From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Woodhouse Subject: Re: [PATCH 0/7] Discard requests, v2 Date: Wed, 13 Aug 2008 13:26:44 +0100 Message-ID: <1218630404.2977.283.camel@pmac.infradead.org> References: <20080812091446.GF20055@kernel.dk> <1218535205.2977.131.camel@pmac.infradead.org> <20080812105428.GG20055@kernel.dk> <1218539797.2977.147.camel@pmac.infradead.org> <1218543544.2977.158.camel@pmac.infradead.org> <20080812125315.GH20055@kernel.dk> <1218546257.2977.167.camel@pmac.infradead.org> <1218556054.2977.188.camel@pmac.infradead.org> <20080812180422.GA21341@shareable.org> <1218622976.2977.257.camel@pmac.infradead.org> <20080813121941.GA8505@shareable.org> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: Jens Axboe , Andrew Morton , Ric Wheeler , linux-fsdevel@vger.kernel.org, gilad@codefidence.com, matthew@wil.cx To: Jamie Lokier Return-path: Received: from bombadil.infradead.org ([18.85.46.34]:45305 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751993AbYHMM06 (ORCPT ); Wed, 13 Aug 2008 08:26:58 -0400 In-Reply-To: <20080813121941.GA8505@shareable.org> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Wed, 2008-08-13 at 13:19 +0100, Jamie Lokier wrote: > You're right, as long as it's not aio, there's no way to sensibly > order it with concurrent writes; I take it back. > > (Aside: You would issue writes to *different* sectors before > BLKDISCARD, and they must be done in that order. That is how you > implement journalling.) Yes, but you implement flushes or barriers to handle that case _anyway_. Normal writes can cross each other too, when they're on different pages. -- David Woodhouse Open Source Technology Centre David.Woodhouse@intel.com Intel Corporation