Closed Bug 1098227 Opened 10 years ago Closed 10 years ago

Android crash in @0x0 | mozilla::layers::Layer::CalculateScissorRect

Categories

(Core :: Graphics: Layers, defect)

All
Android
defect
Not set
critical

Tracking

()

RESOLVED FIXED
Tracking Status
firefox36 + fixed
firefox37 + fixed
firefox38 + fixed
fennec 36+ ---

People

(Reporter: emorley, Assigned: nical)

References

Details

(Keywords: crash, reproducible, topcrash-android-armv7)

Crash Data

This bug was filed from the Socorro interface and is report bp-63aed25a-4cd9-4dab-92cf-a4f742141113. ============================================================= Nexus 5 Crashed whilst trying to scroll around on: http://www.flightradar24.com/data/flights/ua931/
Component: Graphics, Panning and Zooming → Graphics: Layers
Product: Firefox for Android → Core
This has happened on the tesco hudl for me on today's nighly build PS I was scrolling on github
I can still repro in the latest firefox for android nightly using the URL in comment 0, however it takes a bit more effort now, and I don't get the crash reporter, and no entry in about:crashes (Nightly just exits to the home screen).
tracking-fennec: --- → ?
tracking-fennec: ? → ---
ContentHostTexture is not used on Android and this part of the stack makes little sense.
Summary: crash in @0x0 | mozilla::layers::ContentHostTexture::Composite(mozilla::layers::EffectChain&, float, mozilla::gfx::Matrix4x4 const&, mozilla::gfx::Filter const&, mozilla::gfx::RectTyped<mozilla::gfx::UnknownUnits> const&, nsIntRegion const*) → Android crash in @0x0 | mozilla::layers::Layer::CalculateScissorRect
I'm able to reproduce this crash in the latest Nightly using the steps from comment #0 https://crash-stats.mozilla.com/report/index/a0f94485-bb8c-4038-888a-fa1902141128 Galaxy Note 3 (4.4.2)
Coming from the last duplicate, I can reproduce the crash from comment 0 too, but the signature it's a little different. Report ID Date Submitted bp-04ba7f23-ca2e-47ba-83f4-00a1c2141204 12/03/14 21:21 Kindle Fire
tracking-fennec: --- → ?
Keywords: reproducible
Crash Signature: [@ @0x0 | mozilla::layers::ContentHostTexture::Composite(mozilla::layers::EffectChain&, float, mozilla::gfx::Matrix4x4 const&, mozilla::gfx::Filter const&, mozilla::gfx::RectTyped<mozilla::gfx::UnknownUnits> const&, nsIntRegion const*)] → [@ @0x0 | mozilla::layers::ContentHostTexture::Composite(mozilla::layers::EffectChain&, float, mozilla::gfx::Matrix4x4 const&, mozilla::gfx::Filter const&, mozilla::gfx::RectTyped<mozilla::gfx::UnknownUnits> const&, nsIntRegion const*)] [@ OOM | large | …
Nicolas, can you give this a look? Looks like the latest signatures are a little more sane.
Flags: needinfo?(nical.bugzilla)
(In reply to alex_mayorga from comment #7) > Coming from the last duplicate, I can reproduce the crash from comment 0 > too, but the signature it's a little different. > > Report ID Date Submitted > bp-04ba7f23-ca2e-47ba-83f4-00a1c2141204 12/03/14 21:21 > > Kindle Fire The stack makes indeed a bit more sense, though the bottom of it is still somewhat mysterious. Looking at similar crash reports I found this one: https://crash-stats.mozilla.com/report/index/3d5a36f5-1852-45de-b89f-ed6522141211 which has "dlfree" in the stack. Perhaps something is trying to free an invalid address, or maybe it's just part of the random junk that shows up in the bottom of the stack. Not clear what could cause this though. There aren't a lot of crash reports with this signature yet: https://crash-stats.mozilla.com/search/?signature=~mozilla%3A%3Alayers%3A%3AImageHost%3A%3ALock&_facets=signature&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-signature
Flags: needinfo?(nical.bugzilla)
Is bug 1107930 the same as this as well? Together, those ::Composite crashes account for almost 20% of current Aurora 36 crashes on Android.
Crash Signature: , nsIntRegion const*)] [@ OOM | large | mozalloc_abort(char const*) | mozalloc_handle_oom(unsigned int) | moz_xmalloc | mozilla::gl::TexSubImage2DHelper] → , nsIntRegion const*)] [@ mozilla::layers::ContentHostTexture::Composite(mozilla::layers::EffectChain&, float, mozilla::gfx::Matrix4x4 const&, mozilla::gfx::Filter const&, mozilla::gfx::RectTyped<mozilla::gfx::UnknownUnits> const&, nsIntRegion const*)] …
Assignee: nobody → nical.bugzilla
tracking-fennec: ? → 36+
Are all these crashes the same(In reply to alex_mayorga from comment #7) > Coming from the last duplicate, I can reproduce the crash from comment 0 > too, but the signature it's a little different. > > Report ID Date Submitted > bp-04ba7f23-ca2e-47ba-83f4-00a1c2141204 12/03/14 21:21 > > Kindle Fire This is bug 1108449, right?
(In reply to Milan Sreckovic [:milan] from comment #12) > Are all these crashes the same(In reply to alex_mayorga from comment #7) > > Coming from the last duplicate, I can reproduce the crash from comment 0 > > too, but the signature it's a little different. > > > > Report ID Date Submitted > > bp-04ba7f23-ca2e-47ba-83f4-00a1c2141204 12/03/14 21:21 > > > > Kindle Fire > > This is bug 1108449, right? Yup! https://bugzilla.mozilla.org/show_bug.cgi?id=1108449
I suspect this got fixed along with bug 1108449. Can anyone still reproduce this issue?
(In reply to Nicolas Silva [:nical] from comment #14) > I suspect this got fixed along with bug 1108449. Can anyone still reproduce > this issue? There is a lingering patch on that bug that hasn't landed. Status?
(In reply to Aaron Train [:aaronmt] from comment #15) > (In reply to Nicolas Silva [:nical] from comment #14) > > I suspect this got fixed along with bug 1108449. Can anyone still reproduce > > this issue? > > There is a lingering patch on that bug that hasn't landed. Status? Nevermind, I think it's something else, working on a fix.
[Tracking Requested - why for this release]: 19% of Firefox for Android 36 b1 crashes are this signature.
Crash Signature: , nsIntRegion const*)] [@ OOM | large | mozalloc_abort(char const*) | mozalloc_handle_oom(unsigned int) | moz_xmalloc | mozilla::gl::TexSubImage2DHelper] → , nsIntRegion const*)] [@ OOM | large | mozalloc_abort(char const*) | mozalloc_handle_oom(unsigned int) | moz_xmalloc | mozilla::gl::TexSubImage2DHelper] [@ @0x0 | mozilla::layers::ContentHostIncremental::Composite(mozilla::layers::EffectChain&, float, …
A patch that should fix this has landed as part of bug 1119019 and got uplifted to aurora and beta today.
Kairo, can you confirm that it is fixed? Thanks
Flags: needinfo?(kairo)
I have very little time, but we have dedicated QE people on Android. :) Kevin, can you confirm this crash is fixed?
Flags: needinfo?(kairo) → needinfo?(kbrosnan)
AFIK this change is in beta 6. It will be Thursday or Friday before we have good data.
Kevin, do we know if it is fixed now? Thanks
Have one crash of https://bugzilla.mozilla.org/show_bug.cgi?id=1106016 all the other duplicates are not appearing in b6. Looks much better thanks for addressing this.
Status: NEW → RESOLVED
Closed: 10 years ago
Flags: needinfo?(kbrosnan)
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.