> >I hope that GENERIC configuration has the following options.
> >(i.e. almost all PCIBIOS options enabled.)
> 
> Err, the GENERIC configuration has all of the options commented out.
> They don't indicate why they're commend out, so one can only assume
> they are commented out because of instability or bugs (are there open
> PRs?).

PCIBIOS_INTR_FIXUP was disabled because of instability.
Before recent commitment, both fvdl's machine and my machine panics
with PCIBIOS_INTR_FIXUP.
I think the stability becomes better, now.
But apparently we needs more test. (and possibly more fixes.)

I'm not sure why PCIBIOS_BUS_FIXUP is also disabled.
From my limited knowledge, PCIBIOS_BUS_FIXUP doesn't cause any
problem. (But I don't have experience on a machine which has
multiple PCI buses.)
Does anyone know any problem with PCIBIOS_BUS_FIXUP? 

> Incidently, it would be nice to understand whether you were attempting
> to have this ready for 1.5? I think it'd be nice, but I don't have a
> good feeling for the stability.

Yup, I'm suspicious about stability, too.
What I meant is that I'd like to make those options stable enough
for GENERIC kernel of 1.5.
(And I suppose PCIBIOS_BUS_FIXUP is already OK.)

> >Uchiyama-san said PCIBIOS_ADDR_FIXUP may still have problem with
> >pccbb (cardbus bridge), so probably it is a bit early to to enable 
> >this for now.
> 
> Well, GENERIC does not have cardbus support right now. Perhaps
> enabling it in GENERIC but not in CARDBUS?

I'd like to make those options enabled in both GENERIC and CARDBUS
when 1.5 is released.
But of course, we should not enable those options before enough test.

> >(Please correct if there is incorrect or misleading description,
> > also please correct mistakes about English expression.)
> 
> OK. I don't have the best understanding of the pcibios code,
> but I will review the text for clarity and English now.

Thanks. I'll look at your fixes.
--
soda