From a482039889b85c45fc9616e65d560db7a35d4f54 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Mon, 9 Sep 2013 13:19:42 +0800 Subject: Btrfs: allocate the free space by the existed max extent size when ENOSPC By the current code, if the requested size is very large, and all the extents in the free space cache are small, we will waste lots of the cpu time to cut the requested size in half and search the cache again and again until it gets down to the size the allocator can return. In fact, we can know the max extent size in the cache after the first search, so we needn't cut the size in half repeatedly, and just use the max extent size directly. This way can save lots of cpu time and make the performance grow up when there are only fragments in the free space cache. According to my test, if there are only 4KB free space extents in the fs, and the total size of those extents are 256MB, we can reduce the execute time of the following test from 5.4s to 1.4s. dd if=/dev/zero of= bs=1MB count=1 oflag=sync Changelog v2 -> v3: - fix the problem that we skip the block group with the space which is less than we need. Changelog v1 -> v2: - address the problem that we return a wrong start position when searching the free space in a bitmap. Signed-off-by: Miao Xie Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 33 ++++++++++++++++++++++++--------- 1 file changed, 24 insertions(+), 9 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index cfb3cf711b34..2f03181b777f 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6117,10 +6117,13 @@ enum btrfs_loop_type { /* * walks the btree of allocated extents and find a hole of a given size. * The key ins is changed to record the hole: - * ins->objectid == block start + * ins->objectid == start position * ins->flags = BTRFS_EXTENT_ITEM_KEY - * ins->offset == number of blocks + * ins->offset == the size of the hole. * Any available blocks before search_start are skipped. + * + * If there is no suitable free space, we will record the max size of + * the free space extent currently. */ static noinline int find_free_extent(struct btrfs_root *orig_root, u64 num_bytes, u64 empty_size, @@ -6133,6 +6136,7 @@ static noinline int find_free_extent(struct btrfs_root *orig_root, struct btrfs_block_group_cache *block_group = NULL; struct btrfs_block_group_cache *used_block_group; u64 search_start = 0; + u64 max_extent_size = 0; int empty_cluster = 2 * 1024 * 1024; struct btrfs_space_info *space_info; int loop = 0; @@ -6292,7 +6296,10 @@ have_block_group: btrfs_get_block_group(used_block_group); offset = btrfs_alloc_from_cluster(used_block_group, - last_ptr, num_bytes, used_block_group->key.objectid); + last_ptr, + num_bytes, + used_block_group->key.objectid, + &max_extent_size); if (offset) { /* we have a block, we're done */ spin_unlock(&last_ptr->refill_lock); @@ -6355,8 +6362,10 @@ refill_cluster: * cluster */ offset = btrfs_alloc_from_cluster(block_group, - last_ptr, num_bytes, - search_start); + last_ptr, + num_bytes, + search_start, + &max_extent_size); if (offset) { /* we found one, proceed */ spin_unlock(&last_ptr->refill_lock); @@ -6391,13 +6400,18 @@ unclustered_alloc: if (cached && block_group->free_space_ctl->free_space < num_bytes + empty_cluster + empty_size) { + if (block_group->free_space_ctl->free_space > + max_extent_size) + max_extent_size = + block_group->free_space_ctl->free_space; spin_unlock(&block_group->free_space_ctl->tree_lock); goto loop; } spin_unlock(&block_group->free_space_ctl->tree_lock); offset = btrfs_find_space_for_alloc(block_group, search_start, - num_bytes, empty_size); + num_bytes, empty_size, + &max_extent_size); /* * If we didn't find a chunk, and we haven't failed on this * block group before, and this block group is in the middle of @@ -6515,7 +6529,8 @@ loop: ret = 0; } out: - + if (ret == -ENOSPC) + ins->offset = max_extent_size; return ret; } @@ -6573,8 +6588,8 @@ again: flags); if (ret == -ENOSPC) { - if (!final_tried) { - num_bytes = num_bytes >> 1; + if (!final_tried && ins->offset) { + num_bytes = min(num_bytes >> 1, ins->offset); num_bytes = round_down(num_bytes, root->sectorsize); num_bytes = max(num_bytes, min_alloc_size); if (num_bytes == min_alloc_size) -- cgit v1.2.2 From 14575aef4212d2cc02274bf1f5456f3e644f03f8 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 17 Sep 2013 10:48:00 -0400 Subject: Revert "Btrfs: rework the overcommit logic to be based on the total size" This reverts commit 70afa3998c9baed4186df38988246de1abdab56d. It is causing performance issues and wasn't actually correct. There were problems with the way we flushed delalloc and that was the real cause of the early enospc. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 15 +++------------ 1 file changed, 3 insertions(+), 12 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 2f03181b777f..07f7f6942128 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3925,7 +3925,6 @@ static int can_overcommit(struct btrfs_root *root, u64 space_size; u64 avail; u64 used; - u64 to_add; used = space_info->bytes_used + space_info->bytes_reserved + space_info->bytes_pinned + space_info->bytes_readonly; @@ -3959,25 +3958,17 @@ static int can_overcommit(struct btrfs_root *root, BTRFS_BLOCK_GROUP_RAID10)) avail >>= 1; - to_add = space_info->total_bytes; - /* * If we aren't flushing all things, let us overcommit up to * 1/2th of the space. If we can flush, don't let us overcommit * too much, let it overcommit up to 1/8 of the space. */ if (flush == BTRFS_RESERVE_FLUSH_ALL) - to_add >>= 3; + avail >>= 3; else - to_add >>= 1; - - /* - * Limit the overcommit to the amount of free space we could possibly - * allocate for chunks. - */ - to_add = min(avail, to_add); + avail >>= 1; - if (used + bytes < space_info->total_bytes + to_add) + if (used + bytes < space_info->total_bytes + avail) return 1; return 0; } -- cgit v1.2.2 From f0de181c9b48a397c5a2fbe63dcdd2a26a872695 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 17 Sep 2013 10:55:51 -0400 Subject: Btrfs: kill delay_iput arg to the wait_ordered functions This is a left over of how we used to wait for ordered extents, which was to grab the inode and then run filemap flush on it. However if we have an ordered extent then we already are holding a ref on the inode, and we just use btrfs_start_ordered_extent anyway, so there is no reason to have an extra ref on the inode to start work on the ordered extent. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 07f7f6942128..88c6db611de3 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3991,7 +3991,7 @@ static void btrfs_writeback_inodes_sb_nr(struct btrfs_root *root, */ btrfs_start_all_delalloc_inodes(root->fs_info, 0); if (!current->journal_info) - btrfs_wait_all_ordered_extents(root->fs_info, 0); + btrfs_wait_all_ordered_extents(root->fs_info); } } @@ -4021,7 +4021,7 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, if (delalloc_bytes == 0) { if (trans) return; - btrfs_wait_all_ordered_extents(root->fs_info, 0); + btrfs_wait_all_ordered_extents(root->fs_info); return; } @@ -4049,7 +4049,7 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, loops++; if (wait_ordered && !trans) { - btrfs_wait_all_ordered_extents(root->fs_info, 0); + btrfs_wait_all_ordered_extents(root->fs_info); } else { time_left = schedule_timeout_killable(1); if (time_left) -- cgit v1.2.2 From 363e4d354e51558629d13c5ae456f98c4b209576 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 17 Sep 2013 11:02:25 -0400 Subject: Btrfs: remove space_info->reservation_progress This isn't used for anything anymore, just remove it. Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 88c6db611de3..d58bef130a41 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4456,7 +4456,6 @@ static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, space_info->bytes_may_use -= num_bytes; trace_btrfs_space_reservation(fs_info, "space_info", space_info->flags, num_bytes, 0); - space_info->reservation_progress++; spin_unlock(&space_info->lock); } } @@ -4657,7 +4656,6 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info) sinfo->bytes_may_use -= num_bytes; trace_btrfs_space_reservation(fs_info, "space_info", sinfo->flags, num_bytes, 0); - sinfo->reservation_progress++; block_rsv->reserved = block_rsv->size; block_rsv->full = 1; } @@ -5437,7 +5435,6 @@ static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, space_info->bytes_readonly += num_bytes; cache->reserved -= num_bytes; space_info->bytes_reserved -= num_bytes; - space_info->reservation_progress++; } spin_unlock(&cache->lock); spin_unlock(&space_info->lock); -- cgit v1.2.2