Adds proper IPC usage for AudioDevice::GetAudioHardwareBufferSize().

Adds proper IPC usage for AudioDevice::GetAudioHardwareBufferSize().
This method did previously call directly into media/audio.

BUG=none
TEST=WebAudio demo in multi-process mode
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=109388

[henrika (OOO until Aug 14), 2011-11-08 14:29:32]

Avoids an existing crash in  AudioDevice::GetAudioHardwareBufferSize() which
asks for the current hardware sample rate. For Windows 7, this method must be
called from an MTA COM thread and the existing implementation does not fulfill
this condition since it calls directly without using IPC.

This patch adds proper IPC usage.

[tommi (sloooow) - chröme, 2011-11-08 14:31:30]

lgtm

http://codereview.chromium.org/8497014/diff/1/content/common/view_messages.h
File content/common/view_messages.h (right):

http://codereview.chromium.org/8497014/diff/1/content/common/view_messages.h#...
content/common/view_messages.h:686: size_t /* buffer_size */)
fix indent

[henrika (OOO until Aug 14), 2011-11-08 16:29:54]

John and Andrew: I would appreciate a quick check here. Should be trivial I
hope.

The red bots above are not due this change.

Many thanks.

[Raymond Toy (Google), 2011-11-08 18:05:40]

This patch fixes, of course, the crash I was getting.  The couple of
webaudio demos that I tried work.  The audio is kind of buzzy, but that's
an issue with webaudio and the rendering size compared with the buffer size
used by this new code.  I'm working on fixing this issue in webaudio.

Ray


On Tue, Nov 8, 2011 at 8:29 AM, <henrika@chromium.org> wrote:

> John and Andrew: I would appreciate a quick check here. Should be trivial I
> hope.
>
> The red bots above are not due this change.
>
> Many thanks.
>
>
http://codereview.chromium.**org/8497014/<http://codereview.chromium.org/8497...
>

[jam, 2011-11-08 18:31:31]

http://codereview.chromium.org/8497014/diff/7/content/common/view_messages.h
File content/common/view_messages.h (right):

http://codereview.chromium.org/8497014/diff/7/content/common/view_messages.h#...
content/common/view_messages.h:686: size_t /* buffer_size */)
we avoid size_t in message definitions since they are different size in 32/64
bit binaries. in the future it is likely that we will have both binaries talking
to each other.

[scherkus (not reviewing), 2011-11-09 00:56:27]

LGTM but jam@'s comment still stands

[henrika (OOO until Aug 14), 2011-11-09 10:03:07]

John,

please check again. Modified from size_t to unit32.

[tommi (sloooow) - chröme, 2011-11-09 10:04:00]

lgtm

[jam, 2011-11-09 17:43:44]

lgtm