Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(51)

Issue 259533003: WebGL: Fix assertion hit of gpu tests. (Closed)

Created:
6 years, 8 months ago by dshwang
Modified:
6 years, 8 months ago
CC:
blink-reviews, jamesr, krit, dsinclair, jbroman, danakj, Rik, Stephen Chennney, pdr., rwlbuis
Base URL:
https://chromium.googlesource.com/chromium/blink.git@master
Visibility:
Public.

Description

WebGL: Fix assertion hit of gpu tests. This CL removes the invalid assertion in DrawingBuffer::prepareMailbox(); ASSERT(!m_destructionInProgress). It can be hit in the following sequence. 1. WebGL draws something. 2. The compositor begins the frame. 3. Javascript makes a context lost using WEBGL_lose_context extension. 4. Assertion hits. TEST=ContextLost.WebGLContextLostFromGPUProcessExit BUG=365904 Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=172564

Patch Set 1 #

Unified diffs Side-by-side diffs Delta from patch set Stats (+9 lines, -3 lines) Patch
M Source/platform/graphics/gpu/DrawingBuffer.cpp View 1 chunk +9 lines, -3 lines 0 comments Download

Messages

Total messages: 8 (0 generated)
dshwang
Hi, I fix gpu bot failure due to my mistake. Could you review?
6 years, 8 months ago (2014-04-24 19:54:55 UTC) #1
Ken Russell (switch to Gerrit)
Thanks for fixing this. LGTM, but I think there will be remaining failures after this ...
6 years, 8 months ago (2014-04-24 22:05:02 UTC) #2
Ken Russell (switch to Gerrit)
The CQ bit was checked by kbr@chromium.org
6 years, 8 months ago (2014-04-24 22:05:05 UTC) #3
commit-bot: I haz the power
CQ is trying da patch. Follow status at https://chromium-status.appspot.com/cq/dongseong.hwang@intel.com/259533003/1
6 years, 8 months ago (2014-04-24 22:05:56 UTC) #4
Ken Russell (switch to Gerrit)
Thanks again for this patch. I'm mystified by the win_gpu_triggered_tests and mac_gpu_retina_triggered_tests failures in this ...
6 years, 8 months ago (2014-04-24 23:31:35 UTC) #5
commit-bot: I haz the power
Change committed as 172564
6 years, 8 months ago (2014-04-24 23:44:42 UTC) #6
dshwang
On 2014/04/24 23:31:35, Ken Russell wrote: > Thanks again for this patch. > > I'm ...
6 years, 8 months ago (2014-04-25 06:19:40 UTC) #7
dshwang
6 years, 8 months ago (2014-04-25 06:21:46 UTC) #8
Message was sent while issue was closed.
On 2014/04/24 23:31:35, Ken Russell wrote:
> Thanks again for this patch.

Thank you for review.

> I'm mystified by the win_gpu_triggered_tests and
mac_gpu_retina_triggered_tests
> failures in this CL.

I think so.

> The mac_gpu_retina_triggered_tests failure indicates that during one iteration
> of killing the GPU process, the context lost event wasn't delivered to
> JavaScript within the 20 second timeout. I can't reproduce this on my own
> laptop, running exactly the same isolate repeatedly.

Current gpu waterfall shows mac failure is very rare.

> The win_gpu_triggered_tests failure looks like a timeout running the stress
> test. During the ~18th iteration of killing the GPU process, something went
> wrong. I also can't reproduce this locally with a Release build containing
this
> patch.

However, win gpu bot is in serious trouble, now.

The err log;
[4048:3688:0424/133118:ERROR:gpu_process_transport_factory.cc(336)] Lost UI
shared context.
[4048:2940:0424/133120:ERROR:channel.cc(292)] RawChannel fatal error (type 1)
[4048:2940:0424/133120:ERROR:channel.cc(295)] Not shutting down due Windows-only
bug

Powered by Google App Engine
This is Rietveld 408576698