Re: [PATCH] RFC: Set irq thread to RT priority on creation

[Thomas Gleixner]

On Thu, 30 May 2013, Steven Rostedt wrote:

> On Thu, 2013-05-30 at 16:07 +0200, Thomas Gleixner wrote:
> 
> > > This patch solves an issue for me where a device driver is expected to handle an
> > > interrupt immediatly after irq handlers are installed and interrupts enabled.
> > 
> > You miss to explain what kind of issue that is.
> 
> I could envision the case where the interrupt is initialized but doesn't
> go off until much later. If it never ran, then it would still be in
> SCHED_OTHER(), and that first interrupt could have a large delay.

Nope. As Ivo explained it's about an interrupt coming in right away,
i.e. before __setup_irq() reaches:

        if (new->thread)
	   wake_up_process(new->thread);

The ones which come much later do not have that issue as the thread
code already sits in the waiting loop and already adjusted the
priority.

Thanks,

	tglx

--
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/