Closed Bug 1267288 Opened 8 years ago Closed 5 years ago

crash in libsystem_kernel.dylib@0x16f06

Categories

(Core :: Graphics, defect, P3)

45 Branch
x86_64
macOS
defect

Tracking

()

RESOLVED WORKSFORME
Tracking Status
firefox45 --- wontfix
firefox46 --- wontfix
firefox47 + wontfix
firefox48 + wontfix
firefox49 - wontfix
firefox-esr45 --- affected
firefox50 --- affected
firefox51 --- affected
firefox52 --- wontfix
firefox53 --- affected

People

(Reporter: marco, Unassigned)

References

Details

(Keywords: crash, Whiteboard: [gfx-noted])

Crash Data

This bug was filed from the Socorro interface and is 
report bp-e2a22aea-8c10-42fa-b9c8-89f432160419.
=============================================================

Filing this bug because there is no associated report to that signature yet.

There are different frames, for example https://crash-stats.mozilla.com/report/index/d7e79ae8-e176-4b9d-ba47-67abb2160425 or https://crash-stats.mozilla.com/report/index/c1de116d-4c96-428f-9838-2308e2160425.

All crashes are on recent versions of OS X (10.11.4 and 10.11.5).

The volume of crashes started spiking at the end of March, around the time that 10.11.4 got released.

Graphics critical errors in reports include the following:
> GLContext is disabled due to a previous crash.
> Failed to create a SkiaGL DrawTarget, falling back to software.
> Failed to allocate a surface due to invalid size Size(0,0).

Crashes aren't restricted to a single vendor, but most crashes are happening with Intel graphics cards.

There are crashes with the same signature that aren't related to graphics (e.g. a few are related to the Mac AddressBook), but most are indeed related to graphics.

https://crash-stats.mozilla.com/api/SuperSearch/?signature=%3Dlibsystem_kernel.dylib%400x16f06&_facets=platform_version&_facets=graphics_critical_error&_facets=adapter_vendor_id&_facets=proto_signature&_facets=version&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform
This is #43 top crash (with 1648 crashes).
[Tracking Requested - why for this release]: Tracking this for 47 as it is a Mac top crash. 

Yahoo News/Mail is mentioned in a few of the comments, and it looks as if the top 3 urls are facebook, youtube and yahoo.com.
Flags: needinfo?(howareyou322)
Bug 1269472 seems to have potential STR for producing this signature.
See Also: → 1269472
Whiteboard: [gfx-noted]
Crashing for me almost every day with this signature.
(In reply to Gregor Wagner [:gwagner] from comment #5)
> Crashing for me almost every day with this signature.

Gregor, is the any STR to reproduce this issue?
Flags: needinfo?(howareyou322) → needinfo?(anygregor)
(In reply to Peter Chang[:pchang] from comment #6)
> (In reply to Gregor Wagner [:gwagner] from comment #5)
> > Crashing for me almost every day with this signature.
> 
> Gregor, is the any STR to reproduce this issue?

Have you tried looking at the STR in Bug 1269472?
I imagine this is the same as bug 1218070.  The address in kernel for this bug is specific to OS X 10.11 (it seems), and while not all these crashes start in TexSubImage2DHelper, a great number of them does (
See Also: → 1218070
I would say 90% of my crashes are within http://tvthek.orf.at/. It's hard to say when it crashes but usually it crashes when I put my laptop to sleep and go back to a tab that has a video loaded or I navigate away from a tab that has a video loaded from this page.
Flags: needinfo?(anygregor)
[Tracking Requested - why for this release]:
This crash continues to rise and is currently the #3 top Mac crash @ 5.81% (#24 overall), peaking at 598 crashes in a single day on June 21st which is up 30% from a month ago.
Yes, this is mostly tracked in bug 1218070.
Track this for 48 as the crash number is high.
It's really unfortunate that we don't have symbols for these system libraries (bug 951229). This signature seems to correspond to "__pthread_kill on OS X 10.11.4 through 10.11.6", which is both overbroad in including everything that calls abort(), and overspecific in OS X versions.
Depends on: 1288589
Interesting, we don't have (yet?) a single crash with 50
Still no crashes on 50/51. This doesn't look high volume on beta (or really, on release for 48 at this point)

I think we can drop tracking here since we're tracking the high volume crash in bug 1218070.
Crashes for the last week:
* Firefox 51: 8 crashes, #7 on Mac @ 1.73%
* Firefox 50: 5 crashes, #19 on Mac @ 0.44%
* Firefox 49: 31 crashes, #9 on Mac @ 2.14%
* Firefox 48: 860 crashes, #4 on Mac @ 6.61%
* Firefox 47: 1870 crashes, #3 on Mac @ 7.46%
Just add my comment for reproducability.

This bug also happens to me a lot; often when i open/unlock my mac after a sleep, firefox crash directly after.
Crashes reported over the last week:
> 49.*:  10
> 50.*: 140
> 51.*:  12
> 52.*:   5
> 53.*:   4
Overall volume has dropped 85% from peak in late August.

This no longer amounts to a topcrash but is definitely still being reported in current branches.
@merzhin Sorry we missed responding to your comment 17. Is this crash still reproducible for you?
Flags: needinfo?(merzhin)
Seems good ... forgot to update :)

last crash 29/09 so seems good :)
(In reply to Marco Castelluccio [:marco] from comment #0)
> There are crashes with the same signature that aren't related to graphics
> (e.g. a few are related to the Mac AddressBook), but most are indeed related
> to graphics.

This should be Thunderbird related and I filed a bug about this crash a couple of minutes ago as bug 1332246.
Too late for firefox 52, mass-wontfix.

(In reply to Henrik Skupin (:whimboo) [⌚️UTC+1] from comment #21)

...
This should be Thunderbird related and I filed a bug about this crash a
couple of minutes ago as bug 1332246.

Indeed. And whatever this is/was, what's left for Thunderbird isn't worth tracking

Status: NEW → RESOLVED
Closed: 5 years ago
Flags: needinfo?(merzhin)
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.