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] WoV cannot trigger with topic/sof-dev(kernel) and master(sof) branch #1775
this issue exists even before the release of cml-007, I think it is not a duplicate of #1772. I think some code of sof or kernel may not be put upstream. @wwittbrx@lgirdwood
@mrajwa ,Wov shows IO error before trigger it on WHL platform with topic/sof-dev(kernel) and master(sof) branch. Issue still can be reproduced after applying #1768 and #1689
error logger without patch 1768 and 1689
CORE LEVEL COMP_ID TIMESTAMP DELTA FILE_NAME CONTENT
0 1 MEM 746878406.458333 746878400.000000 src/lib/alloc.c:273 error: 171 blocks needed for allocation but only 170 blocks are remaining
error log after applying patch 1768 and 1689
CORE LEVEL COMP_ID TIMESTAMP DELTA FILE_NAME CONTENT
0 1 MEM 941441840.052083 941441856.000000 src/lib/alloc.c:273 error: 171 blocks needed for allocation but only 170 blocks are remaining
0 1 DAI 8.40 941446883.281250 5043.229004 src/audio/dai.c:580 dai_check_for_xrun() error: overrun due to no data available
0 1 COMP 941446889.687500 6.406250 of/audio/component.h:758 comp_overrun() error: dev->comp.id = 40, sink->free = 16, copy_bytes = 0
0 1 PIPE 941446950.104167 60.416668 src/audio/pipeline.c:790 pipeline_copy() error: ret = -61, start->comp.id = 40, dir = 0
0 1 PIPE 8.41 941446957.760417 7.656250 src/audio/pipeline.c:984 pipeline_task(): xrun recover failed! pipeline will be stopped!
0 1 COMP 942077383.125000 630425.375000 rc/audio/component.c:192 comp_set_state() error: wrong state = 5, COMP_TRIGGER_RESET
0 1 COMP 942077395.729167 12.604167 rc/audio/component.c:192 comp_set_state() error: wrong state = 5, COMP_TRIGGER_RESET
Have you applied #1768 and #1689 on top of the master branch? The XRUN issue was caused by #1772 and should be fixed now on master branch. I did some testes yesterday and did not see this issue anymore. However I saw IPC timeout error in dmesg...I think this issue is handled by somebody already.
PR #1773 fix #1772, and #1773 is merged, but we still see xrun with master branch. btw, we are using daily version of kernel and sof. io error is a message of arecord, i guess it it because of wov inability of sof. @mrajwa
Retest it with right command " arecord -Dhw:0,8 -r 16000 -c2 -f s16_le -vvv -i 1.wav -M -N
", I/O error is gone. But still can reproduce the original issue with #1768 and #1689 .
error logger
CORE LEVEL COMP_ID TIMESTAMP DELTA FILE_NAME CONTENT
0 1 MEM 23079106.562500 23079106.000000 src/lib/alloc.c:273 error: 171 blocks needed for allocation but only 170 blocks are remaining
0 1 DAI 8.40 23082850.364583 3743.802002 src/audio/dai.c:580 dai_check_for_xrun() error: overrun due to no data available
0 1 COMP 23082856.770833 6.406250 of/audio/component.h:758 comp_overrun() error: dev->comp.id = 40, sink->free = 16, copy_bytes = 0
0 1 PIPE 23082917.239583 60.468750 src/audio/pipeline.c:790 pipeline_copy() error: ret = -61, start->comp.id = 40, dir = 0
0 1 PIPE 8.41 23082924.791667 7.552083 src/audio/pipeline.c:984 pipeline_task(): xrun recover failed! pipeline will be stopped!
0 1 COMP 30841417.708333 7758493.000000 rc/audio/component.c:192 comp_set_state() error: wrong state = 5, COMP_TRIGGER_RESET
0 1 COMP 30841430.208333 12.500000 rc/audio/component.c:192 comp_set_state() error: wrong state = 5, COMP_TRIGGER_RESET
I have checked this recipe today on my CML platform and it works for me with no issues.
As far as I know this is the only working recipe based on master branches. Please make sure you include all FW PRs and one kernel PR.
The test command I use for testing is sudo arecord -M -N -D hw:0,8 -f s24_le -c 2 -r 16000 -d 10 record_kd.wav
Activity
wwittbrx commentedon Aug 29, 2019
Could you attach sof-logger output? I wonder if this is a duplicate of #1772.
Also I think topic/sof-dev does not have all PR's merged, https://github.com/thesofproject/linux/tree/topic/keyword_detect should work better, not sure though.
aiChaoSONG commentedon Sep 2, 2019
this issue exists even before the release of cml-007, I think it is not a duplicate of #1772. I think some code of sof or kernel may not be put upstream. @wwittbrx @lgirdwood
keqiaozhang commentedon Sep 2, 2019
@mrajwa
Please make sure all WoV-related code or patches are landed on master, we need to align the status with release branch. Thanks.
mrajwa commentedon Sep 2, 2019
@keqiaozhang, @aiChaoSONG, please wait for these two patches #1768 and #1689
aiChaoSONG commentedon Sep 3, 2019
On CML Chrome, after applying #1768 and #1689, still cannot trigger due to some errors. @keqiaozhang @mrajwa @wwittbrx
YvonneYang2 commentedon Sep 3, 2019
@mrajwa ,Wov shows IO error before trigger it on WHL platform with topic/sof-dev(kernel) and master(sof) branch. Issue still can be reproduced after applying #1768 and #1689
error logger without patch 1768 and 1689
error log after applying patch 1768 and 1689
all logs
dmesg after applying patch.log
logger after applying patch.log
logger.log
dmesg.log
mrajwa commentedon Sep 3, 2019
Have you applied #1768 and #1689 on top of the master branch? The XRUN issue was caused by #1772 and should be fixed now on master branch. I did some testes yesterday and did not see this issue anymore. However I saw IPC timeout error in dmesg...I think this issue is handled by somebody already.
mrajwa commentedon Sep 3, 2019
@YvonneYang2, what I/O error are you talking about? All I see is XRUN?
aiChaoSONG commentedon Sep 3, 2019
PR #1773 fix #1772, and #1773 is merged, but we still see xrun with master branch. btw, we are using daily version of kernel and sof. io error is a message of arecord, i guess it it because of wov inability of sof. @mrajwa
YvonneYang2 commentedon Sep 3, 2019
I/O error is showed in console when do arecord

keqiaozhang commentedon Sep 3, 2019
@YvonneYang2
I/O error is caused by the missing "-M -N" options, please retest it. Thanks!
YvonneYang2 commentedon Sep 3, 2019
Retest it with right command " arecord -Dhw:0,8 -r 16000 -c2 -f s16_le -vvv -i 1.wav -M -N
", I/O error is gone. But still can reproduce the original issue with #1768 and #1689 .
error logger
all logs
dmesg.log
error_logger.log
logger.log
wwittbrx commentedon Sep 3, 2019
Please check the following recipe:
FW 90873c5 + #1689 + #1740 + #1768
Kernel f95e4eb + PR1037
I have checked this recipe today on my CML platform and it works for me with no issues.
As far as I know this is the only working recipe based on master branches. Please make sure you include all FW PRs and one kernel PR.
The test command I use for testing is
sudo arecord -M -N -D hw:0,8 -f s24_le -c 2 -r 16000 -d 10 record_kd.wavYvonneYang2 commentedon Sep 4, 2019
Test it with recipe as below, issue cannot reproduce on WHL platform.
FW: master branch 90873c5 + #1689 + #1740 + #1768
Kernel: topic/sof-dev branch f95e4eb + thesofproject/linux#1037
wwittbrx commentedon Sep 4, 2019
Ok, we are on the same page then. Lets wait for PRs to be merged to their respective master branches, retest and then close this issue.
wenqingfu commentedon Sep 25, 2019
All PRs above has been merged. Can this issue be closed now? @keyonjie @mrajwa
aiChaoSONG commentedon Sep 25, 2019
@wenqingfu We will run tests on several platforms today, if everything is fine, then this issue can be closed.
YvonneYang2 commentedon Sep 25, 2019
Issue cannot reproduce on whl-i2s and CML Chrome platforms. Close the bug.
Test recipe:
Firmware: d42c019 https://github.com/thesofproject/sof/commits/master
Kernel: c0f21e6 https://github.com/thesofproject/linux/commits/topic/sof-dev