From mboxrd@z Thu Jan 1 00:00:00 1970 From: Takashi Sakamoto Subject: Re: [RFC] dice: simplifying address registration Date: Sat, 25 Apr 2015 21:25:14 +0900 Message-ID: <553B87AA.60800@sakamocchi.jp> References: <553A3FC5.60809@sakamocchi.jp> <553B5534.1060005@ladisch.de> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from smtp301.phy.lolipop.jp (smtp301.phy.lolipop.jp [210.157.22.84]) by alsa0.perex.cz (Postfix) with ESMTP id 95F6C26155B for ; Sat, 25 Apr 2015 14:25:18 +0200 (CEST) In-Reply-To: <553B5534.1060005@ladisch.de> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: alsa-devel-bounces@alsa-project.org Sender: alsa-devel-bounces@alsa-project.org To: Clemens Ladisch Cc: "alsa-devel@alsa-project.org" List-Id: alsa-devel@alsa-project.org Clemens, Thanks for your comment. On Apr 25 2015 17:49, Clemens Ladisch wrote: > Takashi Sakamoto wrote: >> Currently Dice driver keeps address ranges for each unit instance. But >> there's a way to reuse the same address range. > > This requires an additional check for the packet's source node, which is > essentially a duplicate of the core's address lookup code. > >> I think this idea can simplify driver probe processing > > Yes, but that code is just moved to the module load processing. >>From probe processing, thus address allocation is just one time, not each time IEEE 1394 unit is probed. >> and save resources of host controller. > > When using multiple DICE devices, this saves four bytes of FireWire > address space per card, out of the total address space of 2^48 bytes. > Or what other resources do you mean? Actually, 2^48 bytes are really huge. But there're some rude devices to transfer transaction to a certain address (i.e. Fireworks). For me, it's basically better idea to keep allocated address range as small as possible. ...But seems to be not so practical. OK. I dropped this patch. Thanks Takashi Sakamoto