Re: preempt_disable and sleep()

From: Paul E. McKenney
Date: Thu May 13 2010 - 23:37:40 EST


On Fri, May 14, 2010 at 08:16:07AM +0530, ratheesh k wrote:
> On Fri, May 14, 2010 at 12:13 AM, Sven-Thorsten Dietrich
> <thebigcorporation@xxxxxxxxx> wrote:
>
> > Why would you do this?
> I was reading some linux kernel journel about preempt_disable() . This
> a doubt came to me .
>
>
> > If you had tried this code, you would have seen the corresponding stack
> > dump on the console.
>
>
> Why it so .? could you explain a little more ? .sleep() will schedule
> () , which inturn give back control to kernel .

Sleeping in the kernel with preemption disabled is considered to be a
bug. So the scheduler will print an error and a stack dump when this
happens.

In contrast, it is OK to do the following:

preempt_disable();
do_something();
preempt_enable();
schedule();
preempt_disable();
do_something_else();
preempt_enable();

Thanx, Paul

> Thank,
> Ratheesh
>
> On Fri, May 14, 2010 at 12:13 AM, Sven-Thorsten Dietrich
> <thebigcorporation@xxxxxxxxx> wrote:
> > On Fri, 2010-05-14 at 00:02 +0530, ratheesh k wrote:
> >> I am running RT linux .
> >>
> >> /* code snippet */
> >>
> >> preempt_disable()
> >> ...........
> >> sleep(10);
> >> '............
> >> preempt_enable()
> >>
> >>
> >> will this sleep() will cause reschedule ?.
> >>
> >
> > Why would you do this?
> >
> > If you had tried this code, you would have seen the corresponding stack
> > dump on the console.
> >
> >> if an interrupt comes  while premption is disabled , how it is handled ?.
> >>
> >
> > Depends on the number of (idle) cores, the setting of CONFIG_PREEMPT_[]
> > and the type of interrupt.
> >
> > RT threaded IRQs would not run until preemption is enabled on a single
> > core system.
> >
> >
> >
> >> Thanks,
> >> Ratheesh
> >> --
> >> To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
> >> the body of a message to majordomo@xxxxxxxxxxxxxxx
> >> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >
> >
> >
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/