Prefer to recycle non-RT scratch textures that don't have pending IO

Prefer to recycle non-RT scratch textures that don't have pending IO

BUG=skia:2889
Committed: https://skia.googlesource.com/skia/+/000f829f14a9535a005082731af5de1526284c83

[bsalomon, 2014-10-13 18:50:44]

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

[bsalomon, 2014-10-15 16:26:35]

ping

[bsalomon, 2014-10-15 19:12:55]

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

[bsalomon, 2014-10-15 19:12:55]

Ugh, I sent this to robertphillips@google..com (two dots)

[robertphillips, 2014-10-15 20:06:13]

First pass - it looks good but I would like to make a second pass.

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrContext.cpp
File src/gpu/GrContext.cpp (right):

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrContext.cpp#ne...
src/gpu/GrContext.cpp:477: } else  if (!(desc->fFlags &
kRenderTarget_GrTextureFlagBit)) {
Why not just always pass Prefer if !calledDuringFlush ?

likley ->likely

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
File src/gpu/GrResourceCache2.cpp (right):

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
src/gpu/GrResourceCache2.cpp:75: 
add line break ?

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
src/gpu/GrResourceCache2.cpp:85: }
wihout -> without ?

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2.h
File src/gpu/GrResourceCache2.h (right):

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
src/gpu/GrResourceCache2.h:37: kPreferNoPendingIO_ScratchFlag = 0x1,
has -> have ?

[robertphillips, 2014-10-15 20:35:06]

lgtm - given in person conversation

[bsalomon, 2014-10-16 01:51:14]

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrContext.cpp
File src/gpu/GrContext.cpp (right):

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrContext.cpp#ne...
src/gpu/GrContext.cpp:477: } else  if (!(desc->fFlags &
kRenderTarget_GrTextureFlagBit)) {
On 2014/10/15 20:06:13, robertphillips wrote:
> Why not just always pass Prefer if !calledDuringFlush ?

I think it's better to keep reusing the same texture in the RT case rather than
keeping additional identical scratches alive in the cache.
> 
> likley ->likely

Done.

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
File src/gpu/GrResourceCache2.cpp (right):

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
src/gpu/GrResourceCache2.cpp:75: 
On 2014/10/15 20:06:13, robertphillips wrote:
> add line break ?

Done.

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
src/gpu/GrResourceCache2.cpp:85: }
On 2014/10/15 20:06:13, robertphillips wrote:
> wihout -> without ?

Done.

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2.h
File src/gpu/GrResourceCache2.h (right):

https://codereview.chromium.org/650283002/diff/70001/src/gpu/GrResourceCache2...
src/gpu/GrResourceCache2.h:37: kPreferNoPendingIO_ScratchFlag = 0x1,
On 2014/10/15 20:06:13, robertphillips wrote:
> has -> have ?

Done.

[bsalomon, 2014-10-16 01:51:20]

The CQ bit was checked by bsalomon@google.com

[commit-bot: I haz the power, 2014-10-16 01:58:07]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/650283002/110001

[commit-bot: I haz the power, 2014-10-16 02:04:18]

Committed patchset #4 (id:110001) as 000f829f14a9535a005082731af5de1526284c83