Re: [RFC] Race condition?

[Oliver Neukum]

Am Freitag, 2. August 2002 19:37 schrieb Dave Hansen:
> Oliver Neukum wrote:
> > Am Freitag, 2. August 2002 15:46 schrieb Kasper Dupont:
> >>Is there a race condition in this piece of code from do_fork in
> >
> > It would seem so. Perhaps the BKL was taken previously.
>
> Even if it was, I doubt the code ever knowingly relied upon it.  If I
> know that I'm protected under a lock, I rarely go to the trouble of
> atomic operations.

That depends on where else you need these variables.

> The root of the problem is that the reference count is being relied on
> for the wrong thing.  There is a race on p->user between the
> dup_task_struct() and whenever the atomic_inc(&p->user->__count)
> occcurs.   The user reference count needs to be incremented in
> dup_task_struct(), before the copy occurs.

I don't get you. The user_struct can hardly go away while we are
forking.

IMHO you should add a spinlock to user_struct and take it.
A clear solution that doesn't hurt the common case.

	Regards
		Oliver