Don't background processes with active audio output.

Don't background processes with active audio output.

If a process is actively rendering audio, it should not be put into
the background state.  "Active" is determined by having any stream
in the playing state as considered by AudioRendererHost.

When a tab has active audio, it's already being woken up ~ every
10-20ms, so I expect the power impact of this change to be negligible.

This should reduce glitching under load.

BUG=362294
TEST=Ensure Windows playback under load no longer glitches.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=274020

[DaleCurtis, 2014-05-28 23:00:31]

luken: General review (per cpu@).
jam: content/ approval.

[DaleCurtis, 2014-05-28 23:01:19]

Whoops, scherkus for media/ review. Note: This CL should only be landed with
https://codereview.chromium.org/305533006/

[DaleCurtis, 2014-05-28 23:01:32]

+scherkus for real this time.

[scherkus (not reviewing), 2014-05-29 02:24:42]

what happens when audio starts playing on a already-backgrounded process?

https://codereview.chromium.org/298253004/diff/20001/content/browser/renderer...
File content/browser/renderer_host/media/audio_renderer_host.h (right):

https://codereview.chromium.org/298253004/diff/20001/content/browser/renderer...
content/browser/renderer_host/media/audio_renderer_host.h:173:
base::AtomicRefCount num_playing_streams_;
AtomicRefCount because of IO vs. UI thread?

[jam, 2014-05-29 17:31:19]

lgtm

https://codereview.chromium.org/298253004/diff/20001/content/browser/renderer...
File content/browser/renderer_host/media/audio_renderer_host.h (right):

https://codereview.chromium.org/298253004/diff/20001/content/browser/renderer...
content/browser/renderer_host/media/audio_renderer_host.h:88: // Returns true if
any streams managed by this host are actively playing.
nit: add a comment that this is a thread safe method (since comment at top of
file says that the methods are all IO)

[DaleCurtis, 2014-05-29 18:57:13]

Audio started in the background is no worse off than today without my Window's
"background mode" patch.  With that patch, things will still work, they're just
more likely to glitch under load.

Presumably if the tab is actually playing audio the user wants, they'll switch
back to that tab at some point and avoid future backgrounding.

Handling this case otherwise would be tricky since we'd need the render process
host to have its own MediaObserver impl (which today lives in chrome/) and is
not injected, but rather pulled from GetBrowserClient().

https://codereview.chromium.org/298253004/diff/20001/content/browser/renderer...
File content/browser/renderer_host/media/audio_renderer_host.h (right):

https://codereview.chromium.org/298253004/diff/20001/content/browser/renderer...
content/browser/renderer_host/media/audio_renderer_host.h:88: // Returns true if
any streams managed by this host are actively playing.
On 2014/05/29 17:31:19, jam wrote:
> nit: add a comment that this is a thread safe method (since comment at top of
> file says that the methods are all IO)

Done.

https://codereview.chromium.org/298253004/diff/20001/content/browser/renderer...
content/browser/renderer_host/media/audio_renderer_host.h:173:
base::AtomicRefCount num_playing_streams_;
On 2014/05/29 02:24:42, scherkus wrote:
> AtomicRefCount because of IO vs. UI thread?

Correct.

[scherkus (not reviewing), 2014-05-29 21:49:27]

lgtm

[luken, 2014-05-30 17:57:04]

lgtm

[DaleCurtis, 2014-05-30 17:59:15]

The CQ bit was checked by dalecurtis@chromium.org

[commit-bot: I haz the power, 2014-05-30 18:00:43]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/dalecurtis@chromium.org/298253004/30001

[commit-bot: I haz the power, 2014-05-31 00:09:30]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  android_dbg_triggered_tests on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/android_dbg_triggered...)

[commit-bot: I haz the power, 2014-05-31 01:55:59]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-05-31 01:55:59]

Try jobs failed on following builders:
  android_dbg_triggered_tests on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/android_dbg_triggered...)

[DaleCurtis, 2014-05-31 03:13:54]

The CQ bit was checked by dalecurtis@chromium.org

[commit-bot: I haz the power, 2014-05-31 03:14:31]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/dalecurtis@chromium.org/298253004/30001

[commit-bot: I haz the power, 2014-05-31 04:37:20]

Change committed as 274020