fix the broeken audio for webrtc for precise

It seems PcmReadi() is often gives out more than 1 packets and this destroy the our scheduling for recording.

BUG=153505
TEST=apprtc.appspot.com?debug=loopback, listen to the audio.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=160661

[no longer working on chromium, 2012-10-05 15:42:36]

Hi Andrew,

Could you please take a quick look at this? I need to merge the fix to M23.

[DaleCurtis, 2012-10-05 17:25:39]

Andrew is out travelling.

Won't having PcmReadi provide multiple packets on a regular basis and only
accounting for the first one wreck your timing too? I.e. your timing estimate is
now smaller than it should be.

[xians2, 2012-10-05 19:26:05]

On 2012/10/05 17:25:39, DaleCurtis wrote:
> Andrew is out travelling.
> 
> Won't having PcmReadi provide multiple packets on a regular basis and only
> accounting for the first one wreck your timing too? I.e. your timing estimate
is
> now smaller than it should be.

Normally PcmReadi should only provide one packet at a time then we will schedule
the next read. This happens on ubuntu lucid.
But on precise, I saw it can provide up to 3 packets at a time, then we estimate
the next read should be later than normal case. But this gives a very choppy
audio, I guess somehow slowing on reading makes the OS dropping the data.

This patch schedules the next read to be the normal case, then it checks the OS
to see if there is data available, if there is, it reads it, otherwise it
schedule next read. It does not add much to the complexity.

[DaleCurtis, 2012-10-05 19:26:56]

Thanks for the explanation Xians. LGTM!

[no longer working on chromium, 2012-10-06 21:49:28]

On 2012/10/05 19:26:56, DaleCurtis wrote:
> Thanks for the explanation Xians. LGTM!

Thanks Dale.
FYI, this fix improves the recording audio, but on my machine the audio still
gets a bit choppy when running the loopback for some time. I am not sure if the
problem is real since my machine has proved to have a worse audio than others.
So I am going to land this fix now and let my colleagues to test it. In case the
problem is still there, I might need your help to review another CL.

[commit-bot: I haz the power, 2012-10-06 21:49:43]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/xians@chromium.org/11065037/1