> 
> Worked just fine with 1.3.2, I believe it was. May have been one of the
> custom
> kernels released between 1.3.1 and 1.3.3. Yes, the boot code didn't
> work, but the
> kernel did. Without a hitch. The QDSS would show the normal system
> tests, the
> 1..0.., then the boot loader would go to the serial console, then the
> kernel would
> switch right back to the QDSS. KA650-A v5.3, VMB 2.7, in a BA123 (cause
>
Ok, that make more sense :-) But: (looking in the CVS log files)
I disabled the KA650 QDSS support 1998/06/20 because it caused trouble
with some KA650's. The main reason here was that a real fix for QDSS
should be done for KA650's, and not the hack as before.
If you want to build a kernel with QDSS enabled as console on KA650, just
uncomment the "     } else /* if (vax_boardtype != VAX_BTYP_650) */"
around line 414 in uba/qd.c. Please remember that you _must_ have KA630
compiled into the kernel also.
Another problem may be the 4.3BSD X server. I assume that it won't run
anymore. Unfortuantely I don't have any QDSS to test all this on, otherwise
it would be an easy task to make it work correctly.

> > The mono graphics on 3100 should work just fine as console _but_ you may
> > _not_ have the diag switch at the rear set to "serial console".
> >
> > -- Ragge
> 
> When you say that, do you mean that everything, including the bootloader
> (boot.mopformat, from the 120198 snapshot) will also send it's info to
> the
> graphics console (again, with the console switch set to "graphics")?
> 
The boot loader have no idea about where the console is at all. It just
gets a pointer to a prom routine that does the console in/out. So, if 
the prom decides that the glass tty is console, the boot loader will use
it also.

-- Ragge