906061 - crash in mozalloc_abort(char const*) | mozalloc_handle_oom(unsigned int) | moz_xmalloc | libxul.so@0xe04d44 | libxul.so@0xdf4ad0

Status: RESOLVED INCOMPLETE

(Core :: General, defect)

Description

[Ludovic Hirlimann [:Usul]]

This bug was filed from the Socorro interface and is 
report bp-182ec0dc-0bb0-44a5-97a0-433072130816 .
 ============================================================= 
0 	libmozalloc.so 	mozalloc_abort(char const*) 	memory/mozalloc/mozalloc_abort.cpp
1 	libmozalloc.so 	mozalloc_handle_oom(unsigned int) 	memory/mozalloc/mozalloc_oom.cpp
2 	libmozalloc.so 	moz_xmalloc 	memory/mozalloc/mozalloc.cpp
3 	libxul.so 	libxul.so@0xe04d44 	
4 	libxul.so 	libxul.so@0xdf4ad0

Wondering why it doesn't have more symbols

Comment 1

[Vincent Lefevre]

Apparently the same problem as:

  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709841

(also with moz_xmalloc / mozalloc_handle_oom / mozalloc_abort).

Comment 2

[Wayne Mery (:wsmwk)]

Ludo, have you crashed like this again?


(In reply to Vincent Lefevre from comment #1)
> Apparently the same problem as:
> 
>   http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709841
> 
> (also with moz_xmalloc / mozalloc_handle_oom / mozalloc_abort).

bp-182ec0dc-0bb0-44a5-97a0-433072130816 isn't much of a stack to make a connection to anything.

But http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709841 may be bug 764342 or bug 818839

Comment 3

[Vincent Lefevre]

Yes, I've just changed the forwarded address of http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709841 to Mozilla bug 764342.

Bug 818839 is WFM for 23.0a1, while the crash also occurred with Debian's iceweasel/24.1.0esr-1, so that this is not the same bug (or bug 818839 shouldn't have been closed).

Comment 4

[Wayne Mery (:wsmwk)]

(In reply to Vincent Lefevre from comment #3)
> Yes, I've just changed the forwarded address of
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709841 to Mozilla bug
> 764342.
> 
> Bug 818839 is WFM for 23.0a1, 

Bug 818839 closed based on my testing because the reporter failed to reply.  But it is possible my test was flawed.

Comment 5

[Vincent Lefevre]

Perhaps. Here the crashes occur very randomly.

Comment 6

[Ludovic Hirlimann [:Usul]]

(In reply to Wayne Mery (:wsmwk) from comment #2)
> Ludo, have you crashed like this again?
 nope

Comment 7

[Wayne Mery (:wsmwk)]

(In reply to Vincent Lefevre from comment #3)
> Yes, I've just changed the forwarded address of
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709841 to Mozilla bug
> 764342.
> 
> Bug 818839 is WFM for 23.0a1, while the crash also occurred with Debian's
> iceweasel/24.1.0esr-1, so that this is not the same bug (or bug 818839
> shouldn't have been closed).

Feel free to fix up anything you thing is incorrect. 

Do you still see this one?

Comment 8

[Vincent Lefevre]

(In reply to Wayne Mery (:wsmwk) from comment #7)
> Do you still see this one?

I don't know if this is exactly the same bug. Firefox no longer crashed, but hanged, and I had to kill it with kill -9. See the follow reports:

  bp-42c300da-7bce-4960-83f0-b5c8c2140701
  bp-61bc691d-8e46-473e-b878-b73132140702

Comment 9

[Wayne Mery (:wsmwk)]

(In reply to Vincent Lefevre from comment #8)
> (In reply to Wayne Mery (:wsmwk) from comment #7)
> > Do you still see this one?
> 
> I don't know if this is exactly the same bug. Firefox no longer crashed, but
> hanged, and I had to kill it with kill -9. See the follow reports:
> 
>   bp-42c300da-7bce-4960-83f0-b5c8c2140701
>   bp-61bc691d-8e46-473e-b878-b73132140702

For a hang, you would need to get a hang stacktrace before you kill it.
https://developer.mozilla.org/en-US/docs/How_to_get_a_stacktrace_for_a_bug_report

I suggest if you can reproduce the issue with Firefox started in safe mode, that you file a new bug report

Comment 10

[Vincent Lefevre]

(In reply to Wayne Mery (:wsmwk) from comment #9)
> For a hang, you would need to get a hang stacktrace before you kill it.

This is what I did in bug 1031004.

> https://developer.mozilla.org/en-US/docs/How_to_get_a_stacktrace_for_a_bug_report

This web page doesn't say how one can get a useful backtrace with the official Firefox (the symbols are stripped, but I've never done that!).

> I suggest if you can reproduce the issue with Firefox started in safe mode,
> that you file a new bug report

A crash or hang with "OOM | small" signature always takes several days, so that it is not really possible to reproduce in safe mode.