Create xfer processor backend.

Create xfer processor backend.

This includes:
-Having an actual XP stage at the end of the gl pipeline.
-All Blending work is handled by XP until actually setting GL blend states
-GLPrograms test to test XP

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/4dffc940c430eec66d4707490eace19c9b3f7904

Committed: https://skia.googlesource.com/skia/+/c230414861558ce20b74281b1ca363e56c441832

[egdaniel, 2014-11-26 20:42:54]

egdaniel@google.com changed reviewers:
+ bsalomon@google.com, joshualitt@chromium.org

[egdaniel, 2014-11-26 20:42:54]

This is part 3 and final part (well until all the million follow up cls) of the
XP stuff.

[egdaniel, 2014-12-08 15:33:05]

update to the backend/code emitting part of XPs. Still requires the blending cl
(759713002) to land first.

[joshualitt, 2014-12-08 19:46:40]

joshualitt@google.com changed reviewers:
+ joshualitt@google.com

[joshualitt, 2014-12-08 19:46:41]

mostly nits

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrProcessorU...
File include/gpu/GrProcessorUnitTest.h (right):

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrProcessorU...
include/gpu/GrProcessorUnitTest.h:136: GrTexture* dummyTextures[2])
alignment

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
File src/gpu/effects/GrPorterDuffXferProcessor.cpp (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
src/gpu/effects/GrPorterDuffXferProcessor.cpp:82: 
I'm not saying you HAVE to delete this space, but is this intentional?

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
src/gpu/effects/GrPorterDuffXferProcessor.cpp:130: colorPOI.readsDst() ||
coveragePOI.readsDst());
I thought only xps read dst?

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
src/gpu/effects/GrPorterDuffXferProcessor.cpp:141: // TODO move the gp logic
into the GP base class
remove todo

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLProcessor.h
File src/gpu/gl/GrGLProcessor.h (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLProcessor...
src/gpu/gl/GrGLProcessor.h:129: */
is this supposed to be commented out?

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/builders/GrGL...
File src/gpu/gl/builders/GrGLProgramBuilder.h (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/builders/GrGL...
src/gpu/gl/builders/GrGLProgramBuilder.h:386: 
extra newline

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/builders/GrGL...
src/gpu/gl/builders/GrGLProgramBuilder.h:422: 
extra newline

[bsalomon, 2014-12-08 19:58:59]

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrXferProces...
File include/gpu/GrXferProcessor.h (right):

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrXferProces...
include/gpu/GrXferProcessor.h:36: /** Implemented using
GLFragmentProcessor::GenKey as described in this class's comment. */
Is this comment correct?

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrXferProces...
include/gpu/GrXferProcessor.h:103: virtual bool hasSecondaryOutput() const {
return false; }
Document?

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
File src/gpu/effects/GrPorterDuffXferProcessor.cpp (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
src/gpu/effects/GrPorterDuffXferProcessor.cpp:130: colorPOI.readsDst() ||
coveragePOI.readsDst());
On 2014/12/08 19:46:41, joshualitt wrote:
> I thought only xps read dst?

AFACT that has to wait until XPs are written for the special blend modes.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLProcessor.h
File src/gpu/gl/GrGLProcessor.h (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLProcessor...
src/gpu/gl/GrGLProcessor.h:129: */
On 2014/12/08 19:46:41, joshualitt wrote:
> is this supposed to be commented out?

seems like it can be deleted.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLXferProce...
File src/gpu/gl/GrGLXferProcessor.h (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLXferProce...
src/gpu/gl/GrGLXferProcessor.h:45: /**
Comment seems like it was lifted from somewhere and doesn't apply here

[egdaniel, 2014-12-09 21:10:46]

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrProcessorU...
File include/gpu/GrProcessorUnitTest.h (right):

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrProcessorU...
include/gpu/GrProcessorUnitTest.h:136: GrTexture* dummyTextures[2])
On 2014/12/08 19:46:40, joshualitt wrote:
> alignment

Done.

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrXferProces...
File include/gpu/GrXferProcessor.h (right):

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrXferProces...
include/gpu/GrXferProcessor.h:36: /** Implemented using
GLFragmentProcessor::GenKey as described in this class's comment. */
On 2014/12/08 19:58:59, bsalomon wrote:
> Is this comment correct?

Nope no longer, updated.

https://codereview.chromium.org/764643004/diff/20001/include/gpu/GrXferProces...
include/gpu/GrXferProcessor.h:103: virtual bool hasSecondaryOutput() const {
return false; }
On 2014/12/08 19:58:58, bsalomon wrote:
> Document?

Done.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
File src/gpu/effects/GrPorterDuffXferProcessor.cpp (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
src/gpu/effects/GrPorterDuffXferProcessor.cpp:82: 
On 2014/12/08 19:46:40, joshualitt wrote:
> I'm not saying you HAVE to delete this space, but is this intentional?

Done.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
src/gpu/effects/GrPorterDuffXferProcessor.cpp:130: colorPOI.readsDst() ||
coveragePOI.readsDst());
On 2014/12/08 19:58:59, bsalomon wrote:
> On 2014/12/08 19:46:41, joshualitt wrote:
> > I thought only xps read dst?
> 
> AFACT that has to wait until XPs are written for the special blend modes.

Yes this is correct.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/effects/GrPorter...
src/gpu/effects/GrPorterDuffXferProcessor.cpp:141: // TODO move the gp logic
into the GP base class
On 2014/12/08 19:46:41, joshualitt wrote:
> remove todo

Done.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLProcessor.h
File src/gpu/gl/GrGLProcessor.h (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/GrGLProcessor...
src/gpu/gl/GrGLProcessor.h:129: */
On 2014/12/08 19:58:59, bsalomon wrote:
> On 2014/12/08 19:46:41, joshualitt wrote:
> > is this supposed to be commented out?
> 
> seems like it can be deleted.

Done.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/builders/GrGL...
File src/gpu/gl/builders/GrGLProgramBuilder.h (right):

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/builders/GrGL...
src/gpu/gl/builders/GrGLProgramBuilder.h:386: 
On 2014/12/08 19:46:41, joshualitt wrote:
> extra newline

Done.

https://codereview.chromium.org/764643004/diff/20001/src/gpu/gl/builders/GrGL...
src/gpu/gl/builders/GrGLProgramBuilder.h:422: 
On 2014/12/08 19:46:41, joshualitt wrote:
> extra newline

Done.

[bsalomon, 2014-12-09 21:15:39]

lgtm

[egdaniel, 2014-12-10 15:34:05]

The CQ bit was checked by egdaniel@google.com

[commit-bot: I haz the power, 2014-12-10 15:34:13]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/764643004/100001

[commit-bot: I haz the power, 2014-12-10 15:43:57]

Committed patchset #6 (id:100001) as
https://skia.googlesource.com/skia/+/4dffc940c430eec66d4707490eace19c9b3f7904

[joshualitt, 2014-12-10 22:57:54]

A revert of this CL (patchset #6 id:100001) has been created in
https://codereview.chromium.org/789343002/ by joshualitt@google.com.

The reason for reverting is: CL breaks windows DM.

[egdaniel, 2014-12-11 18:43:03]

So here is the fix to the win8 problems. I've also merged my previous "fix" cls
(more gm ignores and readsDst primaryOutput) all into this CL which are in patch
7, and the win8 fixes are in patch8.

[bsalomon, 2014-12-11 18:52:37]

lgtm

[egdaniel, 2014-12-11 21:06:09]

The CQ bit was checked by egdaniel@google.com

[commit-bot: I haz the power, 2014-12-11 21:06:34]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/764643004/160001