mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Steven Whitehouse <swhiteho@redhat.com>
To: cluster-devel@redhat.com, linux-kernel@vger.kernel.org
Cc: Bob Peterson <rpeterso@redhat.com>,
	Steven Whitehouse <swhiteho@redhat.com>
Subject: [PATCH 20/24] GFS2: speed up delete/unlink performance for large files
Date: Mon, 24 Oct 2011 13:48:34 +0100	[thread overview]
Message-ID: <1319460518-9046-21-git-send-email-swhiteho@redhat.com> (raw)
In-Reply-To: <1319460518-9046-1-git-send-email-swhiteho@redhat.com>

From: Bob Peterson <rpeterso@redhat.com>

This patch improves the performance of delete/unlink
operations in a GFS2 file system where the files are large
by adding a layer of metadata read-ahead for indirect blocks.
Mileage will vary, but on my system, deleting an 8.6G file
dropped from 22 seconds to about 4.5 seconds.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 22ad413..834cd94 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -831,7 +831,7 @@ static int recursive_scan(struct gfs2_inode *ip, struct buffer_head *dibh,
 {
 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
 	struct buffer_head *bh = NULL;
-	__be64 *top, *bottom;
+	__be64 *top, *bottom, *t2;
 	u64 bn;
 	int error;
 	int mh_size = sizeof(struct gfs2_meta_header);
@@ -859,7 +859,27 @@ static int recursive_scan(struct gfs2_inode *ip, struct buffer_head *dibh,
 	if (error)
 		goto out;
 
-	if (height < ip->i_height - 1)
+	if (height < ip->i_height - 1) {
+		struct buffer_head *rabh;
+
+		for (t2 = top; t2 < bottom; t2++, first = 0) {
+			if (!*t2)
+				continue;
+
+			bn = be64_to_cpu(*t2);
+			rabh = gfs2_getbuf(ip->i_gl, bn, CREATE);
+			if (trylock_buffer(rabh)) {
+				if (buffer_uptodate(rabh)) {
+					unlock_buffer(rabh);
+					brelse(rabh);
+					continue;
+				}
+				rabh->b_end_io = end_buffer_read_sync;
+				submit_bh(READA | REQ_META, rabh);
+				continue;
+			}
+			brelse(rabh);
+		}
 		for (; top < bottom; top++, first = 0) {
 			if (!*top)
 				continue;
@@ -871,7 +891,7 @@ static int recursive_scan(struct gfs2_inode *ip, struct buffer_head *dibh,
 			if (error)
 				break;
 		}
-
+	}
 out:
 	brelse(bh);
 	return error;
-- 
1.7.4.4


  parent reply	other threads:[~2011-10-24 12:49 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-10-24 12:48 GFS2: Pre-pull patch posting (merge window) Steven Whitehouse
2011-10-24 12:48 ` [PATCH 01/24] GFS2: Clean up dir hash table reading Steven Whitehouse
2011-10-24 12:48 ` [PATCH 02/24] GFS2: Split data write & wait in fsync Steven Whitehouse
2011-10-24 12:48 ` [PATCH 03/24] GFS2: Fix bug-trap in ail flush code Steven Whitehouse
2011-10-24 12:48 ` [PATCH 04/24] GFS2: Make atime checks more efficient Steven Whitehouse
2011-10-24 12:48 ` [PATCH 05/24] GFS2: Fix inode allocation error path Steven Whitehouse
2011-10-24 12:48 ` [PATCH 06/24] GFS2: Fix bug trap and journaled data fsync Steven Whitehouse
2011-10-24 12:48 ` [PATCH 07/24] GFS2: Use ->dirty_inode() Steven Whitehouse
2011-10-24 12:48 ` [PATCH 08/24] GFS2: Clean up gfs2_create Steven Whitehouse
2011-10-24 12:48 ` [PATCH 09/24] GFS2: Fix lseek after SEEK_DATA, SEEK_HOLE have been added Steven Whitehouse
2011-10-25  7:54   ` Andi Kleen
2011-10-24 12:48 ` [PATCH 10/24] GFS2: Use rbtree for resource groups and clean up bitmap buffer ref count scheme Steven Whitehouse
2011-10-24 12:48 ` [PATCH 11/24] GFS2: Make resource groups "append only" during life of fs Steven Whitehouse
2011-10-24 12:48 ` [PATCH 12/24] GFS2: Cache the most recently used resource group in the inode Steven Whitehouse
2011-10-24 12:48 ` [PATCH 13/24] GFS2: Remove obsolete assert Steven Whitehouse
2011-10-24 12:48 ` [PATCH 14/24] GFS2: Call do_strip() directly from recursive_scan() Steven Whitehouse
2011-10-24 12:48 ` [PATCH 15/24] GFS2: Use cached rgrp in gfs2_rlist_add() Steven Whitehouse
2011-10-24 12:48 ` [PATCH 16/24] GFS2: Fix AIL flush issue during fsync Steven Whitehouse
2011-10-24 12:48 ` [PATCH 17/24] GFS2: Correctly set goal block after allocation Steven Whitehouse
2011-10-24 12:48 ` [PATCH 18/24] GFS2: Clean up ->page_mkwrite Steven Whitehouse
2011-10-24 12:48 ` [PATCH 19/24] GFS2: Fix off-by-one in gfs2_blk2rgrpd Steven Whitehouse
2011-10-24 12:48 ` Steven Whitehouse [this message]
2011-10-24 12:48 ` [PATCH 21/24] GFS2: rewrite fallocate code to write blocks directly Steven Whitehouse
2011-10-24 12:48 ` [PATCH 22/24] GFS2: Misc fixes Steven Whitehouse
2011-10-24 12:48 ` [PATCH 23/24] GFS2: Remove two unused variables Steven Whitehouse
2011-10-24 12:48 ` [PATCH 24/24] GFS2: Move readahead of metadata during deallocation into its own function Steven Whitehouse

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=1319460518-9046-21-git-send-email-swhiteho@redhat.com \
    --to=swhiteho@redhat.com \
    --cc=cluster-devel@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rpeterso@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

all inboxes | Powered by JetHome®