Add SkBarriers to ports.

Add SkBarriers to ports.

This completes a TODO we've had to move our memory-barrier code out of
SkOnce.  I also want to start using sk_acquire_load elsewhere.

BUG=skia:

Committed: http://code.google.com/p/skia/source/detail?r=14970

[bungeman-skia, 2014-05-29 18:02:51]

https://codereview.chromium.org/304593003/diff/40001/include/core/SkOnce.h
File include/core/SkOnce.h (left):

https://codereview.chromium.org/304593003/diff/40001/include/core/SkOnce.h#ol...
include/core/SkOnce.h:74: // TODO(bungeman, mtklein): move all these *barrier*
functions to SkThread when refactoring lands.
yay!!

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_arm.h
File src/ports/SkBarriers_arm.h (right):

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_arm...
src/ports/SkBarriers_arm.h:20: template <typename T>
We should doc that T needs to be atomically read-write (that won't tear).

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_x86.h
File src/ports/SkBarriers_x86.h (right):

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_x86...
src/ports/SkBarriers_x86.h:21: sk_compiler_barrier();    // On x86, all loads
are acquire loads.
keep tsan happy!!!

[mtklein, 2014-05-29 18:04:44]

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_arm.h
File src/ports/SkBarriers_arm.h (right):

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_arm...
src/ports/SkBarriers_arm.h:20: template <typename T>
On 2014/05/29 18:02:52, bungeman1 wrote:
> We should doc that T needs to be atomically read-write (that won't tear).

Done.

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_x86.h
File src/ports/SkBarriers_x86.h (right):

https://codereview.chromium.org/304593003/diff/40001/src/ports/SkBarriers_x86...
src/ports/SkBarriers_x86.h:21: sk_compiler_barrier();    // On x86, all loads
are acquire loads.
On 2014/05/29 18:02:52, bungeman1 wrote:
> keep tsan happy!!!

Coming, in another CL...

[bungeman-skia, 2014-05-29 18:04:48]

lgtm

[mtklein, 2014-05-29 18:05:29]

+reed for include/ refactoring

[reed1, 2014-05-29 18:16:06]

looks much smaller, and therefore gooder.

lgtm

the more arch-specific the files are, the more verbose/wonderful the comments
should be...

[mtklein, 2014-05-29 18:17:58]

On 2014/05/29 18:16:06, reed1 wrote:
> looks much smaller, and therefore gooder.
> 
> lgtm
> 
> the more arch-specific the files are, the more verbose/wonderful the comments
> should be...

Ack.  Uploaded some more and will keep updating them if they reveal themselves
to be more subtle.

[mtklein, 2014-05-29 18:18:28]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-05-29 18:19:24]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/304593003/70001

[commit-bot: I haz the power, 2014-05-29 18:24:58]

Change committed as 14970