Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!rice!sun-spots-request
From: igb@vlsi1.vlsi.columbia.edu (Isidore G. Bendrihem)
Newsgroups: comp.sys.sun
Subject: Re: Problem with SS1
Keywords: SunOS
Message-ID: <6591@brazos.Rice.edu>
Date: 11 Apr 90 15:01:37 GMT
Sender: root@rice.edu
Organization: Sun-Spots
Lines: 34
Approved: Sun-Spots@rice.edu
X-Refs:  Original: v9n116, Replies: v9n117
X-Sun-Spots-Digest: Volume 9, Issue 119, message 3

Thanks to all the people who replied to my "bad silo" problem with one of
our SparcStation.  Many people seem to have this problem. Some of them
have had their motherboards and mice replaced without luck.  According to
a SUN representative, the problem seems to be on 4.0.3c.  This is the
response I got from them:

If you have a support contract, contact the USAC via email at
hotline@sun.com or by calling 1-800-USA-4SUN.

There is a known bug in SunOS 4.0.3c which can result in data fault
panics. The problem deals with an interrupt being received from a serial
port while the system is doing a fork(). If the timing is right, you can
get a data fault panic. Since any shell does a lot of forking...

The backtrace typically looks like:

    _panic(0xf80cb913,0x1e84800,0x1,0xf80cb7b8,0xfffff,0x0) + 7c
    _trap(0x9,0xffffec74,0xff640000,0x80,0x1,0xffffffff) + 180
    fault(0x1,0xfffffff2,0xffffc000,0xffffed20,0xf0,0x1) + 64
    level6(?)
    _bcopy(0xffffc000,0xff640000,0x38a0,0xffffe6a0,0xf810dff8,0x0) + c
    _procdup(0xf810b144,0x94,0x49,0x0,0xffffffff,0xff108c9c) + 134
    _newproc(0x0,0x50,0x14,0xf810ab7c,0xf8108668,0xf810b144) + 494
    _fork1(0x0,0xc,0x1,0xff13e8fc,0xf810ab7c,0xf810b144) + 1a0
    _fork(0xffffefe0,0xf80be528,0xf80be528,0x0,0xffffefb4,0xffffefe0) + 4
    _syscall(0x8000000) + 2d0

Note that you can only get the bug fix if you have a support contract.
Considering how many people seem to be affected by this, I think SUN
should make the fix freely available by posting it to the Rice U.
archives.

Isidore Bendrihem
igb@vlsi.columbia.edu