"Michael L. Hitch" wrote:

> On Thu, 24 Feb 2000, Simon Burge wrote:
> 
> > FWIW, I've always used
> > 
> > 	boot -f rz()netbsd
> > 
> > on my 3100 and 5100 for kernel testing (less typing!) since the
> > controller, device and partition all default to 0.  I haven't tried
> > either of these machines since Toru's recent makebootdev() changes.
> 
>   That will fail with the latest changes.
> 
>   Are the forms of device specification allowed by Ultrix, and (more
> importantly) do we want to still support them?

I can't comment on Ultrix (I don't have Ultrix installed on the
disks on any 3100s or 5100s around here), but the "DECstation
2100/3100 Operator's Guide" says that the controller, unit-number
and partition-number are all optional, and in the given examples
there's no leading slash on any filenames.  From that manual:

	"The file parameter has the format

	dev([controller][,unit-number][,partition-number])[filename]"

Perhaps that last line should really be:

	dev([controller[,unit-number[,partition-number]]])[filename]"

I just tried "boot -f rz(,,)netbsd" and that works ok too at the prom,
but the kernel on that disk has a md root dev so I don't know if the
kernel passed the file specifier ok.  Certainly the first and second
stage bootblocks didn't have a problem with it.

If it counts, my fingers would like to support the "rz()netbsd" style
since that's what they are used to :-)

Simon.

> 
> --
> Michael L. Hitch			mhitch@montana.edu
> Computer Consultant
> Information Technology Center
> Montana State University	Bozeman, MT	USA
>