linux-ext4.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: tytso@mit.edu, darrick.wong@oracle.com
Cc: linux-ext4@vger.kernel.org
Subject: [PATCH 12/47] libext2fs: set interior tree block goal more intelligently
Date: Fri, 07 Nov 2014 13:52:05 -0800	[thread overview]
Message-ID: <20141107215205.883.11850.stgit@birch.djwong.org> (raw)
In-Reply-To: <20141107215042.883.49888.stgit@birch.djwong.org>

When we're splitting an extent node, try to allocate the new interior
tree block just prior to the first extent in the block we're trying to
split.  The previous logic only set a goal block if we had to split
both the current node and its parent, which is somewhat infrequent.
When that would happen, the goal would start at zero, leading to poor
locality.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 lib/ext2fs/extent.c         |   20 +++++++++++++++++++-
 tests/f_extents2/expect.1   |    2 +-
 tests/m_bigjournal/expect.1 |   10 +++++-----
 3 files changed, 25 insertions(+), 7 deletions(-)


diff --git a/lib/ext2fs/extent.c b/lib/ext2fs/extent.c
index b6e4fbd..afb79c8 100644
--- a/lib/ext2fs/extent.c
+++ b/lib/ext2fs/extent.c
@@ -934,6 +934,25 @@ static errcode_t extent_node_split(ext2_extent_handle_t handle,
 	orig_height = info.max_depth - info.curr_level;
 	orig_lblk = extent.e_lblk;
 
+	/* Try to put the index block before the first extent */
+	path = handle->path + handle->level;
+	eh = (struct ext3_extent_header *) path->buf;
+	if (handle->level == handle->max_depth) {
+		struct ext3_extent	*ex;
+
+		ex = EXT_FIRST_EXTENT(eh);
+		goal_blk = ext2fs_le32_to_cpu(ex->ee_start) +
+			((__u64) ext2fs_le16_to_cpu(ex->ee_start_hi) << 32);
+	} else {
+		struct ext3_extent_idx	*ix;
+
+		ix = EXT_FIRST_INDEX(eh);
+		goal_blk = ext2fs_le32_to_cpu(ix->ei_leaf) +
+			((__u64) ext2fs_le16_to_cpu(ix->ei_leaf_hi) << 32);
+	}
+	goal_blk -= EXT2FS_CLUSTER_RATIO(handle->fs);
+	goal_blk &= ~EXT2FS_CLUSTER_MASK(handle->fs);
+
 	/* Is there room in the parent for a new entry? */
 	if (handle->level &&
 			(handle->path[handle->level - 1].entries >=
@@ -947,7 +966,6 @@ static errcode_t extent_node_split(ext2_extent_handle_t handle,
 		retval = ext2fs_extent_get(handle, EXT2_EXTENT_UP, &extent);
 		if (retval)
 			goto done;
-		goal_blk = extent.e_pblk;
 
 		retval = extent_node_split(handle, expand_allowed);
 		if (retval)
diff --git a/tests/f_extents2/expect.1 b/tests/f_extents2/expect.1
index 51e36ff..180568b 100644
--- a/tests/f_extents2/expect.1
+++ b/tests/f_extents2/expect.1
@@ -55,7 +55,7 @@ Pass 2: Checking directory structure
 Pass 3: Checking directory connectivity
 Pass 4: Checking reference counts
 Pass 5: Checking group summary information
-Block bitmap differences:  -(25--34) -(155--199)
+Block bitmap differences:  -(25--33) -(154--199)
 Fix? yes
 
 Free blocks count wrong for group #0 (65535, counted=55).
diff --git a/tests/m_bigjournal/expect.1 b/tests/m_bigjournal/expect.1
index 7246739..61d85f9 100644
--- a/tests/m_bigjournal/expect.1
+++ b/tests/m_bigjournal/expect.1
@@ -52,8 +52,8 @@ Group 0: (Blocks 0-32767)
   Reserved GDT blocks at 2-672
   Block bitmap at 673 (+673), Inode bitmap at 757 (+757)
   Inode table at 841-841 (+841)
-  31836 free blocks, 5 free inodes, 2 directories, 5 unused inodes
-  Free blocks: 932-32767
+  31837 free blocks, 5 free inodes, 2 directories, 5 unused inodes
+  Free blocks: 931-32767
   Free inodes: 12-16
 Group 1: (Blocks 32768-65535) [INODE_UNINIT, BLOCK_UNINIT]
   Backup superblock at 32768, Group descriptors at 32769-32769
@@ -297,11 +297,11 @@ Group 38: (Blocks 1245184-1277951) [INODE_UNINIT, BLOCK_UNINIT]
   32768 free blocks, 16 free inodes, 0 directories, 16 unused inodes
   Free blocks: 1245184-1277951
   Free inodes: 609-624
-Group 39: (Blocks 1277952-1310719) [INODE_UNINIT, BLOCK_UNINIT]
+Group 39: (Blocks 1277952-1310719) [INODE_UNINIT]
   Block bitmap at 712 (bg #0 + 712), Inode bitmap at 796 (bg #0 + 796)
   Inode table at 880-880 (bg #0 + 880)
-  32768 free blocks, 16 free inodes, 0 directories, 16 unused inodes
-  Free blocks: 1277952-1310719
+  32767 free blocks, 16 free inodes, 0 directories, 16 unused inodes
+  Free blocks: 1277952-1310718
   Free inodes: 625-640
 Group 40: (Blocks 1310720-1343487) [INODE_UNINIT]
   Block bitmap at 713 (bg #0 + 713), Inode bitmap at 797 (bg #0 + 797)


  parent reply	other threads:[~2014-11-07 21:52 UTC|newest]

Thread overview: 104+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-11-07 21:50 [PATCH 00/47] e2fsprogs November 2014 patchbomb Darrick J. Wong
2014-11-07 21:50 ` [PATCH 01/47] tests: fix test scripts that don't work on non-Linux systems Darrick J. Wong
2014-11-08  1:55   ` Theodore Ts'o
2014-11-07 21:50 ` [PATCH 02/47] misc: fix compiler warnings and minor build errors Darrick J. Wong
2014-11-08  2:25   ` Theodore Ts'o
2014-11-07 21:51 ` [PATCH 03/47] dumpe2fs: don't crash when the user provides no block device argument Darrick J. Wong
2014-11-08  2:26   ` Theodore Ts'o
2014-11-07 21:51 ` [PATCH 04/47] libext2fs: fix endian handling error; reduce fragmentation some Darrick J. Wong
2014-11-17 21:26   ` Theodore Ts'o
2014-11-07 21:51 ` [PATCH 05/47] libext2fs: set BLOCK_UNINIT for non-last blockgroups if all blocks are free Darrick J. Wong
2014-11-17 22:47   ` Theodore Ts'o
2014-11-07 21:51 ` [PATCH 06/47] libext2fs: don't allow alloc_stats on bad inode/block numbers Darrick J. Wong
2014-11-17 23:00   ` Theodore Ts'o
2014-11-07 21:51 ` [PATCH 07/47] libext2fs: refactor extent head creation Darrick J. Wong
2014-12-03  3:55   ` Theodore Ts'o
2014-11-07 21:51 ` [PATCH 08/47] libext2fs: file IO routines should handle uninit blocks Darrick J. Wong
2014-12-03  3:57   ` Theodore Ts'o
2014-11-07 21:51 ` [PATCH 09/47] libext2fs: use a dynamically sized (or caller-provided) block zeroing buffer Darrick J. Wong
2014-12-12  2:37   ` Theodore Ts'o
2014-12-12  5:02     ` Darrick J. Wong
2014-12-13 16:24       ` Theodore Ts'o
2014-12-13 16:25         ` Theodore Ts'o
2014-12-15 16:09         ` Andreas Dilger
2014-11-07 21:51 ` [PATCH 10/47] libext2fs: support BLKZEROOUT/FALLOC_FL_ZERO_RANGE in ext2fs_zero_blocks Darrick J. Wong
2014-12-13 16:29   ` Theodore Ts'o
2014-12-14  3:08     ` Darrick J. Wong
2014-11-07 21:51 ` [PATCH 11/47] libext2fs: find inode goal when allocating blocks Darrick J. Wong
2014-12-14  1:13   ` Theodore Ts'o
2014-12-14 21:02     ` Darrick J. Wong
2014-11-07 21:52 ` Darrick J. Wong [this message]
2014-12-14  1:17   ` [PATCH 12/47] libext2fs: set interior tree block goal more intelligently Theodore Ts'o
2014-11-07 21:52 ` [PATCH 13/47] libext2fs: add a way to check the theoretical maximum extent tree depth Darrick J. Wong
2014-12-14  2:23   ` Theodore Ts'o
2014-12-14  3:11     ` Darrick J. Wong
2014-12-15  3:48       ` Theodore Ts'o
2014-12-15  6:38         ` Darrick J. Wong
2014-11-07 21:52 ` [PATCH 14/47] libext2fs: ext2fs_new_block2() should call alloc_block hook Darrick J. Wong
2014-11-07 21:52 ` [PATCH 15/47] misc: don't allow mk_hugefiles unless the fs supports extents Darrick J. Wong
2014-12-14  2:51   ` Theodore Ts'o
2014-11-07 21:52 ` [PATCH 16/47] dumpe2fs: 80 column outputs, please Darrick J. Wong
2014-12-14  2:52   ` Theodore Ts'o
2014-11-07 21:52 ` [PATCH 17/47] dumpe2fs: output cleanup Darrick J. Wong
2014-12-14  2:53   ` Theodore Ts'o
2014-11-07 21:52 ` [PATCH 18/47] e2fsck: only complain about no-checksum directory blocks once Darrick J. Wong
2014-12-14  2:55   ` Theodore Ts'o
2014-11-07 21:52 ` [PATCH 19/47] e2fsck: don't complain about root dir csum failures when getting lnf Darrick J. Wong
2014-12-14  2:57   ` Theodore Ts'o
2014-11-07 21:52 ` [PATCH 20/47] tune2fs: warn if extents are not enabled when turning on metadata_csum Darrick J. Wong
2014-12-14  2:58   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 21/47] tune2fs: enable uninit_bg when disabling metadata_csum Darrick J. Wong
2014-12-14  2:58   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 22/47] tests: testcases for enabling/disabling metadata_csum via tune2fs Darrick J. Wong
2014-12-14  3:00   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 23/47] mke2fs: don't zero inode table blocks that are already zeroed Darrick J. Wong
2014-12-14  3:01   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 24/47] resize2fs: don't exit if shrinking sparse_super2 fs to one bg Darrick J. Wong
2014-12-14  3:06   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 25/47] resize2fs: quickly rewrite extent blocks when moving an inode w/ metadata_csum Darrick J. Wong
2014-12-14  3:09   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 26/47] resize2fs: use old_fs to detect per-bg metadata blocks to free Darrick J. Wong
2014-12-15  2:09   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 27/47] resize2fs: don't interpret bitmap shift while crossing flexbg as raid stride Darrick J. Wong
2014-12-15  2:11   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 28/47] resize2fs: set bg flags and unused inode count when resizing Darrick J. Wong
2014-12-15  2:12   ` Theodore Ts'o
2014-11-07 21:53 ` [PATCH 29/47] resize2fs: don't mark unallocated bg metadata blocks when fixing bg flags Darrick J. Wong
2014-12-15  2:17   ` Theodore Ts'o
2014-12-15  7:09     ` Darrick J. Wong
2014-12-19 19:15       ` Darrick J. Wong
2014-11-07 21:54 ` [PATCH 30/47] resize2fs: don't play stupid games with the block count Darrick J. Wong
2014-12-15  3:13   ` Theodore Ts'o
2014-11-07 21:54 ` [PATCH 31/47] libext2fs/e2fsck: provide routines to read-ahead metadata Darrick J. Wong
2014-12-15 17:36   ` Theodore Ts'o
2014-12-15 19:19     ` Darrick J. Wong
2014-11-07 21:54 ` [PATCH 32/47] e2fsck: read-ahead metadata during passes 1, 2, and 4 Darrick J. Wong
2014-12-10 20:27   ` Darrick J. Wong
2014-12-10 22:00     ` Theodore Ts'o
2014-11-07 21:54 ` [PATCH 33/47] e2fsck: rebuild sparse extent trees/convert non-extent ext3 files Darrick J. Wong
2014-12-04 23:20   ` Andreas Dilger
2014-12-04 23:45     ` Darrick J. Wong
2014-12-11 22:05   ` [PATCH v2 " Darrick J. Wong
2014-11-07 21:54 ` [PATCH 34/47] tests: verify proper rebuilding of sparse extent trees and block map file conversion Darrick J. Wong
2014-11-07 21:54 ` [PATCH 35/47] resize2fs: convert fs to and from 64bit mode Darrick J. Wong
2014-12-15  3:32   ` Theodore Ts'o
2014-12-15  6:41     ` Darrick J. Wong
2014-12-15 17:46   ` Theodore Ts'o
2014-11-07 21:54 ` [PATCH 36/47] tests: test resize2fs 32->64 and 64->32bit conversion code Darrick J. Wong
2014-12-15 17:46   ` Theodore Ts'o
2014-11-07 21:54 ` [PATCH 37/47] libext2fs: support allocating uninit blocks in bmap2() Darrick J. Wong
2014-11-07 21:54 ` [PATCH 38/47] libext2fs: find/alloc a range of empty blocks Darrick J. Wong
2014-11-07 21:54 ` [PATCH 39/47] libext2fs: add new hooks to support large allocations Darrick J. Wong
2014-11-07 21:55 ` [PATCH 40/47] libext2fs: implement fallocate Darrick J. Wong
2014-11-07 21:55 ` [PATCH 41/47] libext2fs: use fallocate for creating journals and hugefiles Darrick J. Wong
2014-11-07 21:55 ` [PATCH 42/47] debugfs: implement fallocate Darrick J. Wong
2014-11-07 21:55 ` [PATCH 43/47] tests: test debugfs punch command Darrick J. Wong
2014-11-07 21:55 ` [PATCH 45/47] fuse2fs: translate ACL structures Darrick J. Wong
2014-11-07 21:55 ` [PATCH 46/47] fuse2fs: handle 64-bit dates correctly Darrick J. Wong
2014-11-07 21:55 ` [PATCH 47/47] fuse2fs: implement fallocate Darrick J. Wong
2014-11-12 22:43 ` [PATCH 48/47] misc: fix infinite loop when finding the start of the hugefile start range Darrick J. Wong
2014-12-03  3:06   ` Theodore Ts'o
2014-11-27  0:01 ` [PATCH 49/47] libext2fs: don't report garbage inodes with really large inodes Darrick J. Wong
2014-12-03  3:18   ` Theodore Ts'o
2014-12-04 20:39 ` [PATCH 51/47] e2fsck: force-reread of inode from disk when re-checking a checksum error Darrick J. Wong
2014-12-11 22:49   ` Theodore Ts'o

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=20141107215205.883.11850.stgit@birch.djwong.org \
    --to=darrick.wong@oracle.com \
    --cc=linux-ext4@vger.kernel.org \
    --cc=tytso@mit.edu \
    /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).