Add resource tracking output and command buffer recycling

Add resource tracking output and command buffer recycling

BUG=skia:5042
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115993002

Committed: https://skia.googlesource.com/skia/+/7ec92413307c9da43c013d1e4e15716a44059810

[jvanverth1, 2016-07-01 18:03:22]

Description was changed from

==========
Add resource tracking output and command buffer recycling

BUG=skia:5042
==========

to

==========
Add resource tracking output and command buffer recycling

BUG=skia:5042
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115993002
==========

[jvanverth1, 2016-07-01 19:08:07]

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

[jvanverth1, 2016-07-01 19:08:08]

https://codereview.chromium.org/2115993002/diff/60001/tools/viewer/sk_app/Vul...
File tools/viewer/sk_app/VulkanWindowContext.cpp (left):

https://codereview.chromium.org/2115993002/diff/60001/tools/viewer/sk_app/Vul...
tools/viewer/sk_app/VulkanWindowContext.cpp:413: fContext->abandonContext();
Removing this solves the leaked resources issue in Viewer.

[egdaniel, 2016-07-05 14:18:37]

lgtm with two comments

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
File src/gpu/vk/GrVkResourceProvider.cpp (right):

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
src/gpu/vk/GrVkResourceProvider.cpp:321: }
Do we want an assert in here to make it really obvious that something bad is
happening or do you think the logging will be enough of a signal? Obviously we'd
spew all the dumps first.

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
File src/gpu/vk/GrVkResourceProvider.h (right):

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
src/gpu/vk/GrVkResourceProvider.h:212: SkSTArray<4, GrVkSecondaryCommandBuffer*>
fAvailableSecondaryCommandBuffers;
I wonder if we should make this reserve higher than 4 here? In general I suppose
this should be equal to the number of secondaryCBs we expect to have inside of a
primary command buffer, aka number of RT switchs + 1. For example, 2 offscreen
draws in a single flush gives us 5 secondary command buffers (main RT, offscreen
1, main RT, offscreen 2, main RT). Maybe 8 would be better or higher? I don't
have a good intuition of what we do per flush. In the end we're just storing
pointers here so size wise array shouldn't be too big?

[jvanverth1, 2016-07-06 15:49:23]

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
File src/gpu/vk/GrVkResourceProvider.cpp (right):

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
src/gpu/vk/GrVkResourceProvider.cpp:321: }
On 2016/07/05 14:18:37, egdaniel wrote:
> Do we want an assert in here to make it really obvious that something bad is
> happening or do you think the logging will be enough of a signal? Obviously
we'd
> spew all the dumps first.

Done.

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
File src/gpu/vk/GrVkResourceProvider.h (right):

https://codereview.chromium.org/2115993002/diff/60001/src/gpu/vk/GrVkResource...
src/gpu/vk/GrVkResourceProvider.h:212: SkSTArray<4, GrVkSecondaryCommandBuffer*>
fAvailableSecondaryCommandBuffers;
On 2016/07/05 14:18:37, egdaniel wrote:
> I wonder if we should make this reserve higher than 4 here? In general I
suppose
> this should be equal to the number of secondaryCBs we expect to have inside of
a
> primary command buffer, aka number of RT switchs + 1. For example, 2 offscreen
> draws in a single flush gives us 5 secondary command buffers (main RT,
offscreen
> 1, main RT, offscreen 2, main RT). Maybe 8 would be better or higher? I don't
> have a good intuition of what we do per flush. In the end we're just storing
> pointers here so size wise array shouldn't be too big?

Done.

[jvanverth1, 2016-07-06 15:49:28]

The CQ bit was checked by jvanverth@google.com

[jvanverth1, 2016-07-06 15:49:29]

The patchset sent to the CQ was uploaded after l-g-t-m from egdaniel@google.com
Link to the patchset: https://codereview.chromium.org/2115993002/#ps80001
(title: "Address comments")

[commit-bot: I haz the power, 2016-07-06 15:49:39]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-06 16:24:59]

Description was changed from

==========
Add resource tracking output and command buffer recycling

BUG=skia:5042
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115993002
==========

to

==========
Add resource tracking output and command buffer recycling

BUG=skia:5042
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115993002

Committed:
https://skia.googlesource.com/skia/+/7ec92413307c9da43c013d1e4e15716a44059810
==========

[commit-bot: I haz the power, 2016-07-06 16:25:00]

Committed patchset #5 (id:80001) as
https://skia.googlesource.com/skia/+/7ec92413307c9da43c013d1e4e15716a44059810