Add support building with OpenMAX DL FFT for WebAudio

Add support building with OpenMAX DL FFT for WebAudio

BUG=223172
This just adds a gyp variable that enables or disables the use of the OpenMAX DL FFT routines.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=190019

[Raymond Toy (Google), 2013-02-13 23:22:54]

+ felipeg@

Just for your information.  Not yet ready for commit.

[felipeg, 2013-02-14 12:44:13]

LGTM

Is the new third party repo for ARM FFT ready in the chromium tree ?

[Raymond Toy (Google), 2013-02-14 17:02:01]

On 2013/02/14 12:44:13, felipeg1 wrote:
> LGTM
> 
> Is the new third party repo for ARM FFT ready in the chromium tree ?

Not yet. I'm waiting for
https://code.google.com/p/chromium/issues/detail?id=172998 to get resolved.

[Raymond Toy (Google), 2013-02-22 18:55:38]

On 2013/02/14 17:02:01, rtoy wrote:
> On 2013/02/14 12:44:13, felipeg1 wrote:
> > LGTM
> > 
> > Is the new third party repo for ARM FFT ready in the chromium tree ?
> 
> Not yet. I'm waiting for
> https://code.google.com/p/chromium/issues/detail?id=172998 to get resolved.

It's resolved.  Just need to update DEPS (in a different CL) and we are good to
go.

[Raymond Toy (Google), 2013-02-27 17:51:40]

This CL establishes the use_arm_fft variable needed in
https://bugs.webkit.org/show_bug.cgi?id=109755 to control whether the ARM FFT is
used in webaudio on android.

[Paweł Hajdan Jr., 2013-02-27 17:56:58]

https://codereview.chromium.org/12260023/diff/1/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/12260023/diff/1/build/common.gypi#newcode1167
build/common.gypi:1167: 'use_arm_fft%': '<(use_arm_fft)',
Why so many deep repetitions?

Isn't it sufficient to just put the 0 here?

[Raymond Toy (Google), 2013-02-27 22:36:03]

On 2013/02/27 17:56:58, Paweł Hajdan Jr. wrote:
> https://codereview.chromium.org/12260023/diff/1/build/common.gypi
> File build/common.gypi (right):
> 
> https://codereview.chromium.org/12260023/diff/1/build/common.gypi#newcode1167
> build/common.gypi:1167: 'use_arm_fft%': '<(use_arm_fft)',
> Why so many deep repetitions?
> 
> Isn't it sufficient to just put the 0 here?

With help from rsleevi, the deep nesting has been minimized.

[Ryan Sleevi, 2013-02-28 00:33:41]

https://codereview.chromium.org/12260023/diff/6001/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/12260023/diff/6001/build/common.gypi#newcode1027
build/common.gypi:1027: 'use_arm_fft%': '<use_arm_fft)',
Why isn't this around line 931?

[Raymond Toy (Google), 2013-03-05 19:30:11]

https://codereview.chromium.org/12260023/diff/6001/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/12260023/diff/6001/build/common.gypi#newcode1027
build/common.gypi:1027: 'use_arm_fft%': '<use_arm_fft)',
On 2013/02/28 00:33:41, Ryan Sleevi wrote:
> Why isn't this around line 931?

Moved as requested.

[Raymond Toy (Google), 2013-03-05 20:53:15]

Updated according review comments and also renamed use_arm_fft to
use_openmax_dl_fft, which better reflects that actual usage.

[Ryan Sleevi, 2013-03-05 20:57:54]

Please don't land this until security gives the goahead, but looks fine, mod bug
below

https://codereview.chromium.org/12260023/diff/16001/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/12260023/diff/16001/build/common.gypi#newcode918
build/common.gypi:918: 'use_openmax_dl_fft%': '<use_openmax_dl_fft)',
BUG: <(use_openmax_dl_fft)

[Raymond Toy (Google), 2013-03-05 21:58:05]

On 2013/03/05 20:57:54, Ryan Sleevi wrote:
> Please don't land this until security gives the goahead, but looks fine, mod
bug
> below
> 
> https://codereview.chromium.org/12260023/diff/16001/build/common.gypi
> File build/common.gypi (right):
> 
>
https://codereview.chromium.org/12260023/diff/16001/build/common.gypi#newcode918
> build/common.gypi:918: 'use_openmax_dl_fft%': '<use_openmax_dl_fft)',
> BUG: <(use_openmax_dl_fft)

Typo fixed.  I will wait for the security review before landing this.

[Raymond Toy (Google), 2013-03-22 17:58:56]

On 2013/03/05 21:58:05, rtoy wrote:
> On 2013/03/05 20:57:54, Ryan Sleevi wrote:
> > Please don't land this until security gives the goahead, but looks fine, mod
> bug
> > below
> > 
> > https://codereview.chromium.org/12260023/diff/16001/build/common.gypi
> > File build/common.gypi (right):
> > 
> >
>
https://codereview.chromium.org/12260023/diff/16001/build/common.gypi#newcode918
> > build/common.gypi:918: 'use_openmax_dl_fft%': '<use_openmax_dl_fft)',
> > BUG: <(use_openmax_dl_fft)
> 
> Typo fixed.  I will wait for the security review before landing this.

Security review completed.  The review was at
https://codereview.chromium.org/12317152/

[Ryan Sleevi, 2013-03-22 18:05:22]

Please update the BUG= to refer to a bug that describes why this is being
introduced, so that we can also track what Mstone it's released in, etc.

Otherwise LGTM.

[Raymond Toy (Google), 2013-03-22 18:17:39]

On 2013/03/22 18:05:22, Ryan Sleevi wrote:
> Please update the BUG= to refer to a bug that describes why this is being
> introduced, so that we can also track what Mstone it's released in, etc.
> 
> Otherwise LGTM.

Done.  Description and title updated to better reflect what this is.

[commit-bot: I haz the power, 2013-03-22 19:45:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rtoy@google.com/12260023/22001

[commit-bot: I haz the power, 2013-03-23 14:39:48]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rtoy@google.com/12260023/22001

[commit-bot: I haz the power, 2013-03-23 16:17:10]

Change committed as 190019