From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751155AbWFKLbj (ORCPT ); Sun, 11 Jun 2006 07:31:39 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751156AbWFKLbi (ORCPT ); Sun, 11 Jun 2006 07:31:38 -0400 Received: from pops.net-conex.com ([204.244.176.3]:50399 "EHLO mail.net-conex.com") by vger.kernel.org with ESMTP id S1751155AbWFKLbi (ORCPT ); Sun, 11 Jun 2006 07:31:38 -0400 Date: Sun, 11 Jun 2006 04:31:41 -0700 From: "Robin H. Johnson" To: linux-kernel@vger.kernel.org Cc: "Robin H. Johnson" Subject: Re: [PATCH] tmpfs time granularity fix for [acm]time moving backwards. Message-ID: <20060611113141.GD26475@curie-int.vc.shawcable.net> References: <20060609224936.GA30611@curie-int.vc.shawcable.net> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="a2FkP9tdjPU2nyhF" Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.11 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org --a2FkP9tdjPU2nyhF Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Sun, Jun 11, 2006 at 12:12:41PM +0200, Jan Engelhardt wrote: > >Test output from a filesystem supporting sub-second timestamps (jfs,xfs,= ramfs): > >creat: m=3D1149891452.928796249 c=3D1149891452.928796249 a=3D114989145= 2.928796249 > >futimes: m=3D1149891452.928796249 c=3D1149891452.928796249 a=3D114989145= 2.928796249 > > > >Test output from the tmpfs filesystem with the patch below: > >creat: m=3D1149892086.382150894 c=3D1149892086.382150894 a=3D114989207= 5.473249075 > >futimes: m=3D1149892086.383150885 c=3D1149892086.383150885 a=3D114989208= 6.383150885 > > > Is it normal that creat and futimes match for jfs/xfs? My testcase just seemed to run nice and fast there, if I do some work in the middle or sleep, the numbers will be different more, but otherwise I think = it implies that both operations are taking place during the same time instance= (I have HZ=3D1000 as you may be able to deduce by the 1ms differences in some = of the pairs). I ran it a few more times on ramfs for good measure. creat: m=3D1150025153.933532153 c=3D1150025153.933532153 a=3D1150025151.295= 58289 futimes: m=3D1150025153.933532153 c=3D1150025153.933532153 a=3D1150025153.9= 33532153 =2E.. creat: m=3D1150025156.860505810 c=3D1150025156.860505810 a=3D1150025153.933= 532153 futimes: m=3D1150025156.861505801 c=3D1150025156.861505801 a=3D1150025156.8= 61505801 =2E.. creat: m=3D1150025247.906686387 c=3D1150025247.906686387 a=3D1150025247.459= 690410 futimes: m=3D1150025247.907686378 c=3D1150025247.907686378 a=3D1150025247.9= 07686378 > >Signed-off-by: Robin H. Johnson > Hint: Cc it to the shmem.c maintainer, or it may get missed :) Will do, thanks. --=20 Robin Hugh Johnson E-Mail : robbat2@gentoo.org GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85 --a2FkP9tdjPU2nyhF Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2-ecc0.1.6 (GNU/Linux) Comment: Robbat2 @ Orbis-Terrarum Networks iD8DBQFEi/8dPpIsIjIzwiwRArKTAKCKb3cHLCrqwtbQtPJw45JzRWKZfwCgm+WS cbJZPEqU2uzQjscLJe5qajc= =igy2 -----END PGP SIGNATURE----- --a2FkP9tdjPU2nyhF--