Replace deprecated base::NonThreadSafe in /content/renderer/gpu in favor of ThreadChecker.

Replace deprecated base::NonThreadSafe in /content/renderer/gpu in favor of ThreadChecker.

Note to crash team: This CL is a refactor and has no intended behavior change.

This change was scripted by https://crbug.com/676387#c84.

(amended to use ThreadChecker upon reviewer request)

If you're wondering why there's a second wave of this, the first version of the
script handled "NON_EXPORTED_BASE(public base::NonThreadSafe)" but not
"public NON_EXPORTED_BASE(base::NonThreadSafe)"...

Note-worthy for the reviewer:
 * SequenceChecker enforces thread-safety but not thread-affinity!
   If the classes that were updated are thread-affine (use thread local
   storage or a third-party API that does) they should be migrated to
   ThreadChecker instead.
 * ~NonThreadSafe() used to implicitly check in its destructor
   ~Sequence/ThreadChecker() doesn't by design. To keep this CL a
   no-op, an explicit check was added to the destructor of migrated
   classes.
 * NonThreadSafe used to provide access to subclasses, as such
   the |sequence_checker_| member was made protected rather than
   private where necessary.

BUG=676387
This CL was uploaded by git cl split.

R=kbr@chromium.org

Review-Url: https://codereview.chromium.org/2918083003
Cr-Commit-Position: refs/heads/master@{#476810}
Committed: https://chromium.googlesource.com/chromium/src/+/b0a378fde6de7bd1e63db9b8249600f4ab4feaf9

[gab, 2017-06-02 18:05:04]

The CQ bit was checked by gab@chromium.org to run a CQ dry run

[gab, 2017-06-02 18:05:06]

Hi, this is an automated change made by https://crbug.com/676387#c84 and sharded
across directories for OWNERS.

No human has looked at this specific output yet, see cl description for things
to watch for.

If it LGTY, please CQ.

Thanks!
Gab

[commit-bot: I haz the power, 2017-06-02 18:06:31]

Dry run: CQ is trying da patch.

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

[Ken Russell (switch to Gerrit), 2017-06-02 18:10:44]

kbr@chromium.org changed reviewers:
+ enne@chromium.org

[Ken Russell (switch to Gerrit), 2017-06-02 18:10:45]

enne, could you review this? It actually looks to me like one of the caveats in
the CL description applies -- it looks like this class is supposed to have
thread-affinity.

[enne (OOO), 2017-06-02 18:38:13]

On 2017/06/02 at 18:10:45, kbr wrote:
> enne, could you review this? It actually looks to me like one of the caveats
in the CL description applies -- it looks like this class is supposed to have
thread-affinity.

I agree that it looks like the CompositorExternalBeginFrameSource is supposed to
only be run on the compositor thread.  The CompositorForwardingMessageFilter has
a comment that the handler here is supposed to be run on that thread (and the
filter itself has a ThreadChecker not just a SequenceChecker).

[gab, 2017-06-02 18:45:59]

Description was changed from

==========
Replace deprecated base::NonThreadSafe in /content/renderer/gpu in favor of
SequenceChecker.

Note to crash team: This CL is a refactor and has no intended behavior change.

This change was scripted by https://crbug.com/676387#c84.

If you're wondering why there's a second wave of this, the first version of the
script handled "NON_EXPORTED_BASE(public base::NonThreadSafe)" but not
"public NON_EXPORTED_BASE(base::NonThreadSafe)"...

Note-worthy for the reviewer:
 * SequenceChecker enforces thread-safety but not thread-affinity!
   If the classes that were updated are thread-affine (use thread local
   storage or a third-party API that does) they should be migrated to
   ThreadChecker instead.
 * ~NonThreadSafe() used to implicitly check in its destructor
   ~Sequence/ThreadChecker() doesn't by design. To keep this CL a
   no-op, an explicit check was added to the destructor of migrated
   classes.
 * NonThreadSafe used to provide access to subclasses, as such
   the |sequence_checker_| member was made protected rather than
   private where necessary.

BUG=676387
This CL was uploaded by git cl split.

R=kbr@chromium.org
==========

to

==========
Replace deprecated base::NonThreadSafe in /content/renderer/gpu in favor of
ThreadChecker.

Note to crash team: This CL is a refactor and has no intended behavior change.

This change was scripted by https://crbug.com/676387#c84.

(amended to use ThreadChecker upon reviewer request)

If you're wondering why there's a second wave of this, the first version of the
script handled "NON_EXPORTED_BASE(public base::NonThreadSafe)" but not
"public NON_EXPORTED_BASE(base::NonThreadSafe)"...

Note-worthy for the reviewer:
 * SequenceChecker enforces thread-safety but not thread-affinity!
   If the classes that were updated are thread-affine (use thread local
   storage or a third-party API that does) they should be migrated to
   ThreadChecker instead.
 * ~NonThreadSafe() used to implicitly check in its destructor
   ~Sequence/ThreadChecker() doesn't by design. To keep this CL a
   no-op, an explicit check was added to the destructor of migrated
   classes.
 * NonThreadSafe used to provide access to subclasses, as such
   the |sequence_checker_| member was made protected rather than
   private where necessary.

BUG=676387
This CL was uploaded by git cl split.

R=kbr@chromium.org
==========

[gab, 2017-06-02 18:46:02]

The CQ bit was checked by gab@chromium.org to run a CQ dry run

[gab, 2017-06-02 18:46:08]

On 2017/06/02 18:38:13, enne wrote:
> On 2017/06/02 at 18:10:45, kbr wrote:
> > enne, could you review this? It actually looks to me like one of the caveats
> in the CL description applies -- it looks like this class is supposed to have
> thread-affinity.
> 
> I agree that it looks like the CompositorExternalBeginFrameSource is supposed
to
> only be run on the compositor thread.  The CompositorForwardingMessageFilter
has
> a comment that the handler here is supposed to be run on that thread (and the
> filter itself has a ThreadChecker not just a SequenceChecker).

Done :)

[commit-bot: I haz the power, 2017-06-02 18:46:14]

Dry run: CQ is trying da patch.

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

[enne (OOO), 2017-06-02 18:48:23]

lgtm

[gab, 2017-06-02 18:49:08]

The CQ bit was unchecked by gab@chromium.org

[gab, 2017-06-02 18:49:11]

The CQ bit was checked by gab@chromium.org

[commit-bot: I haz the power, 2017-06-02 18:50:05]

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, 2017-06-02 22:09:40]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1496429351709670,
"parent_rev": "420f4637737c8a0232941c7099b8b9bbb42b5fee", "commit_rev":
"b0a378fde6de7bd1e63db9b8249600f4ab4feaf9"}

[commit-bot: I haz the power, 2017-06-02 22:09:50]

Description was changed from

==========
Replace deprecated base::NonThreadSafe in /content/renderer/gpu in favor of
ThreadChecker.

Note to crash team: This CL is a refactor and has no intended behavior change.

This change was scripted by https://crbug.com/676387#c84.

(amended to use ThreadChecker upon reviewer request)

If you're wondering why there's a second wave of this, the first version of the
script handled "NON_EXPORTED_BASE(public base::NonThreadSafe)" but not
"public NON_EXPORTED_BASE(base::NonThreadSafe)"...

Note-worthy for the reviewer:
 * SequenceChecker enforces thread-safety but not thread-affinity!
   If the classes that were updated are thread-affine (use thread local
   storage or a third-party API that does) they should be migrated to
   ThreadChecker instead.
 * ~NonThreadSafe() used to implicitly check in its destructor
   ~Sequence/ThreadChecker() doesn't by design. To keep this CL a
   no-op, an explicit check was added to the destructor of migrated
   classes.
 * NonThreadSafe used to provide access to subclasses, as such
   the |sequence_checker_| member was made protected rather than
   private where necessary.

BUG=676387
This CL was uploaded by git cl split.

R=kbr@chromium.org
==========

to

==========
Replace deprecated base::NonThreadSafe in /content/renderer/gpu in favor of
ThreadChecker.

Note to crash team: This CL is a refactor and has no intended behavior change.

This change was scripted by https://crbug.com/676387#c84.

(amended to use ThreadChecker upon reviewer request)

If you're wondering why there's a second wave of this, the first version of the
script handled "NON_EXPORTED_BASE(public base::NonThreadSafe)" but not
"public NON_EXPORTED_BASE(base::NonThreadSafe)"...

Note-worthy for the reviewer:
 * SequenceChecker enforces thread-safety but not thread-affinity!
   If the classes that were updated are thread-affine (use thread local
   storage or a third-party API that does) they should be migrated to
   ThreadChecker instead.
 * ~NonThreadSafe() used to implicitly check in its destructor
   ~Sequence/ThreadChecker() doesn't by design. To keep this CL a
   no-op, an explicit check was added to the destructor of migrated
   classes.
 * NonThreadSafe used to provide access to subclasses, as such
   the |sequence_checker_| member was made protected rather than
   private where necessary.

BUG=676387
This CL was uploaded by git cl split.

R=kbr@chromium.org

Review-Url: https://codereview.chromium.org/2918083003
Cr-Commit-Position: refs/heads/master@{#476810}
Committed:
https://chromium.googlesource.com/chromium/src/+/b0a378fde6de7bd1e63db9b82496...
==========

[commit-bot: I haz the power, 2017-06-02 22:09:51]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/chromium/src/+/b0a378fde6de7bd1e63db9b82496...