Status

()

7 years ago
4 years ago

People

(Reporter: glob, Unassigned)

Tracking

12 Branch
Points:
---

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(2 attachments)

(Reporter)

Description

7 years ago
Created attachment 592605 [details]
thread backtrace

i've been seeing periodic hangs of nightly.

i've attached a backtrace of all threads during a hang; roc indicated that thread 28 looks guilty:

Thread 28 (process 80914):
#0  0x00007fff89667e4a in memmove$VARIANT$sse42 ()
#1  0x0000000102264eb6 in js::crash::Ring::push ()


Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:12.0a1) Gecko/20120129 Firefox/12.0a1
(Reporter)

Comment 1

7 years ago
Created attachment 592733 [details]
another set of backtraces

after another hang, here's another set of backtraces (i won't do anymore unless requested).

includes:

Thread 21 (process 96809):
#0  0x00007fff89667e73 in memmove$VARIANT$sse42 ()
#1  0x0000000102264f96 in js::crash::Ring::push ()
When did this start happening? Is it only on a particular site, or just general browsing? Were you visiting sites using web workers (that you know of)?

Also, how did you capture the stack trace? In both threads, breakpad appears to be running. Maybe that's just a symptom of how you captured the stacks, though.

js::crash::Ring::push doesn't really do anything itself besides call memcpy. It's only called at the beginning of a GC. So it's weird that we even end up there on something that's not the main thread. And even if we do, it shouldn't hang there.
(Reporter)

Comment 3

7 years ago
(In reply to Bill McCloskey (:billm) from comment #2)
> When did this start happening?

one to two weeks ago (i think, sorry).

> Is it only on a particular site, or just general browsing?

general browsing; sometimes overnight.

> Were you visiting sites using web workers (that you know of)?

i don't know.

> Also, how did you capture the stack trace? In both threads, breakpad appears
> to be running. Maybe that's just a symptom of how you captured the stacks,
> though.

attached with gdb, then 'thread apply all bt'.
(Assignee)

Updated

4 years ago
Assignee: general → nobody
You need to log in before you can comment on or make changes to this bug.