While running a profiling -current kernel (custom kernel)
and building a GENERIC kernel to generate statistics, I got
this:

	...
cc  -O2 -Wall -Werror -Wstrict-prototypes -Wmissing-prototypes 
-Wpointer-arith -Wno-uninitialized -Wno-main -G 0  -mno-abicalls -mno-half-pic
-EL -I. -I../../../../arch -I../../../.. -nostdinc -DMIPS1 -DMIPS3 -DLKM
-DHZ=256 -DMAXUSERS=64 -D_KERNEL -Dpmax  -c ../../../../kern/kern_xxx.c
cc  -O2 -Wall -Werror -Wstrict-protottrap: address error (load or I-fetch) in
kernel mode
status=0x2, cause=0x10002010, epc=0x80152fa4, vaddr=0x2
pid=1664 cmd=sh usp=0x7fffe6d8 ksp=0xca5e9ed8
Stopped in pid 1664 (sh) at     __mcount+0x104: lw      v1,0(a0)
db> 

Maybe just some anomaly associated with running a profiling
kernel, I don't know.  I had left the box sit idle overnight
as I was curious what sort of statistics I would see from a
long-time idle system (e.g. 8 hours).

-- Ken Wellsch