> >> pci based one in the tree... I'm just unpacking the last -current
> >> tar_files I have (Jun27) to see what might be there...
> >
> >Is the puc driver from -current a good starting point?
> 
> Actually .... yes, because judging from the comments from pucdata.c
> it already knows how to handle the PLX 9050 chip.

Yup.  All the cards I tried (after I peeled the Windows Plug & Pray
sticker off the big chip) had PLX 9050s or 9052(?)s as the PCI interface.


> I bet you could
> just write a description for pucdata.c and it should "just work"
> (finding out what you need to put there might be interesting,
> though).

Yah.

I suggest:

(1) changing the #if 0's to #if 1's in dev/pci/pci.c

(2) expanding your kernel message buffer size (options MSGBUFSIZE=...)

(3) booting with the card in the box, and sending me the output to
grovel over.

(I think I mentioned those in my post to current-users about the
driver.  8-)

It should be fairly obvious what needs to be done, but it sounds like
this card will be slightly different than the others that i've seen:
on the cards that I've seen, each BAR/region on the PLX mapped to
exactly one chip ('com' or 'lpt').  The 'puc' driver may have to
change slightly to accomodate this new board, but that shouldn't be
hard.



cgd