From 581bb050941b4f220f84d3e5ed6dace3d42dd382 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 20 Apr 2011 10:06:11 +0800 Subject: Btrfs: Cache free inode numbers in memory Currently btrfs stores the highest objectid of the fs tree, and it always returns (highest+1) inode number when we create a file, so inode numbers won't be reclaimed when we delete files, so we'll run out of inode numbers as we keep create/delete files in 32bits machines. This fixes it, and it works similarly to how we cache free space in block cgroups. We start a kernel thread to read the file tree. By scanning inode items, we know which chunks of inode numbers are free, and we cache them in an rb-tree. Because we are searching the commit root, we have to carefully handle the cross-transaction case. The rb-tree is a hybrid extent+bitmap tree, so if we have too many small chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram of extents, and a bitmap will be used if we exceed this threshold. The extents threshold is adjusted in runtime. Signed-off-by: Li Zefan --- fs/btrfs/disk-io.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index ef6865c17cd6..d02683b1ee16 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -41,6 +41,7 @@ #include "locking.h" #include "tree-log.h" #include "free-space-cache.h" +#include "inode-map.h" static struct extent_io_ops btree_extent_io_ops; static void end_workqueue_fn(struct btrfs_work *work); @@ -1327,6 +1328,19 @@ again: if (IS_ERR(root)) return root; + root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS); + if (!root->free_ino_ctl) + goto fail; + root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned), + GFP_NOFS); + if (!root->free_ino_pinned) + goto fail; + + btrfs_init_free_ino_ctl(root); + mutex_init(&root->fs_commit_mutex); + spin_lock_init(&root->cache_lock); + init_waitqueue_head(&root->cache_wait); + set_anon_super(&root->anon_super, NULL); if (btrfs_root_refs(&root->root_item) == 0) { @@ -2483,6 +2497,8 @@ int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root) if (btrfs_root_refs(&root->root_item) == 0) synchronize_srcu(&fs_info->subvol_srcu); + __btrfs_remove_free_space_cache(root->free_ino_pinned); + __btrfs_remove_free_space_cache(root->free_ino_ctl); free_fs_root(root); return 0; } @@ -2496,6 +2512,8 @@ static void free_fs_root(struct btrfs_root *root) } free_extent_buffer(root->node); free_extent_buffer(root->commit_root); + kfree(root->free_ino_ctl); + kfree(root->free_ino_pinned); kfree(root->name); kfree(root); } -- cgit v1.2.3 From 82d5902d9c681be37ffa9d70482907f9f0b7ec1f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 20 Apr 2011 10:33:24 +0800 Subject: Btrfs: Support reading/writing on disk free ino cache This is similar to block group caching. We dedicate a special inode in fs tree to save free ino cache. At the very first time we create/delete a file after mount, the free ino cache will be loaded from disk into memory. When the fs tree is commited, the cache will be written back to disk. To keep compatibility, we check the root generation against the generation of the special inode when loading the cache, so the loading will fail if the btrfs filesystem was mounted in an older kernel before. Signed-off-by: Li Zefan --- fs/btrfs/disk-io.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index d02683b1ee16..4f12c30a5470 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2505,6 +2505,7 @@ int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root) static void free_fs_root(struct btrfs_root *root) { + iput(root->cache_inode); WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree)); if (root->anon_super.s_dev) { down_write(&root->anon_super.s_umount); -- cgit v1.2.3 From cfece4db110dacfd6b4b87b912c59e77e6846fc0 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 25 Apr 2011 19:43:52 -0400 Subject: btrfs: add missing spin_unlock to a rare exit path Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index ef6865c17cd6..fe5aec9b3924 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2903,6 +2903,7 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, spin_lock(&delayed_refs->lock); if (delayed_refs->num_entries == 0) { + spin_unlock(&delayed_refs->lock); printk(KERN_INFO "delayed_refs has NO entry\n"); return ret; } -- cgit v1.2.3 From 306e16ce13c0f3d4fc071b45803b5b83c2606011 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 19 Apr 2011 14:29:38 +0200 Subject: btrfs: rename variables clashing with global function names reported by gcc -Wshadow: page_index, page_offset, new_inode, dev_name Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 228cf36ece83..990afa8656a2 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -137,7 +137,7 @@ static const char *btrfs_eb_name[BTRFS_MAX_LEVEL + 1] = { * that covers the entire device */ static struct extent_map *btree_get_extent(struct inode *inode, - struct page *page, size_t page_offset, u64 start, u64 len, + struct page *page, size_t pg_offset, u64 start, u64 len, int create) { struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; -- cgit v1.2.3 From 4891aca2dac612a2f21a3278d9906ade13b55788 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 19 Apr 2011 16:45:00 +0200 Subject: btrfs: fix dereference before check The superblock's ->s_fs_info is properly set in btrfs_fill_super, after a call to open_ctree, which derefereces it before check. Although tree_root is set via btrfs_set_super, let's be defensive and leave the check in place. Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 990afa8656a2..25e4b8f1d0ef 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1611,7 +1611,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, struct btrfs_root *csum_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); struct btrfs_root *tree_root = btrfs_sb(sb); - struct btrfs_fs_info *fs_info = tree_root->fs_info; + struct btrfs_fs_info *fs_info = NULL; struct btrfs_root *chunk_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); struct btrfs_root *dev_root = kzalloc(sizeof(struct btrfs_root), @@ -1623,11 +1623,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, struct btrfs_super_block *disk_super; - if (!extent_root || !tree_root || !fs_info || + if (!extent_root || !tree_root || !tree_root->fs_info || !chunk_root || !dev_root || !csum_root) { err = -ENOMEM; goto fail; } + fs_info = tree_root->fs_info; ret = init_srcu_struct(&fs_info->subvol_srcu); if (ret) { -- cgit v1.2.3 From f993c883ad8e111fb9e9ae603540acbe94f7246c Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 20 Apr 2011 23:35:57 +0200 Subject: btrfs: drop unused argument from extent_io_tree_init all callers pass GFP_NOFS, but the GFP mask argument is not used in the function; GFP_ATOMIC is passed to radix tree initialization and it's the only correct one, since we're using the preload/insert mechanism of radix tree. Let's drop the gfp mask from btrfs function, this will not change behaviour. Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 25e4b8f1d0ef..3ce80f71e98a 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1080,7 +1080,7 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->log_transid = 0; root->last_log_commit = 0; extent_io_tree_init(&root->dirty_log_pages, - fs_info->btree_inode->i_mapping, GFP_NOFS); + fs_info->btree_inode->i_mapping); memset(&root->root_key, 0, sizeof(root->root_key)); memset(&root->root_item, 0, sizeof(root->root_item)); @@ -1712,8 +1712,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node); extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree, - fs_info->btree_inode->i_mapping, - GFP_NOFS); + fs_info->btree_inode->i_mapping); extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree, GFP_NOFS); @@ -1729,9 +1728,9 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->block_group_cache_tree = RB_ROOT; extent_io_tree_init(&fs_info->freed_extents[0], - fs_info->btree_inode->i_mapping, GFP_NOFS); + fs_info->btree_inode->i_mapping); extent_io_tree_init(&fs_info->freed_extents[1], - fs_info->btree_inode->i_mapping, GFP_NOFS); + fs_info->btree_inode->i_mapping); fs_info->pinned_extents = &fs_info->freed_extents[0]; fs_info->do_barriers = 1; -- cgit v1.2.3 From a8067e022ab54fde8953880a64572c3acca644dc Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 21 Apr 2011 00:34:43 +0200 Subject: btrfs: drop unused parameter from extent_map_tree_init the GFP flags are not stored anywhere and all allocations are done via alloc_extent_map(GFP_NOFS). Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 3ce80f71e98a..f2ee584b8efd 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1713,8 +1713,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node); extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree, fs_info->btree_inode->i_mapping); - extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree, - GFP_NOFS); + extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree); BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops; -- cgit v1.2.3 From 172ddd60a662c4d8bf2809462866ddddd6431ea5 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 21 Apr 2011 00:48:27 +0200 Subject: btrfs: drop gfp parameter from alloc_extent_map pass GFP_NOFS directly to kmem_cache_alloc Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index f2ee584b8efd..e1e55679d061 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -154,7 +154,7 @@ static struct extent_map *btree_get_extent(struct inode *inode, } read_unlock(&em_tree->lock); - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) { em = ERR_PTR(-ENOMEM); goto out; -- cgit v1.2.3 From f09d1f60e6aa82fb4cfaa525e21f6287fc1516f4 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 21 Apr 2011 01:08:01 +0200 Subject: btrfs: drop gfp parameter from find_extent_buffer pass GFP_NOFS directly to kmem_cache_alloc Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index e1e55679d061..1c0752e99066 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -963,7 +963,7 @@ struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, struct inode *btree_inode = root->fs_info->btree_inode; struct extent_buffer *eb; eb = find_extent_buffer(&BTRFS_I(btree_inode)->io_tree, - bytenr, blocksize, GFP_NOFS); + bytenr, blocksize); return eb; } @@ -2696,7 +2696,7 @@ int btree_lock_page_hook(struct page *page) goto out; len = page->private >> 2; - eb = find_extent_buffer(io_tree, bytenr, len, GFP_NOFS); + eb = find_extent_buffer(io_tree, bytenr, len); if (!eb) goto out; -- cgit v1.2.3 From ba14419264684b290f0d0b7f48d26eafb11fc0c6 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 21 Apr 2011 01:12:06 +0200 Subject: btrfs: drop gfp parameter from alloc_extent_buffer pass GFP_NOFS directly to kmem_cache_alloc Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 1c0752e99066..4084959b36fd 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -380,7 +380,7 @@ static int csum_dirty_buffer(struct btrfs_root *root, struct page *page) len = page->private >> 2; WARN_ON(len == 0); - eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS); + eb = alloc_extent_buffer(tree, start, len, page); if (eb == NULL) { WARN_ON(1); goto out; @@ -525,7 +525,7 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, len = page->private >> 2; WARN_ON(len == 0); - eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS); + eb = alloc_extent_buffer(tree, start, len, page); if (eb == NULL) { ret = -EIO; goto out; @@ -974,7 +974,7 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, struct extent_buffer *eb; eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree, - bytenr, blocksize, NULL, GFP_NOFS); + bytenr, blocksize, NULL); return eb; } -- cgit v1.2.3 From f2a97a9dbd86eb1ef956bdf20e05c507b32beb96 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 5 May 2011 12:44:41 +0200 Subject: btrfs: remove all unused functions Remove static and global declarations and/or definitions. Reduces size of btrfs.ko by ~3.4kB. text data bss dec hex filename 402081 7464 200 409745 64091 btrfs.ko.base 398620 7144 200 405964 631cc btrfs.ko.remove-all Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 27 --------------------------- 1 file changed, 27 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 4084959b36fd..fa287c551ffc 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -650,12 +650,6 @@ unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info) return 256 * limit; } -int btrfs_congested_async(struct btrfs_fs_info *info, int iodone) -{ - return atomic_read(&info->nr_async_bios) > - btrfs_async_submit_limit(info); -} - static void run_one_async_start(struct btrfs_work *work) { struct async_submit_bio *async; @@ -1283,21 +1277,6 @@ out: return root; } -struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info, - u64 root_objectid) -{ - struct btrfs_root *root; - - if (root_objectid == BTRFS_ROOT_TREE_OBJECTID) - return fs_info->tree_root; - if (root_objectid == BTRFS_EXTENT_TREE_OBJECTID) - return fs_info->extent_root; - - root = radix_tree_lookup(&fs_info->fs_roots_radix, - (unsigned long)root_objectid); - return root; -} - struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info, struct btrfs_key *location) { @@ -1369,11 +1348,6 @@ fail: return ERR_PTR(ret); } -struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info, - struct btrfs_key *location, - const char *name, int namelen) -{ - return btrfs_read_fs_root_no_name(fs_info, location); #if 0 struct btrfs_root *root; int ret; @@ -1402,7 +1376,6 @@ struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info, root->in_sysfs = 1; return root; #endif -} static int btrfs_congested_fn(void *congested_data, int bdi_bits) { -- cgit v1.2.3 From 182608c8294b5fe90d7bbd4b026c82bf0a24b736 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 5 May 2011 13:13:16 +0200 Subject: btrfs: remove old unused commented out code Remove code which has been #if0-ed out for a very long time and does not seem to be related to current codebase anymore. Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 29 ----------------------------- 1 file changed, 29 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index fa287c551ffc..de7b4770ab17 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1348,35 +1348,6 @@ fail: return ERR_PTR(ret); } -#if 0 - struct btrfs_root *root; - int ret; - - root = btrfs_read_fs_root_no_name(fs_info, location); - if (!root) - return NULL; - - if (root->in_sysfs) - return root; - - ret = btrfs_set_root_name(root, name, namelen); - if (ret) { - free_extent_buffer(root->node); - kfree(root); - return ERR_PTR(ret); - } - - ret = btrfs_sysfs_add_root(root); - if (ret) { - free_extent_buffer(root->node); - kfree(root->name); - kfree(root); - return ERR_PTR(ret); - } - root->in_sysfs = 1; - return root; -#endif - static int btrfs_congested_fn(void *congested_data, int bdi_bits) { struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data; -- cgit v1.2.3 From a2de733c78fa7af51ba9670482fa7d392aa67c57 Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Tue, 8 Mar 2011 14:14:00 +0100 Subject: btrfs: scrub This adds an initial implementation for scrub. It works quite straightforward. The usermode issues an ioctl for each device in the fs. For each device, it enumerates the allocated device chunks. For each chunk, the contained extents are enumerated and the data checksums fetched. The extents are read sequentially and the checksums verified. If an error occurs (checksum or EIO), a good copy is searched for. If one is found, the bad copy will be rewritten. All enumerations happen from the commit roots. During a transaction commit, the scrubs get paused and afterwards continue from the new roots. This commit is based on the series originally posted to linux-btrfs with some improvements that resulted from comments from David Sterba, Ilya Dryomov and Jan Schmidt. Signed-off-by: Arne Jansen --- fs/btrfs/disk-io.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index fe5aec9b3924..e48e8095c61f 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1773,6 +1773,17 @@ struct btrfs_root *open_ctree(struct super_block *sb, INIT_LIST_HEAD(&fs_info->ordered_extents); spin_lock_init(&fs_info->ordered_extent_lock); + mutex_init(&fs_info->scrub_lock); + atomic_set(&fs_info->scrubs_running, 0); + atomic_set(&fs_info->scrub_pause_req, 0); + atomic_set(&fs_info->scrubs_paused, 0); + atomic_set(&fs_info->scrub_cancel_req, 0); + init_waitqueue_head(&fs_info->scrub_pause_wait); + init_rwsem(&fs_info->scrub_super_lock); + fs_info->scrub_workers_refcnt = 0; + btrfs_init_workers(&fs_info->scrub_workers, "scrub", + fs_info->thread_pool_size, &fs_info->generic_worker); + sb->s_blocksize = 4096; sb->s_blocksize_bits = blksize_bits(4096); sb->s_bdi = &fs_info->bdi; @@ -2599,6 +2610,7 @@ int close_ctree(struct btrfs_root *root) fs_info->closing = 1; smp_mb(); + btrfs_scrub_cancel(root); btrfs_put_block_group_cache(fs_info); /* -- cgit v1.2.3 From 7a36ddec1003a4e84e79f28ee714a142ed6bc529 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Fri, 6 May 2011 15:33:15 +0200 Subject: btrfs: use printk_ratelimited instead of printk_ratelimit As per printk_ratelimit comment, it should not be used. Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 21 ++++++--------------- 1 file changed, 6 insertions(+), 15 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index de7b4770ab17..cb9d1b8bfe74 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include "compat.h" #include "ctree.h" @@ -254,14 +255,12 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf, memcpy(&found, result, csum_size); read_extent_buffer(buf, &val, 0, csum_size); - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs: %s checksum verify " + printk_ratelimited(KERN_INFO "btrfs: %s checksum verify " "failed on %llu wanted %X found %X " "level %d\n", root->fs_info->sb->s_id, (unsigned long long)buf->start, val, found, btrfs_header_level(buf)); - } if (result != (char *)&inline_result) kfree(result); return 1; @@ -296,13 +295,11 @@ static int verify_parent_transid(struct extent_io_tree *io_tree, ret = 0; goto out; } - if (printk_ratelimit()) { - printk("parent transid verify failed on %llu wanted %llu " + printk_ratelimited("parent transid verify failed on %llu wanted %llu " "found %llu\n", (unsigned long long)eb->start, (unsigned long long)parent_transid, (unsigned long long)btrfs_header_generation(eb)); - } ret = 1; clear_extent_buffer_uptodate(io_tree, eb, &cached_state); out: @@ -533,12 +530,10 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, found_start = btrfs_header_bytenr(eb); if (found_start != start) { - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs bad tree block start " + printk_ratelimited(KERN_INFO "btrfs bad tree block start " "%llu %llu\n", (unsigned long long)found_start, (unsigned long long)eb->start); - } ret = -EIO; goto err; } @@ -550,10 +545,8 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, goto err; } if (check_tree_block_fsid(root, eb)) { - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs bad fsid on block %llu\n", + printk_ratelimited(KERN_INFO "btrfs bad fsid on block %llu\n", (unsigned long long)eb->start); - } ret = -EIO; goto err; } @@ -2108,11 +2101,9 @@ static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate) if (uptodate) { set_buffer_uptodate(bh); } else { - if (printk_ratelimit()) { - printk(KERN_WARNING "lost page write due to " + printk_ratelimited(KERN_WARNING "lost page write due to " "I/O error on %s\n", bdevname(bh->b_bdev, b)); - } /* note, we dont' set_buffer_write_io_error because we have * our own ways of dealing with the IO errors */ -- cgit v1.2.3 From 16cdcec736cd214350cdb591bf1091f8beedefa0 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Fri, 22 Apr 2011 18:12:22 +0800 Subject: btrfs: implement delayed inode items operation Changelog V5 -> V6: - Fix oom when the memory load is high, by storing the delayed nodes into the root's radix tree, and letting btrfs inodes go. Changelog V4 -> V5: - Fix the race on adding the delayed node to the inode, which is spotted by Chris Mason. - Merge Chris Mason's incremental patch into this patch. - Fix deadlock between readdir() and memory fault, which is reported by Itaru Kitayama. Changelog V3 -> V4: - Fix nested lock, which is reported by Itaru Kitayama, by updating space cache inode in time. Changelog V2 -> V3: - Fix the race between the delayed worker and the task which does delayed items balance, which is reported by Tsutomu Itoh. - Modify the patch address David Sterba's comment. - Fix the bug of the cpu recursion spinlock, reported by Chris Mason Changelog V1 -> V2: - break up the global rb-tree, use a list to manage the delayed nodes, which is created for every directory and file, and used to manage the delayed directory name index items and the delayed inode item. - introduce a worker to deal with the delayed nodes. Compare with Ext3/4, the performance of file creation and deletion on btrfs is very poor. the reason is that btrfs must do a lot of b+ tree insertions, such as inode item, directory name item, directory name index and so on. If we can do some delayed b+ tree insertion or deletion, we can improve the performance, so we made this patch which implemented delayed directory name index insertion/deletion and delayed inode update. Implementation: - introduce a delayed root object into the filesystem, that use two lists to manage the delayed nodes which are created for every file/directory. One is used to manage all the delayed nodes that have delayed items. And the other is used to manage the delayed nodes which is waiting to be dealt with by the work thread. - Every delayed node has two rb-tree, one is used to manage the directory name index which is going to be inserted into b+ tree, and the other is used to manage the directory name index which is going to be deleted from b+ tree. - introduce a worker to deal with the delayed operation. This worker is used to deal with the works of the delayed directory name index items insertion and deletion and the delayed inode update. When the delayed items is beyond the lower limit, we create works for some delayed nodes and insert them into the work queue of the worker, and then go back. When the delayed items is beyond the upper bound, we create works for all the delayed nodes that haven't been dealt with, and insert them into the work queue of the worker, and then wait for that the untreated items is below some threshold value. - When we want to insert a directory name index into b+ tree, we just add the information into the delayed inserting rb-tree. And then we check the number of the delayed items and do delayed items balance. (The balance policy is above.) - When we want to delete a directory name index from the b+ tree, we search it in the inserting rb-tree at first. If we look it up, just drop it. If not, add the key of it into the delayed deleting rb-tree. Similar to the delayed inserting rb-tree, we also check the number of the delayed items and do delayed items balance. (The same to inserting manipulation) - When we want to update the metadata of some inode, we cached the data of the inode into the delayed node. the worker will flush it into the b+ tree after dealing with the delayed insertion and deletion. - We will move the delayed node to the tail of the list after we access the delayed node, By this way, we can cache more delayed items and merge more inode updates. - If we want to commit transaction, we will deal with all the delayed node. - the delayed node will be freed when we free the btrfs inode. - Before we log the inode items, we commit all the directory name index items and the delayed inode update. I did a quick test by the benchmark tool[1] and found we can improve the performance of file creation by ~15%, and file deletion by ~20%. Before applying this patch: Create files: Total files: 50000 Total time: 1.096108 Average time: 0.000022 Delete files: Total files: 50000 Total time: 1.510403 Average time: 0.000030 After applying this patch: Create files: Total files: 50000 Total time: 0.932899 Average time: 0.000019 Delete files: Total files: 50000 Total time: 1.215732 Average time: 0.000024 [1] http://marc.info/?l=linux-btrfs&m=128212635122920&q=p3 Many thanks for Kitayama-san's help! Signed-off-by: Miao Xie Reviewed-by: David Sterba Tested-by: Tsutomu Itoh Tested-by: Itaru Kitayama Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 50 +++++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 45 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 228cf36ece83..22c3c9586049 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1058,6 +1058,7 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->name = NULL; root->in_sysfs = 0; root->inode_tree = RB_ROOT; + INIT_RADIX_TREE(&root->delayed_nodes_tree, GFP_ATOMIC); root->block_rsv = NULL; root->orphan_block_rsv = NULL; @@ -1693,6 +1694,13 @@ struct btrfs_root *open_ctree(struct super_block *sb, INIT_LIST_HEAD(&fs_info->ordered_extents); spin_lock_init(&fs_info->ordered_extent_lock); + fs_info->delayed_root = kmalloc(sizeof(struct btrfs_delayed_root), + GFP_NOFS); + if (!fs_info->delayed_root) { + err = -ENOMEM; + goto fail_iput; + } + btrfs_init_delayed_root(fs_info->delayed_root); sb->s_blocksize = 4096; sb->s_blocksize_bits = blksize_bits(4096); @@ -1760,7 +1768,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, bh = btrfs_read_dev_super(fs_devices->latest_bdev); if (!bh) { err = -EINVAL; - goto fail_iput; + goto fail_alloc; } memcpy(&fs_info->super_copy, bh->b_data, sizeof(fs_info->super_copy)); @@ -1772,7 +1780,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, disk_super = &fs_info->super_copy; if (!btrfs_super_root(disk_super)) - goto fail_iput; + goto fail_alloc; /* check FS state, whether FS is broken. */ fs_info->fs_state |= btrfs_super_flags(disk_super); @@ -1788,7 +1796,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, ret = btrfs_parse_options(tree_root, options); if (ret) { err = ret; - goto fail_iput; + goto fail_alloc; } features = btrfs_super_incompat_flags(disk_super) & @@ -1798,7 +1806,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, "unsupported optional features (%Lx).\n", (unsigned long long)features); err = -EINVAL; - goto fail_iput; + goto fail_alloc; } features = btrfs_super_incompat_flags(disk_super); @@ -1814,7 +1822,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, "unsupported option features (%Lx).\n", (unsigned long long)features); err = -EINVAL; - goto fail_iput; + goto fail_alloc; } btrfs_init_workers(&fs_info->generic_worker, @@ -1861,6 +1869,9 @@ struct btrfs_root *open_ctree(struct super_block *sb, &fs_info->generic_worker); btrfs_init_workers(&fs_info->endio_freespace_worker, "freespace-write", 1, &fs_info->generic_worker); + btrfs_init_workers(&fs_info->delayed_workers, "delayed-meta", + fs_info->thread_pool_size, + &fs_info->generic_worker); /* * endios are largely parallel and should have a very @@ -1882,6 +1893,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_start_workers(&fs_info->endio_meta_write_workers, 1); btrfs_start_workers(&fs_info->endio_write_workers, 1); btrfs_start_workers(&fs_info->endio_freespace_worker, 1); + btrfs_start_workers(&fs_info->delayed_workers, 1); fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super); fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages, @@ -2138,6 +2150,9 @@ fail_sb_buffer: btrfs_stop_workers(&fs_info->endio_write_workers); btrfs_stop_workers(&fs_info->endio_freespace_worker); btrfs_stop_workers(&fs_info->submit_workers); + btrfs_stop_workers(&fs_info->delayed_workers); +fail_alloc: + kfree(fs_info->delayed_root); fail_iput: invalidate_inode_pages2(fs_info->btree_inode->i_mapping); iput(fs_info->btree_inode); @@ -2578,6 +2593,7 @@ int close_ctree(struct btrfs_root *root) del_fs_roots(fs_info); iput(fs_info->btree_inode); + kfree(fs_info->delayed_root); btrfs_stop_workers(&fs_info->generic_worker); btrfs_stop_workers(&fs_info->fixup_workers); @@ -2589,6 +2605,7 @@ int close_ctree(struct btrfs_root *root) btrfs_stop_workers(&fs_info->endio_write_workers); btrfs_stop_workers(&fs_info->endio_freespace_worker); btrfs_stop_workers(&fs_info->submit_workers); + btrfs_stop_workers(&fs_info->delayed_workers); btrfs_close_devices(fs_info->fs_devices); btrfs_mapping_tree_free(&fs_info->mapping_tree); @@ -2662,6 +2679,29 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) u64 num_dirty; unsigned long thresh = 32 * 1024 * 1024; + if (current->flags & PF_MEMALLOC) + return; + + btrfs_balance_delayed_items(root); + + num_dirty = root->fs_info->dirty_metadata_bytes; + + if (num_dirty > thresh) { + balance_dirty_pages_ratelimited_nr( + root->fs_info->btree_inode->i_mapping, 1); + } + return; +} + +void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) +{ + /* + * looks as though older kernels can get into trouble with + * this code, they end up stuck in balance_dirty_pages forever + */ + u64 num_dirty; + unsigned long thresh = 32 * 1024 * 1024; + if (current->flags & PF_MEMALLOC) return; -- cgit v1.2.3 From 7a7eaa40a39bde4eefc91aadeb1ce3dc4e6a1252 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 13 Apr 2011 12:54:33 -0400 Subject: Btrfs: take away the num_items argument from btrfs_join_transaction I keep forgetting that btrfs_join_transaction() just ignores the num_items argument, which leads me to sending pointless patches and looking stupid :). So just kill the num_items argument from btrfs_join_transaction and btrfs_start_ioctl_transaction, since neither of them use it. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/disk-io.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 228cf36ece83..9d6c9e332ca3 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1568,7 +1568,7 @@ static int transaction_kthread(void *arg) transid = cur->transid; spin_unlock(&root->fs_info->new_trans_lock); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); if (transid == trans->transid) { ret = btrfs_commit_transaction(trans, root); @@ -2495,13 +2495,13 @@ int btrfs_commit_super(struct btrfs_root *root) down_write(&root->fs_info->cleanup_work_sem); up_write(&root->fs_info->cleanup_work_sem); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); ret = btrfs_commit_transaction(trans, root); BUG_ON(ret); /* run commit again to drop the original snapshot */ - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); btrfs_commit_transaction(trans, root); -- cgit v1.2.3 From a4abeea41adfa3c143c289045f4625dfaeba2212 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 11 Apr 2011 17:25:13 -0400 Subject: Btrfs: kill trans_mutex We use trans_mutex for lots of things, here's a basic list 1) To serialize trans_handles joining the currently running transaction 2) To make sure that no new trans handles are started while we are committing 3) To protect the dead_roots list and the transaction lists Really the serializing trans_handles joining is not too hard, and can really get bogged down in acquiring a reference to the transaction. So replace the trans_mutex with a trans_lock spinlock and use it to do the following 1) Protect fs_info->running_transaction. All trans handles have to do is check this, and then take a reference of the transaction and keep on going. 2) Protect the fs_info->trans_list. This doesn't get used too much, basically it just holds the current transactions, which will usually just be the currently committing transaction and the currently running transaction at most. 3) Protect the dead roots list. This is only ever processed by splicing the list so this is relatively simple. 4) Protect the fs_info->reloc_ctl stuff. This is very lightweight and was using the trans_mutex before, so this is a pretty straightforward change. 5) Protect fs_info->no_trans_join. Because we don't hold the trans_lock over the entirety of the commit we need to have a way to block new people from creating a new transaction while we're doing our work. So we set no_trans_join and in join_transaction we test to see if that is set, and if it is we do a wait_on_commit. 6) Make the transaction use count atomic so we don't need to take locks to modify it when we're dropping references. 7) Add a commit_lock to the transaction to make sure multiple people trying to commit the same transaction don't race and commit at the same time. 8) Make open_ioctl_trans an atomic so we don't have to take any locks for ioctl trans. I have tested this with xfstests, but obviously it is a pretty hairy change so lots of testing is greatly appreciated. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/disk-io.c | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 9d6c9e332ca3..93ef254ec432 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1551,22 +1551,22 @@ static int transaction_kthread(void *arg) vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE); mutex_lock(&root->fs_info->transaction_kthread_mutex); - spin_lock(&root->fs_info->new_trans_lock); + spin_lock(&root->fs_info->trans_lock); cur = root->fs_info->running_transaction; if (!cur) { - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); goto sleep; } now = get_seconds(); if (!cur->blocked && (now < cur->start_time || now - cur->start_time < 30)) { - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); delay = HZ * 5; goto sleep; } transid = cur->transid; - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); @@ -1658,7 +1658,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, INIT_LIST_HEAD(&fs_info->ordered_operations); INIT_LIST_HEAD(&fs_info->caching_block_groups); spin_lock_init(&fs_info->delalloc_lock); - spin_lock_init(&fs_info->new_trans_lock); + spin_lock_init(&fs_info->trans_lock); spin_lock_init(&fs_info->ref_cache_lock); spin_lock_init(&fs_info->fs_roots_radix_lock); spin_lock_init(&fs_info->delayed_iput_lock); @@ -1687,6 +1687,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->sb = sb; fs_info->max_inline = 8192 * 1024; fs_info->metadata_ratio = 0; + fs_info->trans_no_join = 0; fs_info->thread_pool_size = min_t(unsigned long, num_online_cpus() + 2, 8); @@ -1735,7 +1736,6 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->do_barriers = 1; - mutex_init(&fs_info->trans_mutex); mutex_init(&fs_info->ordered_operations_mutex); mutex_init(&fs_info->tree_log_mutex); mutex_init(&fs_info->chunk_mutex); @@ -3006,10 +3006,13 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) WARN_ON(1); - mutex_lock(&root->fs_info->trans_mutex); mutex_lock(&root->fs_info->transaction_kthread_mutex); + spin_lock(&root->fs_info->trans_lock); list_splice_init(&root->fs_info->trans_list, &list); + root->fs_info->trans_no_join = 1; + spin_unlock(&root->fs_info->trans_lock); + while (!list_empty(&list)) { t = list_entry(list.next, struct btrfs_transaction, list); if (!t) @@ -3034,23 +3037,18 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) t->blocked = 0; if (waitqueue_active(&root->fs_info->transaction_wait)) wake_up(&root->fs_info->transaction_wait); - mutex_unlock(&root->fs_info->trans_mutex); - mutex_lock(&root->fs_info->trans_mutex); t->commit_done = 1; if (waitqueue_active(&t->commit_wait)) wake_up(&t->commit_wait); - mutex_unlock(&root->fs_info->trans_mutex); - - mutex_lock(&root->fs_info->trans_mutex); btrfs_destroy_pending_snapshots(t); btrfs_destroy_delalloc_inodes(root); - spin_lock(&root->fs_info->new_trans_lock); + spin_lock(&root->fs_info->trans_lock); root->fs_info->running_transaction = NULL; - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); btrfs_destroy_marked_extents(root, &t->dirty_pages, EXTENT_DIRTY); @@ -3064,8 +3062,10 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) kmem_cache_free(btrfs_transaction_cachep, t); } + spin_lock(&root->fs_info->trans_lock); + root->fs_info->trans_no_join = 0; + spin_unlock(&root->fs_info->trans_lock); mutex_unlock(&root->fs_info->transaction_kthread_mutex); - mutex_unlock(&root->fs_info->trans_mutex); return 0; } -- cgit v1.2.3 From 0956c798ef8dbe0fc215870eb68bd2d8e789f86a Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 18 May 2011 00:11:22 +0000 Subject: BTRFS: Remove unused node_lock 240f62c8756 replaced the node_lock with rcu_read_lock, but forgot to remove the actual lock in the data structure. Remove it here. Signed-off-by: Andi Kleen Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 228cf36ece83..64b289690f9d 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1064,7 +1064,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, INIT_LIST_HEAD(&root->dirty_list); INIT_LIST_HEAD(&root->orphan_list); INIT_LIST_HEAD(&root->root_list); - spin_lock_init(&root->node_lock); spin_lock_init(&root->orphan_lock); spin_lock_init(&root->inode_lock); spin_lock_init(&root->accounting_lock); -- cgit v1.2.3 From c9513edb0079f97749c2ac00c887a22c4ba44792 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Wed, 20 Apr 2011 10:07:30 +0000 Subject: Btrfs: fix the race between reading and updating devices On btrfs_congested_fn and __unplug_io_fn paths, we should hold device_list_mutex to avoid remove/add device path to update fs_devices->devices On __btrfs_close_devices and btrfs_prepare_sprout paths, the devices in fs_devices->devices or fs_devices->devices is updated, so we should hold the mutex to avoid the reader side to reach them Signed-off-by: Xiao Guangrong Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 64b289690f9d..4e53a4fc467f 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1410,6 +1410,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) struct btrfs_device *device; struct backing_dev_info *bdi; + mutex_lock(&info->fs_devices->device_list_mutex); list_for_each_entry(device, &info->fs_devices->devices, dev_list) { if (!device->bdev) continue; @@ -1419,6 +1420,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) break; } } + mutex_unlock(&info->fs_devices->device_list_mutex); return ret; } -- cgit v1.2.3 From 1f78160ce1b1b8e657e2248118c4d91f881763f0 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Wed, 20 Apr 2011 10:09:16 +0000 Subject: Btrfs: using rcu lock in the reader side of devices list fs_devices->devices is only updated on remove and add device paths, so we can use rcu to protect it in the reader side Signed-off-by: Xiao Guangrong Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 4e53a4fc467f..deba3d9c8853 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1410,8 +1410,8 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) struct btrfs_device *device; struct backing_dev_info *bdi; - mutex_lock(&info->fs_devices->device_list_mutex); - list_for_each_entry(device, &info->fs_devices->devices, dev_list) { + rcu_read_lock(); + list_for_each_entry_rcu(device, &info->fs_devices->devices, dev_list) { if (!device->bdev) continue; bdi = blk_get_backing_dev_info(device->bdev); @@ -1420,7 +1420,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) break; } } - mutex_unlock(&info->fs_devices->device_list_mutex); + rcu_read_unlock(); return ret; } @@ -2332,9 +2332,9 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) sb = &root->fs_info->super_for_commit; dev_item = &sb->dev_item; - mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + rcu_read_lock(); head = &root->fs_info->fs_devices->devices; - list_for_each_entry(dev, head, dev_list) { + list_for_each_entry_rcu(dev, head, dev_list) { if (!dev->bdev) { total_errors++; continue; @@ -2367,7 +2367,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) } total_errors = 0; - list_for_each_entry(dev, head, dev_list) { + list_for_each_entry_rcu(dev, head, dev_list) { if (!dev->bdev) continue; if (!dev->in_fs_metadata || !dev->writeable) @@ -2377,7 +2377,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) if (ret) total_errors++; } - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + rcu_read_unlock(); if (total_errors > max_errors) { printk(KERN_ERR "btrfs: %d errors while writing supers\n", total_errors); -- cgit v1.2.3 From 4cb5300bc839b8a943eb19c9f27f25470e22d0ca Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Tue, 24 May 2011 15:35:30 -0400 Subject: Btrfs: add mount -o auto_defrag This will detect small random writes into files and queue the up for an auto defrag process. It isn't well suited to database workloads yet, but works for smaller files such as rpm, sqlite or bdb databases. Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 16d335b342a2..b2588a552658 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1475,6 +1475,7 @@ static int cleaner_kthread(void *arg) btrfs_run_delayed_iputs(root); btrfs_clean_old_snapshots(root); mutex_unlock(&root->fs_info->cleaner_mutex); + btrfs_run_defrag_inodes(root->fs_info); } if (freezing(current)) { @@ -1616,6 +1617,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, spin_lock_init(&fs_info->ref_cache_lock); spin_lock_init(&fs_info->fs_roots_radix_lock); spin_lock_init(&fs_info->delayed_iput_lock); + spin_lock_init(&fs_info->defrag_inodes_lock); init_completion(&fs_info->kobj_unregister); fs_info->tree_root = tree_root; @@ -1638,9 +1640,11 @@ struct btrfs_root *open_ctree(struct super_block *sb, atomic_set(&fs_info->async_delalloc_pages, 0); atomic_set(&fs_info->async_submit_draining, 0); atomic_set(&fs_info->nr_async_bios, 0); + atomic_set(&fs_info->defrag_running, 0); fs_info->sb = sb; fs_info->max_inline = 8192 * 1024; fs_info->metadata_ratio = 0; + fs_info->defrag_inodes = RB_ROOT; fs_info->thread_pool_size = min_t(unsigned long, num_online_cpus() + 2, 8); @@ -2501,6 +2505,14 @@ int close_ctree(struct btrfs_root *root) smp_mb(); btrfs_scrub_cancel(root); + + /* wait for any defraggers to finish */ + wait_event(fs_info->transaction_wait, + (atomic_read(&fs_info->defrag_running) == 0)); + + /* clear out the rbtree of defraggable inodes */ + btrfs_run_defrag_inodes(root->fs_info); + btrfs_put_block_group_cache(fs_info); /* -- cgit v1.2.3 From 174ba50915b08dcfd07c8b5fb795b46a165fa09a Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Fri, 27 May 2011 10:03:58 -0400 Subject: Btrfs: use the device_list_mutex during write_dev_supers write_dev_supers was changed to use RCU to protect the list of devices, but it was then sleeping while it actually wrote the supers. This fixes it to just use the mutex, since we really don't any concurrency in write_dev_supers anyway. Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index b2588a552658..98b6a71decba 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2310,7 +2310,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) sb = &root->fs_info->super_for_commit; dev_item = &sb->dev_item; - rcu_read_lock(); + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); head = &root->fs_info->fs_devices->devices; list_for_each_entry_rcu(dev, head, dev_list) { if (!dev->bdev) { @@ -2355,7 +2355,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) if (ret) total_errors++; } - rcu_read_unlock(); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); if (total_errors > max_errors) { printk(KERN_ERR "btrfs: %d errors while writing supers\n", total_errors); -- cgit v1.2.3 From 632dd772fcbde2ba37c0e8983bd38ef4a1eac906 Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Fri, 10 Jun 2011 12:07:07 +0200 Subject: btrfs: reinitialize scrub workers Scrub starts the workers each time a scrub starts and stops them after it finished. This patch adds an initialization for the workers before each start, otherwise the workers behave strangely. Signed-off-by: Arne Jansen --- fs/btrfs/disk-io.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index a203d363184d..7bbbfebe47e4 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1668,8 +1668,6 @@ struct btrfs_root *open_ctree(struct super_block *sb, init_waitqueue_head(&fs_info->scrub_pause_wait); init_rwsem(&fs_info->scrub_super_lock); fs_info->scrub_workers_refcnt = 0; - btrfs_init_workers(&fs_info->scrub_workers, "scrub", - fs_info->thread_pool_size, &fs_info->generic_worker); sb->s_blocksize = 4096; sb->s_blocksize_bits = blksize_bits(4096); -- cgit v1.2.3 From 5be76758f35ec6578e5b9b150aa513ac26bd9c54 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 9 Jun 2011 10:02:51 +0000 Subject: btrfs: fix unlocked access of delalloc_inodes list_splice_init will make delalloc_inodes empty, but without a spinlock around, this may produce corrupted list head, accessed in many placess, The race window is very tight and nobody seems to have hit it so far. Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index a203d363184d..33b744a5ac03 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2911,9 +2911,8 @@ static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root) INIT_LIST_HEAD(&splice); - list_splice_init(&root->fs_info->delalloc_inodes, &splice); - spin_lock(&root->fs_info->delalloc_lock); + list_splice_init(&root->fs_info->delalloc_inodes, &splice); while (!list_empty(&splice)) { btrfs_inode = list_entry(splice.next, struct btrfs_inode, -- cgit v1.2.3 From ac08aedfa5d3de0dcb3825b598d16c2e57991f54 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Mon, 13 Jun 2011 11:28:50 -0400 Subject: Btrfs: check the return value from set_anon_super Al Viro noticed we weren't checking for set_anon_super failures. This adds the required checks. Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 9f68c6898653..20c111b3fa0d 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1312,7 +1312,9 @@ again: spin_lock_init(&root->cache_lock); init_waitqueue_head(&root->cache_wait); - set_anon_super(&root->anon_super, NULL); + ret = set_anon_super(&root->anon_super, NULL); + if (ret) + goto fail; if (btrfs_root_refs(&root->root_item) == 0) { ret = -ENOENT; -- cgit v1.2.3 From 7585717f304f5ed005cc4ad933a69aab3efbd136 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Mon, 13 Jun 2011 20:00:16 -0400 Subject: Btrfs: fix relocation races The recent commit to get rid of our trans_mutex introduced some races with block group relocation. The problem is that relocation needs to do some record keeping about each root, and it was relying on the transaction mutex to coordinate things in subtle ways. This fix adds a mutex just for the relocation code and makes sure it doesn't have a big impact on normal operations. The race is really fixed in btrfs_record_root_in_trans, which is where we step back and wait for the relocation code to finish accounting setup. Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 20c111b3fa0d..0b2b4b759136 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1620,6 +1620,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, spin_lock_init(&fs_info->fs_roots_radix_lock); spin_lock_init(&fs_info->delayed_iput_lock); spin_lock_init(&fs_info->defrag_inodes_lock); + mutex_init(&fs_info->reloc_mutex); init_completion(&fs_info->kobj_unregister); fs_info->tree_root = tree_root; -- cgit v1.2.3 From 9fe6a50fb764f508dd2de47a66e62e51388791fb Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst Date: Thu, 16 Jun 2011 09:04:57 +0000 Subject: btrfs: Remove unused sysfs code Removes code no longer used. The sysfs file itself is kept, because the btrfs developers expressed interest in putting new entries to sysfs. Signed-off-by: Maarten Lankhorst Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 0b2b4b759136..c25ef5a0ccd6 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1044,7 +1044,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->last_trans = 0; root->highest_objectid = 0; root->name = NULL; - root->in_sysfs = 0; root->inode_tree = RB_ROOT; INIT_RADIX_TREE(&root->delayed_nodes_tree, GFP_ATOMIC); root->block_rsv = NULL; -- cgit v1.2.3 From 35a30d7ce54e087d8025a725d4e5a2fdee723a9f Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 13 Jun 2011 15:18:23 +0000 Subject: btrfs: fix uninitialized return value When allocation fails in btrfs_read_fs_root_no_name, ret is not set although it is returned, holding a garbage value. Signed-off-by: David Sterba Reviewed-by: Li Zefan Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index c25ef5a0ccd6..1ac8db5dc0a3 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1299,12 +1299,12 @@ again: return root; root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS); - if (!root->free_ino_ctl) - goto fail; root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned), GFP_NOFS); - if (!root->free_ino_pinned) + if (!root->free_ino_pinned || !root->free_ino_ctl) { + ret = -ENOMEM; goto fail; + } btrfs_init_free_ino_ctl(root); mutex_init(&root->fs_commit_mutex); -- cgit v1.2.3