[PATCH] sched: remove noop in pre_schedule_rt()

[PATCH] sched: remove noop in pre_schedule_rt()

[Hillf Danton]

Checking task to be RT is removed since only RT task is served here.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
---

--- tip-git/kernel/sched_rt.c	Sun May 22 20:12:01 2011
+++ sched_rt.c	Thu May 26 21:11:19 2011
@@ -1516,7 +1516,7 @@ skip:
 static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
 {
 	/* Try to pull RT tasks here if we lower this rq's prio */
-	if (unlikely(rt_task(prev)) && rq->rt.highest_prio.curr > prev->prio)
+	if (rq->rt.highest_prio.curr > prev->prio)
 		pull_rt_task(rq);
 }

Re: [PATCH] sched: remove noop in pre_schedule_rt()

[Yong Zhang]

On Thu, May 26, 2011 at 9:08 PM, Hillf Danton <dhillf@gmail.com> wrote:
> Checking task to be RT is removed since only RT task is served here.
>
> Signed-off-by: Hillf Danton <dhillf@gmail.com>

Actually this patch has been raised for two times:
https://lkml.org/lkml/2010/12/13/622
http://lkml.org/lkml/2010/2/9/139

Thanks,
Yong

> ---
>
> --- tip-git/kernel/sched_rt.c   Sun May 22 20:12:01 2011
> +++ sched_rt.c  Thu May 26 21:11:19 2011
> @@ -1516,7 +1516,7 @@ skip:
>  static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
>  {
>        /* Try to pull RT tasks here if we lower this rq's prio */
> -       if (unlikely(rt_task(prev)) && rq->rt.highest_prio.curr > prev->prio)
> +       if (rq->rt.highest_prio.curr > prev->prio)
>                pull_rt_task(rq);
>  }
>



-- 
Only stand for myself

Re: [PATCH] sched: remove noop in pre_schedule_rt()

[Steven Rostedt]

On Fri, 2011-05-27 at 17:17 +0800, Yong Zhang wrote:
> On Thu, May 26, 2011 at 9:08 PM, Hillf Danton <dhillf@gmail.com> wrote:
> > Checking task to be RT is removed since only RT task is served here.
> >
> > Signed-off-by: Hillf Danton <dhillf@gmail.com>
> 
> Actually this patch has been raised for two times:
> https://lkml.org/lkml/2010/12/13/622
> http://lkml.org/lkml/2010/2/9/139
> 

Heh, and that's still sitting in my repo on kernel.org. Never been
pulled. I'll ping Ingo about that for the next merge window.

Thanks! I forgot about that pull request.

-- Steve

Re: [PATCH] sched: remove noop in pre_schedule_rt()

[Hillf Danton]

On Fri, May 27, 2011 at 5:17 PM, Yong Zhang <yong.zhang0@gmail.com> wrote:
> On Thu, May 26, 2011 at 9:08 PM, Hillf Danton <dhillf@gmail.com> wrote:
>> Checking task to be RT is removed since only RT task is served here.
>>
>> Signed-off-by: Hillf Danton <dhillf@gmail.com>
>
> Actually this patch has been raised for two times:
> https://lkml.org/lkml/2010/12/13/622
> http://lkml.org/lkml/2010/2/9/139
>

Hm... looks curious why it had been delivered more than once.

Re: [PATCH] sched: remove noop in pre_schedule_rt()

[Steven Rostedt]

On Mon, 2011-06-06 at 16:35 +0800, Hillf Danton wrote:
> On Fri, May 27, 2011 at 5:17 PM, Yong Zhang <yong.zhang0@gmail.com> wrote:
> > On Thu, May 26, 2011 at 9:08 PM, Hillf Danton <dhillf@gmail.com> wrote:
> >> Checking task to be RT is removed since only RT task is served here.
> >>
> >> Signed-off-by: Hillf Danton <dhillf@gmail.com>
> >
> > Actually this patch has been raised for two times:
> > https://lkml.org/lkml/2010/12/13/622
> > http://lkml.org/lkml/2010/2/9/139
> >
> 
> Hm... looks curious why it had been delivered more than once.

What's curious about it?

Yong sent me the patch, and I pulled it into git, and sent a pull
request to Ingo along with some trivial changes of mine. All my pull
requests are posted back to LKML so that we keep an open environment of
what goes into the kernel.

-- Steve

Re: [PATCH] sched: remove noop in pre_schedule_rt()

[Hillf Danton]

On Mon, Jun 6, 2011 at 7:26 PM, Steven Rostedt <rostedt@goodmis.org> wrote:
> On Mon, 2011-06-06 at 16:35 +0800, Hillf Danton wrote:
>> On Fri, May 27, 2011 at 5:17 PM, Yong Zhang <yong.zhang0@gmail.com> wrote:
>> > On Thu, May 26, 2011 at 9:08 PM, Hillf Danton <dhillf@gmail.com> wrote:
>> >> Checking task to be RT is removed since only RT task is served here.
>> >>
>> >> Signed-off-by: Hillf Danton <dhillf@gmail.com>
>> >
>> > Actually this patch has been raised for two times:
>> > https://lkml.org/lkml/2010/12/13/622
>> > http://lkml.org/lkml/2010/2/9/139
>> >
>>
>> Hm... looks curious why it had been delivered more than once.
>
> What's curious about it?
>

Hi Steve

The work was repeated not only by Yong, you see, almost
impossible, which shows RT scheduling was and is
attracting minds, and will be as well, right?

thanks
          Hillf

Re: [PATCH] sched: remove noop in pre_schedule_rt()

[Steven Rostedt]

On Mon, 2011-06-06 at 21:14 +0800, Hillf Danton wrote:

> The work was repeated not only by Yong, you see, almost
> impossible, which shows RT scheduling was and is
> attracting minds, and will be as well, right?

It was an obvious fix, hardly impossible for it to be done twice. This
happens all the time in the kernel. Especially if for some reason a fix
doesn't make it in. This fix didn't make it in simply because it wasn't
one of my normal branches I push to Ingo, and it was at a time that we
all were quite busy and it was simply overlooked.

But yes, the RT scheduler (and the scheduler itself) always attracts
newcomers, and we welcome them. Nothing curious or unusually about it.

-- Steve