In article <Pine.LXG.3.96.990716030029.14766A-100000@kano.cs.utwente.nl>,
Reinoud Zandijk <zandijk@cs.utwente.nl> wrote:
>Hi y'all,
>
>I'm typing to install NetBSD 1.4 on a SPARCstation SLC by netbooting it
>from a Acorn RiscPC, also running NetBSD... Since I was stubborn, I
>managed to get it to boot the NetBSD kernel before I discovered there was
>a `man diskless' (ahum) 
>
>But.... it starts OK, until the it probes `fdc0'. The output is :
>"
>  fdc0 at mainbus0 ioaddr 0xf7200000 ipl 11data fault: pc=0xf0157214
>  addr=0xffd18000 ser=80<INVAL>
>  panic: kernel fault
>  halted
>
>Since a SLC doesn't have a floppy.... it's seems fare to fail, but why
>panic??? Does it require a kernel flag? I both tried the `netbsd.GENERIC'
>and the `netbsd.GENERIC_SCSI3' images. 
>
>Any help would be grateful, since I can't build my own kernel yet! (well,
>how to install a croscompiling gcc :-+ , and is it desireable ).

Well, this bug has been there for a while... I can build a kernel for
you without the fd driver, but it will be elf and /current.
Looks like you die in the line:

 if (*fdc->sc_reg_dor == NE7_RQM) {

Maybe we can use fkbyte() to catch it like the vme driver does?

christos


christos