From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758815AbZCRNdA (ORCPT ); Wed, 18 Mar 2009 09:33:00 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752929AbZCRN2c (ORCPT ); Wed, 18 Mar 2009 09:28:32 -0400 Received: from mx1.redhat.com ([66.187.233.31]:53205 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757707AbZCRN2a (ORCPT ); Wed, 18 Mar 2009 09:28:30 -0400 From: swhiteho@redhat.com To: linux-kernel@vger.kernel.org Cc: cluster-devel@redhat.com, Hisashi Hifumi , Steven Whitehouse Subject: [PATCH 16/18] GFS2: Pagecache usage optimization on GFS2 Date: Wed, 18 Mar 2009 12:23:51 +0000 Message-Id: <1237379033-28095-17-git-send-email-swhiteho@redhat.com> In-Reply-To: <1237379033-28095-16-git-send-email-swhiteho@redhat.com> References: <1237379033-28095-1-git-send-email-swhiteho@redhat.com> <1237379033-28095-2-git-send-email-swhiteho@redhat.com> <1237379033-28095-3-git-send-email-swhiteho@redhat.com> <1237379033-28095-4-git-send-email-swhiteho@redhat.com> <1237379033-28095-5-git-send-email-swhiteho@redhat.com> <1237379033-28095-6-git-send-email-swhiteho@redhat.com> <1237379033-28095-7-git-send-email-swhiteho@redhat.com> <1237379033-28095-8-git-send-email-swhiteho@redhat.com> <1237379033-28095-9-git-send-email-swhiteho@redhat.com> <1237379033-28095-10-git-send-email-swhiteho@redhat.com> <1237379033-28095-11-git-send-email-swhiteho@redhat.com> <1237379033-28095-12-git-send-email-swhiteho@redhat.com> <1237379033-28095-13-git-send-email-swhiteho@redhat.com> <1237379033-28095-14-git-send-email-swhiteho@redhat.com> <1237379033-28095-15-git-send-email-swhiteho@redhat.com> <1237379033-28095-16-git-send-email-swhiteho@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Hisashi Hifumi I introduced "is_partially_uptodate" aops for GFS2. A page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate on pagesize != blocksize environment. This aops checks that all buffers which correspond to a part of a file that we want to read are uptodate. If so, we do not have to issue actual read IO to HDD even if a page is not uptodate because the portion we want to read are uptodate. "block_is_partially_uptodate" function is already used by ext2/3/4. With the following patch random read/write mixed workloads or random read after random write workloads can be optimized and we can get performance improvement. I did a performance test using the sysbench. #sysbench --num-threads=16 --max-requests=200000 --test=fileio --file-num=1 --file-block-size=8K --file-total-size=2G --file-test-mode=rndrw --file-fsync-freq=0 --file-rw-ratio=1 run -2.6.29-rc6 Test execution summary: total time: 202.6389s total number of events: 200000 total time taken by event execution: 2580.0480 per-request statistics: min: 0.0000s avg: 0.0129s max: 49.5852s approx. 95 percentile: 0.0462s -2.6.29-rc6-patched Test execution summary: total time: 177.8639s total number of events: 200000 total time taken by event execution: 2419.0199 per-request statistics: min: 0.0000s avg: 0.0121s max: 52.4306s approx. 95 percentile: 0.0444s arch: ia64 pagesize: 16k blocksize: 4k Signed-off-by: Hisashi Hifumi Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c index a6d00e8..a6dde17 100644 --- a/fs/gfs2/ops_address.c +++ b/fs/gfs2/ops_address.c @@ -1096,6 +1096,7 @@ static const struct address_space_operations gfs2_writeback_aops = { .releasepage = gfs2_releasepage, .direct_IO = gfs2_direct_IO, .migratepage = buffer_migrate_page, + .is_partially_uptodate = block_is_partially_uptodate, }; static const struct address_space_operations gfs2_ordered_aops = { @@ -1111,6 +1112,7 @@ static const struct address_space_operations gfs2_ordered_aops = { .releasepage = gfs2_releasepage, .direct_IO = gfs2_direct_IO, .migratepage = buffer_migrate_page, + .is_partially_uptodate = block_is_partially_uptodate, }; static const struct address_space_operations gfs2_jdata_aops = { @@ -1125,6 +1127,7 @@ static const struct address_space_operations gfs2_jdata_aops = { .bmap = gfs2_bmap, .invalidatepage = gfs2_invalidatepage, .releasepage = gfs2_releasepage, + .is_partially_uptodate = block_is_partially_uptodate, }; void gfs2_set_aops(struct inode *inode) -- 1.6.0.3