Subject: 2.0 install problems on Proliant1600
To: None <>
From: Peter Hufnagel <>
List: netbsd-users
Date: 02/24/2005 13:24:52
I'm successfully running 2.0 (ISO from
on a single CPU box, and an old dual P200 box I had
around, so I figured I'd try it on my Proliant 1600
box...  (dual 600Mhz, 5x4GB drives--for now ;-) ).

It partitions the disk fine, starts extracting the
tarballs, and hits "man.tgz" and pauses about 50%
through, "--stalled--", then picks up again, continues
on, occasionally stalling here and there for 15-20 secs,
and then finally dies (while stalled) in xfont.tgz.
(this is all using sysinst).

Stopped in PID 1435.1 (tar) at 0xc028519f:  repe stosb %es:(%edi)

db>  trace

?(cc083cc4, c03955c0, cd0f99f4, 32000, 0) at 0xc028519f
?(cc0f99f4, 32000, 0, cc083d20, cc083d60) at 0xc0283e34
?(cd0f99f4, 32000, 0, 283, 0) at 0xc0236bf1

.. (excuse it being clipped, but I copied it by hand...)
if I continue it...

db>  cont

the extract fails with "bad header", goes on to make device
nodes, and then Seg-Faults in sysinst.  After that, any cmd

#ls -l
uvm_fault(0xcbf85370,0,0,0,1) -> 0xe
kernel: page fault trap, code=0

Anyone have any clues here?  I read some posts on possible
issues with the smart controllers, but this box doesn't have
one, its just straight scsi bus disks.  I also read a post
on some user having issues on a PL1850, and having to setup
the APIC to a different setting.. but running the smartstart
setup (5.30, from well after the P8 - 6/2/2000 bios revision)
doesnt seem to have any options for APIC.  I'm presuming
maybe it has something to do with dual-cpu and locking...
does anyone have 2.0 running on a PL1600 ?

I'll see if I can get dmsg output tonight.