Stress Test: repetitive unloading and reloading driver can cause kernel panic on BYT using nocodec topology about linux HOT 17 CLOSED

Can you clarify why I see this log in the dmesg files:
[ 210.916884] rt5651 i2c-10EC5651:00: Device with ID register 0xffff8c2d is not rt5651

This means you are not in nodec mode, you are using the ACPI ID to probe the codec. Or you must use the force-nocodec kconfig?

from linux.

the PR #80 and PR #72 can fix this kind of kernel panic issue already.

from linux.

Stress test on driver unload/reload is not priority atm.

@lyakh it seem PR #130 still need to fix some check patch error. Please notify validation here when the PR is merged. Thanks!

from linux.

Firmware load failed after booting. So tests were blocked.
Env:
sof-dev : 108d9cf
sof tool: 78748d4
sof stable-1.2: 7dd4b1d
tplg: sof-byt-nocodec.tplg
Log:
dmesg-byt.log

from linux.

Firmware load failed after booting once. So tests were blocked.
Test env:
kernel: sof-dev 160b45f
sof tool: 57b5212
sof stable-1.2: 7dd4b1d
tplg: sof-byt-nocodec.tplg
dmesg-panic.log

from linux.

@ZhendanYang please paste the relevant dmesg sections .

[   58.688933] gnome-shell[1159]: segfault at 18 ip 00007f751d1b11bc sp 00007ffc174a36e0 error 4 in libmutter-1.so.0.0.0[7f751d0f7000+142000]
[   58.688949] Code: e5 f8 83 c5 0d 89 44 24 28 48 8b 43 30 4c 8d 74 24 30 89 74 24 24 48 8b 7b 20 4c 8d 6c 24 20 89 54 24 2c 4c 89 f2 89 4c 24 20 <8b> 70 18 e8 5c bb fd ff e8 e7 a2 fd ff 85 c0 0f 85 bf 00 00 00 4c 

sorry, I'm not seeing any audio driver bugs here.
Can you retest this console mode only (i.e. don't start GUI) eneter a single user non graphical runlevel.

from linux.

@lgirdwood
After set non graphical runlevel , fimware boot successfully on latest codebase ,but module reload issue still exist.
Test env:
kernel: sof-dev 93853cb
sof tool: 2d7a933
sof stable-1.2: 7dd4b1d
tplg: sof-byt-nocodec.tplg
dmesg log before module reload:
dmesg-byt.log

from linux.

@zhuqingliang ok thanks, I guess you have no log of the oops now ? hard lock up ?

from linux.

@lgirdwood Here is module reload dmesg log
oops.log

from linux.

@stevyan @markyang Could you help to check if this issue is gone with fixing of #144 by PR #311?

from linux.

Summary:
This issue cannot be reproduced on BYT-nocodec using the build on Dec 7.
There is no Kernel Panic.

Test steps:

1. kmod_scripts/sof_bootone.sh
2. aplay -l

**** List of PLAYBACK Hardware Devices ****
card 0: sofnocodec [sof-nocodec], device 0: Passthrough (*) []
  Subdevices: 1/1
  Subdevice #0: subdevice #0

3. arecord -l

**** List of CAPTURE Hardware Devices ****
card 0: sofnocodec [sof-nocodec], device 0: Passthrough (*) []
  Subdevices: 1/1
  Subdevice #0: subdevice #0

Test env:
sof master: b5d6c71 #Dec 7
kernel sof-dev: a71221d #Dec 7
tplg: test-ssp2-mclk-0-I2S-volume-s16le-s16le-48k-19200k-nocodec.tplg

from linux.

@markyang are you using https://github.com/thesofproject/soft/blob/master/kmod_scripts/sof_bootone.sh in you verification?

And here we need to run stress test with multiple unload/reload, So we expect to use soft/kmod_scripts/sof_bootloop.sh

from linux.

@mengdonglin There is an invalid DAI type error in the dmesg, so sof_bootloop.sh doesn't work.
I can update the sof_bootloop.sh script to ignore this error and try later.

from linux.

Summary:
This issue was blocked by /issues/471

from linux.

I haven' t seen any bootloop errors with minnowboard after applying #695

from linux.

@zhuqingliang can we close this or is this still valid?

from linux.

from the last few days' regression report, I think it can conclude it's fixed. Thank you for following up with this bug, @juimonen

from linux.