JDA unittest: add an abort test case.

JDA unittest: add an abort test case.

Elm jpeg driver has a bug that kernel may crash if the decoding is
aborted. Add a testcase to catch it.

BUG=b:35983721
TEST=Run JDA unittest on elm and it can catch the bug.
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2746073002
Cr-Commit-Position: refs/heads/master@{#456630}
Committed: https://chromium.googlesource.com/chromium/src/+/0d263d955d872cc13b45872d6c1e0a40f325840b

[wuchengli, 2017-03-13 08:36:44]

Description was changed from

==========
JDA unittest: add an abort test case.

Elm jpeg driver has a bug that kernel may crash if the decoding is
aborted. Add a testcase to catch it.

BUG=b:35983721
TEST=Run JDA unittest on elm and it can catch the bug.
==========

to

==========
JDA unittest: add an abort test case.

Elm jpeg driver has a bug that kernel may crash if the decoding is
aborted. Add a testcase to catch it.

BUG=b:35983721
TEST=Run JDA unittest on elm and it can catch the bug.
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[wuchengli, 2017-03-13 08:37:20]

wuchengli@chromium.org changed reviewers:
+ henryhsu@chromium.org, jcliang@chromium.org

[wuchengli, 2017-03-13 08:37:21]

PTAL.

[jcliang, 2017-03-13 08:52:35]

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
File media/gpu/jpeg_decode_accelerator_unittest.cc (right):

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
media/gpu/jpeg_decode_accelerator_unittest.cc:536: const int kNumOfJpegToDecode
= 5;
const size_t

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
media/gpu/jpeg_decode_accelerator_unittest.cc:537: for (int j = 0; j <
kNumOfJpegToDecode; j++)
size_t j

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
media/gpu/jpeg_decode_accelerator_unittest.cc:540: // decoding. Then destroy the
first decoder when it is still decoding. The
Are we relying on TestDecode() to destroy the first decoder? Is it guaranteed
that the first decoder is destroyed before the decoding is done?

[wuchengli, 2017-03-13 14:20:08]

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
File media/gpu/jpeg_decode_accelerator_unittest.cc (right):

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
media/gpu/jpeg_decode_accelerator_unittest.cc:536: const int kNumOfJpegToDecode
= 5;
On 2017/03/13 08:52:34, jcliang wrote:
> const size_t
Will fix in next PS.

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
media/gpu/jpeg_decode_accelerator_unittest.cc:537: for (int j = 0; j <
kNumOfJpegToDecode; j++)
On 2017/03/13 08:52:34, jcliang wrote:
> size_t j
Will fix in next PS.

https://codereview.chromium.org/2746073002/diff/1/media/gpu/jpeg_decode_accel...
media/gpu/jpeg_decode_accelerator_unittest.cc:540: // decoding. Then destroy the
first decoder when it is still decoding. The
On 2017/03/13 08:52:34, jcliang wrote:
> Are we relying on TestDecode() to destroy the first decoder? Is it guaranteed
> that the first decoder is destroyed before the decoding is done?
Yes. We are relying on TestDecode() to destroy the first decoder. It's not
guaranteed but the odds is very high.
- After JDA decodes the first frame, we send four more to decode without
waiting. Then destroy the decoder.
- Each decode takes several milliseconds.
We can increase kNumOfJpegToDecode if we want to increase the odds.

[wuchengli, 2017-03-13 14:28:32]

All done. PTAL.

[jcliang, 2017-03-13 16:02:10]

On 2017/03/13 14:28:32, wuchengli wrote:
> All done. PTAL.

I guess the asynchronous nature of chrome JDA makes it hard to guarantee that
the abort happens before decoding is done.
 
lgtm, but I'm not a owner. :/

[henryhsu, 2017-03-14 02:23:44]

lgtm

[wuchengli, 2017-03-14 04:30:51]

On 2017/03/13 16:02:10, jcliang wrote:
> On 2017/03/13 14:28:32, wuchengli wrote:
> > All done. PTAL.
> 
> I guess the asynchronous nature of chrome JDA makes it hard to guarantee that
> the abort happens before decoding is done.
Yeah. If the test calls IOCTL directly, we can guarantee that by
not dequeing the output buffer.
>  
> lgtm, but I'm not a owner. :/

[wuchengli, 2017-03-14 04:31:41]

The CQ bit was checked by wuchengli@chromium.org

[commit-bot: I haz the power, 2017-03-14 04:31:57]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-03-14 05:33:04]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1489465901613440,
"parent_rev": "28fe425e8f26f02b35b14900d19145a8da0bec64", "commit_rev":
"0d263d955d872cc13b45872d6c1e0a40f325840b"}

[commit-bot: I haz the power, 2017-03-14 05:34:06]

Description was changed from

==========
JDA unittest: add an abort test case.

Elm jpeg driver has a bug that kernel may crash if the decoding is
aborted. Add a testcase to catch it.

BUG=b:35983721
TEST=Run JDA unittest on elm and it can catch the bug.
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
JDA unittest: add an abort test case.

Elm jpeg driver has a bug that kernel may crash if the decoding is
aborted. Add a testcase to catch it.

BUG=b:35983721
TEST=Run JDA unittest on elm and it can catch the bug.
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2746073002
Cr-Commit-Position: refs/heads/master@{#456630}
Committed:
https://chromium.googlesource.com/chromium/src/+/0d263d955d872cc13b45872d6c1e...
==========

[commit-bot: I haz the power, 2017-03-14 05:34:07]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/chromium/src/+/0d263d955d872cc13b45872d6c1e...