Closed
Bug 1221411
Opened 9 years ago
Closed 3 years ago
Intermittent test_offscreencanvas_many.html | application crashed [@ nsTArray_Impl<mozilla::gl::GLContext::NamedResource, nsTArrayInfallibleAllocator>::RemoveElementsAt(unsigned long, unsigned long)]
Categories
(Core :: Graphics: Canvas2D, defect, P3)
Core
Graphics: Canvas2D
Tracking
()
RESOLVED
INACTIVE
People
(Reporter: philor, Unassigned)
References
Details
(Keywords: assertion, crash, intermittent-failure, Whiteboard: [gfx-noted])
https://treeherder.mozilla.org/logviewer.html#?job_id=5543182&repo=fx-team
Assertion failure: aCount == 0 || aStart < Length() (Invalid aStart index), at ../../dist/include/nsTArray.h:1648
TEST-UNEXPECTED-FAIL | dom/canvas/test/test_offscreencanvas_many.html | application terminated with exit code 1
PROCESS-CRASH | dom/canvas/test/test_offscreencanvas_many.html | application crashed [@ nsTArray_Impl<mozilla::gl::GLContext::NamedResource, nsTArrayInfallibleAllocator>::RemoveElementsAt(unsigned long, unsigned long)]
Comment 1•9 years ago
|
||
From a look at the crashing code, the most likely way this could happen seems to be that the GLContext is accessed from several thread, and the code that tracks textures is called concurrently. Jeff, does sound possible ? how is the offscreen canvas's GLContext used with respect to the different threads at play, is the main thread touching the GLConext as well?
Flags: needinfo?(jgilbert)
Updated•9 years ago
|
status-firefox46:
--- → affected
Keywords: crash
Updated•9 years ago
|
Whiteboard: [gfx-noted]
Comment 3•8 years ago
|
||
This test still appears to have a lot of issues on OSX. Any chance we can up the priority here a bit?
Flags: needinfo?(milan)
Vincent, any idea what could be going on here?
Flags: needinfo?(milan) → needinfo?(vliu)
Updated•8 years ago
|
Updated•8 years ago
|
Updated•8 years ago
|
Comment 5•8 years ago
|
||
A better summary of the current state of the test_offscreencanvas tests:
https://bugzilla.mozilla.org/buglist.cgi?keywords=intermittent-failure%2C%20&keywords_type=allwords&list_id=13051151&short_desc=test_offscreencanvas&resolution=---&query_format=advanced&short_desc_type=allwordssubstr
Comment 6•8 years ago
|
||
Morris, please help to look at it.
Flags: needinfo?(vliu) → needinfo?(mtseng)
Comment 7•8 years ago
|
||
(In reply to Ryan VanderMeulen [:RyanVM] from comment #5)
> A better summary of the current state of the test_offscreencanvas tests:
> https://bugzilla.mozilla.org/buglist.cgi?keywords=intermittent-
> failure%2C%20&keywords_type=allwords&list_id=13051151&short_desc=test_offscre
> encanvas&resolution=---&query_format=advanced&short_desc_type=allwordssubstr
I'd tried to run the following command on both MAC and Ubuntu many times, but I still can not reproduce any issues about offscreencanvas. :(
./mach mochitest --repeat 20 --tag offscreencanvas
Updated•8 years ago
|
Flags: needinfo?(mtseng)
Comment 8•8 years ago
|
||
Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.
Priority: -- → P3
Updated•7 years ago
|
Blocks: offscreen-canvas
Updated•6 years ago
|
Flags: needinfo?(jgilbert)
Comment 9•3 years ago
|
||
I suspect this is no longer a problem because we always out-of-process WebGL contexts for OffscreenCanvas, and in the compositor process, it always runs on the same thread, thus avoiding a lot of concurrency issues. But without the logs (long since discarded by treeherder), I cannot confirm this. Closing as inactive.
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → INACTIVE
You need to log in
before you can comment on or make changes to this bug.
Description
•