[PATCH] rcu_sched_grace_period: kill the bogus flush_signals()

[PATCH] rcu_sched_grace_period: kill the bogus flush_signals()

[Oleg Nesterov]

As a kernel thread, rcu_sched_grace_period() runs with all signals ignored.
It can never receive a signal even if it sleeps in TASK_INTERRUPTIBLE, it
needs the explicit allow_signal() to be visible for signals.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>

--- PTRACE/kernel/rcupreempt.c~	2009-04-09 00:39:10.000000000 +0200
+++ PTRACE/kernel/rcupreempt.c	2009-05-03 02:17:11.000000000 +0200
@@ -1356,17 +1356,11 @@ static int rcu_sched_grace_period(void *
 
 		rcu_ctrlblk.sched_sleep = rcu_sched_sleeping;
 		spin_unlock_irqrestore(&rcu_ctrlblk.schedlock, flags);
-		ret = 0;
+		ret = 0; /* unused */
 		__wait_event_interruptible(rcu_ctrlblk.sched_wq,
 			rcu_ctrlblk.sched_sleep != rcu_sched_sleeping,
 			ret);
 
-		/*
-		 * Signals would prevent us from sleeping, and we cannot
-		 * do much with them in any case.  So flush them.
-		 */
-		if (ret)
-			flush_signals(current);
 		couldsleepnext = 0;
 
 	} while (!kthread_should_stop());

Re: [PATCH] rcu_sched_grace_period: kill the bogus flush_signals()

[Paul E. McKenney]

On Sun, May 03, 2009 at 11:11:18PM +0200, Oleg Nesterov wrote:
> As a kernel thread, rcu_sched_grace_period() runs with all signals ignored.
> It can never receive a signal even if it sleeps in TASK_INTERRUPTIBLE, it
> needs the explicit allow_signal() to be visible for signals.

Good point!!!

Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>

> Signed-off-by: Oleg Nesterov <oleg@redhat.com>
> 
> --- PTRACE/kernel/rcupreempt.c~	2009-04-09 00:39:10.000000000 +0200
> +++ PTRACE/kernel/rcupreempt.c	2009-05-03 02:17:11.000000000 +0200
> @@ -1356,17 +1356,11 @@ static int rcu_sched_grace_period(void *
> 
>  		rcu_ctrlblk.sched_sleep = rcu_sched_sleeping;
>  		spin_unlock_irqrestore(&rcu_ctrlblk.schedlock, flags);
> -		ret = 0;
> +		ret = 0; /* unused */
>  		__wait_event_interruptible(rcu_ctrlblk.sched_wq,
>  			rcu_ctrlblk.sched_sleep != rcu_sched_sleeping,
>  			ret);
> 
> -		/*
> -		 * Signals would prevent us from sleeping, and we cannot
> -		 * do much with them in any case.  So flush them.
> -		 */
> -		if (ret)
> -			flush_signals(current);
>  		couldsleepnext = 0;
> 
>  	} while (!kthread_should_stop());
> 

[tip:core/rcu] rcu: rcu_sched_grace_period(): kill the bogus flush_signals()

[tip-bot for Oleg Nesterov]

Commit-ID:  41c51c98f588edcdf6141cff1895df738e03ddd4
Gitweb:     http://git.kernel.org/tip/41c51c98f588edcdf6141cff1895df738e03ddd4
Author:     Oleg Nesterov <oleg@redhat.com>
AuthorDate: Sun, 3 May 2009 23:11:18 +0200
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Tue, 5 May 2009 20:28:05 +0200

rcu: rcu_sched_grace_period(): kill the bogus flush_signals()

As a kernel thread, rcu_sched_grace_period() runs with all signals ignored.
It can never receive a signal even if it sleeps in TASK_INTERRUPTIBLE, it
needs the explicit allow_signal() to be visible for signals.

[ Impact: reduce kernel size, remove dead code ]

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
LKML-Reference: <20090503211118.GA22973@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 kernel/rcupreempt.c |    8 +-------
 1 files changed, 1 insertions(+), 7 deletions(-)

diff --git a/kernel/rcupreempt.c b/kernel/rcupreempt.c
index ce97a4d..beb0e65 100644
--- a/kernel/rcupreempt.c
+++ b/kernel/rcupreempt.c
@@ -1356,17 +1356,11 @@ static int rcu_sched_grace_period(void *arg)
 
 		rcu_ctrlblk.sched_sleep = rcu_sched_sleeping;
 		spin_unlock_irqrestore(&rcu_ctrlblk.schedlock, flags);
-		ret = 0;
+		ret = 0; /* unused */
 		__wait_event_interruptible(rcu_ctrlblk.sched_wq,
 			rcu_ctrlblk.sched_sleep != rcu_sched_sleeping,
 			ret);
 
-		/*
-		 * Signals would prevent us from sleeping, and we cannot
-		 * do much with them in any case.  So flush them.
-		 */
-		if (ret)
-			flush_signals(current);
 		couldsleepnext = 0;
 
 	} while (!kthread_should_stop());