From owner-freebsd-bugs  Thu May 11 10:56:59 1995
Return-Path: bugs-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id KAA08557
          for bugs-outgoing; Thu, 11 May 1995 10:56:59 -0700
Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id KAA08551
          for <freebsd-bugs@FreeBSD.org>; Thu, 11 May 1995 10:56:55 -0700
Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id KAA10636; Thu, 11 May 1995 10:56:30 -0700
From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>
Message-Id: <199505111756.KAA10636@gndrsh.aac.dev.com>
Subject: Re: problem for reading old CD-ROM
To: ohki@gssm.otsuka.tsukuba.ac.jp
Date: Thu, 11 May 1995 10:56:30 -0700 (PDT)
Cc: freebsd-bugs@FreeBSD.org
In-Reply-To: <199505111045.TAA00285@smri01.gssm.otsuka.tsukuba.ac.jp> from "ohki@gssm.otsuka.tsukuba.ac.jp" at May 11, 95 07:45:26 pm
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1323      
Sender: bugs-owner@FreeBSD.org
Precedence: bulk

> 
> Dear folks
> 
>  I have a problem to read my old DOS CD-ROMs with FreeBSD 2.0.
> 
>  My CD-ROMs are 512-byte logical block size,
>  and this causes sys/scsi/cd.c to fail reading 512-byte blocks
>  which are not aligned on 2KB boundaries.
>  Beside this, routines in sys/isofs/cd9660 do not work correctly,
>  since my CD-ROMs has directory entries which cross 512-byte block boundaries.
> 
>  I hacked
>   i)  sys/scsi/cd.c to allow unalinged 512-byte block I/O,
>   ii) sys/isofs/cd9660 routines to allow directory entries which
>       cross 512-byte block boundaries if its logical block size is not 2048.
>       This is done mainly with iso_blkatoff() routine.
>       I.e. if its logical block size is not 2048,
>       iso_blkatoff() always read one logical block with
>       the next 512-byte block.
> 
>  Following context diffs are patches I make.
...

I am not sure, I need to go check my iso 9660 spec, but it seems to me
that 512 byte logical blocks are a violation of of the spec.  Changing
the cd9660 code to read non-conformat cd-roms, IMHO, is an ugly hack
at best.

Changing block device code to do unaligned block transfers is even
worse :-(.

-- 
Rod Grimes                                      rgrimes@gndrsh.aac.dev.com
Accurate Automation Company                   Custom computers for FreeBSD