Add multiple video encoders/decoders to video encoder/decoder unittest

Add multiple video encoders/decoders to video encoder/decoder unittest

Some platforms have multiple video decoders and encoders.
Return the first succeed initialized VDA and VEA from all possible platforms.

BUG=chromium:445016
TEST=test on squawks and peach_pit

Committed: https://crrev.com/bcaa081d536e879ece2cb35289a0ac35b74280f8
Cr-Commit-Position: refs/heads/master@{#310775}

[henryhsu, 2015-01-09 12:17:17]

henryhsu@chromium.org changed reviewers:
+ owenlin@chromium.org, posciak@chromium.org, wuchengli@chromium.org

[henryhsu, 2015-01-09 12:17:18]

PTAL

[Pawel Osciak, 2015-01-09 13:09:11]

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
File content/common/gpu/media/video_decode_accelerator_unittest.cc (right):

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:62: //
defined(ARCH_CPU_ARMEL) || (defined(USE_OZONE) && defined(USE_V4L2_CODEC))
Maybe just remove this please, it's actually confusing I think.

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:289:
scoped_ptr<media::VideoDecodeAccelerator> CreateDXVAVDA();
Could these be static and maybe even outside of this class?

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:474: size_t
num_decoders = sizeof(decoders) /
Please use arraysize() and move to the for loop.

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:478: if
(!decoders[i]) continue;
New line please.

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:488:
CHECK(decoder_.get());
Perhaps have decoder_.reset() at the end of each iteration and have this as the
failing check instead of these two lines below?

[Pawel Osciak, 2015-01-09 13:09:42]

Also please say a bit more in the CL message, something similar to what you
wrote in your previous CL for gpu_vea/vda classes.

[henryhsu, 2015-01-09 14:29:23]

PTAL

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
File content/common/gpu/media/video_decode_accelerator_unittest.cc (right):

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:62: //
defined(ARCH_CPU_ARMEL) || (defined(USE_OZONE) && defined(USE_V4L2_CODEC))
On 2015/01/09 13:09:11, Pawel Osciak wrote:
> Maybe just remove this please, it's actually confusing I think.

Done.

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:289:
scoped_ptr<media::VideoDecodeAccelerator> CreateDXVAVDA();
On 2015/01/09 13:09:11, Pawel Osciak wrote:
> Could these be static and maybe even outside of this class?

I think no. First, we have multiple instances of GLRenderingVDAClient in several
test cases. Each instance should have their own VDA. Second, static or global
variables of class type are forbidden.

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:474: size_t
num_decoders = sizeof(decoders) /
On 2015/01/09 13:09:11, Pawel Osciak wrote:
> Please use arraysize() and move to the for loop.

Done.

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:478: if
(!decoders[i]) continue;
On 2015/01/09 13:09:11, Pawel Osciak wrote:
> New line please.

Done.

https://codereview.chromium.org/845663003/diff/1/content/common/gpu/media/vid...
content/common/gpu/media/video_decode_accelerator_unittest.cc:488:
CHECK(decoder_.get());
On 2015/01/09 13:09:11, Pawel Osciak wrote:
> Perhaps have decoder_.reset() at the end of each iteration and have this as
the
> failing check instead of these two lines below?

We don't have to reset decoder_ at the end of each iteration because it will be
replaced by next decoders[i].
And I think CHECK(decoder_.get()) is useless. Once this CHECK is hit, we still
don't know which decoder initial failed.
But CHECK(false) is required to raise error when all decoders cannot be
initialized.

[Pawel Osciak, 2015-01-09 14:45:00]

lgtm

[henryhsu, 2015-01-09 14:48:11]

The CQ bit was checked by henryhsu@chromium.org

[commit-bot: I haz the power, 2015-01-09 14:48:48]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/845663003/20001

[commit-bot: I haz the power, 2015-01-09 15:18:14]

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

[commit-bot: I haz the power, 2015-01-09 15:18:16]

Try jobs failed on following builders:
  win_chromium_rel_ng on tryserver.chromium.win
(http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[henryhsu, 2015-01-09 15:41:14]

The CQ bit was checked by henryhsu@chromium.org

[commit-bot: I haz the power, 2015-01-09 15:42:18]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/845663003/40001

[commit-bot: I haz the power, 2015-01-09 16:24:49]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-01-09 16:27:59]

Patchset 3 (id:??) landed as
https://crrev.com/bcaa081d536e879ece2cb35289a0ac35b74280f8
Cr-Commit-Position: refs/heads/master@{#310775}