Move some of the adding stencil attachment logic of Gpu and into Render Target.

Move some of the adding stencil attachment logic of Gpu and into Render Target.

The new flow of calls for attaching a Stencil looks like:

Client
  rt->attachStencilAttachment()
    gpu->getStencilAttachment()
      glgpu->createStencilAttachment()
    glrt->completeStencilAttachment() //actually attaches

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/ec00d94199fad7723b5987b86c1abef8ddafe2d8

[egdaniel, 2015-09-11 17:59:48]

egdaniel@google.com changed reviewers:
+ bsalomon@google.com, jvanverth@google.com

[egdaniel, 2015-09-11 17:59:49]

https://codereview.chromium.org/1333383002/diff/1/src/gpu/GrGpu.h
File src/gpu/GrGpu.h (right):

https://codereview.chromium.org/1333383002/diff/1/src/gpu/GrGpu.h#newcode387
src/gpu/GrGpu.h:387: virtual void clearStencil(GrRenderTarget* target) = 0;
Since we are not letting the RT handle the attaching, the RT also has to be
responsible for calling clear on the stencil. For future work, maybe this clear
should really be implemented in the RT and not the gpu.

https://codereview.chromium.org/1333383002/diff/1/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (left):

https://codereview.chromium.org/1333383002/diff/1/src/gpu/gl/GrGLGpu.cpp#oldc...
src/gpu/gl/GrGLGpu.cpp:1354: #if 0
I removed this workaround code here. Happy to but it back if you want to keep it
in the code.

https://codereview.chromium.org/1333383002/diff/1/src/gpu/gl/GrGLGpu.h
File src/gpu/gl/GrGLGpu.h (right):

https://codereview.chromium.org/1333383002/diff/1/src/gpu/gl/GrGLGpu.h#newcod...
src/gpu/gl/GrGLGpu.h:105: void invalidateBoundRenderTarget() {
Now needed since we're attaching the stencil in the RT so we need to do a RT
bind. This is really screaming for some shared GPU state object separate from
GrGLGpu

[egdaniel, 2015-09-14 14:11:32]

just a friendly ping since this is block some other work

[jvanverth1, 2015-09-14 18:35:07]

Just a couple of comments.

https://codereview.chromium.org/1333383002/diff/20001/src/gpu/GrRenderTarget.cpp
File src/gpu/GrRenderTarget.cpp (right):

https://codereview.chromium.org/1333383002/diff/20001/src/gpu/GrRenderTarget....
src/gpu/GrRenderTarget.cpp:81: fRenderTarget,
&fRenderTarget->fStencilAttachment);
Indent to (, or at least further over.

https://codereview.chromium.org/1333383002/diff/20001/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (right):

https://codereview.chromium.org/1333383002/diff/20001/src/gpu/gl/GrGLGpu.cpp#...
src/gpu/gl/GrGLGpu.cpp:1351: GrGLStencilAttachment* stencil = new
GrGLStencilAttachment(this,
Where does this get deleted?

[egdaniel, 2015-09-14 18:56:49]

https://codereview.chromium.org/1333383002/diff/20001/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (right):

https://codereview.chromium.org/1333383002/diff/20001/src/gpu/gl/GrGLGpu.cpp#...
src/gpu/gl/GrGLGpu.cpp:1351: GrGLStencilAttachment* stencil = new
GrGLStencilAttachment(this,
On 2015/09/14 18:35:07, jvanverth1 wrote:
> Where does this get deleted?

The caller takes ownership of the ref.

[jvanverth1, 2015-09-14 19:15:19]

lgtm

[egdaniel, 2015-09-14 19:24:25]

updated, ptal

[bsalomon, 2015-09-14 19:34:54]

lgtm

[bsalomon, 2015-09-14 19:35:10]

https://codereview.chromium.org/1333383002/diff/1/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (left):

https://codereview.chromium.org/1333383002/diff/1/src/gpu/gl/GrGLGpu.cpp#oldc...
src/gpu/gl/GrGLGpu.cpp:1354: #if 0
On 2015/09/11 17:59:49, egdaniel wrote:
> I removed this workaround code here. Happy to but it back if you want to keep
it
> in the code.

I think it is ok to delete for now.

https://codereview.chromium.org/1333383002/diff/60001/src/gpu/GrRenderTargetP...
File src/gpu/GrRenderTargetPriv.h (right):

https://codereview.chromium.org/1333383002/diff/60001/src/gpu/GrRenderTargetP...
src/gpu/GrRenderTargetPriv.h:25: * currnetly attached GrStencilAttachment will
be removed if one was previously attached. This
currently

[egdaniel, 2015-09-14 19:49:55]

The CQ bit was checked by egdaniel@google.com

[egdaniel, 2015-09-14 19:49:55]

The patchset sent to the CQ was uploaded after l-g-t-m from
jvanverth@google.com, bsalomon@google.com
Link to the patchset: https://codereview.chromium.org/1333383002/#ps80001
(title: "spelling")

[commit-bot: I haz the power, 2015-09-14 19:50:10]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1333383002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1333383002/80001

[commit-bot: I haz the power, 2015-09-14 19:56:13]

Committed patchset #5 (id:80001) as
https://skia.googlesource.com/skia/+/ec00d94199fad7723b5987b86c1abef8ddafe2d8