Windows arm64 startup crash in [@ webrender::space::SpaceSnapper::set_target_spatial_node]
Categories
(Core :: Graphics: WebRender, defect, P3)
Tracking
()
People
(Reporter: aryx, Unassigned)
References
(Blocks 1 open bug)
Details
(Keywords: crash)
Crash Data
First observed with 83.0a1 20201007155036. At least 2 devices report this. Crash is always EXCEPTION_ACCESS_VIOLATION_WRITE at 0x7ffa32f88fa8.
Crash report: https://crash-stats.mozilla.org/report/index/c7913416-b38e-4d4c-9e02-0c6b80201008
Reason: EXCEPTION_ACCESS_VIOLATION_WRITE
Top 8 frames of crashing thread:
0 xul.dll webrender::space::SpaceSnapper::set_target_spatial_node gfx/wr/webrender/src/space.rs:221
1 xul.dll webrender::scene_building::SceneBuilder::build_item gfx/wr/webrender/src/scene_building.rs:1094
2 xul.dll webrender::scene_building::SceneBuilder::build_item gfx/wr/webrender/src/scene_building.rs:1094
3 xul.dll webrender::scene_building::SceneBuilder::build_all gfx/wr/webrender/src/scene_building.rs:576
4 xul.dll tan
5 xul.dll tan
6 xul.dll tan
7 xul.dll tan
| Reporter | ||
Updated•5 years ago
|
Updated•5 years ago
|
Comment 1•5 years ago
|
||
:jmuizelaer: Not sure, but thought you were looking at arm64 support?
Comment 2•5 years ago
|
||
I don't have ARM64 hardware locally. dmajor do you?
I do. Just going by the minidump, this looks bad. set_target_spatial_node thinks self is some readonly area in libxul. It's possible the pointer was corrupted earlier up the stack though, so it would be super helpful if I could reproduce this live. Do I need to set any prefs and/or does the stack give you any idea as to what I could do to hit this codepath?
Comment 4•5 years ago
|
||
You could try setting gfx.webrender.all=true. WebRender isn't on by default on those machines.
(In reply to Jeff Muizelaar [:jrmuizel] from comment #4)
You could try setting gfx.webrender.all=true. WebRender isn't on by default on those machines.
I set the pref, restarted under a debugger, set a breakpoint on xul!webrender::space::SpaceSnapper::set_target_spatial_node in every process, but it never triggered.
Comment 6•5 years ago
|
||
Tania, would it be possible to you to find someone who can try to reproduce it?
(note: the trap is "Windows on ARM64")
Thanks!
Comment 7•5 years ago
|
||
Hello,
I have tried to reproduce this crash using the latest version of Firefox Nightly 83.0a1 (2020-10-15) (20201015094006) but was unable to have a crash.
Performed tests with the "gfx.webrender.all=true" or false.
Windows was started and Firefox launched but no crash (in some cases navigated to random sites as well); are there any extra steps I need to take into consideration?
I also tested with the 20201007155036 build as well but didn't get any crashes there either.
Please NI me if further assistance is needed.
Comment 8•5 years ago
|
||
Thanks Andrei! Just to remove some potential doubt, you confirm that you tested using Windows on ARM64, right?
Comment 9•5 years ago
|
||
Yes Sylvestre,
I have a Lenovo Yoga with ARM64.
Comment 10•5 years ago
|
||
This hasn't been seen in a week, probably due to bug 1644624.
Description
•