New issue
Advanced search Search tips

Issue 590918 link

Starred by 1 user

Issue metadata

Status: Verified
Owner:
Closed: May 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

NodesChanged signal triggered from cras when output node changes volume

Project Member Reported by jen...@chromium.org, Mar 1 2016

Issue description

Version: R50 and R51
OS: Chrome

What steps will reproduce the problem?
1.Get in chroshell, run following command to monitor dbus signal:
sudo dbus-monitor --system interface=org.chromium.cras.Control
2. Change volume of the audio output.


What is the expected output? What do you see instead?


Please use labels and text to provide additional information.
We should only observe one signal: OutputNodeVolumeChanged

Instead, we observe 2 signals:
OutputNodeVolumeChanged
NodesChanged

The NodesChange bbus signal should be not be sent from cras to chrome, since there is no nodes change happening. 

It does not cause any bugs, but it causes unnecessary handling of the audio nodes each time the volume change happens. Each time when active node changes, chrome code will set volume to the new active node, which triggers unnecessary NodesChanged signal due to this issue. 

 

Comment 1 by ka...@chromium.org, Mar 1 2016

Cc: cychiang@chromium.org
Hi jennyz, would this be related to chameleon audio tests failures, where active node is not as expected:
- issue 588579
- issue 590466
+cychiang@
I am not sure how the failed tests were written. Most likely, it should not cause those issue.

For issue 588579: Does the playback test changes volume during the same test? If not, it should not be related.

For issue 590466: Most likely not related, even though the extra NodesChanged signal is received, chrome code is resilient, and should not select wrong active node. But there is another cras bug(586026) in which cras will unexpectedly turn an active node to be inactive when another audio device was unplugged, that one could cause the active node to be selected incorrectly with R50 code. Does this test involve removing some audio nodes during the test?

Comment 3 by ka...@chromium.org, Mar 1 2016

Not in these cases - they are basic tests where only play/record is done with default volume.
Same failure reasons are observed on  more complex tests where audio node is switched and volume is set during the test one time at the beginning for each node:
https://wmatrix.googleplex.com/unfiltered?tests=audio_AudioNodeSwitch*&platforms=daisy,peach_pi,nyan_big,auron_paine&days_back=7
Status: Assigned (was: Untriaged)
We shouldn't have this redundant signal. Looking into this now.
Status: Fixed (was: Assigned)
This has been fixed by https://chromium-review.googlesource.com/329454
Labels: VerifyIn-54

Comment 7 by ka...@chromium.org, Aug 31 2016

Labels: Bulk-Verified
Status: Verified (was: Fixed)

Sign in to add a comment