On 990538541 seconds since the Beginning of the UNIX epoch
mel kravitz wrote:
>

>FWIW;
>533au2 is really an as1200, labeled a workstation, same motherboard , and
>hardware as the as1200. Same dmesg as 1200 on boot. This is a dual 533 MHZ
>21164A-2 processors, (2) pci bus system.. siop0 scsi for cdrom ,isp0 controller
>for hard drives. In SRM -3 pci slots grouped as pci0 bus contains  a 32 bit vga
>slot , 3 pci and one isa slot grouped with  bridge as pci1.-complex mother to
>say the least. I have tried the following DEC cards with the corresponding
>results:
>1)PBXGA-AA ------> not rcognized by SRM on boot no 'graphics init' on console,
>when placed in vga (32bit slot), in pci1 slot SRM boots -->netbsd kernel
>panics.

I believe that this would be a result of our TGA code expecting
the firmware to do some of the initialization of the PBXGA-[AC]A.
On more modern machines, the SRM does not understand the PBXGA-[AC]A
cards, and so does not initialize them into a state where the tga
driver is comfortable.

The TGA2 however is initialized by the driver rather than the SRM
and is a roughly similar card, so the fix is that the driver should
be generalized to understand how to initialize the earlier model.

This will not solve the fact the firmware will not recognize the
card, but it will at least allow the cards to be used after NetBSD
boots.

This is referenced in GNATS as port-alpha/9328.  The reason that
I have not solved this is that the only PBXGA-AA's that I have are
actually built into the motherboards of my Multias and so I can't
really test code on my PWS433a.

Hmm, actually looking at the PR that I just mentioned, it appears
that your symptoms are a bit different.

>2)PBXGB-AA-------> recognized in SRM in 32bit vga slot-----------> netbsd boot
>hangs, attach fails(?)

This is weird.  Could you try booting up a 1.5 kernel and seeing
if you see the same symptoms?  It would be interesting to see
whether the problems are hardware based, or NetBSD version based.
If they turn out to be hardware based, then can you get into DDB?
We might be able to at least determine where the hang is occuring
and I could try to build you a couple of experimental kernels to
see if we can narrow it down from there.

(If you can break into DDB, then please send a trace of what is
going on to the list.)

It does sound like the au2 is a completely different machine than
those that I've tried this on.  I've seen the card work on an AS200
4/233, a PWS433a, and IIRC an XP1000.

>3)PBXCA-AA(4D20T)-----> recognized in SRM in 32bit slot----------> netbsd boot
>fails with panic.

This is something that I'll be working on shortly.  It is port-alpha/10110.

>4)4D10T (vga permidea2 chip)------> recognized in vga slot------------>
>netbsd1.5V boots dmesg given in previous e-mail.
>-Mel

 == Roland Dowdeswell                      http://www.Imrryr.ORG/~elric/  ==
 == The Unofficial NetBSD Web Pages        http://www.Imrryr.ORG/NetBSD/  ==
 == The NetBSD Project                            http://www.NetBSD.ORG/  ==