Subject: kern/8883: Tekram DC-390 lossage
To: None <gnats-bugs@gnats.netbsd.org>
From: None <nemo@red-bean.com>
List: netbsd-bugs
Date: 11/26/1999 09:27:39
>Number:         8883
>Category:       kern
>Synopsis:       1.4.1 GENERIC crashes with Tekram DC-390 installed
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Nov 26 09:27:00 1999
>Last-Modified:
>Originator:     Joel N. Weber II
>Organization:
Red Bean Software/Gratuitous Organization for Gratuitous Software Enhancement
>Release:        1.4.1; and 1.4 branch from cvs of about a week ago
>Environment:
System: NetBSD duality 1.4.1 NetBSD 1.4.1 (XANTHINE) #0: Thu Nov 25 14:00:04 EST 1999 nemo@duality:/usr/src/syssrc/sys/arch/i386/compile/XANTHINE i386


>Description:
With the 1.4.1 generic i386 kernel, my machine consistently crashes iff my
Tekram DC-390 is installed.  It crashes long after the machine probes the SCSI
controller and disks; when it crashes, the last message printed on the console
is one announcing that the parallel port exists.  When it does not crash
(such as when the DC-390 is removed), the next thing to be printed after
the parallel port is the announcement that, gee, this is a PC that happens
to have a keyboard.

Given where it crashes, I suspect that the machine is crashing in some ISA
device probe that doesn't get along with the DC-390 very well.  It seems
to me that the fact that netbsd has no useful verbosity option that will
announce each probe right before it is attempted is a serious bug; if there
were an appropriate verbose probing option, I could probably more easily
tell exactly where the system is crashing.

I rebuilt a kernel from the 1.4 branch in CVS, stripping all ISA hardware
that I knew I didn't have, adding RAIDframe support, etc.  That kernel
does not crash with the DC-390 installed.  I do not know whether removing
the ISA probes, or whether some bugfix made since 1.4.1, is what fixed
the problem.
>How-To-Repeat:
>Fix:
>Audit-Trail:
>Unformatted: