From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Sandeen Subject: Re: ext4_fallocate Date: Thu, 28 Jun 2012 11:23:14 -0400 Message-ID: <4FEC76E2.1010006@redhat.com> References: <4FE8086F.4070506@zoho.com> <13363B8E-7411-4AA3-8835-F9DEF67B2ABE@dilger.ca> <4FEC744F.6010901@ubuntu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Andreas Dilger , Fredrick , linux-ext4@vger.kernel.org To: Phillip Susi Return-path: Received: from mx1.redhat.com ([209.132.183.28]:26352 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753921Ab2F1PXQ (ORCPT ); Thu, 28 Jun 2012 11:23:16 -0400 In-Reply-To: <4FEC744F.6010901@ubuntu.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 6/28/12 11:12 AM, Phillip Susi wrote: > On 6/25/2012 3:33 AM, Andreas Dilger wrote: >> There was a recent patch series "ext4: add an io-tree to track >> block allocation" that may improve the performance for your case of >> overwrite of uninitialized files, but it hasn't landed yet. > > I'm confused. Why is writing to uninitialized extents slow, and why > would this help? If you have an uninitialized extent, then the blocks > are already allocated, just flagged as containing uninitialized data. > Writing to them should be no different than writing to initialized > extents, save for the step of clearing the uninitialized flag. The other piece is that large uninitialized extents get split into up to 3 extents if you write into the middle of them so there are potentially a lot more metadata updates flying around. - -Eric -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.17 (Darwin) Comment: GPGTools - http://gpgtools.org Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQIcBAEBAgAGBQJP7HbhAAoJECCuFpLhPd7g8BUQAKlVb3EXtDNbziXkmKdkbDWo wf/l4CMol/oxnLbKbvVkRQmRKfhdAsX5qQgmRqJvT3dPFzA7jSZqKDt3BV06fBSn lJqWwdtHwjodeTj2UcoLZm2+8tL3d4AtJw42IntVasRtZRrUGZ5mFi1HMjm8vmK8 xSkmC/YbFtj222WiIpjLyu0XBZtSX9uMP4NZoThmA3MwuEcFr6AVqgi0WW3GVZP+ 2L1ylMJDs6OykE+/Tv4AkFZ7XLtE946ytpqI6rLJvbItWdw1OqWhL6XWnOnGUgYG B4IM5klovqIFOCPkHj0DPLm/RWjhv/+DJMkqEHAJANu4kEyS4O08Acqy/sAf1y4U 0OWCIW6fUHzV/38aMuQlLZeN+enaVmSCMh1kBWlGEeElZ/pRxClpiO+1Rk1Z7x1z f1H+J3NV+wwGa3xiG6XO9biIMo10qdQMy93KLkeh6ndE3SYgvaAxC0AatrTZJS9Q oRiKJLl3uTv1j0h+W+CBsRRmPzf/ad1kUKKPfT0recxy7ggO6X+6fkgsgRo3LRfN 53uAQWELJJu1opM9PsBuvX4uUvKLgbPXmky3JDS+Q3p18I4DfiBiIEPKbkZAkeD/ py0oxuqs0MGTRd9FCCqNODG2AJ7VWnY29hUOaCgWeFT5eG/r9tECJI6l62z4GeBx b5+TFFsdJ/zJVBdjBuIT =WcmZ -----END PGP SIGNATURE-----