From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261569AbVFBCKN (ORCPT ); Wed, 1 Jun 2005 22:10:13 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261572AbVFBCKN (ORCPT ); Wed, 1 Jun 2005 22:10:13 -0400 Received: from main.gmane.org ([80.91.229.2]:3501 "EHLO ciao.gmane.org") by vger.kernel.org with ESMTP id S261569AbVFBCKD (ORCPT ); Wed, 1 Jun 2005 22:10:03 -0400 X-Injected-Via-Gmane: http://gmane.org/ To: linux-kernel@vger.kernel.org From: =?iso-8859-1?q?M=E5ns_Rullg=E5rd?= Subject: Re: OT] Joerg Schilling flames Linux on his Blog Date: Thu, 02 Jun 2005 04:08:39 +0200 Message-ID: References: <26A66BC731DAB741837AF6B2E29C10171E60DE@xmb-hkg-413.apac.cisco.com> <20050530093420.GB15347@hout.vanvergehaald.nl> <429B0683.nail5764GYTVC@burner> <46BE0C64-1246-4259-914B-379071712F01@mac.com> <429C4483.nail5X0215WJQ@burner> <20050531172204.GD17338@voodoo> <429DDA07.nail7BFA4XEC5@burner> <429DEA5B.nail7BFNJVI78@burner> <429DF038.3010503@nortel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Complaints-To: usenet@sea.gmane.org X-Gmane-NNTP-Posting-Host: 76.80-203-227.nextgentel.com User-Agent: Gnus/5.1006 (Gnus v5.10.6) XEmacs/21.4 (Security Through Obscurity, linux) Cancel-Lock: sha1:d6MdPnDg5oJid1LXNLfPRVxbrII= Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Chris Friesen writes: > Joerg Schilling wrote: > >> There is still no new and definitely stable interface that allows me to >> asume that it makes sense to put effort in implementing support for it. > > Why not just accept *any* device node that the user passes in? That's the question I think we're all asking. The answer is that cdrecord *already* does this. In my understanding there are three basic parts involved here: 1) SCSI commands, and the devices that use them, 2) low-level addressing/transport used by the OS to deliver SCSI commands to the devices, and 3) high-level addressing used by userland when talking to the OS. 2 can be more or less anything: traditional b/t/l SCSI, ATAPI, parport, USB, etc, each with it's own unique addressing style. 3 is what applications should be using, and varies across OSes, Unix systems using device nodes, others using other methods I do not know. Through a thorough misunderstanding of the word "portable", Mr. Schilling appears to be attempting to forcibly extend one particular incarnation of 2 to also cover 3. FWIW, I run cdrecord using dev=/dev/cdrw as a regular user, without any suid bits, and close my eyes while the warnings scroll by. Using a small window and the -v flag, they're out of sight before you know it. I have yet to burn a coaster not resulting from bad media or hardware, even on loaded systems, over NFS, or whatnot. -- Måns Rullgård mru@inprovide.com