Re: [BUG] ptraced process waiting on syscall may return kernel internal errnos

[Oleg Nesterov <oleg@tv-sign.ru>]

On 06/15, Rafael J. Wysocki wrote:
>
> +static void freeze_task(struct task_struct *p)
> +{
>  	if (!freezing(p)) {
>  		rmb();
>  		if (!frozen(p)) {
>  			set_freeze_flag(p);
> -			if (p->state == TASK_STOPPED)
> -				force_sig_specific(SIGSTOP, p);
> -			spin_lock_irqsave(&p->sighand->siglock, flags);
> -			signal_wake_up(p, p->state == TASK_STOPPED);
> -			spin_unlock_irqrestore(&p->sighand->siglock, flags);
> +			task_lock(p);
> +			/* We don't want to send signals to kernel threads */
> +			if (p->mm && !(p->flags & PF_BORROWED_MM)) {
> +				task_unlock(p);
> +				send_fake_signal(p);
> +			} else {
> +				task_unlock(p);
> +				wake_up_state(p, TASK_INTERRUPTIBLE);
> +			}

I don't think this is enough. Note that recalc_sigpending() checks freezing().
So a kernel thread still can get TIF_SIGPENDING if it does recalc_sigpending().

> --- linux-2.6.22-rc4-mm2.orig/include/linux/wait.h	2007-06-15 01:05:33.000000000 +0200
> +++ linux-2.6.22-rc4-mm2/include/linux/wait.h	2007-06-15 01:05:41.000000000 +0200
> @@ -240,7 +240,7 @@ do {									\
>  		prepare_to_wait(&wq, &__wait, TASK_INTERRUPTIBLE);	\
>  		if (condition)						\
>  			break;						\
> -		if (!signal_pending(current)) {				\
> +		if (!signal_pending(current) && !freezing(current)) {	\
>  			schedule();					\
>  			continue;					\
>  		}							\

Personally, I think we should not modify wait_event_interruptible() and friends.
If a kernel thread wants to be frozen, it should take care about freezing()
itself.


OK, I guess I was too paranoid and you were right, it is better to ignore this
minor problem for now.

Oleg.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/