Chris Torek paid attention and whapped me with a newspaper- I putzed
my patch about slave-only because I forgot that:

a) it's a root node property

b) it's a mask of sbus slots that are slave-only.

Actually, Chris was much nicer than that, but I'm still an idiot.
Here's a better patch, but I still haven't been able to test it
against an SS1 or SS1+:

Index: sbus.c
===================================================================
RCS file: /cvsroot/src/sys/arch/sparc/dev/sbus.c,v
retrieving revision 1.17
diff -c -r1.17 sbus.c
*** sbus.c	1997/06/01 22:10:39	1.17
--- sbus.c	1998/10/20 17:06:51
***************
*** 85,94 ****
--- 85,100 ----
  	void *args;
  	const char *sbus;
  {
+ 	static char *sl = "slave-only";
  	register struct confargs *ca = args;
  
  	if (sbus)
  		printf("%s at %s", ca->ca_ra.ra_name, sbus);
+ 	/*
+ 	 * Check root node for 'slave-only' property
+ 	 */
+ 	if (getpropint(0, sl, 0) & (1 << ca->ca_slot))
+ 		printf(" %s", sl);
  	printf(" slot %d offset 0x%x", ca->ca_slot, ca->ca_offset);
  	return (UNCONF);
  }