Closed Bug 1723774 Opened 3 years ago Closed 3 years ago

sw-wr: Firefox closes itself without crash reported [@ cs_clip_rectangle_vert::run ]

Categories

(Core :: Graphics: WebRender, defect)

Firefox 90
x86_64
Linux
defect

Tracking

()

RESOLVED DUPLICATE of bug 1689978
Tracking Status
firefox90 --- affected

People

(Reporter: mystiquewolf, Unassigned)

References

(Blocks 1 open bug)

Details

(Keywords: crash)

Crash Data

Attachments

(2 files)

User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:90.0) Gecko/20100101 Firefox/90.0

Steps to reproduce:

Nothing, just leave the laptop unattended for some time and on return Firefox is missing.

Actual results:

Since some days i frequently leave my PC and then when i come back Firefox window is missing. For days i was wondering if i had gone crazy, but today i saw that i'm not crazy yet.

That is, while i was reading a Quora article Firefox just closed itself in front of my eyes. It first became transparent for about 1 second and then the Firefox in the taskbar also disappeared. So i opened Firefox and started writing this bug report. While i was doing this i had to go away from the PC and then when i came back Firefox again was closed on my return.

Expected results:

Firefox shouldn't close itself on its own. Or at least should produce a crash report once it does this.

Attached file Troubleshooting info (deleted) —

The Bugbug bot thinks this bug should belong to the 'Core::Widget: Gtk' component, and is moving the bug to that component. Please revert this change in case you think the bot is wrong.

Component: Untriaged → Widget: Gtk
Product: Firefox → Core

Hooray! Finally a crash report appeared! No idea why they weren't appearing in so many closes of Firefox before.
https://crash-stats.mozilla.org/report/index/88572314-d376-4f1a-8e29-ad9020210803#tab-details

Crash at cs_clip_rectangle_vert::run(cs_clip_rectangle_vert*, char*, unsigned long)

Component: Widget: Gtk → Graphics: WebRender

Firefox frequently freezes. Just now i had a chat with an Amazon support associate and Firefox closed itself during the chat - hopefully i was able to reopen and continue.

I've enabled gfx.webrender.enabled so that it uses the GPU acceleration/implementation. The crash now happened with that config key set to true, but the previous crashes were happening with the software implementation, which for unknown reasons to me was enabled instead of the hardware one.

There wasn't any crash reporter window shown - neither from Firefox nor from KDE not from Apport/Kubuntu. This IMO seems to be a critical issue for these reasons - frequent crashes without a crash backtrace. I don't know what kind of luck i had yesterday to be blessed to actually see a crash report window.

(bug 1721014 could also crash without report. It has been fixed with Firefox 91.)

Crash Signature: [@ cs_clip_rectangle_vert::run ]
Keywords: crash
OS: Unspecified → Linux
Hardware: Unspecified → x86_64
Summary: Firefox closes itself without crash reported → sw-wr: Firefox closes itself without crash reported [@ cs_clip_rectangle_vert::run ]

Well Firefox has crashed at least 5 times today with no crash reports sent. IMO this should be a P1 or P2 and should be uplifted to 90.0.3. Hell, we don't know if the crashes aren't 10 or even 15 times more than the reported ones in Socorro and Bugzilla combined.

Attached image earlyoom (deleted) —

I'm sorry, i've just found out that earlyoom was killing Firefox... Seems like i would have to find out what causes it.

I've measured a Firefox memory report - should i open a new bug or post here?

Flags: needinfo?(stransky)

:lsalzman, can you comment to this bug?

Flags: needinfo?(lsalzman)

I think i haven't seen crashes in a while now (maybe 1-2 days), if it happens again i'll post.

Flags: needinfo?(stransky)
Status: UNCONFIRMED → RESOLVED
Closed: 3 years ago
Flags: needinfo?(lsalzman)
Resolution: --- → DUPLICATE

For clarity, when Firefox was being closed by earlyoom it was because it was using a lot of memory. In the "Main Process" the structured-clone-holder was holding 4388 MB. At one time the "Main Process" was using 5166 MB VmRSS. If it happens again i'll file a separate bug.

You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: