Extract re-usable InitializeMediaFoundation()

Extract re-usable InitializeMediaFoundation()

Allow media::InitializeMediaFoundation() to be used by any Media
Foundation client code.  This benefits existing Chromium code as well as any
Chromium-based browser that adds its own code calling out to Media Foundation.

There are two side effects for DXVAVideoDecodeAccelerator:

  - MFStartup() and MFShutdown() are now called once during the lifetime of the
    GPU process rather than per each accelerator instance.

  - MFStartup() is now called with the MFSTARTUP_LITE flag, which according to
    MSDN doesn't initialize the sockets library.

TEST=media_unittests

Committed: https://crrev.com/c0ab75a86447bc23d45ff7086b2375e2cd83c205
Cr-Commit-Position: refs/heads/master@{#331446}

Committed: https://crrev.com/5105004066c9a728764a1e60ede6f60a44e3ede5
Cr-Commit-Position: refs/heads/master@{#331623}

[wdzierzanowski, 2015-05-21 20:40:47]

wdzierzanowski@opera.com changed reviewers:
+ sandersd@chromium.org

[wdzierzanowski, 2015-05-21 20:41:35]

PTAL

[sandersd (OOO until July 31), 2015-05-22 19:14:55]

https://codereview.chromium.org/1149203002/diff/20001/media/media.gyp
File media/media.gyp (right):

https://codereview.chromium.org/1149203002/diff/20001/media/media.gyp#newcode407
media/media.gyp:407: 'base/win/mf_initializer.cc',
This should be only built on Windows.

Update BUILD.gn as well.

[sandersd (OOO until July 31), 2015-05-22 19:17:57]

https://codereview.chromium.org/1149203002/diff/20001/media/base/win/mf_initi...
File media/base/win/mf_initializer.h (right):

https://codereview.chromium.org/1149203002/diff/20001/media/base/win/mf_initi...
media/base/win/mf_initializer.h:14: MEDIA_EXPORT void
EnsureMediaFoundationInitialized();
I note that for FFmpeg we just call the method InitializeFFmpeg() (and
VtVideoDecodeAccelerator declares content::InitializeVideoToolbox()). It's
probably enough to use that naming scheme.

https://code.google.com/p/chromium/codesearch#chromium/src/media/filters/ffmp...

[wdzierzanowski, 2015-05-25 08:43:02]

https://codereview.chromium.org/1149203002/diff/20001/media/base/win/mf_initi...
File media/base/win/mf_initializer.h (right):

https://codereview.chromium.org/1149203002/diff/20001/media/base/win/mf_initi...
media/base/win/mf_initializer.h:14: MEDIA_EXPORT void
EnsureMediaFoundationInitialized();
On 2015/05/22 19:17:57, sandersd wrote:
> I note that for FFmpeg we just call the method InitializeFFmpeg() (and
> VtVideoDecodeAccelerator declares content::InitializeVideoToolbox()). It's
> probably enough to use that naming scheme.
> 
>
https://code.google.com/p/chromium/codesearch#chromium/src/media/filters/ffmp...

Done.

https://codereview.chromium.org/1149203002/diff/20001/media/media.gyp
File media/media.gyp (right):

https://codereview.chromium.org/1149203002/diff/20001/media/media.gyp#newcode407
media/media.gyp:407: 'base/win/mf_initializer.cc',
On 2015/05/22 19:14:55, sandersd wrote:
> This should be only built on Windows.

I think it already is, courtesy of the win/ subdirectory and the standard
inclusion/exclusion patterns in GYP.

> 
> Update BUILD.gn as well.

Done.

[sandersd (OOO until July 31), 2015-05-26 18:04:54]

lgtm

[wdzierzanowski, 2015-05-26 18:12:48]

The CQ bit was checked by wdzierzanowski@opera.com

[commit-bot: I haz the power, 2015-05-26 18:13:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1149203002/40001

[commit-bot: I haz the power, 2015-05-26 21:56:48]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-05-26 21:58:35]

Patchset 3 (id:??) landed as
https://crrev.com/c0ab75a86447bc23d45ff7086b2375e2cd83c205
Cr-Commit-Position: refs/heads/master@{#331446}

[scottmg, 2015-05-26 22:12:18]

A revert of this CL (patchset #3 id:40001) has been created in
https://codereview.chromium.org/1154273002/ by scottmg@chromium.org.

The reason for reverting is: Broke compile on the Win GN bots:
http://build.chromium.org/p/chromium.win/builders/Win%20x64%20GN%20%28dbg%29/....

[wdzierzanowski, 2015-05-27 08:41:49]

On 2015/05/26 22:12:18, scottmg wrote:
> A revert of this CL (patchset #3 id:40001) has been created in
> https://codereview.chromium.org/1154273002/ by mailto:scottmg@chromium.org.
> 
> The reason for reverting is: Broke compile on the Win GN bots:
>
http://build.chromium.org/p/chromium.win/builders/Win%20x64%20GN%20%28dbg%29/....

Fixed in patchset #4.

[wdzierzanowski, 2015-05-27 09:03:56]

On 2015/05/27 08:41:49, wdzierzanowski wrote:
> On 2015/05/26 22:12:18, scottmg wrote:
> > A revert of this CL (patchset #3 id:40001) has been created in
> > https://codereview.chromium.org/1154273002/ by mailto:scottmg@chromium.org.
> > 
> > The reason for reverting is: Broke compile on the Win GN bots:
> >
>
http://build.chromium.org/p/chromium.win/builders/Win%20x64%20GN%20%28dbg%29/....
> 
> Fixed in patchset #4.

BTW, now that I've looked around a bit, it seems to me Chromium authors prefer
to reland patches through separate reviews, but I'm not sure why. Arguably,
pushing a fixup to the same review makes the reviewer's job easier, because all
the context is still there.

Is it OK if I continue with this review or do you want me to make a new review
containing patchset #4?

[scottmg, 2015-05-27 15:32:52]

On 2015/05/27 09:03:56, wdzierzanowski wrote:
> On 2015/05/27 08:41:49, wdzierzanowski wrote:
> > On 2015/05/26 22:12:18, scottmg wrote:
> > > A revert of this CL (patchset #3 id:40001) has been created in
> > > https://codereview.chromium.org/1154273002/ by
mailto:scottmg@chromium.org.
> > > 
> > > The reason for reverting is: Broke compile on the Win GN bots:
> > >
> >
>
http://build.chromium.org/p/chromium.win/builders/Win%20x64%20GN%20%28dbg%29/....
> > 
> > Fixed in patchset #4.
> 
> BTW, now that I've looked around a bit, it seems to me Chromium authors prefer
> to reland patches through separate reviews, but I'm not sure why. Arguably,
> pushing a fixup to the same review makes the reviewer's job easier, because
all
> the context is still there.
> 
> Is it OK if I continue with this review or do you want me to make a new review
> containing patchset #4?

I personally prefer it as you have done, and as the diff is very small it's fine
to reland this way. lgtm.

[scottmg, 2015-05-27 15:34:15]

On 2015/05/27 15:32:52, scottmg wrote:
> On 2015/05/27 09:03:56, wdzierzanowski wrote:
> > On 2015/05/27 08:41:49, wdzierzanowski wrote:
> > > On 2015/05/26 22:12:18, scottmg wrote:
> > > > A revert of this CL (patchset #3 id:40001) has been created in
> > > > https://codereview.chromium.org/1154273002/ by
> mailto:scottmg@chromium.org.
> > > > 
> > > > The reason for reverting is: Broke compile on the Win GN bots:
> > > >
> > >
> >
>
http://build.chromium.org/p/chromium.win/builders/Win%20x64%20GN%20%28dbg%29/....
> > > 
> > > Fixed in patchset #4.
> > 
> > BTW, now that I've looked around a bit, it seems to me Chromium authors
prefer
> > to reland patches through separate reviews, but I'm not sure why. Arguably,
> > pushing a fixup to the same review makes the reviewer's job easier, because
> all
> > the context is still there.
> > 
> > Is it OK if I continue with this review or do you want me to make a new
review
> > containing patchset #4?
> 
> I personally prefer it as you have done, and as the diff is very small it's
fine
> to reland this way. lgtm.

I added the optional gn trybot runs (not sure why they're optional rather than
part of the main set).

[wdzierzanowski, 2015-05-27 17:06:17]

The CQ bit was checked by wdzierzanowski@opera.com

[wdzierzanowski, 2015-05-27 17:06:18]

The patchset sent to the CQ was uploaded after l-g-t-m from
sandersd@chromium.org
Link to the patchset: https://codereview.chromium.org/1149203002/#ps60001
(title: "Fix linkage of InitializeMediaFoundation() in GN builds")

[commit-bot: I haz the power, 2015-05-27 17:07:30]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1149203002/60001

[sandersd (OOO until July 31), 2015-05-27 17:42:26]

FYI, I also prefer to re-use the same issue for small fixes.

Still lgtm.

[commit-bot: I haz the power, 2015-05-27 19:35:42]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-05-27 19:36:41]

Patchset 4 (id:??) landed as
https://crrev.com/5105004066c9a728764a1e60ede6f60a44e3ede5
Cr-Commit-Position: refs/heads/master@{#331623}