From: swhiteho@redhat.com <swhiteho@redhat.com>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] [PATCH 16/18] GFS2: Pagecache usage optimization on GFS2
Date: Wed, 18 Mar 2009 12:23:51 +0000 [thread overview]
Message-ID: <1237379033-28095-17-git-send-email-swhiteho@redhat.com> (raw)
In-Reply-To: <1237379033-28095-16-git-send-email-swhiteho@redhat.com>
From: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp>
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 <hifumi.hisashi@oss.ntt.co.jp>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
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
next prev parent reply other threads:[~2009-03-18 12:23 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-03-18 12:23 [Cluster-devel] [GFS2] Pre-pull patch posting swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 01/18] GFS2: Fix remount argument parsing swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 02/18] GFS2: Bring back lvb-related stuff to lock_nolock to support quotas swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 03/18] GFS2: change gfs2_quota_scan into a shrinker swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 04/18] GFS2: Remove "double" locking in quota swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 05/18] GFS2: Merge lock_dlm module into GFS2 swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 06/18] GFS2: Remove unused field from glock swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 07/18] GFS2: Fix error path ref counting for root inode swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 08/18] GFS2: Fix deadlock on journal flush swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 09/18] GFS2: Support generation of discard requests swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 10/18] GFS2: Expose UUID via sysfs/uevent swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 11/18] GFS2: Add a "demote a glock" interface to sysfs swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 12/18] GFS2: Fix alignment issue and tidy gfs2_bitfit swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 13/18] GFS2: Support quota/noquota mount arguments swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 14/18] GFS2: fix sparse warnings: constant is so big it is swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 15/18] GFS2: fix sparse warning: Should it be static? swhiteho
2009-03-18 12:23 ` swhiteho [this message]
2009-03-18 12:23 ` [Cluster-devel] [PATCH 17/18] GFS2: Fix locking bug in failed shared to exclusive conversion swhiteho
2009-03-18 12:23 ` [Cluster-devel] [PATCH 18/18] GFS2: Clean up of glops.c swhiteho
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1237379033-28095-17-git-send-email-swhiteho@redhat.com \
--to=swhiteho@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).