From owner-freebsd-questions  Fri May 10 07:50:43 1996
Return-Path: owner-questions
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id HAA23426
          for questions-outgoing; Fri, 10 May 1996 07:50:43 -0700 (PDT)
Received: from who.cdrom.com (who.cdrom.com [204.216.27.3])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id HAA23410
          for <questions@FreeBSD.ORG>; Fri, 10 May 1996 07:50:39 -0700 (PDT)
Received: from hda.com (ip38-max1-fitch.zipnet.net [199.232.245.38])
          by who.cdrom.com (8.6.12/8.6.11) with ESMTP id HAA28219
          for <questions@freebsd.org>; Fri, 10 May 1996 07:24:49 -0700
Received: (from dufault@localhost) by hda.com (8.6.11/8.6.9) id KAA28289; Fri, 10 May 1996 10:29:22 -0400
From: Peter Dufault <dufault@hda.com>
Message-Id: <199605101429.KAA28289@hda.com>
Subject: Re: MO disk support ?
To: luigi@labinfo.iet.unipi.it (Luigi Rizzo)
Date: Fri, 10 May 1996 10:29:22 -0400 (EDT)
Cc: questions@FreeBSD.ORG
In-Reply-To: <199605091841.UAA02564@labinfo.iet.unipi.it> from "Luigi Rizzo" at May 9, 96 08:41:18 pm
Reply-to: hdalog@zipnet.net
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Sender: owner-questions@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

> 
> Hi,
> I have a Sony MO SCSI disk (~120MB) which I'd like to use. Back in 1.1.5
> times, I just modified scsiconf.c by adding the following entry
> 
>         {
>                 T_OPTICAL, T_REMOV, "SONY", "SMO-F331-00"
>                     ,"1.24", "sd", SC_ONE_LU | SC_SHOWME
>         },
> 
> 
> in the knowndevs[] array, and everything went fine. I could use the
> disk as a regular hd, could even boot from it, and change disks when it
> was unmounted.
> 
> With 2.1R, this appears not to work: the disk is assigned the
> "uk" device instead of "sd" as I'd like to.
> 
> scsi_selectdev() returns the correct entry, but then the drive is
> somehow not recognised as an "sd" device. I guess the problem occurs
> somewhere later in the process, such as in scsi_device_lookup()
> or so, but it is not clear to me how to fix this (assuming there is an
> easy way ?).

I think I'm the one who broke this, or at least formalized the breakage
if it was broken when I got my hands on it.

The best fix is probably similar to what is in -current:  Add a
"driver" member to override driver selection with:

> struct scsidevs {
>     u_int32_t type; 
>     u_int32_t driver;       /* normally the same as type */

Add the initialization to the knowndevs table,
duplicating most except specifying the driver you want to use for your entry:

>         {
>                 T_OPTICAL, T_DIRECT, T_REMOV, "SONY", "SMO-F331-00"
>                     ,"1.24", "sd", SC_ONE_LU | SC_SHOWME
>         },
> 

then change the device selection to use the new member at the
bottom of scsi_probedev:

>        *type_p =
>            bestmatch->driver;
>    return bestmatch;

-- 

Peter Dufault               Real-Time Machine Control and Simulation
HD Associates, Inc.         Voice: 508 433 6936
dufault@hda.com             Fax:   508 433 5267