From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 33FD0C77B73 for ; Tue, 30 May 2023 14:22:38 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232645AbjE3OWg (ORCPT ); Tue, 30 May 2023 10:22:36 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:60376 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229566AbjE3OWS (ORCPT ); Tue, 30 May 2023 10:22:18 -0400 Received: from verein.lst.de (verein.lst.de [213.95.11.211]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9FF291720 for ; Tue, 30 May 2023 07:21:13 -0700 (PDT) Received: by verein.lst.de (Postfix, from userid 2407) id 611AB68B05; Tue, 30 May 2023 16:20:10 +0200 (CEST) Date: Tue, 30 May 2023 16:20:10 +0200 From: Christoph Hellwig To: Johannes Thumshirn Cc: Christoph Hellwig , Chris Mason , Josef Bacik , David Sterba , Naohiro Aota , "linux-btrfs@vger.kernel.org" , Matias =?iso-8859-1?Q?Bj=F8rling?= , Damien Le Moal Subject: Re: don't split ordered_extents for zoned writes at I/O submission time Message-ID: <20230530142009.GA8807@lst.de> References: <20230524150317.1767981-1-hch@lst.de> <45d88c8a-4eab-6fc2-8970-6224878940a1@wdc.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <45d88c8a-4eab-6fc2-8970-6224878940a1@wdc.com> User-Agent: Mutt/1.5.17 (2007-11-01) Precedence: bulk List-ID: X-Mailing-List: linux-btrfs@vger.kernel.org On Tue, May 30, 2023 at 01:21:14PM +0000, Johannes Thumshirn wrote: > Btw, as this came up multiple times today, should we also add a patch on top > of this series merging consecutive ordered extents if the on disk extents are > consecutive as well? > > This would reduce fragmentation that unfortunately is a byproduct of the > Zone Append writing scheme. With this series we don't split ordered extents for zoned writes unless they are non-consecutive. So I'm not sure what we should merge, as there isn't any splitting going on unless either a block group didn't have space, or zone append actually reordered.