*BSD News Article 29845


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!agate!overload.lbl.gov!dog.ee.lbl.gov!ihnp4.ucsd.edu!news.service.uci.edu!nemesis.ps.uci.edu!bob
From: bob@nemesis.ps.uci.edu (bob prohaska )
Newsgroups: comp.os.386bsd.questions
Subject: panic during extract
Date: 30 Apr 1994 03:11:31 GMT
Organization: University of California, Irvine
Lines: 59
Message-ID: <2psi93$anl@news.service.uci.edu>
Reply-To: rprohask@uci.edu
NNTP-Posting-Host: nemesis.ps.uci.edu
Summary: Extract dies with a page fault
Keywords: page fault kernel panic
X-Newsreader: TIN [version 1.2 PL2]

Hi all,

Having got 1.1-gamma to run at 33 MHz on a vlb system (with lots of
help!) I'm now trying to make the system work at 50 MHz and running
into trouble:

First of all, the system somtimes crashes on the second floppy 
install disk, usually while copying files to the hard disk.
The only error message displayed so far was "out of swap space".
That seems odd, given that the kernel already knows it has
none!

Eventually the machine got past that point and through all the
reboots, whence a new problem developed:

The system panics, reporting a page fault in kernel mode, usually
while running extract:

Fatal trap 12: page fault while in kernel mode
fault virtual address = 0xf05df24c
fault code = supervisor write, page not present
instruction pointer = 0xf011b7ce
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 114 (tar)
interrupt mask = none
panic page fault

Disks are synced successfully and a crash dump is saved.

The machine came back up, ran for a while and now seems
stuck after much disk activity, though the screen still
scrolls. Restarted it thrashes the disk continuously
during startup and seems stalled now on "starting
network daemons.....inetd"

The system is a 486DX-50, 16M RAM, OPTi 495slc chipset,
256k cache running with BIOS defaults on the motherboard.

The disk controller is a bt445s rev.C, BIOS 4.72
Firmware 3.36 (BusLogic's tech support says this is latest).
It's set for IRQ 12, DMA 5, I/O 330

I've tried turning off the secondary cache, slowing the
system down with the turbo switch and low-level formatting
the hard disk, a seagate st3600n.  None fixes the problem,
though the low level format seems to help the most. The 
trick described in the KNOWNBUGS file of turning switch
10 in block 2 off didn't seem to have any effect. 

Given that all seemed well at 33 MHz I'd like to think
that this is a setup error on my part, but I've no idea
what to look for. Experience suggests that OPTi chipsets
were a poor choice for 1.0 but it seemed that 1.1 solved
the problems. Now I'm not so sure.

thanks for reading!


bob