> > > Having set up dhcpd to serve bootp info, the bootloader then proceeded
> > > to boot the generic kernel from the 200020628 snapshot from
> > > releng.netbsd.org, but after loading the kernel image, the machine
> > > hangs after "entry point XXXXXX"
> > 
> > Sageev George <sageev@bmsrs.usc.edu> reported a similar problem
> > on his 425t back in March.  I think the console is on dca0.
> > 
> > You're not using a serial console?
> 
> That's right, I'm using a Catseye/Topcat 1280x1024/8bit Framebuffer.
> 
> My first suspicion was that it continues booting, but just fails to
> initialize the framebuffer but I have only just now come around to
> connecting a serial console. It does in fact boot the kernel, and then
> complains (rightly) about unconfigured rc.conf and waits for me to
> give it a shell.
> 
> Why it fails to initialize the framebuffer I do not know. But now I at
> least have a console and can build a custom kernel and see how it
> goes.

dcacnattach() tries to test where the console is:

        switch (dca->dca_id) {
#ifdef CONSCODE
        case DCAID0:
        case DCAID1:
#endif
        case DCAREMID0:
        case DCAREMID1:
                break;
        default:
                bus_space_unmap(bst, bsh, DIOCSIZE);
                return (1);
        }

and is getting it wrong on the 400 series.  I've stared at the code
now, and I can't see why it worked before and not now.

> I will see how a custom-made kernel fares.

Try overriding your console with CONSCODE=133 in your kernel
config.
 
> Everything else seems to work fine.

Except the real-time clock... yeah?

	-- Gregory McGarry <g.mcgarry@ieee.org>