Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753438AbZLZOON (ORCPT ); Sat, 26 Dec 2009 09:14:13 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751985AbZLZOOM (ORCPT ); Sat, 26 Dec 2009 09:14:12 -0500 Received: from einhorn.in-berlin.de ([192.109.42.8]:55690 "EHLO einhorn.in-berlin.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751631AbZLZOOL (ORCPT ); Sat, 26 Dec 2009 09:14:11 -0500 X-Envelope-From: stefanr@s5r6.in-berlin.de Message-ID: <4B361A23.8090009@s5r6.in-berlin.de> Date: Sat, 26 Dec 2009 15:13:55 +0100 From: Stefan Richter User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.8.1.23) Gecko/20091025 SeaMonkey/1.1.18 MIME-Version: 1.0 To: Pieter Palmers CC: linux-kernel@vger.kernel.org, Clemens Ladisch , linux1394-devel@lists.sourceforge.net Subject: Re: [PATCH 1/5] firewire: fix use of multiple AV/C devices, allow multiple FCP listeners References: <4B35CACF.2000302@joow.be> <4B35FB02.1040602@s5r6.in-berlin.de> <4B360E42.80208@joow.be> In-Reply-To: <4B360E42.80208@joow.be> X-Enigmail-Version: 0.96.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1522 Lines: 32 Pieter Palmers wrote: [lack of system-wide FCP transaction serialization] > This is not a 'practical' problem because most people (learn to) leave > enough time between starting jackd and starting the mixer. This makes > things usable, but I don't really consider this to be a satisfactory > solution. Ah, so this is yet another of those software flaws which are not talked about because people take the workaround as something inevitable. >>> I think this can only be implemented by tracking which AVC commands have >>> been sent to what device, regardless of the userspace application (or >>> kernel driver) that sent them, and disallowing any transactions that >>> result in scenarios such as the above. BTW, what about AV/C RESERVE? According to the spec, it means that a controller gets exclusive FCP access to a target. At this point, a node-wide FCP serializer will have to decide on a policy: The simpler policy would be to remain oblivious of RESERVE. The more difficult to implement policy would be to keep track of reservations and only allow a single controller process (or user, or...) to perform subsequent transactions for the duration of a reservation. (I for one would keep it simple.) -- Stefan Richter -=====-==--= ==-- ==-=- http://arcgraph.de/sr/ -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/