> So that's in the middle of the PROM's `v2_open' code. What can it
> possibly not like about the argument ("/obio/interrupt") given?

I'm pretty sure it doesn't like the way the mmu is set up somehow.

> Can you verify that that these PROM nodes exist?

/obio/interrupt exists as far as show-devs is concerned, and the same
kernel works if I replace the hypersparc module with a supersparc module.

> What happens if you use the fall-back code, ie. just have opennode()
> return `0' always.

It now gets a kernel fault, at the first call to setpte4m:

data fault: pc=0xf811d624 addr=0xffeca400 sfsr=326<FAV>
panic: kernel fault
Stopped at ...
db> tr
_mem_access_fault4m(0x9, 0x326, 0xffeca400, 0xffeca000, 0xf811d624, 0xf8151c90)
	at _mem_access_fault4m+0x310
memfault_sun4m(0x0, 0x4b, 0xf8002000, 0xfff, 0xf8172000, 0x0)
	at mwmfault_sun4m+0x140
_getprop(0xfe014000, 0xf811d5cc, 0xf8151f00, 0x40, 0x10, 0xc)
	at _getprop+0x8c
_setpte4m(0xfe014000, 0xff14007e, 0xff14007e, 0x0, 0x0, 0x8)
	at _setpte4m+0xc
_bootstrap(...) at _bootstrap+0x200
noplab(...) at noplab+0xc4

--Brian;