1842856 - Assertion failure happened randomly

Status: RESOLVED INCOMPLETE

(Data Platform and Tools :: Glean: SDK, defect)

Description

[Alex]

Attachment: ff-cores.lst

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0

Steps to reproduce:

Since FF 114 it started to fail a few times per day randomly.
In the attachment my coredumpctl info for the last faults.

Comment 1

[BugBot [:suhaib / :marco]]

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

Comment 2

[Martin Stránský [:stransky] (ni? me)]

It's crash at glean_core::metrics::timing_distribution::TimingDistributionMetric::start()

Comment 3

[Chris H-C :chutten]

Thank you for your report. What is your Linux kernel version? This may be bug 1835488 which is correlated with running either Fx114 or Fx115 and kernel 6.4.x for x < 3 (and mayyyybe some later 6.3.y ).

Comment 4

[Alex]

I can't tell for sure what was the kernel version when I got the very first crash, but for now, it's been like a month already since I use 6.4.x (nowadays 6.4.3). However, JFYI, the mentioned CONFIG_PER_VMA_LOCK is n in my .config.

Comment 5

[Chris H-C :chutten]

Hrm. It still seems likely due to the similarity of the environments that this is indeed bug 1835488 (Can't think of anything else in the function that'd SIGILL without resulting in a deeper stack). But it's just dissimilar enough for me that I don't feel right duping this.

(Moving to Glean SDK since that's where the stack ends)

Do you happen to submit crash reports from your Firefox? If so, about:crashes should have links to the submitted ones hosted on crash-stats. We collect a bit more metadata to help with diagnosis that might give us a better chance at figuring this out than the coredump analysis. If you've submitted one or two, linking them here might help us out.

Comment 6

[Jan-Erik Rediger [:janerik]]

We're closing this due to not having enough info to work on (though we still track the other bug of which this might likely be a dupe).
We can reopen when new information arrives though.