Closed Bug 1383742 Opened 7 years ago Closed 7 years ago

Crash in shutdownhang | __psynch_cvwait | <name omitted> | <name omitted> | nsThread::nsChainedEventQueue::GetNormalOrHighPriorityEvent

Categories

(Core :: XPCOM, defect)

Product:
Core
Component:
XPCOM
Platform:
Unspecified
macOS
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla56
Tracking Flags:
Tracking Status
firefox-esr52 --- unaffected
firefox54 --- unaffected
firefox55 --- unaffected
firefox56 --- fixed

People

(Reporter: marcia, Assigned: stone)

References

Details

(Keywords: crash, regression, topcrash-mac)

Crash Data

This bug was filed from the Socorro interface and is report bp-19c24119-3530-4b0b-a411-d0f140170724. ============================================================= Not sure where this belongs, but crashes started on Mac using 20170722100334: http://bit.ly/2uQXTpe. Comments mention: * crash on shutdown * crash while update-restart; Possible regression range based solely on Build ID: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=0faada5c2f308f101ab7c54a87b3dce80b97d0e3&tochange=7ce557b85b611536b69539a7c18d4834ffc92eea
Bug 1351148 landed on 7-21, so maybe that's the regression. Stone, can you take a look? Thanks.
Flags: needinfo?(sshih)
Assignee: nobody → sshih
Flags: needinfo?(sshih)
I have been getting a very similar shutdown hang in bug 1384288. Based on the stacks, these may be dupes?
See Also: → https://bugzilla.mozilla.org/show_bug.cgi?id=1384288
Wondering if bug 1383805 might also be fall-out from this crash?
Might this also be related to the shutdown hangs that are making it difficult/impossible for people to quit Firefox (even with force quit) ?
This signature is ranked #1 in mac top-crashers. There are 133 crashes and they started in nightly 56 with buildid 20170722100334. 92% of the crashes are occuring with Mac OSX 10.12.
Keywords: topcrash-mac
Affected too. Happens when I click the little green "Restart Nightly" button (or whatever it's called). Happened twice in two days. Windows disappear, nothing happens for 5-10 seconds, then the Breakpad dialog appears. Pressing Restart always works and I get my session back. Mine: https://crash-stats.mozilla.com/report/index/98d8054c-b095-4128-9482-1ba7d0170726
Kan-Ru mentioned on the stability list he was seeing these as well, and could try to debug locally.
Crash Signature: [@ shutdownhang | __psynch_cvwait | <name omitted> | <name omitted> | nsThread::nsChainedEventQueue::GetNormalOrHighPriorityEvent] → [@ shutdownhang | __psynch_cvwait | <name omitted> | <name omitted> | nsThread::nsChainedEventQueue::GetNormalOrHighPriorityEvent] [@ shutdownhang | libpthread-2.23.so@0xd360 ]
It becomes intermittent to me. Maybe the issue can only be reproduced after some time of browsing.
Interestingly, on my home machine (also Mac OSX), I just this morning updated through nightly 7/25 to 7/26's builds (from 7/22). This issue did not happen on that machine. Could it perhaps be something other than browser time? I'll compare the configs of these two machines.
I think I'm hitting this, too. My impression is that it only happens after an update has been staged - but that's a sample size of ~5 restarts.
As discussed in the Channel meeting today, these shutdown hangs account for a combined 20% of browser crashes on nightly in the past few days. We are coming up to the nightly merge and we don't want to carry this regression forward. :stone - are you actively working on this and can it be fixed before the merge? Thanks.
I may has misinterpreted but I thought Stone was backing his work in bug 1351148 out (assuming that's the root cause here).
Flags: needinfo?(htsai)
Stone is going to pref off the input event queue on bug 1384850.
Flags: needinfo?(htsai)
Crash Signature: [@ shutdownhang | __psynch_cvwait | <name omitted> | <name omitted> | nsThread::nsChainedEventQueue::GetNormalOrHighPriorityEvent] [@ shutdownhang | libpthread-2.23.so@0xd360 ] → [@ shutdownhang | __psynch_cvwait | <name omitted> | <name omitted> | nsThread::nsChainedEventQueue::GetNormalOrHighPriorityEvent] [@ shutdownhang | libpthread-2.23.so@0xd360 ] [@ shutdownhang | __psynch_cvwait | <name omitted> | <name omitted> | nsThread…
(In reply to Julien Cristau [:jcristau] from comment #16) > bug 1351148 has now been backed out: > https://bugzilla.mozilla.org/show_bug.cgi?id=1351148#c130 No reports after the 7-28 build, so the backout worked.
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
Blocks: 1351148
status-firefox54: --- → unaffected
status-firefox55: --- → unaffected
status-firefox56: affected → fixed
status-firefox-esr52: --- → unaffected
Target Milestone: --- → mozilla56
Someone hit one of the signatures (shutdownhang | __psynch_cvwait | <name omitted> | <name omitted> | nsThread::nsChainedEventQueue::GetNormalOrHighPriorityEvent)using today's nightly build: https://crash-stats.mozilla.com/report/index/8b60b2e6-70e0-4680-9f88-e4c450170816. But the volume has been greatly reduced compared to what it was.
Copying the signatures from bug 1384288 FYI, I've just experienced that with Thunderbird 56.0b3
Crash Signature: nsThread::GetEvent | nsThread::ProcessNextEvent | NS_ProcessNextEvent | mozilla::layers::CompositorThreadHolder::Shutdown] → nsThread::GetEvent | nsThread::ProcessNextEvent | NS_ProcessNextEvent | mozilla::layers::CompositorThreadHolder::Shutdown] [@ shutdownhang | libpthread-2.24.so@0xd510] [@ shutdownhang | libpthread-2.24.so@0xd15f] [@ shutdownhang | libpthread-2.25.so@0…
(In reply to Sylvestre Ledru [:sylvestre] from comment #22) > Copying the signatures from bug 1384288 > > FYI, I've just experienced that with Thunderbird 56.0b3 Crash ID?
Flags: needinfo?(sledru)
You need to log in before you can comment on or make changes to this bug.