Re: [PATCH] race condition in procfs

[Grzegorz Nosek <grzegorz.nosek@gmail.com>]

2005/11/30, Steven Rostedt <rostedt@goodmis.org>:
> (Andrew, this will be the last email that I include you on.  I'm taking
> you off unless you want to stay on this thread, and say so.  I figure
> that you get enough spam without having to read through this. I'll
> obviously add you back if this results in a patch.)

(removed Andrew from the CC as well)

>
> On Wed, 2005-11-30 at 16:29 +0100, Grzegorz Nosek wrote:
> > 2005/11/30, Steven Rostedt <rostedt@goodmis.org>:
> > >
> > > OK, Remove your patches, run the system where you can capture the log,
> > > and provide a full output of the oops.  Make sure you have
> > > CONFIG_KALLSYMS set.
> > >
> >
> > OK, attached an oops from netconsole.
> >
>
> The oops happened at address a01b50eb.  Could you go into the compiled
> directory run gdb on vmlinux and type li *0xa01b50eb and show what you
> get.
>

OK, will send it as soon as I get my hands on it (I'm building a new
kernel at the moment with full debug info). In the meantime, if you
have a copy of fs/proc/array.o handy, have a look at do_task_stat
dissassembly and search for movzbl (%eax), %eax. Regardless of my
kernel config, architecture or whatever, the oops is in that
instruction (clearly a NULL pointer dereference). From some previous
debug build I found out (via objdump -dl) that it's apparently at the
entry point of the get_task_stat inline function.

Best regards,
 Grzegorz Nosek