Re: RT patch acceptance

[Andrea Arcangeli <andrea@suse.de>]

On Wed, Jun 01, 2005 at 09:39:06PM +0200, Ingo Molnar wrote:
> yes. As i said in an earlier mail:
> 
> > > (there are still some ways to introduce latencies into PREEMPT_RT, 
> > > but they are not common and we are working on ways to cover them
> > > all.)
>  
> local_irq_disable() is one way, preempt_disable() is another way. Adding 

btw, I didn't mention preempt_disable because that really is called a
pair of times in the whole drivers.

> asm("cli") to your driver is a third way. In practice these items are 
> not a big issue, so i'm not yet covering them. It's a small detail.  
> Check the amount of local-irq-disable instances in the driver space vs.  
> spinlock use.

It's not as frequent like spin_lock_irq, but it's still a relevant
driver API (unlike preempt_disable). So I think it worth fixing to
really provide the same guarantees as RTAI and rtlinux.