From owner-freebsd-scsi  Mon Jan  1 17:35:51 1996
Return-Path: owner-freebsd-scsi
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id RAA20381
          for freebsd-scsi-outgoing; Mon, 1 Jan 1996 17:35:51 -0800 (PST)
Received: from linus.demon.co.uk (linus.demon.co.uk [158.152.10.220])
          by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id RAA20375
          for <freebsd-scsi@freebsd.org>; Mon, 1 Jan 1996 17:35:46 -0800 (PST)
Received: (from mark@localhost) by linus.demon.co.uk (8.7.3/8.7.3) id BAA00338 for freebsd-scsi@freebsd.org; Tue, 2 Jan 1996 01:35:31 GMT
Message-Id: <199601020135.BAA00338@linus.demon.co.uk>
From: mark@linus.demon.co.uk (Mark Valentine)
Date: Tue, 2 Jan 1996 01:35:30 +0000
X-Mailer: Mail User's Shell (7.2.6 alpha(3) 7/19/95)
To: freebsd-scsi@freebsd.org
Subject: scsi_start_unit() and cd_open()
Sender: owner-freebsd-scsi@freebsd.org
Precedence: bulk

What would be the correct way to add a hook to disable the scsi_start_unit()
in cd_open() for a particular device?  My XM-3401 seems fine without it, and
disabling it allows audio CDs to be controlled in a more flexible manner
(e.g. separate cdplay/cdpause/cdresume/cdstop invocations - and Workman
happily restores its state when restarted).

Even if enabled, shouldn't it only be done on first open?

Incidentally, the call in cd_open() passes CD_START as "flags", which probably
isn't intended (the value appears to be that of SCSI_NOSLEEP in that context).
The CD_STOP/CD_START/CD_EJECT #defines seem to be otherwise unused, and it's
not clear where they were intended to be used (obviously in some context where
"-2" means "eject").

		Mark.