Michael Wolfson wrote:
> 
> I just went through the PR database for macppc and ran across a few
> with could probably be fixed easily enough.

good to check out, indeed. SOme are still marked 1.4.2 :)

>   - Built-in mc0 ethernet does not work (gives timeout errors) on Apple
>     Network Server
>     http://www.netbsd.org/cgi-bin/query-pr-single.pl?number=20608
mc0 seems to be a trouble child... Just today someone reported that his
7300 drops packets and my 9500 has an unusable mc0 too. I wonder if this
patch is ANS specific  ?
 
> The following involve judgement calls, but would be easy to fix (or at
> least close the PR if the decision is to not change anything):
>   - SCSI bus numbering from OF is inappropriate
>     http://www.netbsd.org/cgi-bin/query-pr-single.pl?number=10678
>     this should be a simple matter of moving mesh before the other
> scsibus devices

I fundamentally agree with the guy here. But it is IMHO "tricky".
For example server machines, like some g3's... or upgraded boxes may
vae  an UW scsi card and thus totally ignore the internal slower bus.
The most complicated case is where you have an additional card on a
computer with 2 internal busses... Although I too think that the
internal onboard bus should be scanned first. Since if the box has an
additional card it is either totally unused or has, maybe, attached only
the boot cdrom and thus unlikely to change.

This behavior should be consistent with other architecture/ports, IMHO.
I would get rather confused to see a 68k mac act differently than a ppc
than a sparc...
 
>   - Mouse does not work
>     http://www.netbsd.org/cgi-bin/query-pr-single.pl?number=10240

I never got so far to get X working, so I never noticed and AFAIK we
dont't have something equivalent to framebuffer console and dt on ppc?
(I have those on 68k)

 
my 2 cents...

Riccardo