Commit 4e2f84e6 authored by Liu Bo's avatar Liu Bo Committed by Chris Mason
Browse files

Btrfs: improve fsync by filtering extents that we want 



This is based on Josef's "Btrfs: turbo charge fsync".

The above Josef's patch performs very good in random sync write test,
because we won't have too much extents to merge.

However, it does not performs good on the test:
dd if=/dev/zero of=foobar bs=4k count=12500 oflag=sync

The reason is when we do sequencial sync write, we need to merge the
current extent just with the previous one, so that we can get accumulated
extents to log:

A(4k) --> AA(8k) --> AAA(12k) --> AAAA(16k) ...

So we'll have to flush more and more checksum into log tree, which is the
bottleneck according to my tests.

But we can avoid this by telling fsync the real extents that are needed
to be logged.

With this, I did the above dd sync write test (size=50m),

         w/o (orig)   w/ (josef's)   w/ (this)
SATA      104KB/s       109KB/s       121KB/s
ramdisk   1.5MB/s       1.5MB/s       10.7MB/s (613%)

Signed-off-by: default avatarLiu Bo <bo.li.liu@oracle.com>
parent ca7e70f5

fs/btrfs/extent_map.c

+20 −0
Original line number Original line Diff line number Diff line
@@ -203,6 +203,8 @@ static void try_merge_map(struct extent_map_tree *tree, struct extent_map *em) 
			em->block_start = merge->block_start;

			em->block_start = merge->block_start;

			merge->in_tree = 0;

			merge->in_tree = 0;

			if (merge->generation > em->generation) {

			if (merge->generation > em->generation) {

				em->mod_start = em->start;

				em->mod_len = em->len;

				em->generation = merge->generation;

				em->generation = merge->generation;

				list_move(&em->list, &tree->modified_extents);

				list_move(&em->list, &tree->modified_extents);

			}

			}
@@ -222,6 +224,7 @@ static void try_merge_map(struct extent_map_tree *tree, struct extent_map *em) 
		rb_erase(&merge->rb_node, &tree->map);

		rb_erase(&merge->rb_node, &tree->map);

		merge->in_tree = 0;

		merge->in_tree = 0;

		if (merge->generation > em->generation) {

		if (merge->generation > em->generation) {

			em->mod_len = em->len;

			em->generation = merge->generation;

			em->generation = merge->generation;

			list_move(&em->list, &tree->modified_extents);

			list_move(&em->list, &tree->modified_extents);

		}

		}
@@ -247,6 +250,7 @@ int unpin_extent_cache(struct extent_map_tree *tree, u64 start, u64 len, 
{

{

	int ret = 0;

	int ret = 0;

	struct extent_map *em;

	struct extent_map *em;

	bool prealloc = false;





	write_lock(&tree->lock);

	write_lock(&tree->lock);

	em = lookup_extent_mapping(tree, start, len);

	em = lookup_extent_mapping(tree, start, len);
@@ -259,8 +263,21 @@ int unpin_extent_cache(struct extent_map_tree *tree, u64 start, u64 len, 
	list_move(&em->list, &tree->modified_extents);

	list_move(&em->list, &tree->modified_extents);

	em->generation = gen;

	em->generation = gen;

	clear_bit(EXTENT_FLAG_PINNED, &em->flags);

	clear_bit(EXTENT_FLAG_PINNED, &em->flags);

	em->mod_start = em->start;

	em->mod_len = em->len;



	if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) {

		prealloc = true;

		clear_bit(EXTENT_FLAG_PREALLOC, &em->flags);

	}





	try_merge_map(tree, em);

	try_merge_map(tree, em);



	if (prealloc) {

		em->mod_start = em->start;

		em->mod_len = em->len;

	}



	free_extent_map(em);

	free_extent_map(em);

out:

out:

	write_unlock(&tree->lock);

	write_unlock(&tree->lock);
@@ -298,6 +315,9 @@ int add_extent_mapping(struct extent_map_tree *tree, 
	}

	}

	atomic_inc(&em->refs);

	atomic_inc(&em->refs);





	em->mod_start = em->start;

	em->mod_len = em->len;



	try_merge_map(tree, em);

	try_merge_map(tree, em);

out:

out:

	return ret;

	return ret;

fs/btrfs/extent_map.h

+2 −0
Original line number Original line Diff line number Diff line
@@ -20,6 +20,8 @@ struct extent_map { 
	/* all of these are in bytes */

	/* all of these are in bytes */

	u64 start;

	u64 start;

	u64 len;

	u64 len;

	u64 mod_start;

	u64 mod_len;

	u64 orig_start;

	u64 orig_start;

	u64 block_start;

	u64 block_start;

	u64 block_len;

	u64 block_len;

fs/btrfs/inode.c

+1 −0
Original line number Original line Diff line number Diff line
@@ -1308,6 +1308,7 @@ out_check: 
			em->block_start = disk_bytenr;

			em->block_start = disk_bytenr;

			em->bdev = root->fs_info->fs_devices->latest_bdev;

			em->bdev = root->fs_info->fs_devices->latest_bdev;

			set_bit(EXTENT_FLAG_PINNED, &em->flags);

			set_bit(EXTENT_FLAG_PINNED, &em->flags);

			set_bit(EXTENT_FLAG_PREALLOC, &em->flags);

			while (1) {

			while (1) {

				write_lock(&em_tree->lock);

				write_lock(&em_tree->lock);

				ret = add_extent_mapping(em_tree, em);

				ret = add_extent_mapping(em_tree, em);

fs/btrfs/tree-log.c

+3 −3
Original line number Original line Diff line number Diff line
@@ -2833,8 +2833,8 @@ static int log_one_extent(struct btrfs_trans_handle *trans, 
	struct btrfs_root *log = root->log_root;

	struct btrfs_root *log = root->log_root;

	struct btrfs_file_extent_item *fi;

	struct btrfs_file_extent_item *fi;

	struct btrfs_key key;

	struct btrfs_key key;

	u64 start = em->start;

	u64 start = em->mod_start;

	u64 len = em->len;

	u64 len = em->mod_len;

	u64 num_bytes;

	u64 num_bytes;

	int nritems;

	int nritems;

	int ret;

	int ret;
@@ -2970,7 +2970,7 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, 
		 * sequential then we need to copy the items we have and redo

		 * sequential then we need to copy the items we have and redo

		 * our search

		 * our search

		 */

		 */

		if (args.nr && em->start != args.next_offset) {

		if (args.nr && em->mod_start != args.next_offset) {

			ret = copy_items(trans, log, dst_path, args.src,

			ret = copy_items(trans, log, dst_path, args.src,

					 args.start_slot, args.nr,

					 args.start_slot, args.nr,

					 LOG_INODE_ALL);

					 LOG_INODE_ALL);

CRA Git | Maintained and supported by SUSTech CRA and CCSE