teach TSAN about SkSpinlock, SkRefCnt, and SkOnce

teach TSAN about SkSpinlock, SkRefCnt, and SkOnce

BUG=skia:

Committed: http://code.google.com/p/skia/source/detail?r=14353

NOTRY=true
NOTREECHECKS=true

Committed: http://code.google.com/p/skia/source/detail?r=14354

Committed: http://code.google.com/p/skia/source/detail?r=14390

[bungeman-skia, 2014-04-23 22:37:53]

lgtm

https://codereview.chromium.org/247813005/diff/20001/include/core/SkRefCnt.h
File include/core/SkRefCnt.h (right):

https://codereview.chromium.org/247813005/diff/20001/include/core/SkRefCnt.h#...
include/core/SkRefCnt.h:59: //TODO: issue the barrier.
We do need to actually issue a barrier here on ARM. Should do that at some
point.

[mtklein, 2014-04-23 23:41:47]

https://codereview.chromium.org/247813005/diff/20001/include/core/SkRefCnt.h
File include/core/SkRefCnt.h (right):

https://codereview.chromium.org/247813005/diff/20001/include/core/SkRefCnt.h#...
include/core/SkRefCnt.h:59: //TODO: issue the barrier.
On 2014/04/23 22:37:54, bungeman1 wrote:
> We do need to actually issue a barrier here on ARM. Should do that at some
> point.

Ooh, yes, or as you noted, provide a way to load fRefCnt with that barrier.

[bsalomon, 2014-04-24 13:15:26]

lgtm FWIW (which is not much).

[mtklein, 2014-04-24 13:50:58]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-04-24 13:51:13]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/247813005/20001

[bungeman-skia, 2014-04-24 14:06:56]

https://codereview.chromium.org/247813005/diff/20001/include/core/SkRefCnt.h
File include/core/SkRefCnt.h (right):

https://codereview.chromium.org/247813005/diff/20001/include/core/SkRefCnt.h#...
include/core/SkRefCnt.h:59: //TODO: issue the barrier.
On 2014/04/23 23:41:47, mtklein wrote:
> On 2014/04/23 22:37:54, bungeman1 wrote:
> > We do need to actually issue a barrier here on ARM. Should do that at some
> > point.
> 
> Ooh, yes, or as you noted, provide a way to load fRefCnt with that barrier.

Eh, there's no real reason to provide any barrier to reading fRefCnt. If it
isn't 1 then you can't do anything anyway (and it may already have changed); if
you were testing against 1, then use unique.

[commit-bot: I haz the power, 2014-04-24 15:33:56]

Change committed as 14353

[mtklein, 2014-04-24 15:40:24]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-04-24 15:40:40]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/247813005/40001

[commit-bot: I haz the power, 2014-04-24 15:40:48]

Change committed as 14354

[mtklein, 2014-04-27 19:14:08]

A revert of this CL has been created in
https://codereview.chromium.org/258783005/ by mtklein@google.com.

The reason for reverting is: breaks blink in roll.

[mtklein, 2014-04-27 19:21:17]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-04-27 19:21:45]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/247813005/60001

[commit-bot: I haz the power, 2014-04-27 19:21:57]

Change committed as 14390