[PATCH] RESEND: fix cpus_allowed settings

[PATCH] RESEND: fix cpus_allowed settings

[Gregory Haskins]

Hi Ingo, Steve,
  I sent this patch a few weeks ago along with the migration_disable
series.  I think the controversy with the migration_disable feature
may have resulted in this fix being overlooked.  This patch is against
-rt, but the bug theoretically affects both -rt and sched-devel/mainline.
I can also whip up a sched-devel based patch if you like, but I think it
will apply trivially to both places.

Please consider it for inclusion.

Regards,
-Greg

-------------------------------
Subject: fix cpus_allowed settings

We miss setting nr_cpus_allowed for the kthread case since the normal
set_cpus_allowed() function is not used.

Signed-off-by: Gregory Haskins <ghaskins@novell.com>
---

 kernel/kthread.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/kernel/kthread.c b/kernel/kthread.c
index dcfe724..b193b47 100644
--- a/kernel/kthread.c
+++ b/kernel/kthread.c
@@ -170,6 +170,7 @@ void kthread_bind(struct task_struct *k, unsigned int cpu)
 	wait_task_inactive(k);
 	set_task_cpu(k, cpu);
 	k->cpus_allowed = cpumask_of_cpu(cpu);
+	k->nr_cpus_allowed = 1;
 }
 EXPORT_SYMBOL(kthread_bind);
 

Re: [PATCH] RESEND: fix cpus_allowed settings

[Steven Rostedt]

On Fri, 7 Mar 2008, Gregory Haskins wrote:

> Hi Ingo, Steve,
>   I sent this patch a few weeks ago along with the migration_disable
> series.  I think the controversy with the migration_disable feature
> may have resulted in this fix being overlooked.  This patch is against
> -rt, but the bug theoretically affects both -rt and sched-devel/mainline.
> I can also whip up a sched-devel based patch if you like, but I think it
> will apply trivially to both places.
>
> Please consider it for inclusion.

Thanks, I'll apply it to the -rt queue.

-- Steve

Re: [PATCH] RESEND: fix cpus_allowed settings

[Steven Rostedt]

On Fri, 7 Mar 2008, Gregory Haskins wrote:

> Hi Ingo, Steve,
>   I sent this patch a few weeks ago along with the migration_disable
> series.  I think the controversy with the migration_disable feature
> may have resulted in this fix being overlooked.  This patch is against
> -rt, but the bug theoretically affects both -rt and sched-devel/mainline.
> I can also whip up a sched-devel based patch if you like, but I think it
> will apply trivially to both places.
>
> Please consider it for inclusion.
>

Acked-by: Steven Rostedt <srostedt@redhat.com>

-- Steve