770582 - crash in mozilla::layers::ContainerLayerProperties::ComputeChangeInternal (64bit Windows/Linux)

Status: RESOLVED WORKSFORME

(Core :: Graphics: Layers, defect)

Description

[Robert Kaiser]

This bug was filed from the Socorro interface and is 
report bp-cf7b6d51-0620-49b6-b04c-b3fac2120703 .
============================================================= 

Top frames:

0 	xul.dll 	mozilla::layers::ContainerLayerProperties::ComputeChangeInternal 	gfx/layers/LayerTreeInvalidation.cpp:174
1 	xul.dll 	mozilla::layers::LayerPropertiesBase::OldTransformedBounds 	gfx/layers/LayerTreeInvalidation.cpp:144
2 	xul.dll 	nsXMLHttpRequest::QueryInterface 	content/base/src/nsXMLHttpRequest.cpp:700
3 	xul.dll 	mozilla::layers::LayerPropertiesBase::ComputeChange 	gfx/layers/LayerTreeInvalidation.cpp:117
4 	xul.dll 	nsRegion::SetToElements 	gfx/src/nsRegion.cpp:295
5 	xul.dll 	mozilla::layers::ContainerLayerProperties::ComputeChangeInternal 	gfx/layers/LayerTreeInvalidation.cpp:190
6 	xul.dll 	nsXMLHttpRequest::QueryInterface 	content/base/src/nsXMLHttpRequest.cpp:700
7 	xul.dll 	nsRegion::SubRect 	gfx/src/nsRegion.cpp:1242
8 	xul.dll 	nsXMLHttpRequest::QueryInterface 	content/base/src/nsXMLHttpRequest.cpp:700
9 	xul.dll 	mozilla::layers::LayerPropertiesBase::ComputeChange 	gfx/layers/LayerTreeInvalidation.cpp:117
10 	xul.dll 	nsXMLHttpRequest::QueryInterface 	content/base/src/nsXMLHttpRequest.cpp:700


This started happening on 2012-07-01 with Windows and Linux 64bit Nightly 16.0a1 builds.
See https://crash-stats.mozilla.com/report/list?signature=mozilla%3A%3Alayers%3A%3AContainerLayerProperties%3A%3AComputeChangeInternal%28void%20%28*%29%28mozilla%3A%3Alayers%3A%3AContainerLayer*%2C%20nsIntRegion%20const%26%29%29 and https://crash-stats.mozilla.com/report/list?signature=mozilla%3A%3Alayers%3A%3AContainerLayerProperties%3A%3AComputeChangeInternal for more crashes with those signatures.

Comment 1

[Andrew McCreight [:mccr8]]

Looks like it could be a dlbi regression.

Comment 2

[Jonathan Watt [:jwatt]]

There are no comments on the crash reports, so pretty much all we have to go on here are the stacks. Not much I can do here, Matt, unless you can think of something from that.

Comment 3

[Matt Woodrow (:mattwoodrow)]

I think this is caused by us not holding strong refs to the Layers in the LayerTreeInvalidation structure and then comparing raw pointers.

Holding a strong ref should prevent this from happening.

Comment 4

[Scoobidiver (away)]

There are no crashes over the last four weeks.