Re: powerpc/pic/intr.c IPL_HIGH vs IPL_NONE

[Matt Thomas <matt%3am-software.com@localhost>]

On Jan 3, 2013, at 4:25 AM, Frank Wille wrote:

> On Thu, 03 Jan 2013 19:59:34 +0900 (JST)
> KIYOHARA Takashi <kiyohara%kk.iij4u.or.jp@localhost> wrote:
> 
>> I will commit this change.
>> 
>> Index: intr.c
>> ===================================================================
>> RCS file: /cvsroot/src/sys/arch/powerpc/pic/intr.c,v
>> retrieving revision 1.23
>> diff -u -r1.23 intr.c
>> --- intr.c   31 Aug 2012 13:12:52 -0000      1.23
>> +++ intr.c   3 Jan 2013 10:57:03 -0000
>> @@ -213,7 +213,7 @@
>>      ih->ih_fun = ih_fun;
>>      ih->ih_arg = ih_arg;
>>      ih->ih_next = NULL;
>> -    ih->ih_ipl = ipl;
>> +    ih->ih_ipl = maxipl;
>>      ih->ih_virq = virq;
>>      *p = ih;
> 
> IMHO this is not correct, although it is a workaround for your actual problem.

Actually I think the pic should be entered with IPL_HIGH and let the lower 
priority interrupts (if any) be serviced by the pending int mechanism.  Note 
that any sources with an IPL_HIGH can be serviced directly.

> You would lose the ability that a hwirq is interrupted by the same hwirq
> with a higher priority (from a different device, sharing the same hwirq).

The above allows higher priority interrupts to interrupt lower priority ones.