Restrict ETC1 power-of-two rounding to old IMG drivers.

Restrict ETC1 power-of-two rounding to old IMG drivers.

The ETC1 spec says that any multiple of 4 is allowed for ETC1 size, but
certain old IMG devices (notably the Galaxy Nexus) crash unless the
textures are specifically power-of-two. To avoid wasting RAM on all
devices due to this device-specific problem, this patch plumbs a driver
workaround flag to the thumbnail store (the only user of ETC1 at
present).

We considered adding a workaround to the service side like all of the
others, but because texture size impacts memory budgeting and
coordinates, this one seems like it needs to be in the client code.

NOTRY=true
BUG=150500
Committed: https://crrev.com/d880a3deca8ab1c6c75ef915faa22ab7d40f55ff
Cr-Commit-Position: refs/heads/master@{#291814}

[aelias_OOO_until_Jul13, 2014-08-21 22:11:21]

PTAL.

sievers@ for general review
zmo@ for gpu/
tedchoc@ for content/public/browser/android

[David Trainor- moved to gerrit, 2014-08-21 22:15:01]

https://codereview.chromium.org/470233003/diff/20001/chrome/browser/android/t...
File chrome/browser/android/thumbnail/thumbnail_store.cc (right):

https://codereview.chromium.org/470233003/diff/20001/chrome/browser/android/t...
chrome/browser/android/thumbnail/thumbnail_store.cc:51: size_t
RoundUpMod4(size_t x) {
(x + 3) & ~3 (is what most ETC1 libs use).

[aelias_OOO_until_Jul13, 2014-08-21 22:56:12]

Adding tsepez@ for OWNERS review of content/common/gpu/gpu_messages.h

https://codereview.chromium.org/470233003/diff/20001/chrome/browser/android/t...
File chrome/browser/android/thumbnail/thumbnail_store.cc (right):

https://codereview.chromium.org/470233003/diff/20001/chrome/browser/android/t...
chrome/browser/android/thumbnail/thumbnail_store.cc:51: size_t
RoundUpMod4(size_t x) {
On 2014/08/21 22:15:00, David Trainor wrote:
> (x + 3) & ~3 (is what most ETC1 libs use).

Done.

[Tom Sepez, 2014-08-21 23:00:56]

Messages LGTM.

[no sievers, 2014-08-21 23:02:50]

https://codereview.chromium.org/470233003/diff/20001/content/browser/renderer...
File content/browser/renderer_host/compositor_impl_android.cc (right):

https://codereview.chromium.org/470233003/diff/20001/content/browser/renderer...
content/browser/renderer_host/compositor_impl_android.cc:546: context_provider_
= ContextProviderCommandBuffer::Create(
Is there a way we can pass this through without having to hold on to a context
reference here?
I'm worried it might cause some issues with how things get torn down.
Maybe through LayerTreeSettings?
Or the OutputSurfaceWithoutParent (above) could maybe have a backpointer to
CompositorImpl and pass back the capabilities from BindToClient().

https://codereview.chromium.org/470233003/diff/20001/gpu/config/gpu_driver_bu...
File gpu/config/gpu_driver_bug_list_json.cc (right):

https://codereview.chromium.org/470233003/diff/20001/gpu/config/gpu_driver_bu...
gpu/config/gpu_driver_bug_list_json.cc:22: "version": "7.2",
can you bump this to 7.3?

https://codereview.chromium.org/470233003/diff/20001/gpu/config/gpu_driver_bu...
gpu/config/gpu_driver_bug_list_json.cc:1037: "gl_vendor": {
new syntax :)

"gl_vendor": "Imagination.*",
"gl_renderer": "PowerVR SGX 540.*",

[Ted C, 2014-08-22 17:51:02]

content/public/browser/android - LGTM

[aelias_OOO_until_Jul13, 2014-08-23 01:28:50]

https://codereview.chromium.org/470233003/diff/20001/content/browser/renderer...
File content/browser/renderer_host/compositor_impl_android.cc (right):

https://codereview.chromium.org/470233003/diff/20001/content/browser/renderer...
content/browser/renderer_host/compositor_impl_android.cc:546: context_provider_
= ContextProviderCommandBuffer::Create(
On 2014/08/21 23:02:50, sievers wrote:
> Is there a way we can pass this through without having to hold on to a context
> reference here?
> I'm worried it might cause some issues with how things get torn down.
> Maybe through LayerTreeSettings?
> Or the OutputSurfaceWithoutParent (above) could maybe have a backpointer to
> CompositorImpl and pass back the capabilities from BindToClient().

Done, I took the OutputSurfaceWithoutParent approach.

https://codereview.chromium.org/470233003/diff/20001/gpu/config/gpu_driver_bu...
File gpu/config/gpu_driver_bug_list_json.cc (right):

https://codereview.chromium.org/470233003/diff/20001/gpu/config/gpu_driver_bu...
gpu/config/gpu_driver_bug_list_json.cc:22: "version": "7.2",
On 2014/08/21 23:02:50, sievers wrote:
> can you bump this to 7.3?

Done.

https://codereview.chromium.org/470233003/diff/20001/gpu/config/gpu_driver_bu...
gpu/config/gpu_driver_bug_list_json.cc:1037: "gl_vendor": {
On 2014/08/21 23:02:50, sievers wrote:
> new syntax :)
> 
> "gl_vendor": "Imagination.*",
> "gl_renderer": "PowerVR SGX 540.*",

Done.

[no sievers, 2014-08-25 17:24:50]

lgtm

https://codereview.chromium.org/470233003/diff/80001/content/browser/renderer...
File content/browser/renderer_host/compositor_impl_android.cc (right):

https://codereview.chromium.org/470233003/diff/80001/content/browser/renderer...
content/browser/renderer_host/compositor_impl_android.cc:65: main_thread_ =
base::MessageLoop::current();
use MessageLoopProxy instead of MessagLoop, and then you can hold a
scoped_refptr<> to it

[aelias_OOO_until_Jul13, 2014-08-25 19:10:58]

Done.  Now just need OWNERS review from zmo@

[Zhenyao Mo, 2014-08-25 20:14:26]

LGTM

[aelias_OOO_until_Jul13, 2014-08-25 23:33:51]

The CQ bit was checked by aelias@chromium.org

[commit-bot: I haz the power, 2014-08-25 23:35:06]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/aelias@chromium.org/470233003/160001

[commit-bot: I haz the power, 2014-08-26 00:09:03]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-08-26 00:09:04]

Failed to commit the patch.

[aelias_OOO_until_Jul13, 2014-08-26 00:41:43]

The CQ bit was checked by aelias@chromium.org

[commit-bot: I haz the power, 2014-08-26 00:43:40]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/aelias@chromium.org/470233003/180001

[commit-bot: I haz the power, 2014-08-26 00:46:32]

Committed patchset #10 (180001) as 791d12057856b62b1d6a17ead881fdc94775ee23

[commit-bot: I haz the power, 2014-09-10 02:39:29]

Patchset 10 (id:??) landed as
https://crrev.com/d880a3deca8ab1c6c75ef915faa22ab7d40f55ff
Cr-Commit-Position: refs/heads/master@{#291814}