From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dmitry Monakhov Subject: [PATCH 0/9] ext4: Bunch of DIO/AIO fixes V2 Date: Thu, 13 Sep 2012 19:01:05 +0400 Message-ID: <1347548474-31897-1-git-send-email-dmonakhov@openvz.org> Cc: tytso@mit.edu, jack@suse.cz, wenqing.lz@taobao.com, Dmitry Monakhov To: linux-ext4@vger.kernel.org Return-path: Received: from mail-lb0-f174.google.com ([209.85.217.174]:51413 "EHLO mail-lb0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754448Ab2IMPBU (ORCPT ); Thu, 13 Sep 2012 11:01:20 -0400 Received: by lbbgj3 with SMTP id gj3so2024059lbb.19 for ; Thu, 13 Sep 2012 08:01:19 -0700 (PDT) Sender: linux-ext4-owner@vger.kernel.org List-ID: There are number of races exist caused by lack of synchronization between DIO workers in flight and truncate/fsync/punch_hole routines This patch series try to optimize and fix existing DIO/AIO code. This patch-set survives after all my tests(except fsdefrag) I observe significant performance improvements for big SMP hosts performind DIO for single file. TOC: # first one is simply a cleanup ext4: ext4_inode_info diet # Cleanup and bug fixes ext4: completed_io locking cleanup V2 ext4: serialize dio nonlocked reads with defrag workers V3 ext4: punch_hole should wait for DIO writers ext4: serialize unlocked dio reads with truncate ext4: endless truncate due to nonlocked dio readers V2 ext4: serialize truncate with owerwrite DIO workers V2 # Nasty panch_hole regressions ext4: fix ext_remove_space for punch_hole case ext4: fix ext4_ext_remove_space tree traversal Signed-off-by: Dmitry Monakhov ext4.h | 27 +++++++++++++++---- extents.c | 82 ++++++++++++++++++++++++++++++++-------------------------- fsync.c | 47 +++++++++++---------------------- indirect.c | 17 +++++++++--- inode.c | 49 ++++++++++++++-------------------- move_extent.c | 8 +++++ page-io.c | 76 ++++++++++++++--------------------------------------- super.c | 1 8 files changed, 148 insertions(+), 159 deletions(-)