Closed Bug 1392621 Opened 8 years ago Closed 6 years ago

Crash in mozilla::ipc::MessageChannel::CxxStackFrame::CxxStackFrame | mozilla::ipc::MessageChannel::Send | mozilla::layers::PAPZCTreeManagerChild::SendUpdateZoomConstraints

Categories

(Core :: Panning and Zooming, defect, P3)

57 Branch
All
Windows
defect

Tracking

()

RESOLVED WORKSFORME
Tracking Status
firefox-esr52 --- unaffected
firefox56 --- wontfix
firefox57 --- wontfix
firefox58 --- wontfix
firefox59 --- wontfix
firefox60 --- fix-optional
firefox61 --- fix-optional

People

(Reporter: marcia, Unassigned)

References

(Blocks 1 open bug)

Details

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

Crash Data

This bug was filed from the Socorro interface and is report bp-bd9e189a-0b26-4ab9-a88f-41e570170822. ============================================================= Seen while looking at nightly crash data - crashes started using when nightly was still in 56 using 20170629030206: http://bit.ly/2v2pOUA Not super huge volume, but 13 crashes in 20170821100350.
Component: IPC → Panning and Zooming
Seems like there's a few different signatures [1] where a similar crash happens (moz_crash_reason contains "not on worker thread"). There might be a common root cause. In bug 1331209 comment 7 there's a suggestion this might be because the IPC channel is already torn down. However the messages that are triggering these crashes seem to be messages that fire very early during the content process lifetime, so I'm not sure what to make of that. [1] https://crash-stats.mozilla.com/search/?moz_crash_reason=~not%20on%20worker%20thread&product=Firefox&version=57.0a1&date=%3E%3D2017-08-17T15%3A25%3A00.000Z&date=%3C2017-08-24T15%3A25%3A00.000Z&_sort=-date&_facets=signature&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-signature
Priority: -- → P3
See Also: → 1391458, 1331209
Whiteboard: [gfx-noted]
Version: Trunk → 57 Branch
See Also: → 1443221
Blocks: 1443221
acdedbd4-3e38-4a96-ae94-4fbbd0180410 2018-04-10 18:57:53 Firefox 61.0a1 20180410100115 Windows NT EXCEPTION_BREAKPOINT 0x653a40a3 2018-04-10 18:43:50 042c8ea9-7032-40d4-8970-a08630180412 2018-04-12 07:59:39 Firefox 60.0b7 20180326164103 Windows NT EXCEPTION_BREAKPOINT 0x10990cf4 2018-04-12 07:49:51 18101551-8d13-4b02-98f8-e72ed0180413 2018-04-13 20:16:33 Firefox 60.0b11 20180409184545 Windows NT EXCEPTION_BREAKPOINT 0x7ff9b39137d1 2018-04-11 03:56:35 bcc497da-744a-497c-bd55-8549e0180413 2018-04-13 20:16:32 Firefox 60.0b11 20180409184545 Windows NT EXCEPTION_BREAKPOINT 0x7ff9b39137d1 2018-04-11 03:56:35 36c99b88-5a0d-4c37-ad28-8bf2f0180413 2018-04-13 20:16:31 Firefox 60.0b11 20180409184545 Windows NT EXCEPTION_BREAKPOINT 0x7ff9b39137d1 2018-04-11 03:56:35 e146ae2a-b0f8-42fe-bd1a-97dc70180411 2018-04-11 17:34:40 Firefox 60.0b11 20180409184545 Windows NT EXCEPTION_BREAKPOINT 0x7ff949b10243 2018-04-11 11:49:46 88cf9a96-0ed1-414f-9359-01be00180413 2018-04-13 23:17:58 Firefox 59.0.2 20180323154952 Windows NT EXCEPTION_BREAKPOINT 0x7feeaa9cd13 2018-03-29 04:41:20 e9efd0f6-fb8b-4f21-a49d-518610180413 2018-04-13 21:13:45 Firefox 59.0.2 20180323154952 Windows NT EXCEPTION_BREAKPOINT 0x7fee159cd13 2018-03-28 10:30:29
OS: Windows 8 → Windows
Hardware: Unspecified → All
There seems to be reliable STR in bug 1501285 comment 0.
Reproduced the crash so far 9 times on the latest Nightly on Windows 10 x64. Very intermittent and quite hard to master the timing/STR, here it is: 1. Launch latest Nightly 2. Change the window size manually by dragging the edges -very important* 3. Go to https://rsolomakhin.github.io/pr/us/ (make sure you have browser.search.region set to US or CA) 4. Click on "Buy" and instantly maximize the window *make sure to restart the browser everytime the above STR does not work and repeat it The browser should crash while de Payment Request Widget is still displayed (https://imgur.com/QOR3k0o)
I tried reproducing this a bunch today using the STR in comment 6 but was unable to :(
Btw, bug 1415610 comment 2: Have you guarded against that problem there?
No, I'll do some more testing on Monday with GPU process resets to see if I can trigger this.
Parking with me for now.
Assignee: nobody → kats
Still not able to repro. I can try to just write a IPC guard anyway, but I'll do it in a dep bug.
Assignee: kats → nobody

No crashes in recent release. Resolving as WFM.

Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.