Subject: Re: Another changer, another changer problem
To: Chris Jones <>
From: Gandhi woulda smacked you <>
List: current-users
Date: 10/01/1998 01:11:22
On 29 Sep 1998, Chris Jones wrote:

 * The hack that I mentioned worked.  I'm going to file a pr mentioning
 * it, and including the (two-line) patch I made.  But it still seems odd
 * that it's probing all 8 LUNs.  Any ideas from the peanut gallery?

Is there any specific reason it _shouldn't_ probe all 8 LUNs?

This (my next thought) seems silly, especially given that the number of
LUNs is far from sufficient to handle a modern changer, but it would
be REALLY cool if one could implement a changer driver for a
six-or-fewer changer such that it would recognize each tray as a LUN.
The disadvantage is that our kernel doesn't (yet) do automagic device
detection past the first kernel autoconfig at boot time, i.e., it
doesn't handle auto-offlining of a non-responsive device...

Anyway, back to LUNs.  It seems plausible enough that a SCSI controller
_could_ have 8 LUNs attached to it.

I wish I had a SCSI controller capable of recognising LUNs, actually
-- it would octuple my current drive address space; considering that
a) I'm looking into a CD-R sometime RSN; b) my SCSI bus is full up; and
c) a Sbus FSB/E card is still unreasonably expensive, this would be a
very nice thing to have.

But I digress...

 * Chris


I will not type stupid letters at 1 AM.
I will not type stupid letters at 1 AM.
I will not type stupid letters at 1 AM