From owner-freebsd-hackers  Wed Aug 21 07:06:46 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id HAA17576
          for hackers-outgoing; Wed, 21 Aug 1996 07:06:46 -0700 (PDT)
Received: from tdc.on.ca (tdc.on.ca [204.92.242.39])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id HAA17571
          for <hackers@freebsd.org>; Wed, 21 Aug 1996 07:06:42 -0700 (PDT)
Received: (from martin@localhost) by tdc.on.ca (8.7.5/8.6.6) id KAA02637 for hackers@freebsd.org; Wed, 21 Aug 1996 10:05:59 -0400 (EDT)
From: Martin Renters <martin@tdc.on.ca>
Message-Id: <199608211405.KAA02637@tdc.on.ca>
Subject: CDwriter software
To: hackers@freebsd.org
Date: Wed, 21 Aug 1996 10:05:58 -0400 (EDT)
X-Mailer: ELM [version 2.4 PL24 ME8a]
Content-Type: text
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Is there a particular reason that we've gone with a 'worm' driver
interface for CD writers, as opposed to a 'cdwrite' program like Linux
has?

The reason I ask is that I've just finished writing a 'cdwrite' type
program to burn CDs on a PINNACLE RCD1000/RCD5040 on HPUX 10.x
and the SCSI command set that this drive uses is rather different to that
of the HP/Plasmon writers.  For one, it seems to want to know ahead of
time how much data you want to write in a track and you also have to
supply pregap, and postgap information.  The SCSI write command has a
bit that specifies that more data is to follow, and you're expected
to keep this bit set while writing out your data.  When writing the
pregap/postgap information you need to supply 150 blocks of data, the
buffer for which you'd have to get some somewhere inside the kernel.

This all seems simpler in a 'cdwrite' type program.  Comments?

Martin