Very Computer

My Solaris 2.1 x86 set-up suffers from kernel panics about six hours after
boot, which occur frequently after the first one.

Can anyone tell me what the error message indicates are likely canditates for
the problem.

Other than the base Solaris 2.1, the only thing I am running is INN 1.4 and
drivers for a Digiboard PCX/em

The messade is

BAD TRAP
sh: Page fault
Kernel fault at addr=0xc664 pte=0x0
... ...

A few thoughts on the matter...

1) Do you have any floppies mounted? Floppies tend to be "a bad thing"
   on 2.1 for x86, and there doesn't seem to be any set pattern
   to when they cause system panics.  There is also a possiblity
   it is pcfs, however it seems more related to floppies.

   (mtools supposidly is a better method of accessing floppies)

2) Are you using Virtual Terminals? If you have all of them allocated
   and you try allocating another one, the system will guarenteed panic.

3) Is your motherboard a solid one? You can get some severe weirdness
   if it isn't.

Those are the only reasons I've heard/seen that cause x86 to panic
reliably.

David Holland

--
David Holland

Tonights forcast: Increasing dark, with dark continuing through out the night.

: A few thoughts on the matter...

: 1) Do you have any floppies mounted? Floppies tend to be "a bad thing"
:    on 2.1 for x86, and there doesn't seem to be any set pattern
:    to when they cause system panics.  There is also a possiblity
:    it is pcfs, however it seems more related to floppies.

:    (mtools supposidly is a better method of accessing floppies)

: Those are the only reasons I've heard/seen that cause x86 to panic
: reliably.

I have a x86 kernel panic that I can faithfully reproduce on-demand.
It is, or may be, somewhat related.

I have a 1.3 GB SCSI disk, where the first 300 MB is a DOS partition
and the rest is for Solaris 2.1 x86.  You are supposed to be able to
mount and access the "DOS C: Drive" from Solaris.  

Mounting works...err sort of.  The mount command executes successfully
but anytime I try to access the data on the DOS partition, I get a
kernel panic.  ls(1) works fine, but cat(1) and more(1), etc. (it seems
anything that graps data blocks) are sure fire way to get a panic.

I reported this to Sun, and they have been able to reproduce it, but no
word as to when it will be fixed.

In as much as the subject relates to PCFS I thought I'd throw this in.