2015-02-23 17:31:17

by Fabian Frédérick

[permalink] [raw]
Subject: [PATCH 1/1 net-next] net/atm/signaling.c: replace current->state by __set_current_state()

Use helper functions to access current->state.
Direct assignments are prone to races and therefore buggy.

Thanks to Peter Zijlstra for the exact definition of the problem.

Suggested-By: Peter Zijlstra <[email protected]>
Signed-off-by: Fabian Frederick <[email protected]>
---
net/atm/signaling.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/atm/signaling.c b/net/atm/signaling.c
index 523bce7..0140832 100644
--- a/net/atm/signaling.c
+++ b/net/atm/signaling.c
@@ -40,7 +40,7 @@ static void sigd_put_skb(struct sk_buff *skb)
pr_debug("atmsvc: waiting for signaling daemon...\n");
schedule();
}
- current->state = TASK_RUNNING;
+ __set_current_state(TASK_RUNNING);
remove_wait_queue(&sigd_sleep, &wait);
#else
if (!sigd) {
--
2.1.0


Subject: Re: [PATCH 1/1 net-next] net/atm/signaling.c: replace current->state by __set_current_state()

Instead of fixing code that never gets compiled/test, you could probably
just remove the #if WAIT_FOR_DAEMON code which would get rid of this
raw assignment.

See the #undef around line 268.

On Mon, 23 Feb 2015 18:31:07 +0100
Fabian Frederick <[email protected]> wrote:

> Use helper functions to access current->state.
> Direct assignments are prone to races and therefore buggy.
>
> Thanks to Peter Zijlstra for the exact definition of the problem.
>
> Suggested-By: Peter Zijlstra <[email protected]>
> Signed-off-by: Fabian Frederick <[email protected]>
> ---
> net/atm/signaling.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/net/atm/signaling.c b/net/atm/signaling.c
> index 523bce7..0140832 100644
> --- a/net/atm/signaling.c
> +++ b/net/atm/signaling.c
> @@ -40,7 +40,7 @@ static void sigd_put_skb(struct sk_buff *skb)
> pr_debug("atmsvc: waiting for signaling daemon...\n");
> schedule();
> }
> - current->state = TASK_RUNNING;
> + __set_current_state(TASK_RUNNING);
> remove_wait_queue(&sigd_sleep, &wait);
> #else
> if (!sigd) {

2015-02-23 22:20:44

by David Miller

[permalink] [raw]
Subject: Re: [PATCH 1/1 net-next] net/atm/signaling.c: replace current->state by __set_current_state()

From: chas williams - CONTRACTOR <[email protected]>
Date: Mon, 23 Feb 2015 12:50:25 -0500

> Instead of fixing code that never gets compiled/test, you could probably
> just remove the #if WAIT_FOR_DAEMON code which would get rid of this
> raw assignment.
>
> See the #undef around line 268.

I completely agree.