Date: Fri, 08 Oct 2010 19:34:19 +0300 From: Andriy Gapon <avg@freebsd.org> To: Matthew Jacob <mj@feral.com> Cc: freebsd-scsi@freebsd.org Subject: Re: [HEADSUP] changes to cam_get_device() and cam_open_device() Message-ID: <4CAF480B.3020306@freebsd.org> In-Reply-To: <4CAF3D4A.70001@feral.com> References: <4BCDEBF6.3030609@icyb.net.ua> <4CA30B24.8040707@freebsd.org> <4CAEDF48.1030602@freebsd.org> <201010081221.24584.bruce@cran.org.uk> <4CAF31EE.6060409@freebsd.org> <4CAF3C6D.3020003@freebsd.org> <4CAF3D4A.70001@feral.com>
next in thread | previous in thread | raw e-mail | index | archive | help
on 08/10/2010 18:48 Matthew Jacob said the following: > > > The 'eject on close' variant is a hangover from older stuff. I don't think it is > much used. I believe packages like bacula and amanda just use mtio ioctls directly > on the no-rewind device (haven't worked on tape in a long time, so my recollection > is quite stale). > > I think as long as no-rewind cases are covered, whatever you do that simplifies > things will be a good thing. Matthew, thanks a lot! Here is a new variant of the camlib.c patch: http://people.freebsd.org/~avg/cam_get_device.2.diff I tried to preserve 'n' behavior (and added 'e'), but in a little bit more elaborate fashion. Instead of just checking for a leading 'n' I try to check for a known "non-rewind" peripheral name between the 'n' and unit number. I am aware only of sa(4) driver with such a feature at present. -- Andriy Gapon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4CAF480B.3020306>