Re: [PATCH 0/3] orphaned pgrp fixes

[Oleg Nesterov]

On 03/04, Roland McGrath wrote:
>
> > However, SIGNAL_STOP_STOPPED doesn't relaible with ptrace. I mean, ptracer
> > doesn't clear SIGNAL_STOP_STOPPED. This is btw one of the problems which
> > complicates fixing do_wait(WSTOPPED).
> 
> We can clean that up too if you'd like to bring up the details.

Sorry! I was unclear.

Multiple reasons. Just for example,

	1:~$ strace -o /dev/null perl -e 'kill SIGSTOP,$$; for (;;) {}' &
	[1] 6714
	1:~$ kill -9 $!
	[1]+  Killed                  strace -o /dev/null perl -e 'kill SIGSTOP,$$; for (;;) {}'
	1:~$ grep State: /proc/`pidof perl`/status
	State:  R (running)

but its signal->flags = SIGNAL_STOP_STOPPED.

Please note that unless SIGCONT comes it will exit() with SIGNAL_STOP_STOPPED
set too, this complicates fixing do_wait(WSTOPPED). We can shadow this problem
though, do_wait() can first check TASK_ZOMBIE, and only then SIGNAL_STOP_STOPPED.
I am going to try to cook some patches soon.

> > As for SIGNAL_STOP_STOPPING... I am dreaming to find the way to eliminate
> > this lock-drop in get_signal_to_deliver(). Not sure this is possible, but
> > it is so nasty. For example, SIGNAL_STOP_DEQUEUED is racy, and I don't know
> > how to fix this. The patch we discussed some time ago doesn't really work
> > because dequeue_signal() drops the lock too. The latter is fixable afaics,
> > but needs very ugly changes.
>
> It all certainly deserves more careful thought.

Yes. Somehow the conception of orphaned pgrp doesn't fit in my brain, but I'd
like to at least try to think about this...

Oleg.

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