Re: [PATCH 3/3] signal: do_sigtimedwait() needs retarget_shared_pending()

[Tejun Heo <tj@kernel.org>]

Hello,

On Sat, Apr 23, 2011 at 08:00:00PM +0200, Oleg Nesterov wrote:
> do_sigtimedwait() changes current->blocked and thus it needs
> set_current_bloked()->retarget_shared_pending().
> 
> We could use set_current_bloked() directly. It is fine to change
> ->real_blocked from all-zeroes to ->blocked and vice versa lockless,
> but this is not immediately clear, looks racy, and needs a huge
> comment to explain why this is correct.
> 
> To keep the things simple this patch adds the new static helper,
> __set_task_blocked() which should be called with ->siglock held. This
> way we can change both ->real_blocked and ->blocked atomically under
> ->siglock as the current code does. This is more understandable.
> 
> Signed-off-by: Oleg Nesterov <oleg@redhat.com>

Acked-by: Tejun Heo <tj@kernel.org>

> @@ -2322,7 +2327,8 @@ int do_sigtimedwait(sigset_t *these, sig
>  		/*
>  		 * None ready -- temporarily unblock those we're
>  		 * interested while we are sleeping in so that we'll
> -		 * be awakened when they arrive.
> +		 * be awakened when they arrive. Unblocking is always
> +		 * fine, we can avoid set_current_blocked().
>  		 */
>  		tsk->real_blocked = tsk->blocked;
>  		sigandsets(&tsk->blocked, &tsk->blocked, these);

Maybe it would be a good idea to introduce a new helper which checks /
enforces that the operation indeed is only unblocking?  Also, it can
be a pure preference but I think _locked suffix is better / more
common for APIs which expect the caller to be responsible for locking.

Thanks.

-- 
tejun