Subject: Re: Feb 12 wd.c: lost interrupt - ouch!
To: Dirk Steinberg <>
From: Eric S. Hvozda <>
List: current-users
Date: 02/22/1994 14:22:22
Note this is not a gripe or flame.  It is a statement of a way to start 
attacking the wd.c problem in -current.  NetBSD is a great thing (tm) and I'm 
thankful I had something to leave Linux for...

Dirk Steinberg sez:
> I just want to make other people with IDE drives aware of the problem
> so they know what to look after; maybe *someone* can narrow this bug
> down and supply a fix.

Indeed, we need to start tracking this kind of thing for wd.c.  Perhaps we 
should start with a list of people who have IDE disks working under 0.9 and
current and what the configuration is (who made the disk, where does 
everything live?  root on the IDE?  what about swap? etc...)

If we can identify which disks and configurations work and don't work,
perhaps a basic set of instances can be established for us to start
attacking this problem from.

Example (from me of course :-)

Quantum LP240A
works great

Maxtor 7245AT
works great

Maxtor 7245AT and Quantum LP240A 
root and swap on both, root and swap only on either one
I got lost of stray ISA INTR 7 with this setup, only when the LP240 was
the master.  No problems when it was the slave. wd.c patch made this setup
totally unusable.

Maxtor 7345AT
works great

Maxtor 7245AT and Maxtor 7345AT
root and swap on the 7345AT; only boot blocks on 7245AT.  Works great, no
more stray ISA INTR 7 at all.  It was almost as if the Quantum disk was the
problem to start with.  Still no problems if master and slave are switched.
went back to orginal wd.c; 7245AT is the master.

All of this is with 0.9.  I'm hard up for disk space, and very afraid to 
upgrade to current since I only have IDE disks.  The supposed wd.c patch
created more problems for me than it fixed for me.  I got rid of my problems by
junking the Quantum disk and going back to the original wd.c.

Now I'm willing to keep a database of all people with IDE disks who send me
mail and describe what they have (basically like above).  If some developer
would step forward and like to take a look, or give pointers of where to
start to look, I'm sure it would be appricated.  I'd do it except my machine
is at work and I cannot afford to have it down right now.  Also, it may be
fruitless to attack this from 0.9.  Perhaps I'll be able to do this in the
future, but Uncle Sam wants his taxes :-)