In article <8CC27249-950C-11D7-94E5-000A957650EC@wasabisystems.com>
thorpej@wasabisystems.com wrote:

> I think obj on NFS is necessary to trip the problem; the faulting 
> address on my system is within the pager_map address range, and pageout 
> to FFS vs NFS have definitely behaved differently in the past :-)

Ok, I've tried "make update" in pkgsrc/www/mozilla on nfs and got:

---
trap: TLB miss (load or instr. fetch) in kernel mode
status=0x41003, cause=0x8, epc=0x801a6af8, vaddr=0xc1fc968c
pid=1219 cmd=ld usp=0x7fffcf80 ksp=0xca707db8
Stopped in pid 1219.1 (ld) at   netbsd:bcopy+0xc8:      lb      t3,0(a1)
db> tr
bcopy+c8 (be626c22,c1fc968c,5c8,c1fc9c54) ra 8017c2b0 sz 0
8017c1f4+bc (be626c22,c1fc968c,3,c1fc9c54) ra 8017b420 sz 64
8017b32c+f4 (be626c22,c1fc968c,3,c1fc9c54) ra 8017bdd4 sz 40
snintr+10c (be626c22,c1fc968c,3,c1fc9c54) ra 801761f4 sz 32
apbus_intr_dispatch+54 (be626c22,c1fc968c,3,c1fc9c54) ra 8017ec50 sz 32
8017ec24+2c (be626c22,c1fc968c,3,c1fc9c54) ra 8017eb7c sz 24
news5000_intr+20c (be626c22,c1fc968c,3,c1fc9c54) ra 8017e93c sz 56
cpu_intr+38 (be626c22,c1fc968c,3,c1fc9c54) ra 80165478 sz 24
mips3_UserIntr+b0 (be626c22,c1fc968c,3,c1fc9c54) ra 300dbd00 sz 0
PC 0x300dbd00: not in kernel space
0+300dbd00 (be626c22,c1fc968c,3,c1fc9c54) ra 0 sz 0
User-level: pid 1219.1
db>
---

I'm not sure if it's driver's bug or not, though.
(should we consider to switch MI sonic on newsmips anyway?)
---
Izumi Tsutsui
tsutsui@ceres.dti.ne.jp