Re: RCU question

[George Anzinger <george@mvista.com>]

Zwane Mwaikambo wrote:
> On Fri, 10 Dec 2004, George Anzinger wrote:
> 
> 
>>>Well, softirqs should really be preemptible if you care about RT task
>>>latency.  Ingo's patches have had this for months.  Works great.  Maybe
>>>it's time to push it upstream.
>>
>>Yes, I understand, and soft_irq() does turn on interrupts...
>>I was thinking of something like:
>>
>>	while(softirq_pending()) {
>>		local_irq_enable();
>>		do_softirq();
>>		local_irq_disable();
>>	}
>>		<proceed to idle hlt...>
> 
> 
> But that's a deadlock and if you enable interrupts you race.

Again, I remind you we are in the idle task.  Nothing more important to do.  Or 
do you mean that softirq_pending() will NEVER return false?

The other question is: "Is useful work being done?"

-- 
George Anzinger   george@mvista.com
High-res-timers:  http://sourceforge.net/projects/high-res-timers/