Re: [PATCH 2/2] creds: __task_cred(current) doesn't needrcu_read_lock_held()

[Oleg Nesterov]

On 09/20, David Howells wrote:
>
> Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
>
> > > > -	rcu_dereference_protected(current->cred, 1)
> > >
> > > and:
> > >
> > > > -		rcu_dereference_check(__t->real_cred, 0);		\
> > >
> > > you'll notice they aren't quite the same in one very fundamental way.
> >
> > Do you mean that this patch adds the unnecessary ACCESS_ONCE +
> > smp_read_barrier_depends() to current_cred() or I missed something
> > else?
>
> Something else.  The current_cred() uses ->cred:

Argh!!! I am soooo stupid.

Thanks a lot for the quick NACK!


So. This patch shouldn't touch current_cred().

But this also means, in theory it is not good to assume that
send_signal() can use __task_cred(current) instead of current_cred(),
although I doubt very much that someone can do override_creds() +
kill(SEND_SIG_NOINFO). So the change in __task_cred() is probably not
really good too.

Thanks.

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/