Re: jmide regression on amd64/current

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Sun, Apr 13, 2008 at 09:25:16AM -0700, Scott Ellis wrote:
> After some more experimenting, it doesn't look like it's jmide related 
> at all:  The problem occurs on whatever the first probed device is.  If 
> I disable jmide and attach all 3 drives to piixide (which is still 
> ahcisata), then wd0 gets probed as ST506.  Switching which physical 
> drive is probed first doesn't seem to change anything...the first drive 
> is probed wrong.  This is with an April 2nd -current.  Reverting back to 
> my trusty Feb 11 kernel sees all the drives fine, and the system 
> operates correctly.
> 
> So maybe this is some weird interrupt routing issue that cropped up?  I 
> don't know where to look.  Ideas?

It could be a timing issue that already showed up for some peoples. I've
never been able to reproduce it, but maybe I'll can now. I'll try on my
test box in the next few days.

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--