The latest test result on chell is R54-8719.0.0 and it passed.

https://wmatrix.googleplex.com/platform/unfiltered?hide_missing=True&tests=audio_AudioBasicHeadphone&days_back=20&platforms=chell

The error messages:

[  378.396476] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  378.396515] snd_soc_skl 0000:00:1f.3: ASoC: POST_PMD: media0_in mi event failed: -110
[  378.696468] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  378.696516] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMD: codec0_out mo event failed: -110
[  378.996676] snd_soc_skl 0000:00:1f.3: ipc: unload modules failed :-110
[  378.996714] snd_soc_skl 0000:00:1f.3: Failed to UnLoad module
[  378.996745] snd_soc_skl 0000:00:1f.3: ASoC: POST_PMD: codec0_out mo event failed: -110
[  379.296958] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  383.650271] snd_soc_skl 0000:00:1f.3: skl_is_pipe_mcps_avail: module_id 3 instance 0
[  383.650286] snd_soc_skl 0000:00:1f.3: exceeds ppl mcps available 30000000 > mem -2420000
[  383.650299] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMU: media0_in cpr 0 event failed: -16
[  383.650311] snd_soc_skl 0000:00:1f.3: skl_is_pipe_mcps_avail: module_id 2 instance 0
[  383.650321] snd_soc_skl 0000:00:1f.3: exceeds ppl mcps available 30000000 > mem -2420000
[  383.650332] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMU: codec0_out mo event failed: -16
[  383.650345] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMU: media0_in mi event failed: -22
[  383.952168] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  383.952204] snd_soc_skl 0000:00:1f.3: ASoC: POST_PMD: media0_in mi event failed: -110
[  384.252426] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  384.252471] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMD: codec0_out mo event failed: -110
[  384.552619] snd_soc_skl 0000:00:1f.3: ipc: unload modules failed :-110
[  384.552651] snd_soc_skl 0000:00:1f.3: Failed to UnLoad module
[  384.552677] snd_soc_skl 0000:00:1f.3: ASoC: POST_PMD: codec0_out mo event failed: -110
[  384.852758] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  384.854509] snd_soc_skl 0000:00:1f.3: skl_is_pipe_mcps_avail: module_id 3 instance 0
[  384.854547] snd_soc_skl 0000:00:1f.3: exceeds ppl mcps available 30000000 > mem -3200000
[  384.854582] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMU: media0_in cpr 0 event failed: -16
[  384.854614] snd_soc_skl 0000:00:1f.3: skl_is_pipe_mcps_avail: module_id 2 instance 0
[  384.854652] snd_soc_skl 0000:00:1f.3: exceeds ppl mcps available 30000000 > mem -3200000
[  384.854684] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMU: codec0_out mo event failed: -16
[  384.854719] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMU: media0_in mi event failed: -22
[  405.567976] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  405.568018] snd_soc_skl 0000:00:1f.3: ASoC: POST_PMD: media0_in mi event failed: -110
[  405.868366] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden
[  405.868412] snd_soc_skl 0000:00:1f.3: ASoC: PRE_PMD: codec0_out mo event failed: -110
[  406.168536] snd_soc_skl 0000:00:1f.3: ipc: unload modules failed :-110
[  406.168576] snd_soc_skl 0000:00:1f.3: Failed to UnLoad module
[  406.168608] snd_soc_skl 0000:00:1f.3: ASoC: POST_PMD: codec0_out mo event failed: -110
[  406.468677] snd_soc_skl 0000:00:1f.3: ipc: bind/unbind faileden


There must be something wrong between 8719 and 8730.
However it is not obvious on the blamelist.
I will try some versions between them.

Weird... I can reproduce the issue on 8719.0.0 (which passed the test on dashboard)
It persists even after reboot.

The issue is only triggered by internal speaker.
When internal speaker is used and issue repros, headphone will get effected too.
If I only use headphone then it works fine.

The internal speaker test dashboard shows that 8715 was fine, and 8719 was bad.
https://wmatrix.googleplex.com/matrix/unfiltered?tests=audio_AudioBasicInternalSpeaker&days_back=20&hide_missing=True

I checked the version between them and found that 8717.0 is good and 8718.0 is bad.

In the blamelist
https://crosland.corp.google.com/log/8717.0.0..8718.0.0

The only one that looks suspicious is https://chromium-review.googlesource.com/#/c/366590/
which change the audio firmware on SKL.

Updated on https://code.google.com/p/chrome-os-partner/issues/detail?id=53260#c47.
I will revert it if there is no proper fix.

+Kalin
Hi Kalin, I am not sure what happened to chell in the lab(it is still pingable), but from the dashboard https://wmatrix.googleplex.com/matrix/unfiltered?tests=audio_AudioBasicInternalSpeaker&days_back=20&hide_missing=True, sometimes the test does not run.
Could you please check what might cause the skip of the test ?

There are ongoing effort like https://code.google.com/p/chrome-os-partner/issues/detail?id=53260 for skylake so it will be quite important to make sure test suite is up and running for chell. For example in #5 we can identify the bad version using test result of 8715 and 8719 and that is really helpful.

Thanks a lot!!

Hi Jimmy,
Thanks for the detailed follow up.

Specifically for the builds R54-8717.0.0 and R54-8718.0.0, I suspect scheduler skipped them by some reason. I am seeing R54-8715.0.0 running on 8/18, and no tests were run for half a day. After which tests against R54-8719.0.0 start running.
Screenshot: https://screenshot.googleplex.com/TK1FX2DKAdc

Unfortunately there is no good chameleon coverage on chell or other SKY devices at this time, to free up the load from the single audio box, this test requires as environment. This will change for the better once supplies get sufficient.

Revert posted https://chromium-review.googlesource.com/#/c/374301/1

Re#8 Hi Kalin, There are two set for chell.
chromeos1-row5-rack7-host1 is in audio box
chromeos1-row2-rack4-host5 is with audio board only.

I thought with these two set that should be enough to cover audio test suite?
Not sure why they are skipped, but I found that they are in Ready state most of the time, so there might be reason other than being occupied.
Let keep an eye on it and see when will the test be scheduled again.
Thanks!

Same way DUT at chromeos1-row2-rack4-host5 jumped from R54-8715.0.0 to R54-8719.0.0. - https://screenshot.googleplex.com/4xR0rUvzd3E

Till this morning, I believe there were issues with PFQ and builds infra health -https://screenshot.googleplex.com/fmLCxSJQosE , hence the limited amount of boards for bulk of test suites being tested.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/linux-firmware/+/7b6d9301cebd52486e8f27d7175a236c710f3856

commit 7b6d9301cebd52486e8f27d7175a236c710f3856
Author: Cheng-Yi Chiang <cychiang@chromium.org>
Date: Wed Aug 24 03:28:16 2016

Revert "FROMLIST: linux-firmware: intel: Update Skylake audio firmware"

This reverts commit 733affbe694f9c000eac5233182803ee34c490d8.
New firmeware dsp_fw_release_v958 makes chell fail to play audio on internal
speaker. Details and error messages are in  crbug.com/640048 .

BUG= chromium:640048 
TEST=with dsp_fw_release_v951, internal speaker works fine on chell.

Signed-off-by: Cheng-Yi Chiang <cychiang@chromium.org>
Change-Id: Ia228ce6b825c260f25929177d12afe5744d562be
Reviewed-on: https://chromium-review.googlesource.com/375579

[modify] https://crrev.com/7b6d9301cebd52486e8f27d7175a236c710f3856/intel/dsp_fw_release.bin
[modify] https://crrev.com/7b6d9301cebd52486e8f27d7175a236c710f3856/WHENCE
[delete] https://crrev.com/c626a39683577d582c489d785521b5d8f06941ba/intel/dsp_fw_release_v958.bin

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/linux-firmware/+/98be180943f3b2c5938da49f22d54bea4bdd32bc

commit 98be180943f3b2c5938da49f22d54bea4bdd32bc
Author: Cheng-Yi Chiang <cychiang@chromium.org>
Date: Wed Aug 24 03:28:16 2016

Revert "FROMLIST: linux-firmware: intel: Update Skylake audio firmware"

This reverts commit 733affbe694f9c000eac5233182803ee34c490d8.
New firmeware dsp_fw_release_v958 makes chell fail to play audio on internal
speaker. Details and error messages are in  crbug.com/640048 .

BUG= chromium:640048 
TEST=with dsp_fw_release_v951, internal speaker works fine on chell.

Signed-off-by: Cheng-Yi Chiang <cychiang@chromium.org>
Change-Id: Ia228ce6b825c260f25929177d12afe5744d562be
Reviewed-on: https://chromium-review.googlesource.com/374301
Reviewed-by: Hsinyu Chao <hychao@chromium.org>

[modify] https://crrev.com/98be180943f3b2c5938da49f22d54bea4bdd32bc/intel/dsp_fw_release.bin
[modify] https://crrev.com/98be180943f3b2c5938da49f22d54bea4bdd32bc/WHENCE
[delete] https://crrev.com/c626a39683577d582c489d785521b5d8f06941ba/intel/dsp_fw_release_v958.bin

Marking as fixed to get this verified 

verified on 8740.0.0/54.0.2837.0 using Ted, youtube. 

Hi,
 The issue with 958 FW is because the ADI FW also needs update.
So instead of revert of 
https://chromium-review.googlesource.com/#/c/366590/


The right fix is it take the ADI FW update :
https://chrome-internal-review.googlesource.com/#/c/273645/

New FW dspv958 needs the above ADI FW.

Hi Sathya, thank you for the explanation.
Please reland 366590 with a CQ-DEPEND=273645.
That will make sure they are merged at the same image version.
I noticed that in 273645, you have CQ-DEPEND=CL:366590, but that only constraint CL 273645 to depend on CL 366590, which had been merged and reverted.
You will need to update commit message of CL 273645 to have a new CQ-DEPEND value pointing to the reland of 366590.
Thanks!

Hi,
 Please find the reland of 366590:
https://chromium-review.googlesource.com/377259

I have updated the CQ-DEPEND of 273645.
Please merge as required.
Thanks a lot.