From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 16 Aug 2002 06:45:52 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 16 Aug 2002 06:45:52 -0400 Received: from pc2-cwma1-5-cust12.swa.cable.ntl.com ([80.5.121.12]:56048 "EHLO irongate.swansea.linux.org.uk") by vger.kernel.org with ESMTP id ; Fri, 16 Aug 2002 06:45:51 -0400 Subject: Re: ide-2.4.19-ac4.11.patch, late but stable From: Alan Cox To: Andre Hedrick Cc: Markus Plail , linux-kernel@vger.kernel.org In-Reply-To: References: Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.3 (1.0.3-6) Date: 16 Aug 2002 11:48:03 +0100 Message-Id: <1029494883.31487.34.camel@irongate.swansea.linux.org.uk> Mime-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 2002-08-16 at 08:01, Andre Hedrick wrote: > As much as I hate to concept of a DMA mempool, it looks like the direction > to follow. Games such as HOST<>DEVICE || feast<>famine of buffer streams > appear to be the norm to push vast amounts of atapi-dma. The alternative > is to have device level request queues and have the queues carry the SG or > PRD list for that portion. We have mempool in 2.5 and backported so we can preallocate pools of memory to meet some streaming requirement and then discard the entire pool later. What do you actually need ?