon Feb 18, 2020
Seems a Firmware regression issue
I changed firmware to 20200110(commit:c6baa979),error not seen
Liviali155 commentedon Feb 18, 2020
Bi-sect result:
caused by Firmware commit 91fb594
Liviali155 commentedon Feb 18, 2020
Related #2280
lgirdwood commentedon Feb 18, 2020
@Liviali155 is this fatal ? Does audio continue or do you need to restart aplay ?
@slawblauciak This looks like a race that's easy to reproduce with slower clock ?
Liviali155 commentedon Feb 19, 2020
After error occurred arecord stop,but if arecord again,it can work normally
lgirdwood commentedon Feb 21, 2020
@slawblauciak @mmaka1 can we make the 400M clock the default for CML etc, the low power clock can then be optionally selected by user who need it until the switching is dynamic at runtime.
[-][BUG][CML Mantis]I/O error when do Capture-pause-release test using (0,0)[/-][+][BUG][CML Mantis]I/O error when do Capture-pause-release test using headset(0,0)[/+]slawblauciak commentedon Feb 25, 2020
Hmm, the FW log doesn't appear complete. I can see many pause/release operations being done from the dmesg, but only one in the FW trace. Is the trace you provided cut in any way? Does it always end up like that?
4 remaining items
juimonen commentedon Feb 26, 2020
@Liviali155 ok thanks, at least the error trace seems to be consistent so xrun (no data), recover fail, and pipeline stop.
[-][BUG][CML Mantis]I/O error when do Capture-pause-release test using headset(0,0)[/-][+][BUG][CML]I/O error when do Capture-pause-release test using headset(0,0)[/+]tlauda commentedon Mar 3, 2020
@lgirdwood @Liviali155 Basically the same problem as with #2467. If L1SEN is disabled, then it looks like some kind of pause/release sequence problem with this HDA codec. In the FW log we can clearly see, that the new data hasn't arrived on HDA link after releasing the stream. I'm suggesting codec issue, since we haven't seen this kind of error on CI for a long long time (since the HDA DAI sequence has been fixed in kernel).
lgirdwood commentedon Mar 4, 2020
[-][BUG][CML]I/O error when do Capture-pause-release test using headset(0,0)[/-][+][BUG][CML]" ipc error for 0x60060000 size 12"when do Capture-pause-release test using headset(0,0)[/+]Liviali155 commentedon Mar 5, 2020
On CFL-S RVP also can reproduce this issue
1.cd sof-test(as attachment)
2.cd test-case
3.export TPLG=/lib/firmware/intel/sof-tplg/sof-hda-generic-4ch.tplg
4.check-pause-resume.sh -c 100 -m capture
I/O error when pause/release arecord(0,0)(headset) ,dmesg show "ipc error for 0x60060000 size "
Platform: CFL-S RVP with onboard codec ALC700 in HDA mode
dmesg_error.log
sof_errortrace.log
dmesg.txt
etrace.txt
slogger.txt
sof-test.zip
tlauda commentedon Mar 10, 2020
@Liviali155 Please verify with #2512.
Liviali155 commentedon Mar 10, 2020
Tried with PR #2512 ,error not seen
Platform: CML laptop with onboard codec ALC3204 in HDA mode
TGL-U RVP with onboard codec ALC711 in HDA mode
tlauda commentedon Mar 11, 2020
Fixed with #2512.