On Wed, 18 Feb 1998, Dave Huang wrote:

> On Wed, 18 Feb 1998, Dave Huang wrote:
> > A kernel built from my normal config crashes with a vm_fault() after
> > probing mc0 (MACE internal ethernet). Normally, it'd probe esp0 after
> > that... The ddb stack trace says something about config_search() ->
> > mapply() -> faultstkadj(), then trap/panic stuff.
> 
> Hmmm.... how weird and freaky. My normal kernel config crashes, but if I
> just take out "options NETATALK", the kernel works. (And a kernel
> without mc0 or netatalk hangs early on; a kernel without mc0 but with
> netatalk won't even link). All of this seems too flaky for it to really
> be netatalk's fault... any chance the recent pmap changes introduced a
> bug? 

This is with a make clean between them?

Do you have any other network interfaces besides mc0? NETATALK probably
implicitly assumes there's a network interface around...

Take care,

Bill