Re: [PATCH v3] exec: do not sleep in TASK_TRACED under ->cred_guard_mutex

[Oleg Nesterov <oleg@redhat.com>]

On 09/09, Chuck Ebbert wrote:
>
> On Fri, 4 Sep 2009 23:33:37 +0200
> Oleg Nesterov <oleg@redhat.com> wrote:
>
> > On 09/04, Andrew Morton wrote:
> > >
> > > On Fri, 4 Sep 2009 19:26:48 +0200
> > > Oleg Nesterov <oleg@redhat.com> wrote:
> > >
> > > > Tom Horsley reports that his debugger hangs when it tries to read
> > > > /proc/pid_of_tracee/maps, this happens since
> > > >
> > > > 	"mm_for_maps: take ->cred_guard_mutex to fix the race with exec"
> > > > 	04b836cbf19e885f8366bccb2e4b0474346c02d
> > > >
> > > > commit in 2.6.31.
>
> >
> > > I see a Cc:stable in the mail headers, but not in the changelog.  I
> > > don't think the patch is applicable to -stable unless we miss 2.6.31.
> >
> > -stable has the same problems but I agree, it can live without this fix.
> >
>
> That patch was backported in 2.6.30.5 -- are you sure -stable doesn't need
> this fix?

No, I am not sure, that is why I cc'ed -stable.

But otoh the problem is minor, both tracer and tracee can be killed.

I dunno. I don't know if gdb or any other "important" application
read /proc/pid/maps after PTRACE_EVENT_EXEC...

Oleg.