Crash in [@ hsw::lowp::bilerp_clamp_8888]
Categories
(Core :: Graphics, defect, P3)
Tracking
()
Tracking | Status | |
---|---|---|
firefox81 | --- | affected |
People
(Reporter: gsvelto, Unassigned)
Details
(Keywords: crash, Whiteboard: [tbird crash])
Crash Data
This bug is for crash report bp-682567bc-e68e-4ac6-812a-f48570200810.
Top 10 frames of crashing thread:
0 xul.dll hsw::lowp::bilerp_clamp_8888 gfx/skia/skia/src/opts/SkRasterPipeline_opts.h:4290
1 xul.dll hsw::lowp::start_pipeline gfx/skia/skia/src/opts/SkRasterPipeline_opts.h:3141
2 xul.dll std::_Func_impl_no_alloc<`lambda at /builds/worker/checkouts/gecko/gfx/skia/skia/src/core/SkRasterPipeline.cpp:355:12', void, unsigned long long, unsigned long long, unsigned long long, unsigned long long>::_Do_call
3 xul.dll SkRasterPipelineBlitter::blitAntiH gfx/skia/skia/src/core/SkRasterPipelineBlitter.cpp:326
4 xul.dll SkAAClipBlitter::blitRect gfx/skia/skia/src/core/SkAAClip.cpp:1972
5 xul.dll SkRectClipBlitter::blitAntiRect gfx/skia/skia/src/core/SkBlitter.cpp:450
6 xul.dll static SkScan::AAAFillPath gfx/skia/skia/src/core/SkScan_AAAPath.cpp:1988
7 xul.dll static SkScan::AntiFillPath gfx/skia/skia/src/core/SkScan_AntiPath.cpp:792
8 xul.dll static SkScan::AntiFillPath gfx/skia/skia/src/core/SkScan_AntiPath.cpp:836
9 xul.dll SkDraw::drawDevPath const gfx/skia/skia/src/core/SkDraw.cpp:885
Not sure what to make of this but the crashes seem consistent across multiple reports so it seems like a valid issue. Many of the nightly crashes appear to be on Facebook.
Updated•4 years ago
|
Comment 1•4 years ago
|
||
@Lee: looks like a crash in Skia, could you take a look?
Updated•4 years ago
|
Comment 2•2 years ago
|
||
(In reply to Gabriele Svelto [:gsvelto] from comment #0)
Not sure what to make of this but the crashes seem consistent across multiple reports so it seems like a valid issue. Many of the nightly crashes appear to be on Facebook.
in the last 6 months there are essentially no beta or nightly crashes.
in the last 6 months No significant change in crash rate IMO https://crash-stats.mozilla.org/signature/?product=Firefox&signature=hsw%3A%3Alowp%3A%3Abilerp_clamp_8888&date=%3E%3D2021-09-05T15%3A13%3A00.000Z&date=%3C2022-03-05T15%3A13%3A00.000Z#graphs
Updated•2 years ago
|
Reporter | ||
Comment 4•7 months ago
|
||
No, looking at the few recent crashes there's a significant chance that they're being caused by flaky hardware (many have triggered our bit-flip detection logic). Zooming out looking at similar signatures with a larger volume - such as [@ hsw::lowp::gather]
this pattern holds: most crashes are probably caused by bad memory. So this was likely a flaky-hardware issue all along.
Description
•