Add GrProgramElement base class for GrEffect with deferred exec ref.

Add GrProgramElement base class for GrEffect with deferred exec ref.

BUG=skia:2889
Committed: https://skia.googlesource.com/skia/+/95740981c36266e4595ddde2264aa38e3c7e2d02

[bsalomon, 2014-09-04 15:41:24]

bsalomon@google.com changed reviewers:
+ robertphillips@google.com

[bsalomon, 2014-09-04 15:41:25]

Also adds auto-unref/smartptr helper classes that allow converting normal refs
to deferred reads/writes (for resources) or deferred executions (for program
elements).

[robertphillips, 2014-09-04 17:39:13]

first pass

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
File include/gpu/GrProgramElement.h (right):

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:20: *
All GrGpuResource objects owned by a GrProgramResource ?

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:21: * All GrGpuResource objects owned by a
GrProgramResource or derived classes (either directly or
owned via GrProgramResource ?

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:46: --fRefCnt;
Extra space here ?

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:71: private:
convertRefToPendeingExecution - extra 'e' ?

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:82: mutable int32_t fRefCnt;
// comment ?

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramRes...
File include/gpu/GrProgramResource.h (right):

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramRes...
include/gpu/GrProgramResource.h:15: /**
Should these be over a space ?

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramRes...
include/gpu/GrProgramResource.h:57: /** Called when the program element/draw
state is no longer owned by GrDrawTarget-client code.
Should these be over a space ?

https://codereview.chromium.org/537773004/diff/60001/src/gpu/GrProgramElement...
File src/gpu/GrProgramElementRef.h (right):

https://codereview.chromium.org/537773004/diff/60001/src/gpu/GrProgramElement...
src/gpu/GrProgramElementRef.h:30: SkASSERT(!fOwnPendingExec);
SkRefCnt_SafeAssign ?

[bsalomon, 2014-09-04 18:09:42]

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
File include/gpu/GrProgramElement.h (right):

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:20: *
On 2014/09/04 17:39:12, robertphillips wrote:
> All GrGpuResource objects owned by a GrProgramResource ?

changed to ".. owned by a GrProgramElement"

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:21: * All GrGpuResource objects owned by a
GrProgramResource or derived classes (either directly or
On 2014/09/04 17:39:12, robertphillips wrote:
> owned via GrProgramResource ?

changed to "... wrapped in a GrProgramResource"

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:46: --fRefCnt;
On 2014/09/04 17:39:12, robertphillips wrote:
> Extra space here ?

Done.

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:71: private:
On 2014/09/04 17:39:12, robertphillips wrote:
> convertRefToPendeingExecution - extra 'e' ?

Done.

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramEle...
include/gpu/GrProgramElement.h:82: mutable int32_t fRefCnt;
On 2014/09/04 17:39:12, robertphillips wrote:
> // comment ?

Done.

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramRes...
File include/gpu/GrProgramResource.h (right):

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramRes...
include/gpu/GrProgramResource.h:15: /**
On 2014/09/04 17:39:12, robertphillips wrote:
> Should these be over a space ?

Done.

https://codereview.chromium.org/537773004/diff/60001/include/gpu/GrProgramRes...
include/gpu/GrProgramResource.h:57: /** Called when the program element/draw
state is no longer owned by GrDrawTarget-client code.
On 2014/09/04 17:39:12, robertphillips wrote:
> Should these be over a space ?

Done.

https://codereview.chromium.org/537773004/diff/60001/src/gpu/GrProgramElement...
File src/gpu/GrProgramElementRef.h (right):

https://codereview.chromium.org/537773004/diff/60001/src/gpu/GrProgramElement...
src/gpu/GrProgramElementRef.h:30: SkASSERT(!fOwnPendingExec);
On 2014/09/04 17:39:12, robertphillips wrote:
> SkRefCnt_SafeAssign ?

SkRefCnt_SafeAssign takes a ref, we're adopting a ref.

[robertphillips, 2014-09-04 19:48:57]

lgtm + suggestions

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
File include/gpu/GrProgramElement.h (right):

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
include/gpu/GrProgramElement.h:16: /**
and GrGeometryProcessor ?

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
include/gpu/GrProgramElement.h:66: */
Why rr ? pr ?

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
File include/gpu/GrProgramElementRef.h (right):

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
include/gpu/GrProgramElementRef.h:14: /**
GrProgramElement -> GrProgramElement-derived class ?

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramRe...
File include/gpu/GrProgramResource.h (right):

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramRe...
include/gpu/GrProgramResource.h:57: /** Called when the program element/draw
state is no longer owned by GrDrawTarget-client code.
code _will_ no longer ?

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramElemen...
File src/gpu/GrProgramElement.cpp (right):

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramElemen...
src/gpu/GrProgramElement.cpp:14: // underlying GrGpuResource if there are any
refs to the GrProgramElement and a single
single pending read ?

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramElemen...
src/gpu/GrProgramElement.cpp:15: // pending read if there are any pending
executions of the GrProgramElement. The
pending read ?

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramResour...
File src/gpu/GrProgramResource.cpp (right):

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramResour...
src/gpu/GrProgramResource.cpp:65: fOwnRef = false;
fIOType = kNone_IOType; ?

[bsalomon, 2014-09-04 20:03:17]

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
File include/gpu/GrProgramElement.h (right):

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
include/gpu/GrProgramElement.h:16: /**
On 2014/09/04 19:48:57, robertphillips wrote:
> and GrGeometryProcessor ?

Done.

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
include/gpu/GrProgramElement.h:66: */
On 2014/09/04 19:48:57, robertphillips wrote:
> Why rr ? pr ?

oops... was called a "readresource" in a prev iteration. Done.

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
File include/gpu/GrProgramElementRef.h (right):

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramEl...
include/gpu/GrProgramElementRef.h:14: /**
On 2014/09/04 19:48:57, robertphillips wrote:
> GrProgramElement -> GrProgramElement-derived class ?

Done.

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramRe...
File include/gpu/GrProgramResource.h (right):

https://codereview.chromium.org/537773004/diff/140001/include/gpu/GrProgramRe...
include/gpu/GrProgramResource.h:57: /** Called when the program element/draw
state is no longer owned by GrDrawTarget-client code.
On 2014/09/04 19:48:57, robertphillips wrote:
> code _will_ no longer ?

Done.

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramElemen...
File src/gpu/GrProgramElement.cpp (right):

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramElemen...
src/gpu/GrProgramElement.cpp:14: // underlying GrGpuResource if there are any
refs to the GrProgramElement and a single
On 2014/09/04 19:48:57, robertphillips wrote:
> single pending read ?

changed to read/write

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramElemen...
src/gpu/GrProgramElement.cpp:15: // pending read if there are any pending
executions of the GrProgramElement. The
On 2014/09/04 19:48:57, robertphillips wrote:
> pending read ?
changed to read/write

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramResour...
File src/gpu/GrProgramResource.cpp (right):

https://codereview.chromium.org/537773004/diff/140001/src/gpu/GrProgramResour...
src/gpu/GrProgramResource.cpp:65: fOwnRef = false;
On 2014/09/04 19:48:57, robertphillips wrote:
> fIOType = kNone_IOType; ?

Done (here and in reset()).

[bsalomon, 2014-09-04 20:03:20]

The CQ bit was checked by bsalomon@google.com

[commit-bot: I haz the power, 2014-09-04 20:03:55]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/bsalomon@google.com/537773004/150001

[commit-bot: I haz the power, 2014-09-04 20:12:40]

Committed patchset #9 (id:150001) as 95740981c36266e4595ddde2264aa38e3c7e2d02