ext4: Fix small file fragmentation

For small file block allocations, mballoc uses per cpu prealloc
space.  Use goal block when searching for the right prealloc
space.  Also make sure ext4_da_writepages tries to write
all the pages for small files in single attempt

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

2 files changed, 61 insertions, 13 deletions

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index d1906d9..7e91913e 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -2282,13 +2282,12 @@ static int ext4_da_writepages_trans_blocks(struct inode *inode)
 static int ext4_da_writepages(struct address_space *mapping,
 			      struct writeback_control *wbc)
 {
-	struct inode *inode = mapping->host;
 	handle_t *handle = NULL;
-	int needed_blocks;
-	int ret = 0;
-	long to_write;
 	loff_t range_start = 0;
-	long pages_skipped = 0;
+	struct inode *inode = mapping->host;
+	int needed_blocks, ret = 0, nr_to_writebump = 0;
+	long to_write, pages_skipped = 0;
+	struct ext4_sb_info *sbi = EXT4_SB(mapping->host->i_sb);
 
 	/*
 	 * No pages to write? This is mainly a kludge to avoid starting
@@ -2297,6 +2296,16 @@ static int ext4_da_writepages(struct address_space *mapping,
 	 */
 	if (!mapping->nrpages || !mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
 		return 0;
+	/*
+	 * Make sure nr_to_write is >= sbi->s_mb_stream_request
+	 * This make sure small files blocks are allocated in
+	 * single attempt. This ensure that small files
+	 * get less fragmented.
+	 */
+	if (wbc->nr_to_write < sbi->s_mb_stream_request) {
+		nr_to_writebump = sbi->s_mb_stream_request - wbc->nr_to_write;
+		wbc->nr_to_write = sbi->s_mb_stream_request;
+	}
 
 	if (!wbc->range_cyclic)
 		/*
@@ -2377,7 +2386,7 @@ restart_loop:
 	}
 
 out_writepages:
-	wbc->nr_to_write = to_write;
+	wbc->nr_to_write = to_write - nr_to_writebump;
 	wbc->range_start = range_start;
 	return ret;
 }
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 865e9dd..e0e3a5e 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -3282,6 +3282,35 @@ static void ext4_mb_use_group_pa(struct ext4_allocation_context *ac,
 }
 
 /*
+ * Return the prealloc space that have minimal distance
+ * from the goal block. @cpa is the prealloc
+ * space that is having currently known minimal distance
+ * from the goal block.
+ */
+static struct ext4_prealloc_space *
+ext4_mb_check_group_pa(ext4_fsblk_t goal_block,
+			struct ext4_prealloc_space *pa,
+			struct ext4_prealloc_space *cpa)
+{
+	ext4_fsblk_t cur_distance, new_distance;
+
+	if (cpa == NULL) {
+		atomic_inc(&pa->pa_count);
+		return pa;
+	}
+	cur_distance = abs(goal_block - cpa->pa_pstart);
+	new_distance = abs(goal_block - pa->pa_pstart);
+
+	if (cur_distance < new_distance)
+		return cpa;
+
+	/* drop the previous reference */
+	atomic_dec(&cpa->pa_count);
+	atomic_inc(&pa->pa_count);
+	return pa;
+}
+
+/*
  * search goal blocks in preallocated space
  */
 static noinline_for_stack int
@@ -3290,7 +3319,8 @@ ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
 	int order, i;
 	struct ext4_inode_info *ei = EXT4_I(ac->ac_inode);
 	struct ext4_locality_group *lg;
-	struct ext4_prealloc_space *pa;
+	struct ext4_prealloc_space *pa, *cpa = NULL;
+	ext4_fsblk_t goal_block;
 
 	/* only data can be preallocated */
 	if (!(ac->ac_flags & EXT4_MB_HINT_DATA))
@@ -3333,6 +3363,13 @@ ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
 		/* The max size of hash table is PREALLOC_TB_SIZE */
 		order = PREALLOC_TB_SIZE - 1;
 
+	goal_block = ac->ac_g_ex.fe_group * EXT4_BLOCKS_PER_GROUP(ac->ac_sb) +
+		     ac->ac_g_ex.fe_start +
+		     le32_to_cpu(EXT4_SB(ac->ac_sb)->s_es->s_first_data_block);
+	/*
+	 * search for the prealloc space that is having
+	 * minimal distance from the goal block.
+	 */
 	for (i = order; i < PREALLOC_TB_SIZE; i++) {
 		rcu_read_lock();
 		list_for_each_entry_rcu(pa, &lg->lg_prealloc_list[i],
@@ -3340,17 +3377,19 @@ ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
 			spin_lock(&pa->pa_lock);
 			if (pa->pa_deleted == 0 &&
 					pa->pa_free >= ac->ac_o_ex.fe_len) {
-				atomic_inc(&pa->pa_count);
-				ext4_mb_use_group_pa(ac, pa);
-				spin_unlock(&pa->pa_lock);
-				ac->ac_criteria = 20;
-				rcu_read_unlock();
-				return 1;
+
+				cpa = ext4_mb_check_group_pa(goal_block,
+								pa, cpa);
 			}
 			spin_unlock(&pa->pa_lock);
 		}
 		rcu_read_unlock();
 	}
+	if (cpa) {
+		ext4_mb_use_group_pa(ac, cpa);
+		ac->ac_criteria = 20;
+		return 1;
+	}
 	return 0;
 }