From owner-freebsd-hackers Sat Jan 28 10:28:39 1995 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id KAA06214 for hackers-outgoing; Sat, 28 Jan 1995 10:28:39 -0800 Received: from cs.weber.edu (cs.weber.edu [137.190.16.16]) by freefall.cdrom.com (8.6.9/8.6.6) with SMTP id KAA06208 for ; Sat, 28 Jan 1995 10:28:37 -0800 Received: by cs.weber.edu (4.1/SMI-4.1.1) id AA07320; Sat, 28 Jan 95 11:22:06 MST From: terry@cs.weber.edu (Terry Lambert) Message-Id: <9501281822.AA07320@cs.weber.edu> Subject: Re: SyQuest works with FreeBSD 2.0R ! To: julian@tfs.com (Julian Elischer) Date: Sat, 28 Jan 95 11:22:05 MST Cc: peter@bonkers.taronga.com, dufault@hda.com, john@pyromania.apana.org.au, hackers@FreeBSD.org In-Reply-To: from "Julian Elischer" at Jan 27, 95 11:26:34 am X-Mailer: ELM [version 2.4dev PL52] Sender: hackers-owner@FreeBSD.org Precedence: bulk > > The real solution to removable media is to put some sort of ident in the > > disklabel (including time and system ID) and track that. If you have a > > mounted filesystem and get a UNIT ATTENTION then you can see whether > > it's been swapped or not. > > At the moment, a unit-attention invalidates all accesses, until > all have been closed. (the device becomes 'locked'. > when the reference count has gone down to 0, the lock is released.. With respect, by the time you get a unit attention, it's too late to flush your in core buffers that haven't been written yet, unless you can throw the bits at the disk as it's walking down the hallway. Asking for the disk *back* seems reasonable to me. Terry Lambert terry@cs.weber.edu --- Any opinions in this posting are my own and not those of my present or previous employers.