2007-11-18 09:46:27

by Scott James Remnant

[permalink] [raw]
Subject: [PATCH] wait_task_stopped: pass correct exit_code to wait_noreap_copyout

In wait_task_stopped() exit_code already contains the right value for
the si_status member of siginfo, and this is simply set in the non
WNOWAIT case.

Pass it unchanged to wait_noreap_copyout(); we would only need to
shift it and add 0x7f if we were returning it in the user status field
and that isn't used for any function that permits WNOWAIT.

Signed-off-by: Scott James Remnant <[email protected]>
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Roland McGrath <[email protected]>

--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -1389,7 +1389,7 @@ static int wait_task_stopped(struct task_struct
*p, int delayed_group_leader,
if (unlikely(!exit_code) || unlikely(p->exit_state))
goto bail_ref;
return wait_noreap_copyout(p, pid, uid,
- why, (exit_code << 8) | 0x7f,
+ why, exit_code,
infop, ru);
}



Attachments:
signature.asc (189.00 B)
This is a digitally signed message part

2007-11-20 06:44:07

by Andrew Morton

[permalink] [raw]
Subject: Re: [PATCH] wait_task_stopped: pass correct exit_code to wait_noreap_copyout

On Sun, 18 Nov 2007 09:13:24 +0000 Scott James Remnant <[email protected]> wrote:

> In wait_task_stopped() exit_code already contains the right value for
> the si_status member of siginfo, and this is simply set in the non
> WNOWAIT case.
>
> Pass it unchanged to wait_noreap_copyout(); we would only need to
> shift it and add 0x7f if we were returning it in the user status field
> and that isn't used for any function that permits WNOWAIT.
>
> Signed-off-by: Scott James Remnant <[email protected]>
> Signed-off-by: Oleg Nesterov <[email protected]>
> Signed-off-by: Roland McGrath <[email protected]>
>
> --- a/kernel/exit.c
> +++ b/kernel/exit.c
> @@ -1389,7 +1389,7 @@ static int wait_task_stopped(struct task_struct
> *p, int delayed_group_leader,
> if (unlikely(!exit_code) || unlikely(p->exit_state))
> goto bail_ref;
> return wait_noreap_copyout(p, pid, uid,
> - why, (exit_code << 8) | 0x7f,
> + why, exit_code,
> infop, ru);
> }

Is this bug visible to userspace? If so, I'm surprised that none of the
various testsuites (which like to exercise this sort of interface) has
detected it.

2007-11-20 07:55:38

by Scott James Remnant

[permalink] [raw]
Subject: Re: [PATCH] wait_task_stopped: pass correct exit_code to wait_noreap_copyout

On Mon, 2007-11-19 at 22:43 -0800, Andrew Morton wrote:
> On Sun, 18 Nov 2007 09:13:24 +0000 Scott James Remnant <[email protected]> wrote:
>
> > In wait_task_stopped() exit_code already contains the right value for
> > the si_status member of siginfo, and this is simply set in the non
> > WNOWAIT case.
> >
> > Pass it unchanged to wait_noreap_copyout(); we would only need to
> > shift it and add 0x7f if we were returning it in the user status field
> > and that isn't used for any function that permits WNOWAIT.
> >
> Is this bug visible to userspace? If so, I'm surprised that none of the
> various testsuites (which like to exercise this sort of interface) has
> detected it.
>
Absolutely; if you call waitid() with a stopped or traced process,
you'll get the signal in siginfo.si_status as expected -- however if you
call waitid(WNOWAIT) at the same time, you'll get the signal << 8 | 0x7f

Scott
--
Scott James Remnant
[email protected]


Attachments:
signature.asc (189.00 B)
This is a digitally signed message part

2007-11-20 08:16:53

by Andrew Morton

[permalink] [raw]
Subject: Re: [PATCH] wait_task_stopped: pass correct exit_code to wait_noreap_copyout

On Tue, 20 Nov 2007 07:55:22 +0000 Scott James Remnant <[email protected]> wrote:

> On Mon, 2007-11-19 at 22:43 -0800, Andrew Morton wrote:
> > On Sun, 18 Nov 2007 09:13:24 +0000 Scott James Remnant <[email protected]> wrote:
> >
> > > In wait_task_stopped() exit_code already contains the right value for
> > > the si_status member of siginfo, and this is simply set in the non
> > > WNOWAIT case.
> > >
> > > Pass it unchanged to wait_noreap_copyout(); we would only need to
> > > shift it and add 0x7f if we were returning it in the user status field
> > > and that isn't used for any function that permits WNOWAIT.
> > >
> > Is this bug visible to userspace? If so, I'm surprised that none of the
> > various testsuites (which like to exercise this sort of interface) has
> > detected it.
> >
> Absolutely; if you call waitid() with a stopped or traced process,
> you'll get the signal in siginfo.si_status as expected -- however if you
> call waitid(WNOWAIT) at the same time, you'll get the signal << 8 | 0x7f
>

hm, OK. Well I guess I'll stick a for-2.6.23 tag on this as well as
queueing it for 2.6.24.