From owner-freebsd-current@FreeBSD.ORG Tue Mar 29 08:48:35 2005 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5B52616A4CE for ; Tue, 29 Mar 2005 08:48:35 +0000 (GMT) Received: from harmony.village.org (rover.village.org [168.103.84.182]) by mx1.FreeBSD.org (Postfix) with ESMTP id D608443D64 for ; Tue, 29 Mar 2005 08:48:34 +0000 (GMT) (envelope-from imp@bsdimp.com) Received: from localhost (warner@rover2.village.org [10.0.0.1]) by harmony.village.org (8.13.3/8.13.1) with ESMTP id j2T8k0Zp096581; Tue, 29 Mar 2005 01:46:00 -0700 (MST) (envelope-from imp@bsdimp.com) Date: Tue, 29 Mar 2005 01:46:02 -0700 (MST) Message-Id: <20050329.014602.66168889.imp@bsdimp.com> To: phk@phk.freebsd.dk From: "M. Warner Losh" In-Reply-To: <39953.1112084492@critter.freebsd.dk> References: <20050329.011148.69987814.imp@bsdimp.com> <39953.1112084492@critter.freebsd.dk> X-Mailer: Mew version 3.3 on Emacs 21.3 / Mule 5.0 (SAKAKI) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit cc: current@freebsd.org cc: vova@fbsd.ru cc: mdodd@freebsd.org cc: freebsd-mobile@freebsd.org cc: julian@elischer.org cc: ticso@cicely.de Subject: Re: Reattach/redetect allways connected umass device - is it possible ? X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 29 Mar 2005 08:48:35 -0000 In message: <39953.1112084492@critter.freebsd.dk> "Poul-Henning Kamp" writes: : I will maintain, based on what little that I have read about user : interfaces, that the logical and intuitive behaviour for a computing : system is to notice immediately (< 2s) that a storage media has : been removed. I'm willing to be a tad more tolerant about inserts, : but not much more than 5 seconds. Then we'd have to poll every second in a sane way to accomplish that. And finding the sane way that doesn't interfere with other bus usage may be tricky. Unless we're going to give events to the actual user (meaning userland entities that inform the user in a friendly way), I'd maintain that there's no difference between knowing that the media is ejected immediately, and the time of next use. The user experience will be the same either way. : I don't really care if our hardware is lame and brain-dead and needs : to be polled in Shakespearan english by actors in victorian custumes, : our job is to hide all that crap and give the system what to the : users look like a consistent and predictable behaviour. Assuming, of course, that it can be done with a reasonable level of performance for the rest of the system, and that unintended consequences of the polling don't happen (like keeping a disk spinning all the time, when it would have spun down). However, other systems do effect polling in some fashion without spinning things up, so it would be just a matter of finding these things. In the short run, however, adding a few checks to critical parts of the path, like daopen, would make the user experience much better. Warner