Musca segfaults with error 4

Bug #367736 reported by Inxsible
2
Affects Status Importance Assigned to Milestone
Musca
In Progress
High
Sean Pringle

Bug Description

Musca seems to randomly crash at times. This has happened to me while using different apps, mostly heavier ones like firefox or while watching a movie using cvlc etc. I have attached the core dump and the musca binary for you to have a look.

I am using musca 0.9.18. However, I had seen this happen to me in earlier versions as well. Let me know if you need any more information.

Revision history for this message
Inxsible (inxsible) wrote :
Changed in musca:
assignee: nobody → sean-pringle
Revision history for this message
Sean Pringle (sean-pringle) wrote :

[sean@vela tmp]$ file musca
musca: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.16, stripped

Would you mind rebuilding with symbols unstripped (we compile with gcc -g by default), and capture the core again with this new binary, please? No symbols makes it hard to dig around with gdb. Attach new core and binary as before.

If this is a segfault occuring as a result of us not checking EINTR (error 4) correctly somewhere, then you can also try adding the following line to the top of musca.c:

#define _GNU_SOURCE

This should make the default EINTR behavior be to retry the interrupted system call. However, since this might be a heisenbug, only make this change *after* you've triggered the fault at least once with the new binary.

Revision history for this message
Inxsible (inxsible) wrote : Re: [Bug 367736] Re: Musca segfaults with error 4

I am currently running version 20, with unstripped symbols. I guess we will
have to wait until I get a segfault again.

Thanks for your help so far.

Murtuza Akthari

On Mon, Apr 27, 2009 at 11:41 PM, Sean Pringle <email address hidden>wrote:

> [sean@vela tmp]$ file musca
> musca: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
> dynamically linked (uses shared libs), for GNU/Linux 2.6.16, stripped
>
> Would you mind rebuilding with symbols unstripped (we compile with gcc
> -g by default), and capture the core again with this new binary, please?
> No symbols makes it hard to dig around with gdb. Attach new core and
> binary as before.
>
> If this is a segfault occuring as a result of us not checking EINTR
> (error 4) correctly somewhere, then you can also try adding the
> following line to the top of musca.c:
>
> #define _GNU_SOURCE
>
> This should make the default EINTR behavior be to retry the interrupted
> system call. However, since this might be a heisenbug, only make this
> change *after* you've triggered the fault at least once with the new
> binary.
>
> --
> Musca segfaults with error 4
> https://bugs.launchpad.net/bugs/367736
> You received this bug notification because you are a direct subscriber
> of the bug.
>

--
---INXS

Revision history for this message
Sean Pringle (sean-pringle) wrote :

Keep us posted...

Changed in musca:
importance: Undecided → High
status: New → In Progress
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.