Beware of revision 218075

[Marcel Moolenaar]

On Jan 29, 2011, at 3:27 PM, Nathan Whitehorn wrote:
>>> Index: intr_machdep.c
>>> ===================================================================
>>> --- intr_machdep.c      (revision 218074)
>>> +++ intr_machdep.c      (revision 218075)
>>> @@ -373,6 +426,9 @@
>>>                    i->pol != INTR_POLARITY_CONFORM)
>>>                        PIC_CONFIG(i->pic, i->intline, i->trig, i->pol);
>>> 
>>> +               if (i != NULL&&  i->pic == root_pic)
>>> +                       PIC_BIND(i->pic, i->intline, i->cpu);
>>> +
>>>                if (i->event != NULL)
>>>                        PIC_ENABLE(i->pic, i->intline, vector);
>>>        }
>> 
>> That can be reverted without harm AFAICT. One of my earlier
>> changes removed the SYSINIT to do the binding and instead do
>> it here. I partially reverted that (i.e. the SYSINIT is
>> still there, but the change above remained), because I didn't
>> know the consequence of such a change.
> I've reverted it in r218081. Everything seems OK now.

Thanks. I was just about to revert the change myself after
realizing that I may have responded in a way that would
put the burden on you.

>>> The "max IRQ = 128" thing also likely breaks Cell systems, like the PS3, where the PIC has 256 interrupts, but I haven't looked into that in detail yet.
>> This is where we need to implement multiple passes. I tried to
>> do that in the same commit, but ended up with a lot more
>> churn than I was comfortable handling. But ideally, we probe
>> busses and interrupt controllers first, so that we don't have
>> this problem -- all interrupt controllers will have registered
>> before we need to map from PIC+pin to IRQ.
>> 
>> There's probably a quick fix in case it's broken. If the pin
>> is higher than the number of IRQs we have recorded for the PIC,
>> we can simply bump the number of irqs the PIC has (i.e. extend
>> the IRQ range assigned to the PIC). Unless we have other PICs
>> recorded this can always be done -- since the PS3 only has 1
>> PIC, this should be a quick kluge to get it working again.
> 
> Ah, OK. My PS3 seems to boot fine, so don't worry about it. Good luck navigating through the rest of this minefield!

I actually think I've just passed through it. This commit was
perceived to be the most dangerous and it seems I sustained
nothing more than a bruise or scratch :-)

-- 
Marcel Moolenaar
xcllnt at mac.com