Crash in [@ nsIFrame::HasAnyStateBits]
Categories
(Core :: Layout, defect)
Tracking
()
People
(Reporter: aryx, Unassigned)
References
Details
(Keywords: crash)
Crash Data
This crash signature preexisted but it got more frequent with Firefox 118: already 181 crash reports from 177 installations compared to 96 from 90 for the whole v117 release cycle.
Crash report: https://crash-stats.mozilla.org/report/index/c5eb3222-7c10-4cdc-9605-526660231005
Reason: EXCEPTION_ACCESS_VIOLATION_READ
Top 10 frames of crashing thread:
0 xul.dll nsIFrame::HasAnyStateBits const layout/generic/nsIFrame.h:2311
0 xul.dll nsStyleTransformMatrix::TransformReferenceBox::EnsureDimensionsAreCached layout/style/nsStyleTransformMatrix.cpp:113
1 xul.dll nsStyleTransformMatrix::TransformReferenceBox::Height layout/style/nsStyleTransformMatrix.h:112
1 xul.dll nsStyleTransformMatrix::Convert2DPosition::<lambda_3>::operator const layout/style/nsStyleTransformMatrix.cpp:626
1 xul.dll mozilla::StyleLengthPercentageUnion::ResolveToCSSPixelsWith const layout/style/ServoStyleConstsInlines.h:707
1 xul.dll nsStyleTransformMatrix::Convert2DPosition layout/style/nsStyleTransformMatrix.cpp:622
2 xul.dll mozilla::MotionPathUtils::ResolveMotionPath layout/base/MotionPathUtils.cpp:524
3 xul.dll mozilla::nsDisplayTransform::FrameTransformProperties::FrameTransformProperties layout/painting/nsDisplayList.cpp:6176
3 xul.dll mozilla::nsDisplayTransform::GetResultingTransformMatrix layout/painting/nsDisplayList.cpp:6193
4 xul.dll nsIFrame::GetTransformMatrix const layout/generic/nsIFrame.cpp:7335
Comment 1•2 years ago
|
||
The severity field is not set for this bug.
:tlouw, could you have a look please?
For more information, please visit BugBot documentation.
Comment 2•2 years ago
|
||
Boris do you think this might be related to work you've been doing?
Comment 3•2 years ago
•
|
||
The crash report mention we are trying to access a null pointer when calling nsIFrame::HasAnyStateBits. However, the frame pointer should be valid when calling nsDisplayTransform::GetResultingTransformMatrix(this, ...). Not sure what happened.
Also, I didn't update here in 117 and 118., so probably this is regressed from others. I'd keep monitoring this bug.
Updated•2 years ago
|
Comment 5•2 years ago
|
||
About 1/3 of the crashes here (538/1591) over the last 3 months have some value for "Possible Bit Flips Max Confidence".
There are 538 crashes where that field exists out of 1,591 total crashes. Perusing through some of those, I'm seeing crash addresses that end with
So, I think it's reasonable to assume that hardware issues are responsible for at least a portion of the crash volume here.
Comment 6•2 years ago
|
||
That pattern continues; 148/339 (44%) of crashes in the past month have some nonzero value for "Possible Bit Flips Max Confidence".
Comment 7•2 years ago
|
||
Bug 1906768 has a fuzz testcase that crashes with this same signature.
Comment 8•2 years ago
|
||
Thanks, good find!
That one is a recent regression, though, so it's not likely to be related to the historical crash volume here. (Unfortunately HasStateBits() isn't an especially specific crash signature; a crash with that signature just means some code is dealing with a bogus nsIFrame pointer, essentially.)
Still: good call to link them up with see-also, given the common signature.
| Comment hidden (Intermittent Failures Robot) |
Description
•