Subject: Re: drivers for 5380/9224 on KA410/KA43 (rather long)
To: None <port-vax@NetBSD.ORG>
From: Michael Sokolov <sokolov@alpha.CES.CWRU.Edu>
List: port-vax
Date: 03/03/1998 13:45:21
   Dear Allison,
   
   You wrote:
> From what I know they are the most different from each other of any 3100
> and 2000.
   
   Pretty much.
   
> Owing to the 3100m76 is not supported by ultrix I suspect it is also
> different from the rest of the 3100s.
   
   If by "the rest of the 3100s" you mean all VS3100 models and all MV3100
models lower than 30, then to some degree yes. VS3100 M30/38/40/48 uses
KA42 and MV3100 M10/10e/20/20e uses KA41. As we have learned from Antonio,
they are quite different physically, even a cursory look at either VMS or
Ultrix indicates that they are indistinguishable for software. KA43, the
system board of VS3100 M76, has minor differences from KA42/41. Yes, these
differences are minor, and people often perceive them as great only because
all previous VS3100 and MV3100 system boards had been indistinguishable for
software and because some manager has refused to allocate 30 minutes to
make Ultrix support it.
   
> Or one controller shared by both.
   
   Do you mean the external DMA controller?
   
> No surprize as the vms and ultrix engineering gang tend to get real upset
> if the same device is moved around by vax hardware engineering.  It's
> (was?) also an internal design spec point.
   
   It's very understandable why they tend to get upset! If the NCR
controller addresses were different on different system boards, Ultrix
would need many more "controller scsi<n>" lines in the GENERIC kernel
configuration file, and the way the rz and tz devices are configured makes
this very problematic.
   
> Smaller and specific is always
> better than larger more general.
   
   So you like duplicating a huge source file to change one line, don't
you?
   
   Sincerely,
   Michael Sokolov
   Phone: 440-449-0299
   ARPA Internet SMTP mail: sokolov@alpha.ces.cwru.edu