From 8c48f7e88e293b9dd422bd8884842aea85d30b22 Subject: [PATCH] ext4: optimize ext4_find_delalloc_range() in nodelalloc mode We found performance regression when using bigalloc with "nodelalloc" (1MB cluster size): 1. mke2fs -C 1048576 -O ^has_journal,bigalloc /dev/sda 2. mount -o nodelalloc /dev/sda /test/ 3. time dd if=/dev/zero of=/test/io bs=1048576 count=1024 The "dd" will cost about 2 seconds to finish, but if we mke2fs without "bigalloc", "dd" will only cost less than 1 second. The reason is: when using ext4 with "nodelalloc", it will call ext4_find_delalloc_cluster() nearly everytime it call ext4_ext_map_blocks(), and ext4_find_delalloc_range() will also scan all pages in cluster because no buffer is "delayed". A cluster has 256 pages (1MB cluster), so it will scan 256 * 256k pags when creating a 1G file. That severely hurts the performance. Therefore, we return immediately from ext4_find_delalloc_range() in nodelalloc mode, since by definition there can't be any delalloc pages. Signed-off-by: Robin Dong Signed-off-by: "Theodore Ts'o" --- fs/ext4/extents.c | 3 +++ 1 file changed, 3 insertions(+) Index: linux-stage/fs/ext4/extents.c =================================================================== --- linux-stage.orig/fs/ext4/extents.c +++ linux-stage/fs/ext4/extents.c @@ -3909,6 +3909,9 @@ int ext4_find_delalloc_range(struct inod { struct extent_status es; + if (!test_opt(inode->i_sb, DELALLOC)) + return 0; + ext4_es_find_delayed_extent_range(inode, lblk_start, lblk_end, &es); if (es.es_len == 0) return 0; /* there is no delay extent in this tree */ @@ -5115,6 +5118,9 @@ static int ext4_find_delayed_extent(stru struct extent_status es; ext4_lblk_t block, next_del; + if (!test_opt(inode->i_sb, DELALLOC)) + return 0; + if (newes->es_pblk == 0) { ext4_es_find_delayed_extent_range(inode, newes->es_lblk, newes->es_lblk + newes->es_len - 1, &es);