public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] remove sleep_on from sunrpc
@ 2004-01-17 18:43 Manfred Spraul
  2004-01-18  1:45 ` Peter Osterlund
  0 siblings, 1 reply; 2+ messages in thread
From: Manfred Spraul @ 2004-01-17 18:43 UTC (permalink / raw)
  To: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 400 bytes --]

sunrpc uses sleep_on to wait until the rpciod thread has died. This
could race if rpciod_down is called without lock_kernel. Right now all
calls in the main tree are under lock_kernel, but rpciod is an exported
function and sleep_on should die.
Additionally, I've replaced spin_lock_irqsave() with spin_lock_irq():
rpciod_down sleeps, interrupts are guaranteed to be enabled.

--
        Manfred





[-- Attachment #2: patch-sleep_on_sunrpc --]
[-- Type: text/plain, Size: 1100 bytes --]

--- 2.6/net/sunrpc/sched.c	2004-01-17 12:19:50.000000000 +0100
+++ build-2.6/net/sunrpc/sched.c	2004-01-17 12:52:33.000000000 +0100
@@ -1060,7 +1060,7 @@
 void
 rpciod_down(void)
 {
-	unsigned long flags;
+	DECLARE_WAITQUEUE(wait, current);
 
 	down(&rpciod_sema);
 	dprintk("rpciod_down pid %d sema %d\n", rpciod_pid, rpciod_users);
@@ -1085,17 +1085,22 @@
 	/*
 	 * Display a message if we're going to wait longer.
 	 */
-	while (rpciod_pid) {
+	add_wait_queue(&rpciod_killer, &wait);
+	for (;;) {
+		set_current_state(TASK_INTERRUPTIBLE);
+		if (rpciod_pid == 0)
+			break;
 		dprintk("rpciod_down: waiting for pid %d to exit\n", rpciod_pid);
 		if (signalled()) {
 			dprintk("rpciod_down: caught signal\n");
 			break;
 		}
-		interruptible_sleep_on(&rpciod_killer);
+		schedule();
 	}
-	spin_lock_irqsave(&current->sighand->siglock, flags);
+	remove_wait_queue(&rpciod_killer, &wait);
+	spin_lock_irq(&current->sighand->siglock);
 	recalc_sigpending();
-	spin_unlock_irqrestore(&current->sighand->siglock, flags);
+	spin_unlock_irq(&current->sighand->siglock);
 out:
 	up(&rpciod_sema);
 }





^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] remove sleep_on from sunrpc
  2004-01-17 18:43 [PATCH] remove sleep_on from sunrpc Manfred Spraul
@ 2004-01-18  1:45 ` Peter Osterlund
  0 siblings, 0 replies; 2+ messages in thread
From: Peter Osterlund @ 2004-01-18  1:45 UTC (permalink / raw)
  To: Manfred Spraul; +Cc: linux-kernel

Manfred Spraul <manfred@colorfullife.com> writes:

> -	while (rpciod_pid) {
> +	add_wait_queue(&rpciod_killer, &wait);
> +	for (;;) {
> +		set_current_state(TASK_INTERRUPTIBLE);
> +		if (rpciod_pid == 0)
> +			break;
>  		dprintk("rpciod_down: waiting for pid %d to exit\n", rpciod_pid);
>  		if (signalled()) {
>  			dprintk("rpciod_down: caught signal\n");
>  			break;
>  		}
> -		interruptible_sleep_on(&rpciod_killer);
> +		schedule();
>  	}
> -	spin_lock_irqsave(&current->sighand->siglock, flags);
> +	remove_wait_queue(&rpciod_killer, &wait);
> +	spin_lock_irq(&current->sighand->siglock);
>  	recalc_sigpending();
> -	spin_unlock_irqrestore(&current->sighand->siglock, flags);
> +	spin_unlock_irq(&current->sighand->siglock);
>  out:
>  	up(&rpciod_sema);
>  }

Aren't you forgetting to set_current_state(TASK_RUNNING) after the
loop?

-- 
Peter Osterlund - petero2@telia.com
http://w1.894.telia.com/~u89404340

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2004-01-18  1:42 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-01-17 18:43 [PATCH] remove sleep_on from sunrpc Manfred Spraul
2004-01-18  1:45 ` Peter Osterlund

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox