diff options
Diffstat (limited to 'fs/btrfs')
52 files changed, 8971 insertions, 5801 deletions
diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile index 31610ea73aec..9b72dcf1cd25 100644 --- a/fs/btrfs/Makefile +++ b/fs/btrfs/Makefile @@ -7,4 +7,4 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \ extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \ extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \ export.o tree-log.o acl.o free-space-cache.o zlib.o lzo.o \ - compression.o delayed-ref.o relocation.o + compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c index 9c949348510b..f66fc9959733 100644 --- a/fs/btrfs/acl.c +++ b/fs/btrfs/acl.c @@ -170,7 +170,7 @@ static int btrfs_xattr_acl_set(struct dentry *dentry, const char *name, int ret; struct posix_acl *acl = NULL; - if (!is_owner_or_cap(dentry->d_inode)) + if (!inode_owner_or_capable(dentry->d_inode)) return -EPERM; if (!IS_POSIXACL(dentry->d_inode)) @@ -178,16 +178,18 @@ static int btrfs_xattr_acl_set(struct dentry *dentry, const char *name, if (value) { acl = posix_acl_from_xattr(value, size); - if (acl == NULL) { - value = NULL; - size = 0; - } else if (IS_ERR(acl)) { + if (IS_ERR(acl)) return PTR_ERR(acl); + + if (acl) { + ret = posix_acl_valid(acl); + if (ret) + goto out; } } ret = btrfs_set_acl(NULL, dentry->d_inode, acl, type); - +out: posix_acl_release(acl); return ret; @@ -286,7 +288,7 @@ int btrfs_acl_chmod(struct inode *inode) return 0; acl = btrfs_get_acl(inode, ACL_TYPE_ACCESS); - if (IS_ERR(acl) || !acl) + if (IS_ERR_OR_NULL(acl)) return PTR_ERR(acl); clone = posix_acl_clone(acl, GFP_KERNEL); diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index ccc991c542df..93b1aa932014 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -22,6 +22,7 @@ #include "extent_map.h" #include "extent_io.h" #include "ordered-data.h" +#include "delayed-inode.h" /* in memory btrfs inode */ struct btrfs_inode { @@ -136,9 +137,8 @@ struct btrfs_inode { * items we think we'll end up using, and reserved_extents is the number * of extent items we've reserved metadata for. */ - spinlock_t accounting_lock; atomic_t outstanding_extents; - int reserved_extents; + atomic_t reserved_extents; /* * ordered_data_close is set by truncate when a file that used @@ -153,20 +153,34 @@ struct btrfs_inode { unsigned ordered_data_close:1; unsigned orphan_meta_reserved:1; unsigned dummy_inode:1; + unsigned in_defrag:1; /* * always compress this one file */ unsigned force_compress:4; + struct btrfs_delayed_node *delayed_node; + struct inode vfs_inode; }; +extern unsigned char btrfs_filetype_table[]; + static inline struct btrfs_inode *BTRFS_I(struct inode *inode) { return container_of(inode, struct btrfs_inode, vfs_inode); } +static inline u64 btrfs_ino(struct inode *inode) +{ + u64 ino = BTRFS_I(inode)->location.objectid; + + if (ino <= BTRFS_FIRST_FREE_OBJECTID) + ino = inode->i_ino; + return ino; +} + static inline void btrfs_i_size_write(struct inode *inode, u64 size) { i_size_write(inode, size); diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 4d2110eafe29..bfe42b03eaf9 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -125,9 +125,10 @@ static int check_compressed_csum(struct inode *inode, kunmap_atomic(kaddr, KM_USER0); if (csum != *cb_sum) { - printk(KERN_INFO "btrfs csum failed ino %lu " + printk(KERN_INFO "btrfs csum failed ino %llu " "extent %llu csum %u " - "wanted %u mirror %d\n", inode->i_ino, + "wanted %u mirror %d\n", + (unsigned long long)btrfs_ino(inode), (unsigned long long)disk_start, csum, *cb_sum, cb->mirror_num); ret = -EIO; @@ -332,7 +333,7 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, struct compressed_bio *cb; unsigned long bytes_left; struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; - int page_index = 0; + int pg_index = 0; struct page *page; u64 first_byte = disk_start; struct block_device *bdev; @@ -340,6 +341,8 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, WARN_ON(start & ((u64)PAGE_CACHE_SIZE - 1)); cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS); + if (!cb) + return -ENOMEM; atomic_set(&cb->pending_bios, 0); cb->errors = 0; cb->inode = inode; @@ -354,14 +357,18 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev; bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS); + if(!bio) { + kfree(cb); + return -ENOMEM; + } bio->bi_private = cb; bio->bi_end_io = end_compressed_bio_write; atomic_inc(&cb->pending_bios); /* create and submit bios for the compressed pages */ bytes_left = compressed_len; - for (page_index = 0; page_index < cb->nr_pages; page_index++) { - page = compressed_pages[page_index]; + for (pg_index = 0; pg_index < cb->nr_pages; pg_index++) { + page = compressed_pages[pg_index]; page->mapping = inode->i_mapping; if (bio->bi_size) ret = io_tree->ops->merge_bio_hook(page, 0, @@ -426,7 +433,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, struct compressed_bio *cb) { unsigned long end_index; - unsigned long page_index; + unsigned long pg_index; u64 last_offset; u64 isize = i_size_read(inode); int ret; @@ -450,13 +457,13 @@ static noinline int add_ra_bio_pages(struct inode *inode, end_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT; while (last_offset < compressed_end) { - page_index = last_offset >> PAGE_CACHE_SHIFT; + pg_index = last_offset >> PAGE_CACHE_SHIFT; - if (page_index > end_index) + if (pg_index > end_index) break; rcu_read_lock(); - page = radix_tree_lookup(&mapping->page_tree, page_index); + page = radix_tree_lookup(&mapping->page_tree, pg_index); rcu_read_unlock(); if (page) { misses++; @@ -470,7 +477,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, if (!page) break; - if (add_to_page_cache_lru(page, mapping, page_index, + if (add_to_page_cache_lru(page, mapping, pg_index, GFP_NOFS)) { page_cache_release(page); goto next; @@ -554,7 +561,7 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, unsigned long uncompressed_len = bio->bi_vcnt * PAGE_CACHE_SIZE; unsigned long compressed_len; unsigned long nr_pages; - unsigned long page_index; + unsigned long pg_index; struct page *page; struct block_device *bdev; struct bio *comp_bio; @@ -607,10 +614,10 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev; - for (page_index = 0; page_index < nr_pages; page_index++) { - cb->compressed_pages[page_index] = alloc_page(GFP_NOFS | + for (pg_index = 0; pg_index < nr_pages; pg_index++) { + cb->compressed_pages[pg_index] = alloc_page(GFP_NOFS | __GFP_HIGHMEM); - if (!cb->compressed_pages[page_index]) + if (!cb->compressed_pages[pg_index]) goto fail2; } cb->nr_pages = nr_pages; @@ -628,8 +635,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, comp_bio->bi_end_io = end_compressed_bio_read; atomic_inc(&cb->pending_bios); - for (page_index = 0; page_index < nr_pages; page_index++) { - page = cb->compressed_pages[page_index]; + for (pg_index = 0; pg_index < nr_pages; pg_index++) { + page = cb->compressed_pages[pg_index]; page->mapping = inode->i_mapping; page->index = em_start >> PAGE_CACHE_SHIFT; @@ -657,8 +664,9 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, atomic_inc(&cb->pending_bios); if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) { - btrfs_lookup_bio_sums(root, inode, comp_bio, - sums); + ret = btrfs_lookup_bio_sums(root, inode, + comp_bio, sums); + BUG_ON(ret); } sums += (comp_bio->bi_size + root->sectorsize - 1) / root->sectorsize; @@ -683,8 +691,10 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0); BUG_ON(ret); - if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) - btrfs_lookup_bio_sums(root, inode, comp_bio, sums); + if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) { + ret = btrfs_lookup_bio_sums(root, inode, comp_bio, sums); + BUG_ON(ret); + } ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0); BUG_ON(ret); @@ -693,8 +703,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, return 0; fail2: - for (page_index = 0; page_index < nr_pages; page_index++) - free_page((unsigned long)cb->compressed_pages[page_index]); + for (pg_index = 0; pg_index < nr_pages; pg_index++) + free_page((unsigned long)cb->compressed_pages[pg_index]); kfree(cb->compressed_pages); fail1: @@ -936,7 +946,7 @@ void btrfs_exit_compress(void) int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, unsigned long total_out, u64 disk_start, struct bio_vec *bvec, int vcnt, - unsigned long *page_index, + unsigned long *pg_index, unsigned long *pg_offset) { unsigned long buf_offset; @@ -945,7 +955,7 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, unsigned long working_bytes = total_out - buf_start; unsigned long bytes; char *kaddr; - struct page *page_out = bvec[*page_index].bv_page; + struct page *page_out = bvec[*pg_index].bv_page; /* * start byte is the first byte of the page we're currently @@ -986,11 +996,11 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, /* check if we need to pick another page */ if (*pg_offset == PAGE_CACHE_SIZE) { - (*page_index)++; - if (*page_index >= vcnt) + (*pg_index)++; + if (*pg_index >= vcnt) return 0; - page_out = bvec[*page_index].bv_page; + page_out = bvec[*pg_index].bv_page; *pg_offset = 0; start_byte = page_offset(page_out) - disk_start; diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h index 51000174b9d7..a12059f4f0fd 100644 --- a/fs/btrfs/compression.h +++ b/fs/btrfs/compression.h @@ -37,7 +37,7 @@ int btrfs_decompress(int type, unsigned char *data_in, struct page *dest_page, int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, unsigned long total_out, u64 disk_start, struct bio_vec *bvec, int vcnt, - unsigned long *page_index, + unsigned long *pg_index, unsigned long *pg_offset); int btrfs_submit_compressed_write(struct inode *inode, u64 start, diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index b5baff0dccfe..b0e18d986e0a 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -38,11 +38,6 @@ static int balance_node_right(struct btrfs_trans_handle *trans, struct extent_buffer *src_buf); static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, int level, int slot); -static int setup_items_for_insert(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - u32 total_data, u32 total_size, int nr); - struct btrfs_path *btrfs_alloc_path(void) { @@ -107,7 +102,7 @@ void btrfs_free_path(struct btrfs_path *p) { if (!p) return; - btrfs_release_path(NULL, p); + btrfs_release_path(p); kmem_cache_free(btrfs_path_cachep, p); } @@ -117,7 +112,7 @@ void btrfs_free_path(struct btrfs_path *p) * * It is safe to call this on paths that no locks or extent buffers held. */ -noinline void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p) +noinline void btrfs_release_path(struct btrfs_path *p) { int i; @@ -147,10 +142,11 @@ noinline void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p) struct extent_buffer *btrfs_root_node(struct btrfs_root *root) { struct extent_buffer *eb; - spin_lock(&root->node_lock); - eb = root->node; + + rcu_read_lock(); + eb = rcu_dereference(root->node); extent_buffer_get(eb); - spin_unlock(&root->node_lock); + rcu_read_unlock(); return eb; } @@ -165,14 +161,8 @@ struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root) while (1) { eb = btrfs_root_node(root); btrfs_tree_lock(eb); - - spin_lock(&root->node_lock); - if (eb == root->node) { - spin_unlock(&root->node_lock); + if (eb == root->node) break; - } - spin_unlock(&root->node_lock); - btrfs_tree_unlock(eb); free_extent_buffer(eb); } @@ -458,10 +448,8 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, else parent_start = 0; - spin_lock(&root->node_lock); - root->node = cow; extent_buffer_get(cow); - spin_unlock(&root->node_lock); + rcu_assign_pointer(root->node, cow); btrfs_free_tree_block(trans, root, buf, parent_start, last_ref); @@ -542,6 +530,9 @@ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans, ret = __btrfs_cow_block(trans, root, buf, parent, parent_slot, cow_ret, search_start, 0); + + trace_btrfs_cow_block(root, buf, *cow_ret); + return ret; } @@ -686,6 +677,8 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, if (!cur) { cur = read_tree_block(root, blocknr, blocksize, gen); + if (!cur) + return -EIO; } else if (!uptodate) { btrfs_read_buffer(cur, gen); } @@ -732,122 +725,6 @@ static inline unsigned int leaf_data_end(struct btrfs_root *root, return btrfs_item_offset_nr(leaf, nr - 1); } -/* - * extra debugging checks to make sure all the items in a key are - * well formed and in the proper order - */ -static int check_node(struct btrfs_root *root, struct btrfs_path *path, - int level) -{ - struct extent_buffer *parent = NULL; - struct extent_buffer *node = path->nodes[level]; - struct btrfs_disk_key parent_key; - struct btrfs_disk_key node_key; - int parent_slot; - int slot; - struct btrfs_key cpukey; - u32 nritems = btrfs_header_nritems(node); - - if (path->nodes[level + 1]) - parent = path->nodes[level + 1]; - - slot = path->slots[level]; - BUG_ON(nritems == 0); - if (parent) { - parent_slot = path->slots[level + 1]; - btrfs_node_key(parent, &parent_key, parent_slot); - btrfs_node_key(node, &node_key, 0); - BUG_ON(memcmp(&parent_key, &node_key, - sizeof(struct btrfs_disk_key))); - BUG_ON(btrfs_node_blockptr(parent, parent_slot) != - btrfs_header_bytenr(node)); - } - BUG_ON(nritems > BTRFS_NODEPTRS_PER_BLOCK(root)); - if (slot != 0) { - btrfs_node_key_to_cpu(node, &cpukey, slot - 1); - btrfs_node_key(node, &node_key, slot); - BUG_ON(comp_keys(&node_key, &cpukey) <= 0); - } - if (slot < nritems - 1) { - btrfs_node_key_to_cpu(node, &cpukey, slot + 1); - btrfs_node_key(node, &node_key, slot); - BUG_ON(comp_keys(&node_key, &cpukey) >= 0); - } - return 0; -} - -/* - * extra checking to make sure all the items in a leaf are - * well formed and in the proper order - */ -static int check_leaf(struct btrfs_root *root, struct btrfs_path *path, - int level) -{ - struct extent_buffer *leaf = path->nodes[level]; - struct extent_buffer *parent = NULL; - int parent_slot; - struct btrfs_key cpukey; - struct btrfs_disk_key parent_key; - struct btrfs_disk_key leaf_key; - int slot = path->slots[0]; - - u32 nritems = btrfs_header_nritems(leaf); - - if (path->nodes[level + 1]) - parent = path->nodes[level + 1]; - - if (nritems == 0) - return 0; - - if (parent) { - parent_slot = path->slots[level + 1]; - btrfs_node_key(parent, &parent_key, parent_slot); - btrfs_item_key(leaf, &leaf_key, 0); - - BUG_ON(memcmp(&parent_key, &leaf_key, - sizeof(struct btrfs_disk_key))); - BUG_ON(btrfs_node_blockptr(parent, parent_slot) != - btrfs_header_bytenr(leaf)); - } - if (slot != 0 && slot < nritems - 1) { - btrfs_item_key(leaf, &leaf_key, slot); - btrfs_item_key_to_cpu(leaf, &cpukey, slot - 1); - if (comp_keys(&leaf_key, &cpukey) <= 0) { - btrfs_print_leaf(root, leaf); - printk(KERN_CRIT "slot %d offset bad key\n", slot); - BUG_ON(1); - } - if (btrfs_item_offset_nr(leaf, slot - 1) != - btrfs_item_end_nr(leaf, slot)) { - btrfs_print_leaf(root, leaf); - printk(KERN_CRIT "slot %d offset bad\n", slot); - BUG_ON(1); - } - } - if (slot < nritems - 1) { - btrfs_item_key(leaf, &leaf_key, slot); - btrfs_item_key_to_cpu(leaf, &cpukey, slot + 1); - BUG_ON(comp_keys(&leaf_key, &cpukey) >= 0); - if (btrfs_item_offset_nr(leaf, slot) != - btrfs_item_end_nr(leaf, slot + 1)) { - btrfs_print_leaf(root, leaf); - printk(KERN_CRIT "slot %d offset bad\n", slot); - BUG_ON(1); - } - } - BUG_ON(btrfs_item_offset_nr(leaf, 0) + - btrfs_item_size_nr(leaf, 0) != BTRFS_LEAF_DATA_SIZE(root)); - return 0; -} - -static noinline int check_block(struct btrfs_root *root, - struct btrfs_path *path, int level) -{ - return 0; - if (level == 0) - return check_leaf(root, path, level); - return check_node(root, path, level); -} /* * search for key in the extent_buffer. The items start at offset p, @@ -1046,9 +923,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, goto enospc; } - spin_lock(&root->node_lock); - root->node = child; - spin_unlock(&root->node_lock); + rcu_assign_pointer(root->node, child); add_root_to_dirty_list(root); btrfs_tree_unlock(child); @@ -1188,7 +1063,6 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, } } /* double check we haven't messed things up */ - check_block(root, path, level); if (orig_ptr != btrfs_node_blockptr(path->nodes[level], path->slots[level])) BUG(); @@ -1449,7 +1323,7 @@ static noinline int reada_for_balance(struct btrfs_root *root, ret = -EAGAIN; /* release the whole path */ - btrfs_release_path(root, path); + btrfs_release_path(path); /* read the blocks */ if (block1) @@ -1596,7 +1470,7 @@ read_block_for_search(struct btrfs_trans_handle *trans, return 0; } free_extent_buffer(tmp); - btrfs_release_path(NULL, p); + btrfs_release_path(p); return -EIO; } } @@ -1615,7 +1489,7 @@ read_block_for_search(struct btrfs_trans_handle *trans, if (p->reada) reada_for_search(root, p, level, slot, key->objectid); - btrfs_release_path(NULL, p); + btrfs_release_path(p); ret = -EAGAIN; tmp = read_tree_block(root, blocknr, blocksize, 0); @@ -1684,7 +1558,7 @@ setup_nodes_for_search(struct btrfs_trans_handle *trans, } b = p->nodes[level]; if (!b) { - btrfs_release_path(NULL, p); + btrfs_release_path(p); goto again; } BUG_ON(btrfs_header_nritems(b) == 1); @@ -1798,12 +1672,6 @@ cow_done: if (!cow) btrfs_unlock_up_safe(p, level + 1); - ret = check_block(root, p, level); - if (ret) { - ret = -1; - goto done; - } - ret = bin_search(b, key, level, &slot); if (level != 0) { @@ -1880,7 +1748,7 @@ done: if (!p->leave_spinning) btrfs_set_path_blocking(p); if (ret < 0) - btrfs_release_path(root, p); + btrfs_release_path(p); return ret; } @@ -2130,10 +1998,8 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(c); - spin_lock(&root->node_lock); old = root->node; - root->node = c; - spin_unlock(&root->node_lock); + rcu_assign_pointer(root->node, c); /* the super has an extra ref to root->node */ free_extent_buffer(old); @@ -3155,7 +3021,7 @@ static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans, struct btrfs_file_extent_item); extent_len = btrfs_file_extent_num_bytes(leaf, fi); } - btrfs_release_path(root, path); + btrfs_release_path(path); path->keep_locks = 1; path->search_for_split = 1; @@ -3345,7 +3211,6 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, struct btrfs_path *path, u32 new_size, int from_end) { - int ret = 0; int slot; struct extent_buffer *leaf; struct btrfs_item *item; @@ -3443,12 +3308,11 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, btrfs_set_item_size(leaf, item, new_size); btrfs_mark_buffer_dirty(leaf); - ret = 0; if (btrfs_leaf_free_space(root, leaf) < 0) { btrfs_print_leaf(root, leaf); BUG(); } - return ret; + return 0; } /* @@ -3458,7 +3322,6 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, u32 data_size) { - int ret = 0; int slot; struct extent_buffer *leaf; struct btrfs_item *item; @@ -3523,12 +3386,11 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, btrfs_set_item_size(leaf, item, old_size + data_size); btrfs_mark_buffer_dirty(leaf); - ret = 0; if (btrfs_leaf_free_space(root, leaf) < 0) { btrfs_print_leaf(root, leaf); BUG(); } - return ret; + return 0; } /* @@ -3688,11 +3550,10 @@ out: * to save stack depth by doing the bulk of the work in a function * that doesn't call btrfs_search_slot */ -static noinline_for_stack int -setup_items_for_insert(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - u32 total_data, u32 total_size, int nr) +int setup_items_for_insert(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *cpu_key, u32 *data_size, + u32 total_data, u32 total_size, int nr) { struct btrfs_item *item; int i; @@ -3776,7 +3637,6 @@ setup_items_for_insert(struct btrfs_trans_handle *trans, ret = 0; if (slot == 0) { - struct btrfs_disk_key disk_key; btrfs_cpu_key_to_disk(&disk_key, cpu_key); ret = fixup_low_keys(trans, root, path, &disk_key, 1); } @@ -3840,7 +3700,8 @@ int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root unsigned long ptr; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size); if (!ret) { leaf = path->nodes[0]; @@ -4077,7 +3938,7 @@ int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path) else return 1; - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) return ret; @@ -4201,7 +4062,7 @@ find_next_key: sret = btrfs_find_next_key(root, path, min_key, level, cache_only, min_trans); if (sret == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } else { goto out; @@ -4217,6 +4078,7 @@ find_next_key: } btrfs_set_path_blocking(path); cur = read_node_slot(root, cur, slot); + BUG_ON(!cur); btrfs_tree_lock(cur); @@ -4279,7 +4141,7 @@ next: btrfs_node_key_to_cpu(c, &cur_key, slot); orig_lowest = path->lowest_level; - btrfs_release_path(root, path); + btrfs_release_path(path); path->lowest_level = level; ret = btrfs_search_slot(NULL, root, &cur_key, path, 0, 0); @@ -4356,7 +4218,7 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path) again: level = 1; next = NULL; - btrfs_release_path(root, path); + btrfs_release_path(path); path->keep_locks = 1; @@ -4412,7 +4274,7 @@ again: goto again; if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto done; } @@ -4451,7 +4313,7 @@ again: goto again; if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto done; } diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 2c98b3af6052..6c093fa98f61 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -23,15 +23,18 @@ #include <linux/mm.h> #include <linux/highmem.h> #include <linux/fs.h> +#include <linux/rwsem.h> #include <linux/completion.h> #include <linux/backing-dev.h> #include <linux/wait.h> #include <linux/slab.h> #include <linux/kobject.h> +#include <trace/events/btrfs.h> #include <asm/kmap_types.h> #include "extent_io.h" #include "extent_map.h" #include "async-thread.h" +#include "ioctl.h" struct btrfs_trans_handle; struct btrfs_transaction; @@ -40,6 +43,7 @@ extern struct kmem_cache *btrfs_trans_handle_cachep; extern struct kmem_cache *btrfs_transaction_cachep; extern struct kmem_cache *btrfs_bit_radix_cachep; extern struct kmem_cache *btrfs_path_cachep; +extern struct kmem_cache *btrfs_free_space_cachep; struct btrfs_ordered_sum; #define BTRFS_MAGIC "_BHRfS_M" @@ -103,6 +107,12 @@ struct btrfs_ordered_sum; /* For storing free space cache */ #define BTRFS_FREE_SPACE_OBJECTID -11ULL +/* + * The inode number assigned to the special inode for sotring + * free ino cache + */ +#define BTRFS_FREE_INO_OBJECTID -12ULL + /* dummy objectid represents multiple objectids */ #define BTRFS_MULTIPLE_OBJECTIDS -255ULL @@ -185,7 +195,6 @@ struct btrfs_mapping_tree { struct extent_map_tree map_tree; }; -#define BTRFS_UUID_SIZE 16 struct btrfs_dev_item { /* the internal btrfs device id */ __le64 devid; @@ -292,7 +301,6 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes) sizeof(struct btrfs_stripe) * (num_stripes - 1); } -#define BTRFS_FSID_SIZE 16 #define BTRFS_HEADER_FLAG_WRITTEN (1ULL << 0) #define BTRFS_HEADER_FLAG_RELOC (1ULL << 1) @@ -508,6 +516,12 @@ struct btrfs_extent_item_v0 { /* use full backrefs for extent pointers in the block */ #define BTRFS_BLOCK_FLAG_FULL_BACKREF (1ULL << 8) +/* + * this flag is only used internally by scrub and may be changed at any time + * it is only declared here to avoid collisions + */ +#define BTRFS_EXTENT_FLAG_SUPER (1ULL << 48) + struct btrfs_tree_block_info { struct btrfs_disk_key key; u8 level; @@ -716,7 +730,7 @@ struct btrfs_space_info { u64 total_bytes; /* total bytes in the space, this doesn't take mirrors into account */ u64 bytes_used; /* total bytes used, - this does't take mirrors into account */ + this doesn't take mirrors into account */ u64 bytes_pinned; /* total bytes pinned, will be freed when the transaction finishes */ u64 bytes_reserved; /* total bytes the allocator has reserved for @@ -729,10 +743,21 @@ struct btrfs_space_info { u64 disk_total; /* total bytes on disk, takes mirrors into account */ - int full; /* indicates that we cannot allocate any more + /* + * we bump reservation progress every time we decrement + * bytes_reserved. This way people waiting for reservations + * know something good has happened and they can check + * for progress. The number here isn't to be trusted, it + * just shows reclaim activity + */ + unsigned long reservation_progress; + + unsigned int full:1; /* indicates that we cannot allocate any more chunks for this space */ - int force_alloc; /* set if we need to force a chunk alloc for - this space */ + unsigned int chunk_alloc:1; /* set if we are allocating a chunk */ + + unsigned int force_alloc; /* set if we need to force a chunk + alloc for this space */ struct list_head list; @@ -773,9 +798,6 @@ struct btrfs_free_cluster { /* first extent starting offset */ u64 window_start; - /* if this cluster simply points at a bitmap in the block group */ - bool points_to_bitmap; - struct btrfs_block_group_cache *block_group; /* * when a cluster is allocated from a block group, we put the @@ -820,9 +842,6 @@ struct btrfs_block_group_cache { u64 bytes_super; u64 flags; u64 sectorsize; - int extents_thresh; - int free_extents; - int total_bitmaps; unsigned int ro:1; unsigned int dirty:1; unsigned int iref:1; @@ -837,9 +856,7 @@ struct btrfs_block_group_cache { struct btrfs_space_info *space_info; /* free space cache stuff */ - spinlock_t tree_lock; - struct rb_root free_space_offset; - u64 free_space; + struct btrfs_free_space_ctl *free_space_ctl; /* block group cache stuff */ struct rb_node cache_node; @@ -859,6 +876,7 @@ struct btrfs_block_group_cache { struct reloc_control; struct btrfs_device; struct btrfs_fs_devices; +struct btrfs_delayed_root; struct btrfs_fs_info { u8 fsid[BTRFS_FSID_SIZE]; u8 chunk_tree_uuid[BTRFS_UUID_SIZE]; @@ -885,7 +903,10 @@ struct btrfs_fs_info { /* logical->physical extent mapping */ struct btrfs_mapping_tree mapping_tree; - /* block reservation for extent, checksum and root tree */ + /* + * block reservation for extent, checksum, root tree and + * delayed dir index item + */ struct btrfs_block_rsv global_block_rsv; /* block reservation for delay allocation */ struct btrfs_block_rsv delalloc_block_rsv; @@ -1012,6 +1033,7 @@ struct btrfs_fs_info { * for the sys_munmap function call path */ struct btrfs_workers fixup_workers; + struct btrfs_workers delayed_workers; struct task_struct *transaction_kthread; struct task_struct *cleaner_kthread; int thread_pool_size; @@ -1052,6 +1074,11 @@ struct btrfs_fs_info { /* all metadata allocations go through this cluster */ struct btrfs_free_cluster meta_alloc_cluster; + /* auto defrag inodes go here */ + spinlock_t defrag_inodes_lock; + struct rb_root defrag_inodes; + atomic_t defrag_running; + spinlock_t ref_cache_lock; u64 total_ref_cache_size; @@ -1067,8 +1094,21 @@ struct btrfs_fs_info { void *bdev_holder; + /* private scrub information */ + struct mutex scrub_lock; + atomic_t scrubs_running; + atomic_t scrub_pause_req; + atomic_t scrubs_paused; + atomic_t scrub_cancel_req; + wait_queue_head_t scrub_pause_wait; + struct rw_semaphore scrub_super_lock; + int scrub_workers_refcnt; + struct btrfs_workers scrub_workers; + /* filesystem state */ u64 fs_state; + + struct btrfs_delayed_root *delayed_root; }; /* @@ -1078,9 +1118,6 @@ struct btrfs_fs_info { struct btrfs_root { struct extent_buffer *node; - /* the node lock is held while changing the node pointer */ - spinlock_t node_lock; - struct extent_buffer *commit_root; struct btrfs_root *log_root; struct btrfs_root *reloc_root; @@ -1097,6 +1134,16 @@ struct btrfs_root { spinlock_t accounting_lock; struct btrfs_block_rsv *block_rsv; + /* free ino cache stuff */ + struct mutex fs_commit_mutex; + struct btrfs_free_space_ctl *free_ino_ctl; + enum btrfs_caching_type cached; + spinlock_t cache_lock; + wait_queue_head_t cache_wait; + struct btrfs_free_space_ctl *free_ino_pinned; + u64 cache_progress; + struct inode *cache_inode; + struct mutex log_mutex; wait_queue_head_t log_writer_wait; wait_queue_head_t log_commit_wait[2]; @@ -1152,12 +1199,49 @@ struct btrfs_root { struct rb_root inode_tree; /* + * radix tree that keeps track of delayed nodes of every inode, + * protected by inode_lock + */ + struct radix_tree_root delayed_nodes_tree; + /* * right now this just gets used so that a root has its own devid * for stat. It may be used for more later */ struct super_block anon_super; }; +struct btrfs_ioctl_defrag_range_args { + /* start of the defrag operation */ + __u64 start; + + /* number of bytes to defrag, use (u64)-1 to say all */ + __u64 len; + + /* + * flags for the operation, which can include turning + * on compression for this one defrag + */ + __u64 flags; + + /* + * any extent bigger than this will be considered + * already defragged. Use 0 to take the kernel default + * Use 1 to say every single extent must be rewritten + */ + __u32 extent_thresh; + + /* + * which compression method to use if turning on compression + * for this defrag operation. If unspecified, zlib will + * be used + */ + __u32 compress_type; + + /* spare for later */ + __u32 unused[4]; +}; + + /* * inode items have the data typically returned from stat and store other * info about object characteristics. There is one for every file and dir in @@ -1254,6 +1338,8 @@ struct btrfs_root { #define BTRFS_MOUNT_SPACE_CACHE (1 << 12) #define BTRFS_MOUNT_CLEAR_CACHE (1 << 13) #define BTRFS_MOUNT_USER_SUBVOL_RM_ALLOWED (1 << 14) +#define BTRFS_MOUNT_ENOSPC_DEBUG (1 << 15) +#define BTRFS_MOUNT_AUTO_DEFRAG (1 << 16) #define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt) #define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt) @@ -1273,6 +1359,9 @@ struct btrfs_root { #define BTRFS_INODE_NODUMP (1 << 8) #define BTRFS_INODE_NOATIME (1 << 9) #define BTRFS_INODE_DIRSYNC (1 << 10) +#define BTRFS_INODE_COMPRESS (1 << 11) + +#define BTRFS_INODE_ROOT_ITEM_INIT (1 << 31) /* some macros to generate set/get funcs for the struct fields. This * assumes there is a lefoo_to_cpu for every type, so lets make a simple @@ -1426,26 +1515,12 @@ static inline u64 btrfs_stripe_offset_nr(struct extent_buffer *eb, return btrfs_stripe_offset(eb, btrfs_stripe_nr(c, nr)); } -static inline void btrfs_set_stripe_offset_nr(struct extent_buffer *eb, - struct btrfs_chunk *c, int nr, - u64 val) -{ - btrfs_set_stripe_offset(eb, btrfs_stripe_nr(c, nr), val); -} - static inline u64 btrfs_stripe_devid_nr(struct extent_buffer *eb, struct btrfs_chunk *c, int nr) { return btrfs_stripe_devid(eb, btrfs_stripe_nr(c, nr)); } -static inline void btrfs_set_stripe_devid_nr(struct extent_buffer *eb, - struct btrfs_chunk *c, int nr, - u64 val) -{ - btrfs_set_stripe_devid(eb, btrfs_stripe_nr(c, nr), val); -} - /* struct btrfs_block_group_item */ BTRFS_SETGET_STACK_FUNCS(block_group_used, struct btrfs_block_group_item, used, 64); @@ -1503,14 +1578,6 @@ btrfs_inode_ctime(struct btrfs_inode_item *inode_item) return (struct btrfs_timespec *)ptr; } -static inline struct btrfs_timespec * -btrfs_inode_otime(struct btrfs_inode_item *inode_item) -{ - unsigned long ptr = (unsigned long)inode_item; - ptr += offsetof(struct btrfs_inode_item, otime); - return (struct btrfs_timespec *)ptr; -} - BTRFS_SETGET_FUNCS(timespec_sec, struct btrfs_timespec, sec, 64); BTRFS_SETGET_FUNCS(timespec_nsec, struct btrfs_timespec, nsec, 32); @@ -1861,33 +1928,6 @@ static inline u8 *btrfs_header_chunk_tree_uuid(struct extent_buffer *eb) return (u8 *)ptr; } -static inline u8 *btrfs_super_fsid(struct extent_buffer *eb) -{ - unsigned long ptr = offsetof(struct btrfs_super_block, fsid); - return (u8 *)ptr; -} - -static inline u8 *btrfs_header_csum(struct extent_buffer *eb) -{ - unsigned long ptr = offsetof(struct btrfs_header, csum); - return (u8 *)ptr; -} - -static inline struct btrfs_node *btrfs_buffer_node(struct extent_buffer *eb) -{ - return NULL; -} - -static inline struct btrfs_leaf *btrfs_buffer_leaf(struct extent_buffer *eb) -{ - return NULL; -} - -static inline struct btrfs_header *btrfs_buffer_header(struct extent_buffer *eb) -{ - return NULL; -} - static inline int btrfs_is_leaf(struct extent_buffer *eb) { return btrfs_header_level(eb) == 0; @@ -2041,22 +2081,6 @@ static inline struct btrfs_root *btrfs_sb(struct super_block *sb) return sb->s_fs_info; } -static inline int btrfs_set_root_name(struct btrfs_root *root, - const char *name, int len) -{ - /* if we already have a name just free it */ - kfree(root->name); - - root->name = kmalloc(len+1, GFP_KERNEL); - if (!root->name) - return -ENOMEM; - - memcpy(root->name, name, len); - root->name[len] = '\0'; - - return 0; -} - static inline u32 btrfs_level_size(struct btrfs_root *root, int level) { if (level == 0) @@ -2085,6 +2109,13 @@ static inline bool btrfs_mixed_space_info(struct btrfs_space_info *space_info) } /* extent-tree.c */ +static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root, + int num_items) +{ + return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) * + 3 * num_items; +} + void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); @@ -2094,12 +2125,9 @@ int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, u64 num_bytes, u64 *refs, u64 *flags); int btrfs_pin_extent(struct btrfs_root *root, u64 bytenr, u64 num, int reserved); -int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct extent_buffer *leaf); int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 objectid, u64 offset, u64 bytenr); -int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy); struct btrfs_block_group_cache *btrfs_lookup_block_group( struct btrfs_fs_info *info, u64 bytenr); @@ -2147,6 +2175,8 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, u64 root_objectid, u64 owner, u64 offset); int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len); +int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve, int sinfo); int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, @@ -2217,8 +2247,12 @@ u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo); int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end); int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr, - u64 num_bytes); + u64 num_bytes, u64 *actual_bytes); +int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, + struct btrfs_root *root, u64 type); +int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range); +int btrfs_init_space_info(struct btrfs_fs_info *fs_info); /* ctree.c */ int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key, int level, int *slot); @@ -2270,10 +2304,12 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *parent, int start_slot, int cache_only, u64 *last_ret, struct btrfs_key *progress); -void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p); +void btrfs_release_path(struct btrfs_path *p); struct btrfs_path *btrfs_alloc_path(void); void btrfs_free_path(struct btrfs_path *p); void btrfs_set_path_blocking(struct btrfs_path *p); +void btrfs_clear_path_blocking(struct btrfs_path *p, + struct extent_buffer *held); void btrfs_unlock_up_safe(struct btrfs_path *p, int level); int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2285,13 +2321,12 @@ static inline int btrfs_del_item(struct btrfs_trans_handle *trans, return btrfs_del_items(trans, root, path, path->slots[0], 1); } +int setup_items_for_insert(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *cpu_key, u32 *data_size, + u32 total_data, u32 total_size, int nr); int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_key *key, void *data, u32 data_size); -int btrfs_insert_some_items(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - int nr); int btrfs_insert_empty_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, @@ -2337,16 +2372,16 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root *item); int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct btrfs_root_item *item, struct btrfs_key *key); -int btrfs_search_root(struct btrfs_root *root, u64 search_start, - u64 *found_objectid); int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid); int btrfs_find_orphan_roots(struct btrfs_root *tree_root); int btrfs_set_root_node(struct btrfs_root_item *item, struct extent_buffer *node); +void btrfs_check_and_init_root_item(struct btrfs_root_item *item); + /* dir-item.c */ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, - int name_len, u64 dir, + int name_len, struct inode *dir, struct btrfs_key *location, u8 type, u64 index); struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2380,6 +2415,9 @@ struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans, struct btrfs_path *path, u64 dir, const char *name, u16 name_len, int mod); +int verify_dir_item(struct btrfs_root *root, + struct extent_buffer *leaf, + struct btrfs_dir_item *dir_item); /* orphan.c */ int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans, @@ -2388,12 +2426,6 @@ int btrfs_del_orphan_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 offset); int btrfs_find_orphan_item(struct btrfs_root *root, u64 offset); -/* inode-map.c */ -int btrfs_find_free_objectid(struct btrfs_trans_handle *trans, - struct btrfs_root *fs_root, - u64 dirid, u64 *objectid); -int btrfs_find_highest_inode(struct btrfs_root *fs_root, u64 *objectid); - /* inode-item.c */ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2438,8 +2470,6 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans, struct btrfs_ordered_sum *sums); int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode, struct bio *bio, u64 file_start, int contig); -int btrfs_csum_file_bytes(struct btrfs_root *root, struct inode *inode, - u64 start, unsigned long len); struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, @@ -2447,8 +2477,8 @@ struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans, int btrfs_csum_truncate(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, u64 isize); -int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, - u64 end, struct list_head *list); +int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, + struct list_head *list, int search_commit); /* inode.c */ /* RHEL and EL kernels have a patch that renames PG_checked to FsMisc */ @@ -2477,8 +2507,6 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, u32 min_type); int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput); -int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput, - int sync); int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end, struct extent_state **cached_state); int btrfs_writepages(struct address_space *mapping, @@ -2495,9 +2523,8 @@ unsigned long btrfs_force_ra(struct address_space *mapping, int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf); int btrfs_readpage(struct file *file, struct page *page); void btrfs_evict_inode(struct inode *inode); -void btrfs_put_inode(struct inode *inode); int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc); -void btrfs_dirty_inode(struct inode *inode); +void btrfs_dirty_inode(struct inode *inode, int flags); struct inode *btrfs_alloc_inode(struct super_block *sb); void btrfs_destroy_inode(struct inode *inode); int btrfs_drop_inode(struct inode *inode); @@ -2506,17 +2533,15 @@ void btrfs_destroy_cachep(void); long btrfs_ioctl_trans_end(struct file *file); struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, struct btrfs_root *root, int *was_new); -int btrfs_commit_write(struct file *file, struct page *page, - unsigned from, unsigned to); struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page, - size_t page_offset, u64 start, u64 end, + size_t pg_offset, u64 start, u64 end, int create); int btrfs_update_inode(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode); int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode); -void btrfs_orphan_cleanup(struct btrfs_root *root); +int btrfs_orphan_cleanup(struct btrfs_root *root); void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending, u64 *bytes_to_reserve); @@ -2524,7 +2549,7 @@ void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending); void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, struct btrfs_root *root); -int btrfs_cont_expand(struct inode *inode, loff_t size); +int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size); int btrfs_invalidate_inodes(struct btrfs_root *root); void btrfs_add_delayed_iput(struct inode *inode); void btrfs_run_delayed_iputs(struct btrfs_root *root); @@ -2541,18 +2566,27 @@ extern const struct dentry_operations btrfs_dentry_operations; long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg); void btrfs_update_iflags(struct inode *inode); void btrfs_inherit_iflags(struct inode *inode, struct inode *dir); - +int btrfs_defrag_file(struct inode *inode, struct file *file, + struct btrfs_ioctl_defrag_range_args *range, + u64 newer_than, unsigned long max_pages); /* file.c */ +int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, + struct inode *inode); +int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info); int btrfs_sync_file(struct file *file, int datasync); int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, int skip_pinned); -int btrfs_check_file(struct btrfs_root *root, struct inode *inode); extern const struct file_operations btrfs_file_operations; int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, u64 start, u64 end, u64 *hint_byte, int drop_cache); int btrfs_mark_extent_written(struct btrfs_trans_handle *trans, struct inode *inode, u64 start, u64 end); int btrfs_release_file(struct inode *inode, struct file *file); +void btrfs_drop_pages(struct page **pages, size_t num_pages); +int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode, + struct page **pages, size_t num_pages, + loff_t pos, size_t write_bytes, + struct extent_state **cached); /* tree-defrag.c */ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans, @@ -2561,10 +2595,6 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans, /* sysfs.c */ int btrfs_init_sysfs(void); void btrfs_exit_sysfs(void); -int btrfs_sysfs_add_super(struct btrfs_fs_info *fs); -int btrfs_sysfs_add_root(struct btrfs_root *root); -void btrfs_sysfs_del_root(struct btrfs_root *root); -void btrfs_sysfs_del_super(struct btrfs_fs_info *root); /* xattr.c */ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size); @@ -2607,4 +2637,18 @@ void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans, u64 *bytes_to_reserve); void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending); + +/* scrub.c */ +int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, + struct btrfs_scrub_progress *progress, int readonly); +int btrfs_scrub_pause(struct btrfs_root *root); +int btrfs_scrub_pause_super(struct btrfs_root *root); +int btrfs_scrub_continue(struct btrfs_root *root); +int btrfs_scrub_continue_super(struct btrfs_root *root); +int btrfs_scrub_cancel(struct btrfs_root *root); +int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev); +int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid); +int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, + struct btrfs_scrub_progress *progress); + #endif diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c new file mode 100644 index 000000000000..01e29503a54b --- /dev/null +++ b/fs/btrfs/delayed-inode.c @@ -0,0 +1,1695 @@ +/* + * Copyright (C) 2011 Fujitsu. All rights reserved. + * Written by Miao Xie <miaox@cn.fujitsu.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/slab.h> +#include "delayed-inode.h" +#include "disk-io.h" +#include "transaction.h" + +#define BTRFS_DELAYED_WRITEBACK 400 +#define BTRFS_DELAYED_BACKGROUND 100 + +static struct kmem_cache *delayed_node_cache; + +int __init btrfs_delayed_inode_init(void) +{ + delayed_node_cache = kmem_cache_create("delayed_node", + sizeof(struct btrfs_delayed_node), + 0, + SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, + NULL); + if (!delayed_node_cache) + return -ENOMEM; + return 0; +} + +void btrfs_delayed_inode_exit(void) +{ + if (delayed_node_cache) + kmem_cache_destroy(delayed_node_cache); +} + +static inline void btrfs_init_delayed_node( + struct btrfs_delayed_node *delayed_node, + struct btrfs_root *root, u64 inode_id) +{ + delayed_node->root = root; + delayed_node->inode_id = inode_id; + atomic_set(&delayed_node->refs, 0); + delayed_node->count = 0; + delayed_node->in_list = 0; + delayed_node->inode_dirty = 0; + delayed_node->ins_root = RB_ROOT; + delayed_node->del_root = RB_ROOT; + mutex_init(&delayed_node->mutex); + delayed_node->index_cnt = 0; + INIT_LIST_HEAD(&delayed_node->n_list); + INIT_LIST_HEAD(&delayed_node->p_list); + delayed_node->bytes_reserved = 0; +} + +static inline int btrfs_is_continuous_delayed_item( + struct btrfs_delayed_item *item1, + struct btrfs_delayed_item *item2) +{ + if (item1->key.type == BTRFS_DIR_INDEX_KEY && + item1->key.objectid == item2->key.objectid && + item1->key.type == item2->key.type && + item1->key.offset + 1 == item2->key.offset) + return 1; + return 0; +} + +static inline struct btrfs_delayed_root *btrfs_get_delayed_root( + struct btrfs_root *root) +{ + return root->fs_info->delayed_root; +} + +static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node( + struct inode *inode) +{ + struct btrfs_delayed_node *node; + struct btrfs_inode *btrfs_inode = BTRFS_I(inode); + struct btrfs_root *root = btrfs_inode->root; + u64 ino = btrfs_ino(inode); + int ret; + +again: + node = ACCESS_ONCE(btrfs_inode->delayed_node); + if (node) { + atomic_inc(&node->refs); /* can be accessed */ + return node; + } + + spin_lock(&root->inode_lock); + node = radix_tree_lookup(&root->delayed_nodes_tree, ino); + if (node) { + if (btrfs_inode->delayed_node) { + spin_unlock(&root->inode_lock); + goto again; + } + btrfs_inode->delayed_node = node; + atomic_inc(&node->refs); /* can be accessed */ + atomic_inc(&node->refs); /* cached in the inode */ + spin_unlock(&root->inode_lock); + return node; + } + spin_unlock(&root->inode_lock); + + node = kmem_cache_alloc(delayed_node_cache, GFP_NOFS); + if (!node) + return ERR_PTR(-ENOMEM); + btrfs_init_delayed_node(node, root, ino); + + atomic_inc(&node->refs); /* cached in the btrfs inode */ + atomic_inc(&node->refs); /* can be accessed */ + + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); + if (ret) { + kmem_cache_free(delayed_node_cache, node); + return ERR_PTR(ret); + } + + spin_lock(&root->inode_lock); + ret = radix_tree_insert(&root->delayed_nodes_tree, ino, node); + if (ret == -EEXIST) { + kmem_cache_free(delayed_node_cache, node); + spin_unlock(&root->inode_lock); + radix_tree_preload_end(); + goto again; + } + btrfs_inode->delayed_node = node; + spin_unlock(&root->inode_lock); + radix_tree_preload_end(); + + return node; +} + +/* + * Call it when holding delayed_node->mutex + * + * If mod = 1, add this node into the prepared list. + */ +static void btrfs_queue_delayed_node(struct btrfs_delayed_root *root, + struct btrfs_delayed_node *node, + int mod) +{ + spin_lock(&root->lock); + if (node->in_list) { + if (!list_empty(&node->p_list)) + list_move_tail(&node->p_list, &root->prepare_list); + else if (mod) + list_add_tail(&node->p_list, &root->prepare_list); + } else { + list_add_tail(&node->n_list, &root->node_list); + list_add_tail(&node->p_list, &root->prepare_list); + atomic_inc(&node->refs); /* inserted into list */ + root->nodes++; + node->in_list = 1; + } + spin_unlock(&root->lock); +} + +/* Call it when holding delayed_node->mutex */ +static void btrfs_dequeue_delayed_node(struct btrfs_delayed_root *root, + struct btrfs_delayed_node *node) +{ + spin_lock(&root->lock); + if (node->in_list) { + root->nodes--; + atomic_dec(&node->refs); /* not in the list */ + list_del_init(&node->n_list); + if (!list_empty(&node->p_list)) + list_del_init(&node->p_list); + node->in_list = 0; + } + spin_unlock(&root->lock); +} + +struct btrfs_delayed_node *btrfs_first_delayed_node( + struct btrfs_delayed_root *delayed_root) +{ + struct list_head *p; + struct btrfs_delayed_node *node = NULL; + + spin_lock(&delayed_root->lock); + if (list_empty(&delayed_root->node_list)) + goto out; + + p = delayed_root->node_list.next; + node = list_entry(p, struct btrfs_delayed_node, n_list); + atomic_inc(&node->refs); +out: + spin_unlock(&delayed_root->lock); + + return node; +} + +struct btrfs_delayed_node *btrfs_next_delayed_node( + struct btrfs_delayed_node *node) +{ + struct btrfs_delayed_root *delayed_root; + struct list_head *p; + struct btrfs_delayed_node *next = NULL; + + delayed_root = node->root->fs_info->delayed_root; + spin_lock(&delayed_root->lock); + if (!node->in_list) { /* not in the list */ + if (list_empty(&delayed_root->node_list)) + goto out; + p = delayed_root->node_list.next; + } else if (list_is_last(&node->n_list, &delayed_root->node_list)) + goto out; + else + p = node->n_list.next; + + next = list_entry(p, struct btrfs_delayed_node, n_list); + atomic_inc(&next->refs); +out: + spin_unlock(&delayed_root->lock); + + return next; +} + +static void __btrfs_release_delayed_node( + struct btrfs_delayed_node *delayed_node, + int mod) +{ + struct btrfs_delayed_root *delayed_root; + + if (!delayed_node) + return; + + delayed_root = delayed_node->root->fs_info->delayed_root; + + mutex_lock(&delayed_node->mutex); + if (delayed_node->count) + btrfs_queue_delayed_node(delayed_root, delayed_node, mod); + else + btrfs_dequeue_delayed_node(delayed_root, delayed_node); + mutex_unlock(&delayed_node->mutex); + + if (atomic_dec_and_test(&delayed_node->refs)) { + struct btrfs_root *root = delayed_node->root; + spin_lock(&root->inode_lock); + if (atomic_read(&delayed_node->refs) == 0) { + radix_tree_delete(&root->delayed_nodes_tree, + delayed_node->inode_id); + kmem_cache_free(delayed_node_cache, delayed_node); + } + spin_unlock(&root->inode_lock); + } +} + +static inline void btrfs_release_delayed_node(struct btrfs_delayed_node *node) +{ + __btrfs_release_delayed_node(node, 0); +} + +struct btrfs_delayed_node *btrfs_first_prepared_delayed_node( + struct btrfs_delayed_root *delayed_root) +{ + struct list_head *p; + struct btrfs_delayed_node *node = NULL; + + spin_lock(&delayed_root->lock); + if (list_empty(&delayed_root->prepare_list)) + goto out; + + p = delayed_root->prepare_list.next; + list_del_init(p); + node = list_entry(p, struct btrfs_delayed_node, p_list); + atomic_inc(&node->refs); +out: + spin_unlock(&delayed_root->lock); + + return node; +} + +static inline void btrfs_release_prepared_delayed_node( + struct btrfs_delayed_node *node) +{ + __btrfs_release_delayed_node(node, 1); +} + +struct btrfs_delayed_item *btrfs_alloc_delayed_item(u32 data_len) +{ + struct btrfs_delayed_item *item; + item = kmalloc(sizeof(*item) + data_len, GFP_NOFS); + if (item) { + item->data_len = data_len; + item->ins_or_del = 0; + item->bytes_reserved = 0; + item->block_rsv = NULL; + item->delayed_node = NULL; + atomic_set(&item->refs, 1); + } + return item; +} + +/* + * __btrfs_lookup_delayed_item - look up the delayed item by key + * @delayed_node: pointer to the delayed node + * @key: the key to look up + * @prev: used to store the prev item if the right item isn't found + * @next: used to store the next item if the right item isn't found + * + * Note: if we don't find the right item, we will return the prev item and + * the next item. + */ +static struct btrfs_delayed_item *__btrfs_lookup_delayed_item( + struct rb_root *root, + struct btrfs_key *key, + struct btrfs_delayed_item **prev, + struct btrfs_delayed_item **next) +{ + struct rb_node *node, *prev_node = NULL; + struct btrfs_delayed_item *delayed_item = NULL; + int ret = 0; + + node = root->rb_node; + + while (node) { + delayed_item = rb_entry(node, struct btrfs_delayed_item, + rb_node); + prev_node = node; + ret = btrfs_comp_cpu_keys(&delayed_item->key, key); + if (ret < 0) + node = node->rb_right; + else if (ret > 0) + node = node->rb_left; + else + return delayed_item; + } + + if (prev) { + if (!prev_node) + *prev = NULL; + else if (ret < 0) + *prev = delayed_item; + else if ((node = rb_prev(prev_node)) != NULL) { + *prev = rb_entry(node, struct btrfs_delayed_item, + rb_node); + } else + *prev = NULL; + } + + if (next) { + if (!prev_node) + *next = NULL; + else if (ret > 0) + *next = delayed_item; + else if ((node = rb_next(prev_node)) != NULL) { + *next = rb_entry(node, struct btrfs_delayed_item, + rb_node); + } else + *next = NULL; + } + return NULL; +} + +struct btrfs_delayed_item *__btrfs_lookup_delayed_insertion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item; + + item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key, + NULL, NULL); + return item; +} + +struct btrfs_delayed_item *__btrfs_lookup_delayed_deletion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item; + + item = __btrfs_lookup_delayed_item(&delayed_node->del_root, key, + NULL, NULL); + return item; +} + +struct btrfs_delayed_item *__btrfs_search_delayed_insertion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item, *next; + + item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key, + NULL, &next); + if (!item) + item = next; + + return item; +} + +struct btrfs_delayed_item *__btrfs_search_delayed_deletion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item, *next; + + item = __btrfs_lookup_delayed_item(&delayed_node->del_root, key, + NULL, &next); + if (!item) + item = next; + + return item; +} + +static int __btrfs_add_delayed_item(struct btrfs_delayed_node *delayed_node, + struct btrfs_delayed_item *ins, + int action) +{ + struct rb_node **p, *node; + struct rb_node *parent_node = NULL; + struct rb_root *root; + struct btrfs_delayed_item *item; + int cmp; + + if (action == BTRFS_DELAYED_INSERTION_ITEM) + root = &delayed_node->ins_root; + else if (action == BTRFS_DELAYED_DELETION_ITEM) + root = &delayed_node->del_root; + else + BUG(); + p = &root->rb_node; + node = &ins->rb_node; + + while (*p) { + parent_node = *p; + item = rb_entry(parent_node, struct btrfs_delayed_item, + rb_node); + + cmp = btrfs_comp_cpu_keys(&item->key, &ins->key); + if (cmp < 0) + p = &(*p)->rb_right; + else if (cmp > 0) + p = &(*p)->rb_left; + else + return -EEXIST; + } + + rb_link_node(node, parent_node, p); + rb_insert_color(node, root); + ins->delayed_node = delayed_node; + ins->ins_or_del = action; + + if (ins->key.type == BTRFS_DIR_INDEX_KEY && + action == BTRFS_DELAYED_INSERTION_ITEM && + ins->key.offset >= delayed_node->index_cnt) + delayed_node->index_cnt = ins->key.offset + 1; + + delayed_node->count++; + atomic_inc(&delayed_node->root->fs_info->delayed_root->items); + return 0; +} + +static int __btrfs_add_delayed_insertion_item(struct btrfs_delayed_node *node, + struct btrfs_delayed_item *item) +{ + return __btrfs_add_delayed_item(node, item, + BTRFS_DELAYED_INSERTION_ITEM); +} + +static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node, + struct btrfs_delayed_item *item) +{ + return __btrfs_add_delayed_item(node, item, + BTRFS_DELAYED_DELETION_ITEM); +} + +static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) +{ + struct rb_root *root; + struct btrfs_delayed_root *delayed_root; + + delayed_root = delayed_item->delayed_node->root->fs_info->delayed_root; + + BUG_ON(!delayed_root); + BUG_ON(delayed_item->ins_or_del != BTRFS_DELAYED_DELETION_ITEM && + delayed_item->ins_or_del != BTRFS_DELAYED_INSERTION_ITEM); + + if (delayed_item->ins_or_del == BTRFS_DELAYED_INSERTION_ITEM) + root = &delayed_item->delayed_node->ins_root; + else + root = &delayed_item->delayed_node->del_root; + + rb_erase(&delayed_item->rb_node, root); + delayed_item->delayed_node->count--; + atomic_dec(&delayed_root->items); + if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND && + waitqueue_active(&delayed_root->wait)) + wake_up(&delayed_root->wait); +} + +static void btrfs_release_delayed_item(struct btrfs_delayed_item *item) +{ + if (item) { + __btrfs_remove_delayed_item(item); + if (atomic_dec_and_test(&item->refs)) + kfree(item); + } +} + +struct btrfs_delayed_item *__btrfs_first_delayed_insertion_item( + struct btrfs_delayed_node *delayed_node) +{ + struct rb_node *p; + struct btrfs_delayed_item *item = NULL; + + p = rb_first(&delayed_node->ins_root); + if (p) + item = rb_entry(p, struct btrfs_delayed_item, rb_node); + + return item; +} + +struct btrfs_delayed_item *__btrfs_first_delayed_deletion_item( + struct btrfs_delayed_node *delayed_node) +{ + struct rb_node *p; + struct btrfs_delayed_item *item = NULL; + + p = rb_first(&delayed_node->del_root); + if (p) + item = rb_entry(p, struct btrfs_delayed_item, rb_node); + + return item; +} + +struct btrfs_delayed_item *__btrfs_next_delayed_item( + struct btrfs_delayed_item *item) +{ + struct rb_node *p; + struct btrfs_delayed_item *next = NULL; + + p = rb_next(&item->rb_node); + if (p) + next = rb_entry(p, struct btrfs_delayed_item, rb_node); + + return next; +} + +static inline struct btrfs_delayed_node *btrfs_get_delayed_node( + struct inode *inode) +{ + struct btrfs_inode *btrfs_inode = BTRFS_I(inode); + struct btrfs_delayed_node *delayed_node; + + delayed_node = btrfs_inode->delayed_node; + if (delayed_node) + atomic_inc(&delayed_node->refs); + + return delayed_node; +} + +static inline struct btrfs_root *btrfs_get_fs_root(struct btrfs_root *root, + u64 root_id) +{ + struct btrfs_key root_key; + + if (root->objectid == root_id) + return root; + + root_key.objectid = root_id; + root_key.type = BTRFS_ROOT_ITEM_KEY; + root_key.offset = (u64)-1; + return btrfs_read_fs_root_no_name(root->fs_info, &root_key); +} + +static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_delayed_item *item) +{ + struct btrfs_block_rsv *src_rsv; + struct btrfs_block_rsv *dst_rsv; + u64 num_bytes; + int ret; + + if (!trans->bytes_reserved) + return 0; + + src_rsv = trans->block_rsv; + dst_rsv = &root->fs_info->global_block_rsv; + + num_bytes = btrfs_calc_trans_metadata_size(root, 1); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + if (!ret) { + item->bytes_reserved = num_bytes; + item->block_rsv = dst_rsv; + } + + return ret; +} + +static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, + struct btrfs_delayed_item *item) +{ + if (!item->bytes_reserved) + return; + + btrfs_block_rsv_release(root, item->block_rsv, + item->bytes_reserved); +} + +static int btrfs_delayed_inode_reserve_metadata( + struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_block_rsv *src_rsv; + struct btrfs_block_rsv *dst_rsv; + u64 num_bytes; + int ret; + + if (!trans->bytes_reserved) + return 0; + + src_rsv = trans->block_rsv; + dst_rsv = &root->fs_info->global_block_rsv; + + num_bytes = btrfs_calc_trans_metadata_size(root, 1); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + if (!ret) + node->bytes_reserved = num_bytes; + + return ret; +} + +static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_block_rsv *rsv; + + if (!node->bytes_reserved) + return; + + rsv = &root->fs_info->global_block_rsv; + btrfs_block_rsv_release(root, rsv, + node->bytes_reserved); + node->bytes_reserved = 0; +} + +/* + * This helper will insert some continuous items into the same leaf according + * to the free space of the leaf. + */ +static int btrfs_batch_insert_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_item *item) +{ + struct btrfs_delayed_item *curr, *next; + int free_space; + int total_data_size = 0, total_size = 0; + struct extent_buffer *leaf; + char *data_ptr; + struct btrfs_key *keys; + u32 *data_size; + struct list_head head; + int slot; + int nitems; + int i; + int ret = 0; + + BUG_ON(!path->nodes[0]); + + leaf = path->nodes[0]; + free_space = btrfs_leaf_free_space(root, leaf); + INIT_LIST_HEAD(&head); + + next = item; + + /* + * count the number of the continuous items that we can insert in batch + */ + while (total_size + next->data_len + sizeof(struct btrfs_item) <= + free_space) { + total_data_size += next->data_len; + total_size += next->data_len + sizeof(struct btrfs_item); + list_add_tail(&next->tree_list, &head); + nitems++; + + curr = next; + next = __btrfs_next_delayed_item(curr); + if (!next) + break; + + if (!btrfs_is_continuous_delayed_item(curr, next)) + break; + } + + if (!nitems) { + ret = 0; + goto out; + } + + /* + * we need allocate some memory space, but it might cause the task + * to sleep, so we set all locked nodes in the path to blocking locks + * first. + */ + btrfs_set_path_blocking(path); + + keys = kmalloc(sizeof(struct btrfs_key) * nitems, GFP_NOFS); + if (!keys) { + ret = -ENOMEM; + goto out; + } + + data_size = kmalloc(sizeof(u32) * nitems, GFP_NOFS); + if (!data_size) { + ret = -ENOMEM; + goto error; + } + + /* get keys of all the delayed items */ + i = 0; + list_for_each_entry(next, &head, tree_list) { + keys[i] = next->key; + data_size[i] = next->data_len; + i++; + } + + /* reset all the locked nodes in the patch to spinning locks. */ + btrfs_clear_path_blocking(path, NULL); + + /* insert the keys of the items */ + ret = setup_items_for_insert(trans, root, path, keys, data_size, + total_data_size, total_size, nitems); + if (ret) + goto error; + + /* insert the dir index items */ + slot = path->slots[0]; + list_for_each_entry_safe(curr, next, &head, tree_list) { + data_ptr = btrfs_item_ptr(leaf, slot, char); + write_extent_buffer(leaf, &curr->data, + (unsigned long)data_ptr, + curr->data_len); + slot++; + + btrfs_delayed_item_release_metadata(root, curr); + + list_del(&curr->tree_list); + btrfs_release_delayed_item(curr); + } + +error: + kfree(data_size); + kfree(keys); +out: + return ret; +} + +/* + * This helper can just do simple insertion that needn't extend item for new + * data, such as directory name index insertion, inode insertion. + */ +static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_item *delayed_item) +{ + struct extent_buffer *leaf; + struct btrfs_item *item; + char *ptr; + int ret; + + ret = btrfs_insert_empty_item(trans, root, path, &delayed_item->key, + delayed_item->data_len); + if (ret < 0 && ret != -EEXIST) + return ret; + + leaf = path->nodes[0]; + + item = btrfs_item_nr(leaf, path->slots[0]); + ptr = btrfs_item_ptr(leaf, path->slots[0], char); + + write_extent_buffer(leaf, delayed_item->data, (unsigned long)ptr, + delayed_item->data_len); + btrfs_mark_buffer_dirty(leaf); + + btrfs_delayed_item_release_metadata(root, delayed_item); + return 0; +} + +/* + * we insert an item first, then if there are some continuous items, we try + * to insert those items into the same leaf. + */ +static int btrfs_insert_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_path *path, + struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_delayed_item *curr, *prev; + int ret = 0; + +do_again: + mutex_lock(&node->mutex); + curr = __btrfs_first_delayed_insertion_item(node); + if (!curr) + goto insert_end; + + ret = btrfs_insert_delayed_item(trans, root, path, curr); + if (ret < 0) { + btrfs_release_path(path); + goto insert_end; + } + + prev = curr; + curr = __btrfs_next_delayed_item(prev); + if (curr && btrfs_is_continuous_delayed_item(prev, curr)) { + /* insert the continuous items into the same leaf */ + path->slots[0]++; + btrfs_batch_insert_items(trans, root, path, curr); + } + btrfs_release_delayed_item(prev); + btrfs_mark_buffer_dirty(path->nodes[0]); + + btrfs_release_path(path); + mutex_unlock(&node->mutex); + goto do_again; + +insert_end: + mutex_unlock(&node->mutex); + return ret; +} + +static int btrfs_batch_delete_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_item *item) +{ + struct btrfs_delayed_item *curr, *next; + struct extent_buffer *leaf; + struct btrfs_key key; + struct list_head head; + int nitems, i, last_item; + int ret = 0; + + BUG_ON(!path->nodes[0]); + + leaf = path->nodes[0]; + + i = path->slots[0]; + last_item = btrfs_header_nritems(leaf) - 1; + if (i > last_item) + return -ENOENT; /* FIXME: Is errno suitable? */ + + next = item; + INIT_LIST_HEAD(&head); + btrfs_item_key_to_cpu(leaf, &key, i); + nitems = 0; + /* + * count the number of the dir index items that we can delete in batch + */ + while (btrfs_comp_cpu_keys(&next->key, &key) == 0) { + list_add_tail(&next->tree_list, &head); + nitems++; + + curr = next; + next = __btrfs_next_delayed_item(curr); + if (!next) + break; + + if (!btrfs_is_continuous_delayed_item(curr, next)) + break; + + i++; + if (i > last_item) + break; + btrfs_item_key_to_cpu(leaf, &key, i); + } + + if (!nitems) + return 0; + + ret = btrfs_del_items(trans, root, path, path->slots[0], nitems); + if (ret) + goto out; + + list_for_each_entry_safe(curr, next, &head, tree_list) { + btrfs_delayed_item_release_metadata(root, curr); + list_del(&curr->tree_list); + btrfs_release_delayed_item(curr); + } + +out: + return ret; +} + +static int btrfs_delete_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_path *path, + struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_delayed_item *curr, *prev; + int ret = 0; + +do_again: + mutex_lock(&node->mutex); + curr = __btrfs_first_delayed_deletion_item(node); + if (!curr) + goto delete_fail; + + ret = btrfs_search_slot(trans, root, &curr->key, path, -1, 1); + if (ret < 0) + goto delete_fail; + else if (ret > 0) { + /* + * can't find the item which the node points to, so this node + * is invalid, just drop it. + */ + prev = curr; + curr = __btrfs_next_delayed_item(prev); + btrfs_release_delayed_item(prev); + ret = 0; + btrfs_release_path(path); + if (curr) + goto do_again; + else + goto delete_fail; + } + + btrfs_batch_delete_items(trans, root, path, curr); + btrfs_release_path(path); + mutex_unlock(&node->mutex); + goto do_again; + +delete_fail: + btrfs_release_path(path); + mutex_unlock(&node->mutex); + return ret; +} + +static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node) +{ + struct btrfs_delayed_root *delayed_root; + + if (delayed_node && delayed_node->inode_dirty) { + BUG_ON(!delayed_node->root); + delayed_node->inode_dirty = 0; + delayed_node->count--; + + delayed_root = delayed_node->root->fs_info->delayed_root; + atomic_dec(&delayed_root->items); + if (atomic_read(&delayed_root->items) < + BTRFS_DELAYED_BACKGROUND && + waitqueue_active(&delayed_root->wait)) + wake_up(&delayed_root->wait); + } +} + +static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_node *node) +{ + struct btrfs_key key; + struct btrfs_inode_item *inode_item; + struct extent_buffer *leaf; + int ret; + + mutex_lock(&node->mutex); + if (!node->inode_dirty) { + mutex_unlock(&node->mutex); + return 0; + } + + key.objectid = node->inode_id; + btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); + key.offset = 0; + ret = btrfs_lookup_inode(trans, root, path, &key, 1); + if (ret > 0) { + btrfs_release_path(path); + mutex_unlock(&node->mutex); + return -ENOENT; + } else if (ret < 0) { + mutex_unlock(&node->mutex); + return ret; + } + + btrfs_unlock_up_safe(path, 1); + leaf = path->nodes[0]; + inode_item = btrfs_item_ptr(leaf, path->slots[0], + struct btrfs_inode_item); + write_extent_buffer(leaf, &node->inode_item, (unsigned long)inode_item, + sizeof(struct btrfs_inode_item)); + btrfs_mark_buffer_dirty(leaf); + btrfs_release_path(path); + + btrfs_delayed_inode_release_metadata(root, node); + btrfs_release_delayed_inode(node); + mutex_unlock(&node->mutex); + + return 0; +} + +/* Called when committing the transaction. */ +int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root) +{ + struct btrfs_delayed_root *delayed_root; + struct btrfs_delayed_node *curr_node, *prev_node; + struct btrfs_path *path; + int ret = 0; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + path->leave_spinning = 1; + + delayed_root = btrfs_get_delayed_root(root); + + curr_node = btrfs_first_delayed_node(delayed_root); + while (curr_node) { + root = curr_node->root; + ret = btrfs_insert_delayed_items(trans, path, root, + curr_node); + if (!ret) + ret = btrfs_delete_delayed_items(trans, path, root, + curr_node); + if (!ret) + ret = btrfs_update_delayed_inode(trans, root, path, + curr_node); + if (ret) { + btrfs_release_delayed_node(curr_node); + break; + } + + prev_node = curr_node; + curr_node = btrfs_next_delayed_node(curr_node); + btrfs_release_delayed_node(prev_node); + } + + btrfs_free_path(path); + return ret; +} + +static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_delayed_node *node) +{ + struct btrfs_path *path; + int ret; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + path->leave_spinning = 1; + + ret = btrfs_insert_delayed_items(trans, path, node->root, node); + if (!ret) + ret = btrfs_delete_delayed_items(trans, path, node->root, node); + if (!ret) + ret = btrfs_update_delayed_inode(trans, node->root, path, node); + btrfs_free_path(path); + + return ret; +} + +int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, + struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); + int ret; + + if (!delayed_node) + return 0; + + mutex_lock(&delayed_node->mutex); + if (!delayed_node->count) { + mutex_unlock(&delayed_node->mutex); + btrfs_release_delayed_node(delayed_node); + return 0; + } + mutex_unlock(&delayed_node->mutex); + + ret = __btrfs_commit_inode_delayed_items(trans, delayed_node); + btrfs_release_delayed_node(delayed_node); + return ret; +} + +void btrfs_remove_delayed_node(struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node; + + delayed_node = ACCESS_ONCE(BTRFS_I(inode)->delayed_node); + if (!delayed_node) + return; + + BTRFS_I(inode)->delayed_node = NULL; + btrfs_release_delayed_node(delayed_node); +} + +struct btrfs_async_delayed_node { + struct btrfs_root *root; + struct btrfs_delayed_node *delayed_node; + struct btrfs_work work; +}; + +static void btrfs_async_run_delayed_node_done(struct btrfs_work *work) +{ + struct btrfs_async_delayed_node *async_node; + struct btrfs_trans_handle *trans; + struct btrfs_path *path; + struct btrfs_delayed_node *delayed_node = NULL; + struct btrfs_root *root; + unsigned long nr = 0; + int need_requeue = 0; + int ret; + + async_node = container_of(work, struct btrfs_async_delayed_node, work); + + path = btrfs_alloc_path(); + if (!path) + goto out; + path->leave_spinning = 1; + + delayed_node = async_node->delayed_node; + root = delayed_node->root; + + trans = btrfs_join_transaction(root, 0); + if (IS_ERR(trans)) + goto free_path; + + ret = btrfs_insert_delayed_items(trans, path, root, delayed_node); + if (!ret) + ret = btrfs_delete_delayed_items(trans, path, root, + delayed_node); + + if (!ret) + btrfs_update_delayed_inode(trans, root, path, delayed_node); + + /* + * Maybe new delayed items have been inserted, so we need requeue + * the work. Besides that, we must dequeue the empty delayed nodes + * to avoid the race between delayed items balance and the worker. + * The race like this: + * Task1 Worker thread + * count == 0, needn't requeue + * also needn't insert the + * delayed node into prepare + * list again. + * add lots of delayed items + * queue the delayed node + * already in the list, + * and not in the prepare + * list, it means the delayed + * node is being dealt with + * by the worker. + * do delayed items balance + * the delayed node is being + * dealt with by the worker + * now, just wait. + * the worker goto idle. + * Task1 will sleep until the transaction is commited. + */ + mutex_lock(&delayed_node->mutex); + if (delayed_node->count) + need_requeue = 1; + else + btrfs_dequeue_delayed_node(root->fs_info->delayed_root, + delayed_node); + mutex_unlock(&delayed_node->mutex); + + nr = trans->blocks_used; + + btrfs_end_transaction_dmeta(trans, root); + __btrfs_btree_balance_dirty(root, nr); +free_path: + btrfs_free_path(path); +out: + if (need_requeue) + btrfs_requeue_work(&async_node->work); + else { + btrfs_release_prepared_delayed_node(delayed_node); + kfree(async_node); + } +} + +static int btrfs_wq_run_delayed_node(struct btrfs_delayed_root *delayed_root, + struct btrfs_root *root, int all) +{ + struct btrfs_async_delayed_node *async_node; + struct btrfs_delayed_node *curr; + int count = 0; + +again: + curr = btrfs_first_prepared_delayed_node(delayed_root); + if (!curr) + return 0; + + async_node = kmalloc(sizeof(*async_node), GFP_NOFS); + if (!async_node) { + btrfs_release_prepared_delayed_node(curr); + return -ENOMEM; + } + + async_node->root = root; + async_node->delayed_node = curr; + + async_node->work.func = btrfs_async_run_delayed_node_done; + async_node->work.flags = 0; + + btrfs_queue_worker(&root->fs_info->delayed_workers, &async_node->work); + count++; + + if (all || count < 4) + goto again; + + return 0; +} + +void btrfs_balance_delayed_items(struct btrfs_root *root) +{ + struct btrfs_delayed_root *delayed_root; + + delayed_root = btrfs_get_delayed_root(root); + + if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) + return; + + if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) { + int ret; + ret = btrfs_wq_run_delayed_node(delayed_root, root, 1); + if (ret) + return; + + wait_event_interruptible_timeout( + delayed_root->wait, + (atomic_read(&delayed_root->items) < + BTRFS_DELAYED_BACKGROUND), + HZ); + return; + } + + btrfs_wq_run_delayed_node(delayed_root, root, 0); +} + +int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, const char *name, + int name_len, struct inode *dir, + struct btrfs_disk_key *disk_key, u8 type, + u64 index) +{ + struct btrfs_delayed_node *delayed_node; + struct btrfs_delayed_item *delayed_item; + struct btrfs_dir_item *dir_item; + int ret; + + delayed_node = btrfs_get_or_create_delayed_node(dir); + if (IS_ERR(delayed_node)) + return PTR_ERR(delayed_node); + + delayed_item = btrfs_alloc_delayed_item(sizeof(*dir_item) + name_len); + if (!delayed_item) { + ret = -ENOMEM; + goto release_node; + } + + ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item); + /* + * we have reserved enough space when we start a new transaction, + * so reserving metadata failure is impossible + */ + BUG_ON(ret); + + delayed_item->key.objectid = btrfs_ino(dir); + btrfs_set_key_type(&delayed_item->key, BTRFS_DIR_INDEX_KEY); + delayed_item->key.offset = index; + + dir_item = (struct btrfs_dir_item *)delayed_item->data; + dir_item->location = *disk_key; + dir_item->transid = cpu_to_le64(trans->transid); + dir_item->data_len = 0; + dir_item->name_len = cpu_to_le16(name_len); + dir_item->type = type; + memcpy((char *)(dir_item + 1), name, name_len); + + mutex_lock(&delayed_node->mutex); + ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item); + if (unlikely(ret)) { + printk(KERN_ERR "err add delayed dir index item(name: %s) into " + "the insertion tree of the delayed node" + "(root id: %llu, inode id: %llu, errno: %d)\n", + name, + (unsigned long long)delayed_node->root->objectid, + (unsigned long long)delayed_node->inode_id, + ret); + BUG(); + } + mutex_unlock(&delayed_node->mutex); + +release_node: + btrfs_release_delayed_node(delayed_node); + return ret; +} + +static int btrfs_delete_delayed_insertion_item(struct btrfs_root *root, + struct btrfs_delayed_node *node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item; + + mutex_lock(&node->mutex); + item = __btrfs_lookup_delayed_insertion_item(node, key); + if (!item) { + mutex_unlock(&node->mutex); + return 1; + } + + btrfs_delayed_item_release_metadata(root, item); + btrfs_release_delayed_item(item); + mutex_unlock(&node->mutex); + return 0; +} + +int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *dir, + u64 index) +{ + struct btrfs_delayed_node *node; + struct btrfs_delayed_item *item; + struct btrfs_key item_key; + int ret; + + node = btrfs_get_or_create_delayed_node(dir); + if (IS_ERR(node)) + return PTR_ERR(node); + + item_key.objectid = btrfs_ino(dir); + btrfs_set_key_type(&item_key, BTRFS_DIR_INDEX_KEY); + item_key.offset = index; + + ret = btrfs_delete_delayed_insertion_item(root, node, &item_key); + if (!ret) + goto end; + + item = btrfs_alloc_delayed_item(0); + if (!item) { + ret = -ENOMEM; + goto end; + } + + item->key = item_key; + + ret = btrfs_delayed_item_reserve_metadata(trans, root, item); + /* + * we have reserved enough space when we start a new transaction, + * so reserving metadata failure is impossible. + */ + BUG_ON(ret); + + mutex_lock(&node->mutex); + ret = __btrfs_add_delayed_deletion_item(node, item); + if (unlikely(ret)) { + printk(KERN_ERR "err add delayed dir index item(index: %llu) " + "into the deletion tree of the delayed node" + "(root id: %llu, inode id: %llu, errno: %d)\n", + (unsigned long long)index, + (unsigned long long)node->root->objectid, + (unsigned long long)node->inode_id, + ret); + BUG(); + } + mutex_unlock(&node->mutex); +end: + btrfs_release_delayed_node(node); + return ret; +} + +int btrfs_inode_delayed_dir_index_count(struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node = BTRFS_I(inode)->delayed_node; + int ret = 0; + + if (!delayed_node) + return -ENOENT; + + /* + * Since we have held i_mutex of this directory, it is impossible that + * a new directory index is added into the delayed node and index_cnt + * is updated now. So we needn't lock the delayed node. + */ + if (!delayed_node->index_cnt) + return -EINVAL; + + BTRFS_I(inode)->index_cnt = delayed_node->index_cnt; + return ret; +} + +void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, + struct list_head *del_list) +{ + struct btrfs_delayed_node *delayed_node; + struct btrfs_delayed_item *item; + + delayed_node = btrfs_get_delayed_node(inode); + if (!delayed_node) + return; + + mutex_lock(&delayed_node->mutex); + item = __btrfs_first_delayed_insertion_item(delayed_node); + while (item) { + atomic_inc(&item->refs); + list_add_tail(&item->readdir_list, ins_list); + item = __btrfs_next_delayed_item(item); + } + + item = __btrfs_first_delayed_deletion_item(delayed_node); + while (item) { + atomic_inc(&item->refs); + list_add_tail(&item->readdir_list, del_list); + item = __btrfs_next_delayed_item(item); + } + mutex_unlock(&delayed_node->mutex); + /* + * This delayed node is still cached in the btrfs inode, so refs + * must be > 1 now, and we needn't check it is going to be freed + * or not. + * + * Besides that, this function is used to read dir, we do not + * insert/delete delayed items in this period. So we also needn't + * requeue or dequeue this delayed node. + */ + atomic_dec(&delayed_node->refs); +} + +void btrfs_put_delayed_items(struct list_head *ins_list, + struct list_head *del_list) +{ + struct btrfs_delayed_item *curr, *next; + + list_for_each_entry_safe(curr, next, ins_list, readdir_list) { + list_del(&curr->readdir_list); + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + } + + list_for_each_entry_safe(curr, next, del_list, readdir_list) { + list_del(&curr->readdir_list); + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + } +} + +int btrfs_should_delete_dir_index(struct list_head *del_list, + u64 index) +{ + struct btrfs_delayed_item *curr, *next; + int ret; + + if (list_empty(del_list)) + return 0; + + list_for_each_entry_safe(curr, next, del_list, readdir_list) { + if (curr->key.offset > index) + break; + + list_del(&curr->readdir_list); + ret = (curr->key.offset == index); + + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + + if (ret) + return 1; + else + continue; + } + return 0; +} + +/* + * btrfs_readdir_delayed_dir_index - read dir info stored in the delayed tree + * + */ +int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent, + filldir_t filldir, + struct list_head *ins_list) +{ + struct btrfs_dir_item *di; + struct btrfs_delayed_item *curr, *next; + struct btrfs_key location; + char *name; + int name_len; + int over = 0; + unsigned char d_type; + + if (list_empty(ins_list)) + return 0; + + /* + * Changing the data of the delayed item is impossible. So + * we needn't lock them. And we have held i_mutex of the + * directory, nobody can delete any directory indexes now. + */ + list_for_each_entry_safe(curr, next, ins_list, readdir_list) { + list_del(&curr->readdir_list); + + if (curr->key.offset < filp->f_pos) { + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + continue; + } + + filp->f_pos = curr->key.offset; + + di = (struct btrfs_dir_item *)curr->data; + name = (char *)(di + 1); + name_len = le16_to_cpu(di->name_len); + + d_type = btrfs_filetype_table[di->type]; + btrfs_disk_key_to_cpu(&location, &di->location); + + over = filldir(dirent, name, name_len, curr->key.offset, + location.objectid, d_type); + + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + + if (over) + return 1; + } + return 0; +} + +BTRFS_SETGET_STACK_FUNCS(stack_inode_generation, struct btrfs_inode_item, + generation, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_sequence, struct btrfs_inode_item, + sequence, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_transid, struct btrfs_inode_item, + transid, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_size, struct btrfs_inode_item, size, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_nbytes, struct btrfs_inode_item, + nbytes, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_block_group, struct btrfs_inode_item, + block_group, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_nlink, struct btrfs_inode_item, nlink, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_uid, struct btrfs_inode_item, uid, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_gid, struct btrfs_inode_item, gid, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_mode, struct btrfs_inode_item, mode, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_rdev, struct btrfs_inode_item, rdev, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_flags, struct btrfs_inode_item, flags, 64); + +BTRFS_SETGET_STACK_FUNCS(stack_timespec_sec, struct btrfs_timespec, sec, 64); +BTRFS_SETGET_STACK_FUNCS(stack_timespec_nsec, struct btrfs_timespec, nsec, 32); + +static void fill_stack_inode_item(struct btrfs_trans_handle *trans, + struct btrfs_inode_item *inode_item, + struct inode *inode) +{ + btrfs_set_stack_inode_uid(inode_item, inode->i_uid); + btrfs_set_stack_inode_gid(inode_item, inode->i_gid); + btrfs_set_stack_inode_size(inode_item, BTRFS_I(inode)->disk_i_size); + btrfs_set_stack_inode_mode(inode_item, inode->i_mode); + btrfs_set_stack_inode_nlink(inode_item, inode->i_nlink); + btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode)); + btrfs_set_stack_inode_generation(inode_item, + BTRFS_I(inode)->generation); + btrfs_set_stack_inode_sequence(inode_item, BTRFS_I(inode)->sequence); + btrfs_set_stack_inode_transid(inode_item, trans->transid); + btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev); + btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags); + btrfs_set_stack_inode_block_group(inode_item, + BTRFS_I(inode)->block_group); + + btrfs_set_stack_timespec_sec(btrfs_inode_atime(inode_item), + inode->i_atime.tv_sec); + btrfs_set_stack_timespec_nsec(btrfs_inode_atime(inode_item), + inode->i_atime.tv_nsec); + + btrfs_set_stack_timespec_sec(btrfs_inode_mtime(inode_item), + inode->i_mtime.tv_sec); + btrfs_set_stack_timespec_nsec(btrfs_inode_mtime(inode_item), + inode->i_mtime.tv_nsec); + + btrfs_set_stack_timespec_sec(btrfs_inode_ctime(inode_item), + inode->i_ctime.tv_sec); + btrfs_set_stack_timespec_nsec(btrfs_inode_ctime(inode_item), + inode->i_ctime.tv_nsec); +} + +int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node; + int ret; + + delayed_node = btrfs_get_or_create_delayed_node(inode); + if (IS_ERR(delayed_node)) + return PTR_ERR(delayed_node); + + mutex_lock(&delayed_node->mutex); + if (delayed_node->inode_dirty) { + fill_stack_inode_item(trans, &delayed_node->inode_item, inode); + goto release_node; + } + + ret = btrfs_delayed_inode_reserve_metadata(trans, root, delayed_node); + /* + * we must reserve enough space when we start a new transaction, + * so reserving metadata failure is impossible + */ + BUG_ON(ret); + + fill_stack_inode_item(trans, &delayed_node->inode_item, inode); + delayed_node->inode_dirty = 1; + delayed_node->count++; + atomic_inc(&root->fs_info->delayed_root->items); +release_node: + mutex_unlock(&delayed_node->mutex); + btrfs_release_delayed_node(delayed_node); + return ret; +} + +static void __btrfs_kill_delayed_node(struct btrfs_delayed_node *delayed_node) +{ + struct btrfs_root *root = delayed_node->root; + struct btrfs_delayed_item *curr_item, *prev_item; + + mutex_lock(&delayed_node->mutex); + curr_item = __btrfs_first_delayed_insertion_item(delayed_node); + while (curr_item) { + btrfs_delayed_item_release_metadata(root, curr_item); + prev_item = curr_item; + curr_item = __btrfs_next_delayed_item(prev_item); + btrfs_release_delayed_item(prev_item); + } + + curr_item = __btrfs_first_delayed_deletion_item(delayed_node); + while (curr_item) { + btrfs_delayed_item_release_metadata(root, curr_item); + prev_item = curr_item; + curr_item = __btrfs_next_delayed_item(prev_item); + btrfs_release_delayed_item(prev_item); + } + + if (delayed_node->inode_dirty) { + btrfs_delayed_inode_release_metadata(root, delayed_node); + btrfs_release_delayed_inode(delayed_node); + } + mutex_unlock(&delayed_node->mutex); +} + +void btrfs_kill_delayed_inode_items(struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node; + + delayed_node = btrfs_get_delayed_node(inode); + if (!delayed_node) + return; + + __btrfs_kill_delayed_node(delayed_node); + btrfs_release_delayed_node(delayed_node); +} + +void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) +{ + u64 inode_id = 0; + struct btrfs_delayed_node *delayed_nodes[8]; + int i, n; + + while (1) { + spin_lock(&root->inode_lock); + n = radix_tree_gang_lookup(&root->delayed_nodes_tree, + (void **)delayed_nodes, inode_id, + ARRAY_SIZE(delayed_nodes)); + if (!n) { + spin_unlock(&root->inode_lock); + break; + } + + inode_id = delayed_nodes[n - 1]->inode_id + 1; + + for (i = 0; i < n; i++) + atomic_inc(&delayed_nodes[i]->refs); + spin_unlock(&root->inode_lock); + + for (i = 0; i < n; i++) { + __btrfs_kill_delayed_node(delayed_nodes[i]); + btrfs_release_delayed_node(delayed_nodes[i]); + } + } +} diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h new file mode 100644 index 000000000000..eb7d240aa648 --- /dev/null +++ b/fs/btrfs/delayed-inode.h @@ -0,0 +1,141 @@ +/* + * Copyright (C) 2011 Fujitsu. All rights reserved. + * Written by Miao Xie <miaox@cn.fujitsu.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#ifndef __DELAYED_TREE_OPERATION_H +#define __DELAYED_TREE_OPERATION_H + +#include <linux/rbtree.h> +#include <linux/spinlock.h> +#include <linux/mutex.h> +#include <linux/list.h> +#include <linux/wait.h> +#include <asm/atomic.h> + +#include "ctree.h" + +/* types of the delayed item */ +#define BTRFS_DELAYED_INSERTION_ITEM 1 +#define BTRFS_DELAYED_DELETION_ITEM 2 + +struct btrfs_delayed_root { + spinlock_t lock; + struct list_head node_list; + /* + * Used for delayed nodes which is waiting to be dealt with by the + * worker. If the delayed node is inserted into the work queue, we + * drop it from this list. + */ + struct list_head prepare_list; + atomic_t items; /* for delayed items */ + int nodes; /* for delayed nodes */ + wait_queue_head_t wait; +}; + +struct btrfs_delayed_node { + u64 inode_id; + u64 bytes_reserved; + struct btrfs_root *root; + /* Used to add the node into the delayed root's node list. */ + struct list_head n_list; + /* + * Used to add the node into the prepare list, the nodes in this list + * is waiting to be dealt with by the async worker. + */ + struct list_head p_list; + struct rb_root ins_root; + struct rb_root del_root; + struct mutex mutex; + struct btrfs_inode_item inode_item; + atomic_t refs; + u64 index_cnt; + bool in_list; + bool inode_dirty; + int count; +}; + +struct btrfs_delayed_item { + struct rb_node rb_node; + struct btrfs_key key; + struct list_head tree_list; /* used for batch insert/delete items */ + struct list_head readdir_list; /* used for readdir items */ + u64 bytes_reserved; + struct btrfs_block_rsv *block_rsv; + struct btrfs_delayed_node *delayed_node; + atomic_t refs; + int ins_or_del; + u32 data_len; + char data[0]; +}; + +static inline void btrfs_init_delayed_root( + struct btrfs_delayed_root *delayed_root) +{ + atomic_set(&delayed_root->items, 0); + delayed_root->nodes = 0; + spin_lock_init(&delayed_root->lock); + init_waitqueue_head(&delayed_root->wait); + INIT_LIST_HEAD(&delayed_root->node_list); + INIT_LIST_HEAD(&delayed_root->prepare_list); +} + +int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, const char *name, + int name_len, struct inode *dir, + struct btrfs_disk_key *disk_key, u8 type, + u64 index); + +int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *dir, + u64 index); + +int btrfs_inode_delayed_dir_index_count(struct inode *inode); + +int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root); + +void btrfs_balance_delayed_items(struct btrfs_root *root); + +int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, + struct inode *inode); +/* Used for evicting the inode. */ +void btrfs_remove_delayed_node(struct inode *inode); +void btrfs_kill_delayed_inode_items(struct inode *inode); + + +int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode); + +/* Used for drop dead root */ +void btrfs_kill_all_delayed_nodes(struct btrfs_root *root); + +/* Used for readdir() */ +void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, + struct list_head *del_list); +void btrfs_put_delayed_items(struct list_head *ins_list, + struct list_head *del_list); +int btrfs_should_delete_dir_index(struct list_head *del_list, + u64 index); +int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent, + filldir_t filldir, + struct list_head *ins_list); + +/* for init */ +int __init btrfs_delayed_inode_init(void); +void btrfs_delayed_inode_exit(void); +#endif diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index e807b143b857..125cf76fcd08 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -281,44 +281,6 @@ again: } /* - * This checks to see if there are any delayed refs in the - * btree for a given bytenr. It returns one if it finds any - * and zero otherwise. - * - * If it only finds a head node, it returns 0. - * - * The idea is to use this when deciding if you can safely delete an - * extent from the extent allocation tree. There may be a pending - * ref in the rbtree that adds or removes references, so as long as this - * returns one you need to leave the BTRFS_EXTENT_ITEM in the extent - * allocation tree. - */ -int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr) -{ - struct btrfs_delayed_ref_node *ref; - struct btrfs_delayed_ref_root *delayed_refs; - struct rb_node *prev_node; - int ret = 0; - - delayed_refs = &trans->transaction->delayed_refs; - spin_lock(&delayed_refs->lock); - - ref = find_ref_head(&delayed_refs->root, bytenr, NULL); - if (ref) { - prev_node = rb_prev(&ref->rb_node); - if (!prev_node) - goto out; - ref = rb_entry(prev_node, struct btrfs_delayed_ref_node, - rb_node); - if (ref->bytenr == bytenr) - ret = 1; - } -out: - spin_unlock(&delayed_refs->lock); - return ret; -} - -/* * helper function to update an extent delayed ref in the * rbtree. existing and update must both have the same * bytenr and parent @@ -483,6 +445,8 @@ static noinline int add_delayed_ref_head(struct btrfs_trans_handle *trans, INIT_LIST_HEAD(&head_ref->cluster); mutex_init(&head_ref->mutex); + trace_btrfs_delayed_ref_head(ref, head_ref, action); + existing = tree_insert(&delayed_refs->root, &ref->rb_node); if (existing) { @@ -537,6 +501,8 @@ static noinline int add_delayed_tree_ref(struct btrfs_trans_handle *trans, } full_ref->level = level; + trace_btrfs_delayed_tree_ref(ref, full_ref, action); + existing = tree_insert(&delayed_refs->root, &ref->rb_node); if (existing) { @@ -591,6 +557,8 @@ static noinline int add_delayed_data_ref(struct btrfs_trans_handle *trans, full_ref->objectid = owner; full_ref->offset = offset; + trace_btrfs_delayed_data_ref(ref, full_ref, action); + existing = tree_insert(&delayed_refs->root, &ref->rb_node); if (existing) { @@ -741,79 +709,3 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr) return btrfs_delayed_node_to_head(ref); return NULL; } - -/* - * add a delayed ref to the tree. This does all of the accounting required - * to make sure the delayed ref is eventually processed before this - * transaction commits. - * - * The main point of this call is to add and remove a backreference in a single - * shot, taking the lock only once, and only searching for the head node once. - * - * It is the same as doing a ref add and delete in two separate calls. - */ -#if 0 -int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans, - u64 bytenr, u64 num_bytes, u64 orig_parent, - u64 parent, u64 orig_ref_root, u64 ref_root, - u64 orig_ref_generation, u64 ref_generation, - u64 owner_objectid, int pin) -{ - struct btrfs_delayed_ref *ref; - struct btrfs_delayed_ref *old_ref; - struct btrfs_delayed_ref_head *head_ref; - struct btrfs_delayed_ref_root *delayed_refs; - int ret; - - ref = kmalloc(sizeof(*ref), GFP_NOFS); - if (!ref) - return -ENOMEM; - - old_ref = kmalloc(sizeof(*old_ref), GFP_NOFS); - if (!old_ref) { - kfree(ref); - return -ENOMEM; - } - - /* - * the parent = 0 case comes from cases where we don't actually - * know the parent yet. It will get updated later via a add/drop - * pair. - */ - if (parent == 0) - parent = bytenr; - if (orig_parent == 0) - orig_parent = bytenr; - - head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS); - if (!head_ref) { - kfree(ref); - kfree(old_ref); - return -ENOMEM; - } - delayed_refs = &trans->transaction->delayed_refs; - spin_lock(&delayed_refs->lock); - - /* - * insert both the head node and the new ref without dropping - * the spin lock - */ - ret = __btrfs_add_delayed_ref(trans, &head_ref->node, bytenr, num_bytes, - (u64)-1, 0, 0, 0, - BTRFS_UPDATE_DELAYED_HEAD, 0); - BUG_ON(ret); - - ret = __btrfs_add_delayed_ref(trans, &ref->node, bytenr, num_bytes, - parent, ref_root, ref_generation, - owner_objectid, BTRFS_ADD_DELAYED_REF, 0); - BUG_ON(ret); - - ret = __btrfs_add_delayed_ref(trans, &old_ref->node, bytenr, num_bytes, - orig_parent, orig_ref_root, - orig_ref_generation, owner_objectid, - BTRFS_DROP_DELAYED_REF, pin); - BUG_ON(ret); - spin_unlock(&delayed_refs->lock); - return 0; -} -#endif diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h index 50e3cf92fbda..e287e3b0eab0 100644 --- a/fs/btrfs/delayed-ref.h +++ b/fs/btrfs/delayed-ref.h @@ -166,12 +166,6 @@ int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_head * btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr); -int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr); -int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans, - u64 bytenr, u64 num_bytes, u64 orig_parent, - u64 parent, u64 orig_ref_root, u64 ref_root, - u64 orig_ref_generation, u64 ref_generation, - u64 owner_objectid, int pin); int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_head *head); int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans, diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c index f0cad5ae5be7..685f2593c4f0 100644 --- a/fs/btrfs/dir-item.c +++ b/fs/btrfs/dir-item.c @@ -50,7 +50,6 @@ static struct btrfs_dir_item *insert_with_overflow(struct btrfs_trans_handle if (di) return ERR_PTR(-EEXIST); ret = btrfs_extend_item(trans, root, path, data_size); - WARN_ON(ret > 0); } if (ret < 0) return ERR_PTR(ret); @@ -124,8 +123,9 @@ int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans, * to use for the second index (if one is created). */ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root - *root, const char *name, int name_len, u64 dir, - struct btrfs_key *location, u8 type, u64 index) + *root, const char *name, int name_len, + struct inode *dir, struct btrfs_key *location, + u8 type, u64 index) { int ret = 0; int ret2 = 0; @@ -137,13 +137,17 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root struct btrfs_disk_key disk_key; u32 data_size; - key.objectid = dir; + key.objectid = btrfs_ino(dir); btrfs_set_key_type(&key, BTRFS_DIR_ITEM_KEY); key.offset = btrfs_name_hash(name, name_len); path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; path->leave_spinning = 1; + btrfs_cpu_key_to_disk(&disk_key, location); + data_size = sizeof(*dir_item) + name_len; dir_item = insert_with_overflow(trans, root, path, &key, data_size, name, name_len); @@ -151,11 +155,10 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root ret = PTR_ERR(dir_item); if (ret == -EEXIST) goto second_insert; - goto out; + goto out_free; } leaf = path->nodes[0]; - btrfs_cpu_key_to_disk(&disk_key, location); btrfs_set_dir_item_key(leaf, dir_item, &disk_key); btrfs_set_dir_type(leaf, dir_item, type); btrfs_set_dir_data_len(leaf, dir_item, 0); @@ -170,29 +173,13 @@ second_insert: /* FIXME, use some real flag for selecting the extra index */ if (root == root->fs_info->tree_root) { ret = 0; - goto out; + goto out_free; } - btrfs_release_path(root, path); + btrfs_release_path(path); - btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY); - key.offset = index; - dir_item = insert_with_overflow(trans, root, path, &key, data_size, - name, name_len); - if (IS_ERR(dir_item)) { - ret2 = PTR_ERR(dir_item); - goto out; - } - leaf = path->nodes[0]; - btrfs_cpu_key_to_disk(&disk_key, location); - btrfs_set_dir_item_key(leaf, dir_item, &disk_key); - btrfs_set_dir_type(leaf, dir_item, type); - btrfs_set_dir_data_len(leaf, dir_item, 0); - btrfs_set_dir_name_len(leaf, dir_item, name_len); - btrfs_set_dir_transid(leaf, dir_item, trans->transid); - name_ptr = (unsigned long)(dir_item + 1); - write_extent_buffer(leaf, name, name_ptr, name_len); - btrfs_mark_buffer_dirty(leaf); -out: + ret2 = btrfs_insert_delayed_dir_index(trans, root, name, name_len, dir, + &disk_key, type, index); +out_free: btrfs_free_path(path); if (ret) return ret; @@ -377,6 +364,9 @@ struct btrfs_dir_item *btrfs_match_dir_item_name(struct btrfs_root *root, leaf = path->nodes[0]; dir_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dir_item); + if (verify_dir_item(root, leaf, dir_item)) + return NULL; + total_len = btrfs_item_size_nr(leaf, path->slots[0]); while (cur < total_len) { this_len = sizeof(*dir_item) + @@ -429,3 +419,35 @@ int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans, } return ret; } + +int verify_dir_item(struct btrfs_root *root, + struct extent_buffer *leaf, + struct btrfs_dir_item *dir_item) +{ + u16 namelen = BTRFS_NAME_LEN; + u8 type = btrfs_dir_type(leaf, dir_item); + + if (type >= BTRFS_FT_MAX) { + printk(KERN_CRIT "btrfs: invalid dir item type: %d\n", + (int)type); + return 1; + } + + if (type == BTRFS_FT_XATTR) + namelen = XATTR_NAME_MAX; + + if (btrfs_dir_name_len(leaf, dir_item) > namelen) { + printk(KERN_CRIT "btrfs: invalid dir item name len: %u\n", + (unsigned)btrfs_dir_data_len(leaf, dir_item)); + return 1; + } + + /* BTRFS_MAX_XATTR_SIZE is the same for all dir items */ + if (btrfs_dir_data_len(leaf, dir_item) > BTRFS_MAX_XATTR_SIZE(root)) { + printk(KERN_CRIT "btrfs: invalid dir item data len: %u\n", + (unsigned)btrfs_dir_data_len(leaf, dir_item)); + return 1; + } + + return 0; +} diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index fdce8799b98d..98b6a71decba 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -29,6 +29,8 @@ #include <linux/crc32c.h> #include <linux/slab.h> #include <linux/migrate.h> +#include <linux/ratelimit.h> +#include <asm/unaligned.h> #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -40,6 +42,7 @@ #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); @@ -136,7 +139,7 @@ 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; @@ -153,7 +156,7 @@ 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; @@ -198,7 +201,7 @@ u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len) void btrfs_csum_final(u32 crc, char *result) { - *(__le32 *)result = ~cpu_to_le32(crc); + put_unaligned_le32(~crc, result); } /* @@ -253,14 +256,12 @@ 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; @@ -295,13 +296,11 @@ 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: @@ -323,6 +322,7 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, int num_copies = 0; int mirror_num = 0; + clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags); io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree; while (1) { ret = read_extent_buffer_pages(io_tree, eb, start, 1, @@ -331,6 +331,14 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, !verify_parent_transid(io_tree, eb, parent_transid)) return ret; + /* + * This buffer's crc is fine, but its contents are corrupted, so + * there is no reason to read the other copies, they won't be + * any less wrong. + */ + if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags)) + return ret; + num_copies = btrfs_num_copies(&root->fs_info->mapping_tree, eb->start, eb->len); if (num_copies == 1) @@ -359,14 +367,18 @@ static int csum_dirty_buffer(struct btrfs_root *root, struct page *page) tree = &BTRFS_I(page->mapping->host)->io_tree; - if (page->private == EXTENT_PAGE_PRIVATE) + if (page->private == EXTENT_PAGE_PRIVATE) { + WARN_ON(1); goto out; - if (!page->private) + } + if (!page->private) { + WARN_ON(1); goto out; + } 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; @@ -415,6 +427,73 @@ static int check_tree_block_fsid(struct btrfs_root *root, return ret; } +#define CORRUPT(reason, eb, root, slot) \ + printk(KERN_CRIT "btrfs: corrupt leaf, %s: block=%llu," \ + "root=%llu, slot=%d\n", reason, \ + (unsigned long long)btrfs_header_bytenr(eb), \ + (unsigned long long)root->objectid, slot) + +static noinline int check_leaf(struct btrfs_root *root, + struct extent_buffer *leaf) +{ + struct btrfs_key key; + struct btrfs_key leaf_key; + u32 nritems = btrfs_header_nritems(leaf); + int slot; + + if (nritems == 0) + return 0; + + /* Check the 0 item */ + if (btrfs_item_offset_nr(leaf, 0) + btrfs_item_size_nr(leaf, 0) != + BTRFS_LEAF_DATA_SIZE(root)) { + CORRUPT("invalid item offset size pair", leaf, root, 0); + return -EIO; + } + + /* + * Check to make sure each items keys are in the correct order and their + * offsets make sense. We only have to loop through nritems-1 because + * we check the current slot against the next slot, which verifies the + * next slot's offset+size makes sense and that the current's slot + * offset is correct. + */ + for (slot = 0; slot < nritems - 1; slot++) { + btrfs_item_key_to_cpu(leaf, &leaf_key, slot); + btrfs_item_key_to_cpu(leaf, &key, slot + 1); + + /* Make sure the keys are in the right order */ + if (btrfs_comp_cpu_keys(&leaf_key, &key) >= 0) { + CORRUPT("bad key order", leaf, root, slot); + return -EIO; + } + + /* + * Make sure the offset and ends are right, remember that the + * item data starts at the end of the leaf and grows towards the + * front. + */ + if (btrfs_item_offset_nr(leaf, slot) != + btrfs_item_end_nr(leaf, slot + 1)) { + CORRUPT("slot offset bad", leaf, root, slot); + return -EIO; + } + + /* + * Check to make sure that we don't point outside of the leaf, + * just incase all the items are consistent to eachother, but + * all point outside of the leaf. + */ + if (btrfs_item_end_nr(leaf, slot) > + BTRFS_LEAF_DATA_SIZE(root)) { + CORRUPT("slot end outside of leaf", leaf, root, slot); + return -EIO; + } + } + + return 0; +} + #ifdef CONFIG_DEBUG_LOCK_ALLOC void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, int level) { @@ -444,7 +523,7 @@ 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; @@ -452,12 +531,10 @@ 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; } @@ -469,10 +546,8 @@ 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; } @@ -481,8 +556,20 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, btrfs_set_buffer_lockdep_class(eb, found_level); ret = csum_tree_block(root, eb, 1); - if (ret) + if (ret) { ret = -EIO; + goto err; + } + + /* + * If this is a leaf block and it is corrupt, set the corrupt bit so + * that we don't try and read the other copies of this block, just + * return -EIO. + */ + if (found_level == 0 && check_leaf(root, eb)) { + set_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags); + ret = -EIO; + } end = min_t(u64, eb->len, PAGE_CACHE_SIZE); end = eb->start + end - 1; @@ -557,12 +644,6 @@ 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; @@ -843,7 +924,6 @@ static const struct address_space_operations btree_aops = { .writepages = btree_writepages, .releasepage = btree_releasepage, .invalidatepage = btree_invalidatepage, - .sync_page = block_sync_page, #ifdef CONFIG_MIGRATION .migratepage = btree_migratepage, #endif @@ -871,7 +951,7 @@ 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; } @@ -882,7 +962,7 @@ 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; } @@ -966,13 +1046,13 @@ 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; 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); @@ -988,7 +1068,7 @@ 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)); @@ -1156,7 +1236,10 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, root, fs_info, location->objectid); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + kfree(root); + return ERR_PTR(-ENOMEM); + } ret = btrfs_search_slot(NULL, tree_root, location, path, 0, 0); if (ret == 0) { l = path->nodes[0]; @@ -1180,27 +1263,14 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, root->commit_root = btrfs_root_node(root); BUG_ON(!root->node); out: - if (location->objectid != BTRFS_TREE_LOG_OBJECTID) + if (location->objectid != BTRFS_TREE_LOG_OBJECTID) { root->ref_cows = 1; + btrfs_check_and_init_root_item(&root->root_item); + } 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) { @@ -1229,6 +1299,19 @@ 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) { @@ -1272,41 +1355,6 @@ 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; - - 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; @@ -1314,7 +1362,8 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) struct btrfs_device *device; struct backing_dev_info *bdi; - 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); @@ -1323,86 +1372,11 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) break; } } + rcu_read_unlock(); return ret; } /* - * this unplugs every device on the box, and it is only used when page - * is null - */ -static void __unplug_io_fn(struct backing_dev_info *bdi, struct page *page) -{ - struct btrfs_device *device; - struct btrfs_fs_info *info; - - info = (struct btrfs_fs_info *)bdi->unplug_io_data; - list_for_each_entry(device, &info->fs_devices->devices, dev_list) { - if (!device->bdev) - continue; - - bdi = blk_get_backing_dev_info(device->bdev); - if (bdi->unplug_io_fn) - bdi->unplug_io_fn(bdi, page); - } -} - -static void btrfs_unplug_io_fn(struct backing_dev_info *bdi, struct page *page) -{ - struct inode *inode; - struct extent_map_tree *em_tree; - struct extent_map *em; - struct address_space *mapping; - u64 offset; - - /* the generic O_DIRECT read code does this */ - if (1 || !page) { - __unplug_io_fn(bdi, page); - return; - } - - /* - * page->mapping may change at any time. Get a consistent copy - * and use that for everything below - */ - smp_mb(); - mapping = page->mapping; - if (!mapping) - return; - - inode = mapping->host; - - /* - * don't do the expensive searching for a small number of - * devices - */ - if (BTRFS_I(inode)->root->fs_info->fs_devices->open_devices <= 2) { - __unplug_io_fn(bdi, page); - return; - } - - offset = page_offset(page); - - em_tree = &BTRFS_I(inode)->extent_tree; - read_lock(&em_tree->lock); - em = lookup_extent_mapping(em_tree, offset, PAGE_CACHE_SIZE); - read_unlock(&em_tree->lock); - if (!em) { - __unplug_io_fn(bdi, page); - return; - } - - if (em->block_start >= EXTENT_MAP_LAST_BYTE) { - free_extent_map(em); - __unplug_io_fn(bdi, page); - return; - } - offset = offset - em->start; - btrfs_unplug_page(&BTRFS_I(inode)->root->fs_info->mapping_tree, - em->block_start + offset, page); - free_extent_map(em); -} - -/* * If this fails, caller must call bdi_destroy() to get rid of the * bdi again. */ @@ -1416,8 +1390,6 @@ static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi) return err; bdi->ra_pages = default_backing_dev_info.ra_pages; - bdi->unplug_io_fn = btrfs_unplug_io_fn; - bdi->unplug_io_data = info; bdi->congested_fn = btrfs_congested_fn; bdi->congested_data = info; return 0; @@ -1503,6 +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)) { @@ -1592,7 +1565,7 @@ 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), @@ -1604,11 +1577,12 @@ 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) { @@ -1628,6 +1602,8 @@ struct btrfs_root *open_ctree(struct super_block *sb, goto fail_bdi; } + fs_info->btree_inode->i_mapping->flags &= ~__GFP_FS; + INIT_RADIX_TREE(&fs_info->fs_roots_radix, GFP_ATOMIC); INIT_LIST_HEAD(&fs_info->trans_list); INIT_LIST_HEAD(&fs_info->dead_roots); @@ -1641,6 +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; @@ -1663,15 +1640,35 @@ 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); 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); + + 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); @@ -1690,10 +1687,8 @@ 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); - extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree, - GFP_NOFS); + fs_info->btree_inode->i_mapping); + 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; @@ -1707,9 +1702,9 @@ 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; @@ -1739,7 +1734,7 @@ 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)); @@ -1751,17 +1746,23 @@ 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); btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY); + /* + * In the long term, we'll store the compression type in the super + * block, and it'll be used for per file compression control. + */ + fs_info->compress_type = BTRFS_COMPRESS_ZLIB; + ret = btrfs_parse_options(tree_root, options); if (ret) { err = ret; - goto fail_iput; + goto fail_alloc; } features = btrfs_super_incompat_flags(disk_super) & @@ -1771,7 +1772,7 @@ 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); @@ -1787,7 +1788,7 @@ 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, @@ -1834,6 +1835,9 @@ 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 @@ -1855,6 +1859,7 @@ 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, @@ -1963,6 +1968,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->metadata_alloc_profile = (u64)-1; fs_info->system_alloc_profile = fs_info->metadata_alloc_profile; + ret = btrfs_init_space_info(fs_info); + if (ret) { + printk(KERN_ERR "Failed to initial space info: %d\n", ret); + goto fail_block_groups; + } + ret = btrfs_read_block_groups(extent_root); if (ret) { printk(KERN_ERR "Failed to read block groups: %d\n", ret); @@ -2054,9 +2065,14 @@ struct btrfs_root *open_ctree(struct super_block *sb, if (!(sb->s_flags & MS_RDONLY)) { down_read(&fs_info->cleanup_work_sem); - btrfs_orphan_cleanup(fs_info->fs_root); - btrfs_orphan_cleanup(fs_info->tree_root); + err = btrfs_orphan_cleanup(fs_info->fs_root); + if (!err) + err = btrfs_orphan_cleanup(fs_info->tree_root); up_read(&fs_info->cleanup_work_sem); + if (err) { + close_ctree(tree_root); + return ERR_PTR(err); + } } return tree_root; @@ -2100,6 +2116,9 @@ 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); @@ -2127,11 +2146,9 @@ 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 */ @@ -2295,7 +2312,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) mutex_lock(&root->fs_info->fs_devices->device_list_mutex); 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; @@ -2328,7 +2345,7 @@ 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) @@ -2366,12 +2383,15 @@ 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; } 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); @@ -2379,6 +2399,8 @@ 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); } @@ -2431,8 +2453,12 @@ int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info) root_objectid = gang[ret - 1]->root_key.objectid + 1; for (i = 0; i < ret; i++) { + int err; + root_objectid = gang[i]->root_key.objectid; - btrfs_orphan_cleanup(gang[i]); + err = btrfs_orphan_cleanup(gang[i]); + if (err) + return err; } root_objectid++; } @@ -2478,6 +2504,15 @@ int close_ctree(struct btrfs_root *root) fs_info->closing = 1; 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); /* @@ -2489,7 +2524,7 @@ int close_ctree(struct btrfs_root *root) * ERROR state on disk. * * 2. when btrfs flips readonly just in btrfs_commit_super, - * and in such case, btrfs cannnot write sb via btrfs_commit_super, + * and in such case, btrfs cannot write sb via btrfs_commit_super, * and since fs_state has been set BTRFS_SUPER_FLAG_ERROR flag, * btrfs will cleanup all FS resources first and write sb then. */ @@ -2536,6 +2571,7 @@ 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); @@ -2547,6 +2583,7 @@ 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); @@ -2623,6 +2660,29 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) 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; + num_dirty = root->fs_info->dirty_metadata_bytes; if (num_dirty > thresh) { @@ -2655,7 +2715,7 @@ 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; @@ -2782,6 +2842,7 @@ 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; } @@ -2943,7 +3004,10 @@ static int btrfs_destroy_pinned_extent(struct btrfs_root *root, break; /* opt_discard */ - ret = btrfs_error_discard_extent(root, start, end + 1 - start); + if (btrfs_test_opt(root, DISCARD)) + ret = btrfs_error_discard_extent(root, start, + end + 1 - start, + NULL); clear_extent_dirty(unpin, start, end, GFP_NOFS); btrfs_error_unpin_extent_range(root, start, end); @@ -3012,7 +3076,7 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) btrfs_destroy_pinned_extent(root, root->fs_info->pinned_extents); - t->use_count = 0; + atomic_set(&t->use_count, 0); list_del_init(&t->list); memset(t, 0, sizeof(*t)); kmem_cache_free(btrfs_transaction_cachep, t); diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index 07b20dc2fd95..a0b610a67aae 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -55,35 +55,20 @@ int btrfs_commit_super(struct btrfs_root *root); int btrfs_error_commit_super(struct btrfs_root *root); struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize); -struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info, - u64 root_objectid); -struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info, - struct btrfs_key *location, - const char *name, int namelen); struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, struct btrfs_key *location); struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info, struct btrfs_key *location); int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info); -int btrfs_insert_dev_radix(struct btrfs_root *root, - struct block_device *bdev, - u64 device_id, - u64 block_start, - u64 num_blocks); void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); +void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root); void btrfs_mark_buffer_dirty(struct extent_buffer *buf); -void btrfs_mark_buffer_dirty_nonblocking(struct extent_buffer *buf); int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid); int btrfs_set_buffer_uptodate(struct extent_buffer *buf); -int wait_on_tree_block_writeback(struct btrfs_root *root, - struct extent_buffer *buf); int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid); u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len); void btrfs_csum_final(u32 crc, char *result); -int btrfs_open_device(struct btrfs_device *dev); -int btrfs_verify_block_csum(struct btrfs_root *root, - struct extent_buffer *buf); int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio, int metadata); int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode, @@ -91,8 +76,6 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode, unsigned long bio_flags, u64 bio_offset, extent_submit_bio_hook_t *submit_bio_start, extent_submit_bio_hook_t *submit_bio_done); - -int btrfs_congested_async(struct btrfs_fs_info *info, int iodone); unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info); int btrfs_write_tree_block(struct extent_buffer *buf); int btrfs_wait_tree_block_writeback(struct extent_buffer *buf); diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c index ff27d7a477b2..1b8dc33778f9 100644 --- a/fs/btrfs/export.c +++ b/fs/btrfs/export.c @@ -21,14 +21,18 @@ static int btrfs_encode_fh(struct dentry *dentry, u32 *fh, int *max_len, int len = *max_len; int type; - if ((len < BTRFS_FID_SIZE_NON_CONNECTABLE) || - (connectable && len < BTRFS_FID_SIZE_CONNECTABLE)) + if (connectable && (len < BTRFS_FID_SIZE_CONNECTABLE)) { + *max_len = BTRFS_FID_SIZE_CONNECTABLE; return 255; + } else if (len < BTRFS_FID_SIZE_NON_CONNECTABLE) { + *max_len = BTRFS_FID_SIZE_NON_CONNECTABLE; + return 255; + } len = BTRFS_FID_SIZE_NON_CONNECTABLE; type = FILEID_BTRFS_WITHOUT_PARENT; - fid->objectid = inode->i_ino; + fid->objectid = btrfs_ino(inode); fid->root_objectid = BTRFS_I(inode)->root->objectid; fid->gen = inode->i_generation; @@ -174,13 +178,13 @@ static struct dentry *btrfs_get_parent(struct dentry *child) if (!path) return ERR_PTR(-ENOMEM); - if (dir->i_ino == BTRFS_FIRST_FREE_OBJECTID) { + if (btrfs_ino(dir) == BTRFS_FIRST_FREE_OBJECTID) { key.objectid = root->root_key.objectid; key.type = BTRFS_ROOT_BACKREF_KEY; key.offset = (u64)-1; root = root->fs_info->tree_root; } else { - key.objectid = dir->i_ino; + key.objectid = btrfs_ino(dir); key.type = BTRFS_INODE_REF_KEY; key.offset = (u64)-1; } @@ -240,6 +244,7 @@ static int btrfs_get_name(struct dentry *parent, char *name, struct btrfs_key key; int name_len; int ret; + u64 ino; if (!dir || !inode) return -EINVAL; @@ -247,19 +252,21 @@ static int btrfs_get_name(struct dentry *parent, char *name, if (!S_ISDIR(dir->i_mode)) return -EINVAL; + ino = btrfs_ino(inode); + path = btrfs_alloc_path(); if (!path) return -ENOMEM; path->leave_spinning = 1; - if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) { + if (ino == BTRFS_FIRST_FREE_OBJECTID) { key.objectid = BTRFS_I(inode)->root->root_key.objectid; key.type = BTRFS_ROOT_BACKREF_KEY; key.offset = (u64)-1; root = root->fs_info->tree_root; } else { - key.objectid = inode->i_ino; - key.offset = dir->i_ino; + key.objectid = ino; + key.offset = btrfs_ino(dir); key.type = BTRFS_INODE_REF_KEY; } @@ -268,7 +275,7 @@ static int btrfs_get_name(struct dentry *parent, char *name, btrfs_free_path(path); return ret; } else if (ret > 0) { - if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) { + if (ino == BTRFS_FIRST_FREE_OBJECTID) { path->slots[0]--; } else { btrfs_free_path(path); @@ -277,11 +284,11 @@ static int btrfs_get_name(struct dentry *parent, char *name, } leaf = path->nodes[0]; - if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) { - rref = btrfs_item_ptr(leaf, path->slots[0], + if (ino == BTRFS_FIRST_FREE_OBJECTID) { + rref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); - name_ptr = (unsigned long)(rref + 1); - name_len = btrfs_root_ref_name_len(leaf, rref); + name_ptr = (unsigned long)(rref + 1); + name_len = btrfs_root_ref_name_len(leaf, rref); } else { iref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_ref); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 4e7e012ad667..169bd62ce776 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -33,11 +33,28 @@ #include "locking.h" #include "free-space-cache.h" +/* control flags for do_chunk_alloc's force field + * CHUNK_ALLOC_NO_FORCE means to only allocate a chunk + * if we really need one. + * + * CHUNK_ALLOC_FORCE means it must try to allocate one + * + * CHUNK_ALLOC_LIMITED means to only try and allocate one + * if we have very few chunks already allocated. This is + * used as part of the clustering code to help make sure + * we have a good pool of storage to cluster in, without + * filling the FS with empty chunks + * + */ +enum { + CHUNK_ALLOC_NO_FORCE = 0, + CHUNK_ALLOC_FORCE = 1, + CHUNK_ALLOC_LIMITED = 2, +}; + static int update_block_group(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, int alloc); -static int update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo); static int __btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, u64 parent, @@ -77,7 +94,7 @@ static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits) return (cache->flags & bits) == bits; } -void btrfs_get_block_group(struct btrfs_block_group_cache *cache) +static void btrfs_get_block_group(struct btrfs_block_group_cache *cache) { atomic_inc(&cache->count); } @@ -88,6 +105,7 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache) WARN_ON(cache->pinned > 0); WARN_ON(cache->reserved > 0); WARN_ON(cache->reserved_pinned > 0); + kfree(cache->free_space_ctl); kfree(cache); } } @@ -362,7 +380,7 @@ again: break; caching_ctl->progress = last; - btrfs_release_path(extent_root, path); + btrfs_release_path(path); up_read(&fs_info->extent_commit_sem); mutex_unlock(&caching_ctl->mutex); if (btrfs_transaction_in_commit(fs_info)) @@ -442,7 +460,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, * allocate blocks for the tree root we can't do the fast caching since * we likely hold important locks. */ - if (!trans->transaction->in_commit && + if (trans && (!trans->transaction->in_commit) && (root && root != root->fs_info->tree_root)) { spin_lock(&cache->lock); if (cache->cached != BTRFS_CACHE_NO) { @@ -471,7 +489,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, if (load_cache_only) return 0; - caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_KERNEL); + caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); BUG_ON(!caching_ctl); INIT_LIST_HEAD(&caching_ctl->list); @@ -737,8 +755,12 @@ again: atomic_inc(&head->node.refs); spin_unlock(&delayed_refs->lock); - btrfs_release_path(root->fs_info->extent_root, path); + btrfs_release_path(path); + /* + * Mutex was contended, block until it's released and try + * again + */ mutex_lock(&head->mutex); mutex_unlock(&head->mutex); btrfs_put_delayed_ref(&head->node); @@ -917,7 +939,7 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans, break; } } - btrfs_release_path(root, path); + btrfs_release_path(path); if (owner < BTRFS_FIRST_FREE_OBJECTID) new_size += sizeof(*bi); @@ -930,7 +952,6 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans, BUG_ON(ret); ret = btrfs_extend_item(trans, root, path, new_size); - BUG_ON(ret); leaf = path->nodes[0]; item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); @@ -1025,7 +1046,7 @@ again: return 0; #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 key.type = BTRFS_EXTENT_REF_V0_KEY; - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret < 0) { err = ret; @@ -1063,7 +1084,7 @@ again: if (match_extent_data_ref(leaf, ref, root_objectid, owner, offset)) { if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } err = 0; @@ -1124,7 +1145,7 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans, if (match_extent_data_ref(leaf, ref, root_objectid, owner, offset)) break; - btrfs_release_path(root, path); + btrfs_release_path(path); key.offset++; ret = btrfs_insert_empty_item(trans, root, path, &key, size); @@ -1150,7 +1171,7 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(leaf); ret = 0; fail: - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } @@ -1276,7 +1297,7 @@ static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans, ret = -ENOENT; #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 if (ret == -ENOENT && parent) { - btrfs_release_path(root, path); + btrfs_release_path(path); key.type = BTRFS_EXTENT_REF_V0_KEY; ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret > 0) @@ -1305,7 +1326,7 @@ static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans, } ret = btrfs_insert_empty_item(trans, root, path, &key, 0); - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } @@ -1538,7 +1559,6 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans, size = btrfs_extent_inline_ref_size(type); ret = btrfs_extend_item(trans, root, path, size); - BUG_ON(ret); ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); refs = btrfs_extent_refs(leaf, ei); @@ -1591,7 +1611,7 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans, if (ret != -ENOENT) return ret; - btrfs_release_path(root, path); + btrfs_release_path(path); *ref_ret = NULL; if (owner < BTRFS_FIRST_FREE_OBJECTID) { @@ -1667,7 +1687,6 @@ int update_inline_extent_backref(struct btrfs_trans_handle *trans, end - ptr - size); item_size -= size; ret = btrfs_truncate_item(trans, root, path, item_size, 1); - BUG_ON(ret); } btrfs_mark_buffer_dirty(leaf); return 0; @@ -1740,39 +1759,45 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, return ret; } -static void btrfs_issue_discard(struct block_device *bdev, +static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len) { - blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL, 0); + return blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0); } static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, - u64 num_bytes) + u64 num_bytes, u64 *actual_bytes) { int ret; - u64 map_length = num_bytes; + u64 discarded_bytes = 0; struct btrfs_multi_bio *multi = NULL; - if (!btrfs_test_opt(root, DISCARD)) - return 0; /* Tell the block device(s) that the sectors can be discarded */ - ret = btrfs_map_block(&root->fs_info->mapping_tree, READ, - bytenr, &map_length, &multi, 0); + ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD, + bytenr, &num_bytes, &multi, 0); if (!ret) { struct btrfs_bio_stripe *stripe = multi->stripes; int i; - if (map_length > num_bytes) - map_length = num_bytes; for (i = 0; i < multi->num_stripes; i++, stripe++) { - btrfs_issue_discard(stripe->dev->bdev, - stripe->physical, - map_length); + ret = btrfs_issue_discard(stripe->dev->bdev, + stripe->physical, + stripe->length); + if (!ret) + discarded_bytes += stripe->length; + else if (ret != -EOPNOTSUPP) + break; } kfree(multi); } + if (discarded_bytes && ret == -EOPNOTSUPP) + ret = 0; + + if (actual_bytes) + *actual_bytes = discarded_bytes; + return ret; } @@ -1839,7 +1864,7 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, __run_delayed_extent_op(extent_op, leaf, item); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root->fs_info->extent_root, path); + btrfs_release_path(path); path->reada = 1; path->leave_spinning = 1; @@ -2274,6 +2299,10 @@ again: atomic_inc(&ref->refs); spin_unlock(&delayed_refs->lock); + /* + * Mutex was contended, block until it's + * released and try again + */ mutex_lock(&head->mutex); mutex_unlock(&head->mutex); @@ -2338,8 +2367,12 @@ static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, atomic_inc(&head->node.refs); spin_unlock(&delayed_refs->lock); - btrfs_release_path(root->fs_info->extent_root, path); + btrfs_release_path(path); + /* + * Mutex was contended, block until it's released and let + * caller try again + */ mutex_lock(&head->mutex); mutex_unlock(&head->mutex); btrfs_put_delayed_ref(&head->node); @@ -2487,126 +2520,6 @@ out: return ret; } -#if 0 -int btrfs_cache_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct extent_buffer *buf, u32 nr_extents) -{ - struct btrfs_key key; - struct btrfs_file_extent_item *fi; - u64 root_gen; - u32 nritems; - int i; - int level; - int ret = 0; - int shared = 0; - - if (!root->ref_cows) - return 0; - - if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) { - shared = 0; - root_gen = root->root_key.offset; - } else { - shared = 1; - root_gen = trans->transid - 1; - } - - level = btrfs_header_level(buf); - nritems = btrfs_header_nritems(buf); - - if (level == 0) { - struct btrfs_leaf_ref *ref; - struct btrfs_extent_info *info; - - ref = btrfs_alloc_leaf_ref(root, nr_extents); - if (!ref) { - ret = -ENOMEM; - goto out; - } - - ref->root_gen = root_gen; - ref->bytenr = buf->start; - ref->owner = btrfs_header_owner(buf); - ref->generation = btrfs_header_generation(buf); - ref->nritems = nr_extents; - info = ref->extents; - - for (i = 0; nr_extents > 0 && i < nritems; i++) { - u64 disk_bytenr; - btrfs_item_key_to_cpu(buf, &key, i); - if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY) - continue; - fi = btrfs_item_ptr(buf, i, - struct btrfs_file_extent_item); - if (btrfs_file_extent_type(buf, fi) == - BTRFS_FILE_EXTENT_INLINE) - continue; - disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi); - if (disk_bytenr == 0) - continue; - - info->bytenr = disk_bytenr; - info->num_bytes = - btrfs_file_extent_disk_num_bytes(buf, fi); - info->objectid = key.objectid; - info->offset = key.offset; - info++; - } - - ret = btrfs_add_leaf_ref(root, ref, shared); - if (ret == -EEXIST && shared) { - struct btrfs_leaf_ref *old; - old = btrfs_lookup_leaf_ref(root, ref->bytenr); - BUG_ON(!old); - btrfs_remove_leaf_ref(root, old); - btrfs_free_leaf_ref(root, old); - ret = btrfs_add_leaf_ref(root, ref, shared); - } - WARN_ON(ret); - btrfs_free_leaf_ref(root, ref); - } -out: - return ret; -} - -/* when a block goes through cow, we update the reference counts of - * everything that block points to. The internal pointers of the block - * can be in just about any order, and it is likely to have clusters of - * things that are close together and clusters of things that are not. - * - * To help reduce the seeks that come with updating all of these reference - * counts, sort them by byte number before actual updates are done. - * - * struct refsort is used to match byte number to slot in the btree block. - * we sort based on the byte number and then use the slot to actually - * find the item. - * - * struct refsort is smaller than strcut btrfs_item and smaller than - * struct btrfs_key_ptr. Since we're currently limited to the page size - * for a btree block, there's no way for a kmalloc of refsorts for a - * single node to be bigger than a page. - */ -struct refsort { - u64 bytenr; - u32 slot; -}; - -/* - * for passing into sort() - */ -static int refsort_cmp(const void *a_void, const void *b_void) -{ - const struct refsort *a = a_void; - const struct refsort *b = b_void; - - if (a->bytenr < b->bytenr) - return -1; - if (a->bytenr > b->bytenr) - return 1; - return 0; -} -#endif - static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf, @@ -2709,7 +2622,7 @@ static int write_one_cache_group(struct btrfs_trans_handle *trans, bi = btrfs_item_ptr_offset(leaf, path->slots[0]); write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item)); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); fail: if (ret) return ret; @@ -2762,7 +2675,7 @@ again: inode = lookup_free_space_inode(root, block_group, path); if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { ret = PTR_ERR(inode); - btrfs_release_path(root, path); + btrfs_release_path(path); goto out; } @@ -2831,7 +2744,7 @@ again: out_put: iput(inode); out_free: - btrfs_release_path(root, path); + btrfs_release_path(path); out: spin_lock(&block_group->lock); block_group->disk_cache_state = dcs; @@ -3015,7 +2928,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->bytes_readonly = 0; found->bytes_may_use = 0; found->full = 0; - found->force_alloc = 0; + found->force_alloc = CHUNK_ALLOC_NO_FORCE; + found->chunk_alloc = 0; *space_info = found; list_add_rcu(&found->list, &info->space_info); atomic_set(&found->caching_threads, 0); @@ -3120,7 +3034,8 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes) /* make sure bytes are sectorsize aligned */ bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1); - if (root == root->fs_info->tree_root) { + if (root == root->fs_info->tree_root || + BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) { alloc_chunk = 0; committed = 1; } @@ -3146,7 +3061,7 @@ again: if (!data_sinfo->full && alloc_chunk) { u64 alloc_target; - data_sinfo->force_alloc = 1; + data_sinfo->force_alloc = CHUNK_ALLOC_FORCE; spin_unlock(&data_sinfo->lock); alloc: alloc_target = btrfs_get_alloc_profile(root, 1); @@ -3156,7 +3071,8 @@ alloc: ret = do_chunk_alloc(trans, root->fs_info->extent_root, bytes + 2 * 1024 * 1024, - alloc_target, 0); + alloc_target, + CHUNK_ALLOC_NO_FORCE); btrfs_end_transaction(trans, root); if (ret < 0) { if (ret != -ENOSPC) @@ -3186,18 +3102,6 @@ commit_trans: goto again; } -#if 0 /* I hope we never need this code again, just in case */ - printk(KERN_ERR "no space left, need %llu, %llu bytes_used, " - "%llu bytes_reserved, " "%llu bytes_pinned, " - "%llu bytes_readonly, %llu may use %llu total\n", - (unsigned long long)bytes, - (unsigned long long)data_sinfo->bytes_used, - (unsigned long long)data_sinfo->bytes_reserved, - (unsigned long long)data_sinfo->bytes_pinned, - (unsigned long long)data_sinfo->bytes_readonly, - (unsigned long long)data_sinfo->bytes_may_use, - (unsigned long long)data_sinfo->total_bytes); -#endif return -ENOSPC; } data_sinfo->bytes_may_use += bytes; @@ -3235,31 +3139,56 @@ static void force_metadata_allocation(struct btrfs_fs_info *info) rcu_read_lock(); list_for_each_entry_rcu(found, head, list) { if (found->flags & BTRFS_BLOCK_GROUP_METADATA) - found->force_alloc = 1; + found->force_alloc = CHUNK_ALLOC_FORCE; } rcu_read_unlock(); } static int should_alloc_chunk(struct btrfs_root *root, - struct btrfs_space_info *sinfo, u64 alloc_bytes) + struct btrfs_space_info *sinfo, u64 alloc_bytes, + int force) { u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly; + u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved; u64 thresh; - if (sinfo->bytes_used + sinfo->bytes_reserved + - alloc_bytes + 256 * 1024 * 1024 < num_bytes) + if (force == CHUNK_ALLOC_FORCE) + return 1; + + /* + * in limited mode, we want to have some free space up to + * about 1% of the FS size. + */ + if (force == CHUNK_ALLOC_LIMITED) { + thresh = btrfs_super_total_bytes(&root->fs_info->super_copy); + thresh = max_t(u64, 64 * 1024 * 1024, + div_factor_fine(thresh, 1)); + + if (num_bytes - num_allocated < thresh) + return 1; + } + + /* + * we have two similar checks here, one based on percentage + * and once based on a hard number of 256MB. The idea + * is that if we have a good amount of free + * room, don't allocate a chunk. A good mount is + * less than 80% utilized of the chunks we have allocated, + * or more than 256MB free + */ + if (num_allocated + alloc_bytes + 256 * 1024 * 1024 < num_bytes) return 0; - if (sinfo->bytes_used + sinfo->bytes_reserved + - alloc_bytes < div_factor(num_bytes, 8)) + if (num_allocated + alloc_bytes < div_factor(num_bytes, 8)) return 0; thresh = btrfs_super_total_bytes(&root->fs_info->super_copy); + + /* 256MB or 5% of the FS */ thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 5)); if (num_bytes > thresh && sinfo->bytes_used < div_factor(num_bytes, 3)) return 0; - return 1; } @@ -3269,10 +3198,9 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, { struct btrfs_space_info *space_info; struct btrfs_fs_info *fs_info = extent_root->fs_info; + int wait_for_alloc = 0; int ret = 0; - mutex_lock(&fs_info->chunk_mutex); - flags = btrfs_reduce_alloc_profile(extent_root, flags); space_info = __find_space_info(extent_root->fs_info, flags); @@ -3283,21 +3211,40 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, } BUG_ON(!space_info); +again: spin_lock(&space_info->lock); if (space_info->force_alloc) - force = 1; + force = space_info->force_alloc; if (space_info->full) { spin_unlock(&space_info->lock); - goto out; + return 0; } - if (!force && !should_alloc_chunk(extent_root, space_info, - alloc_bytes)) { + if (!should_alloc_chunk(extent_root, space_info, alloc_bytes, force)) { spin_unlock(&space_info->lock); - goto out; + return 0; + } else if (space_info->chunk_alloc) { + wait_for_alloc = 1; + } else { + space_info->chunk_alloc = 1; } + spin_unlock(&space_info->lock); + mutex_lock(&fs_info->chunk_mutex); + + /* + * The chunk_mutex is held throughout the entirety of a chunk + * allocation, so once we've acquired the chunk_mutex we know that the + * other guy is done and we need to recheck and see if we should + * allocate. + */ + if (wait_for_alloc) { + mutex_unlock(&fs_info->chunk_mutex); + wait_for_alloc = 0; + goto again; + } + /* * If we have mixed data/metadata chunks we want to make sure we keep * allocating mixed chunks instead of individual chunks. @@ -3323,9 +3270,10 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, space_info->full = 1; else ret = 1; - space_info->force_alloc = 0; + + space_info->force_alloc = CHUNK_ALLOC_NO_FORCE; + space_info->chunk_alloc = 0; spin_unlock(&space_info->lock); -out: mutex_unlock(&extent_root->fs_info->chunk_mutex); return ret; } @@ -3342,19 +3290,24 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans, u64 max_reclaim; u64 reclaimed = 0; long time_left; - int pause = 1; int nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT; int loops = 0; + unsigned long progress; block_rsv = &root->fs_info->delalloc_block_rsv; space_info = block_rsv->space_info; smp_mb(); reserved = space_info->bytes_reserved; + progress = space_info->reservation_progress; if (reserved == 0) return 0; + /* nothing to shrink - nothing to reclaim */ + if (root->fs_info->delalloc_bytes == 0) + return 0; + max_reclaim = min(reserved, to_reclaim); while (loops < 1024) { @@ -3365,31 +3318,36 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans, writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages); spin_lock(&space_info->lock); - if (reserved > space_info->bytes_reserved) { - loops = 0; + if (reserved > space_info->bytes_reserved) reclaimed += reserved - space_info->bytes_reserved; - } else { - loops++; - } reserved = space_info->bytes_reserved; spin_unlock(&space_info->lock); + loops++; + if (reserved == 0 || reclaimed >= max_reclaim) break; if (trans && trans->transaction->blocked) return -EAGAIN; - __set_current_state(TASK_INTERRUPTIBLE); - time_left = schedule_timeout(pause); + time_left = schedule_timeout_interruptible(1); /* We were interrupted, exit */ if (time_left) break; - pause <<= 1; - if (pause > HZ / 10) - pause = HZ / 10; + /* we've kicked the IO a few times, if anything has been freed, + * exit. There is no sense in looping here for a long time + * when we really need to commit the transaction, or there are + * just too many writers without enough free space + */ + + if (loops > 3) { + smp_mb(); + if (progress != space_info->reservation_progress) + break; + } } return reclaimed >= to_reclaim; @@ -3576,8 +3534,8 @@ static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv, spin_unlock(&block_rsv->lock); } -void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, - struct btrfs_block_rsv *dest, u64 num_bytes) +static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, + struct btrfs_block_rsv *dest, u64 num_bytes) { struct btrfs_space_info *space_info = block_rsv->space_info; @@ -3612,6 +3570,7 @@ void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, if (num_bytes) { spin_lock(&space_info->lock); space_info->bytes_reserved -= num_bytes; + space_info->reservation_progress++; spin_unlock(&space_info->lock); } } @@ -3779,23 +3738,7 @@ static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) u64 meta_used; u64 data_used; int csum_size = btrfs_super_csum_size(&fs_info->super_copy); -#if 0 - /* - * per tree used space accounting can be inaccuracy, so we - * can't rely on it. - */ - spin_lock(&fs_info->extent_root->accounting_lock); - num_bytes = btrfs_root_used(&fs_info->extent_root->root_item); - spin_unlock(&fs_info->extent_root->accounting_lock); - - spin_lock(&fs_info->csum_root->accounting_lock); - num_bytes += btrfs_root_used(&fs_info->csum_root->root_item); - spin_unlock(&fs_info->csum_root->accounting_lock); - spin_lock(&fs_info->tree_root->accounting_lock); - num_bytes += btrfs_root_used(&fs_info->tree_root->root_item); - spin_unlock(&fs_info->tree_root->accounting_lock); -#endif sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); spin_lock(&sinfo->lock); data_used = sinfo->bytes_used; @@ -3844,13 +3787,11 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info) if (block_rsv->reserved >= block_rsv->size) { num_bytes = block_rsv->reserved - block_rsv->size; sinfo->bytes_reserved -= num_bytes; + sinfo->reservation_progress++; block_rsv->reserved = block_rsv->size; block_rsv->full = 1; } -#if 0 - printk(KERN_INFO"global block rsv size %llu reserved %llu\n", - block_rsv->size, block_rsv->reserved); -#endif + spin_unlock(&sinfo->lock); spin_unlock(&block_rsv->lock); } @@ -3896,12 +3837,6 @@ static void release_global_block_rsv(struct btrfs_fs_info *fs_info) WARN_ON(fs_info->chunk_block_rsv.reserved > 0); } -static u64 calc_trans_metadata_size(struct btrfs_root *root, int num_items) -{ - return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) * - 3 * num_items; -} - int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans, struct btrfs_root *root, int num_items) @@ -3912,7 +3847,7 @@ int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans, if (num_items == 0 || root->fs_info->chunk_root == root) return 0; - num_bytes = calc_trans_metadata_size(root, num_items); + num_bytes = btrfs_calc_trans_metadata_size(root, num_items); ret = btrfs_block_rsv_add(trans, root, &root->fs_info->trans_block_rsv, num_bytes); if (!ret) { @@ -3951,14 +3886,14 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, * If all of the metadata space is used, we can commit * transaction and use space it freed. */ - u64 num_bytes = calc_trans_metadata_size(root, 4); + u64 num_bytes = btrfs_calc_trans_metadata_size(root, 4); return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); } void btrfs_orphan_release_metadata(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; - u64 num_bytes = calc_trans_metadata_size(root, 4); + u64 num_bytes = btrfs_calc_trans_metadata_size(root, 4); btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes); } @@ -3972,7 +3907,7 @@ int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans, * two for root back/forward refs, two for directory entries * and one for root of the snapshot. */ - u64 num_bytes = calc_trans_metadata_size(root, 5); + u64 num_bytes = btrfs_calc_trans_metadata_size(root, 5); dst_rsv->space_info = src_rsv->space_info; return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); } @@ -3988,6 +3923,7 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv; u64 to_reserve; int nr_extents; + int reserved_extents; int ret; if (btrfs_transaction_in_commit(root->fs_info)) @@ -3995,26 +3931,24 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) num_bytes = ALIGN(num_bytes, root->sectorsize); - spin_lock(&BTRFS_I(inode)->accounting_lock); nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents) + 1; - if (nr_extents > BTRFS_I(inode)->reserved_extents) { - nr_extents -= BTRFS_I(inode)->reserved_extents; - to_reserve = calc_trans_metadata_size(root, nr_extents); + reserved_extents = atomic_read(&BTRFS_I(inode)->reserved_extents); + + if (nr_extents > reserved_extents) { + nr_extents -= reserved_extents; + to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); } else { nr_extents = 0; to_reserve = 0; } - spin_unlock(&BTRFS_I(inode)->accounting_lock); to_reserve += calc_csum_metadata_size(inode, num_bytes); ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1); if (ret) return ret; - spin_lock(&BTRFS_I(inode)->accounting_lock); - BTRFS_I(inode)->reserved_extents += nr_extents; + atomic_add(nr_extents, &BTRFS_I(inode)->reserved_extents); atomic_inc(&BTRFS_I(inode)->outstanding_extents); - spin_unlock(&BTRFS_I(inode)->accounting_lock); block_rsv_add_bytes(block_rsv, to_reserve, 1); @@ -4029,24 +3963,34 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) struct btrfs_root *root = BTRFS_I(inode)->root; u64 to_free; int nr_extents; + int reserved_extents; num_bytes = ALIGN(num_bytes, root->sectorsize); atomic_dec(&BTRFS_I(inode)->outstanding_extents); WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents) < 0); - spin_lock(&BTRFS_I(inode)->accounting_lock); - nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents); - if (nr_extents < BTRFS_I(inode)->reserved_extents) { - nr_extents = BTRFS_I(inode)->reserved_extents - nr_extents; - BTRFS_I(inode)->reserved_extents -= nr_extents; - } else { - nr_extents = 0; - } - spin_unlock(&BTRFS_I(inode)->accounting_lock); + reserved_extents = atomic_read(&BTRFS_I(inode)->reserved_extents); + do { + int old, new; + + nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents); + if (nr_extents >= reserved_extents) { + nr_extents = 0; + break; + } + old = reserved_extents; + nr_extents = reserved_extents - nr_extents; + new = reserved_extents - nr_extents; + old = atomic_cmpxchg(&BTRFS_I(inode)->reserved_extents, + reserved_extents, new); + if (likely(old == reserved_extents)) + break; + reserved_extents = old; + } while (1); to_free = calc_csum_metadata_size(inode, num_bytes); if (nr_extents > 0) - to_free += calc_trans_metadata_size(root, nr_extents); + to_free += btrfs_calc_trans_metadata_size(root, nr_extents); btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv, to_free); @@ -4133,6 +4077,7 @@ static int update_block_group(struct btrfs_trans_handle *trans, btrfs_set_block_group_used(&cache->item, old_val); cache->reserved -= num_bytes; cache->space_info->bytes_reserved -= num_bytes; + cache->space_info->reservation_progress++; cache->space_info->bytes_used += num_bytes; cache->space_info->disk_used += num_bytes * factor; spin_unlock(&cache->lock); @@ -4184,6 +4129,7 @@ static int pin_down_extent(struct btrfs_root *root, if (reserved) { cache->reserved -= num_bytes; cache->space_info->bytes_reserved -= num_bytes; + cache->space_info->reservation_progress++; } spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); @@ -4214,8 +4160,8 @@ int btrfs_pin_extent(struct btrfs_root *root, * update size of reserved extents. this function may return -EAGAIN * if 'reserve' is true or 'sinfo' is false. */ -static int update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo) +int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve, int sinfo) { int ret = 0; if (sinfo) { @@ -4234,6 +4180,7 @@ static int 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); @@ -4353,7 +4300,9 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, if (ret) break; - ret = btrfs_discard_extent(root, start, end + 1 - start); + if (btrfs_test_opt(root, DISCARD)) + ret = btrfs_discard_extent(root, start, + end + 1 - start, NULL); clear_extent_dirty(unpin, start, end, GFP_NOFS); unpin_extent_range(root, start, end); @@ -4450,7 +4399,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, NULL, refs_to_drop, is_data); BUG_ON(ret); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); path->leave_spinning = 1; key.objectid = bytenr; @@ -4489,7 +4438,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, owner_objectid, 0); BUG_ON(ret < 0); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); path->leave_spinning = 1; key.objectid = bytenr; @@ -4559,7 +4508,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = btrfs_del_items(trans, extent_root, path, path->slots[0], num_to_del); BUG_ON(ret); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); if (is_data) { ret = btrfs_del_csums(trans, root, bytenr, num_bytes); @@ -4694,10 +4643,10 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)); btrfs_add_free_space(cache, buf->start, buf->len); - ret = update_reserved_bytes(cache, buf->len, 0, 0); + ret = btrfs_update_reserved_bytes(cache, buf->len, 0, 0); if (ret == -EAGAIN) { /* block group became read-only */ - update_reserved_bytes(cache, buf->len, 0, 1); + btrfs_update_reserved_bytes(cache, buf->len, 0, 1); goto out; } @@ -4712,6 +4661,7 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, if (ret) { spin_lock(&cache->space_info->lock); cache->space_info->bytes_reserved -= buf->len; + cache->space_info->reservation_progress++; spin_unlock(&cache->space_info->lock); } goto out; @@ -4733,6 +4683,11 @@ pin: } } out: + /* + * Deleting the buffer, clear the corrupt flag since it doesn't matter + * anymore. + */ + clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags); btrfs_put_block_group(cache); } @@ -4796,7 +4751,7 @@ wait_block_group_cache_progress(struct btrfs_block_group_cache *cache, return 0; wait_event(caching_ctl->wait, block_group_cache_done(cache) || - (cache->free_space >= num_bytes)); + (cache->free_space_ctl->free_space >= num_bytes)); put_caching_control(caching_ctl); return 0; @@ -5180,7 +5135,7 @@ checks: search_start - offset); BUG_ON(offset > search_start); - ret = update_reserved_bytes(block_group, num_bytes, 1, + ret = btrfs_update_reserved_bytes(block_group, num_bytes, 1, (data & BTRFS_BLOCK_GROUP_DATA)); if (ret == -EAGAIN) { btrfs_add_free_space(block_group, offset, num_bytes); @@ -5271,11 +5226,13 @@ loop: if (allowed_chunk_alloc) { ret = do_chunk_alloc(trans, root, num_bytes + - 2 * 1024 * 1024, data, 1); + 2 * 1024 * 1024, data, + CHUNK_ALLOC_LIMITED); allowed_chunk_alloc = 0; done_chunk_alloc = 1; - } else if (!done_chunk_alloc) { - space_info->force_alloc = 1; + } else if (!done_chunk_alloc && + space_info->force_alloc == CHUNK_ALLOC_NO_FORCE) { + space_info->force_alloc = CHUNK_ALLOC_LIMITED; } if (loop < LOOP_NO_EMPTY_SIZE) { @@ -5361,7 +5318,8 @@ again: */ if (empty_size || root->ref_cows) ret = do_chunk_alloc(trans, root->fs_info->extent_root, - num_bytes + 2 * 1024 * 1024, data, 0); + num_bytes + 2 * 1024 * 1024, data, + CHUNK_ALLOC_NO_FORCE); WARN_ON(num_bytes < root->sectorsize); ret = find_free_extent(trans, root, num_bytes, empty_size, @@ -5373,10 +5331,10 @@ again: num_bytes = num_bytes & ~(root->sectorsize - 1); num_bytes = max(num_bytes, min_alloc_size); do_chunk_alloc(trans, root->fs_info->extent_root, - num_bytes, data, 1); + num_bytes, data, CHUNK_ALLOC_FORCE); goto again; } - if (ret == -ENOSPC) { + if (ret == -ENOSPC && btrfs_test_opt(root, ENOSPC_DEBUG)) { struct btrfs_space_info *sinfo; sinfo = __find_space_info(root->fs_info, data); @@ -5386,6 +5344,8 @@ again: dump_space_info(sinfo, num_bytes, 1); } + trace_btrfs_reserved_extent_alloc(root, ins->objectid, ins->offset); + return ret; } @@ -5401,12 +5361,15 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) return -ENOSPC; } - ret = btrfs_discard_extent(root, start, len); + if (btrfs_test_opt(root, DISCARD)) + ret = btrfs_discard_extent(root, start, len, NULL); btrfs_add_free_space(cache, start, len); - update_reserved_bytes(cache, len, 0, 1); + btrfs_update_reserved_bytes(cache, len, 0, 1); btrfs_put_block_group(cache); + trace_btrfs_reserved_extent_free(root, start, len); + return ret; } @@ -5433,7 +5396,8 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; path->leave_spinning = 1; ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, @@ -5603,7 +5567,7 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, put_caching_control(caching_ctl); } - ret = update_reserved_bytes(block_group, ins->offset, 1, 1); + ret = btrfs_update_reserved_bytes(block_group, ins->offset, 1, 1); BUG_ON(ret); btrfs_put_block_group(block_group); ret = alloc_reserved_file_extent(trans, root, 0, root_objectid, @@ -6036,6 +6000,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans, if (reada && level == 1) reada_walk_down(trans, root, wc, path); next = read_tree_block(root, bytenr, blocksize, generation); + if (!next) + return -EIO; btrfs_tree_lock(next); btrfs_set_lock_blocking(next); } @@ -6372,7 +6338,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, trans->block_rsv = block_rsv; } } - btrfs_release_path(root, path); + btrfs_release_path(path); BUG_ON(err); ret = btrfs_del_root(trans, tree_root, &root->root_key); @@ -6427,10 +6393,14 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; wc = kzalloc(sizeof(*wc), GFP_NOFS); - BUG_ON(!wc); + if (!wc) { + btrfs_free_path(path); + return -ENOMEM; + } btrfs_assert_tree_locked(parent); parent_level = btrfs_header_level(parent); @@ -6472,1495 +6442,6 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, return ret; } -#if 0 -static unsigned long calc_ra(unsigned long start, unsigned long last, - unsigned long nr) -{ - return min(last, start + nr - 1); -} - -static noinline int relocate_inode_pages(struct inode *inode, u64 start, - u64 len) -{ - u64 page_start; - u64 page_end; - unsigned long first_index; - unsigned long last_index; - unsigned long i; - struct page *page; - struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; - struct file_ra_state *ra; - struct btrfs_ordered_extent *ordered; - unsigned int total_read = 0; - unsigned int total_dirty = 0; - int ret = 0; - - ra = kzalloc(sizeof(*ra), GFP_NOFS); - if (!ra) - return -ENOMEM; - - mutex_lock(&inode->i_mutex); - first_index = start >> PAGE_CACHE_SHIFT; - last_index = (start + len - 1) >> PAGE_CACHE_SHIFT; - - /* make sure the dirty trick played by the caller work */ - ret = invalidate_inode_pages2_range(inode->i_mapping, - first_index, last_index); - if (ret) - goto out_unlock; - - file_ra_state_init(ra, inode->i_mapping); - - for (i = first_index ; i <= last_index; i++) { - if (total_read % ra->ra_pages == 0) { - btrfs_force_ra(inode->i_mapping, ra, NULL, i, - calc_ra(i, last_index, ra->ra_pages)); - } - total_read++; -again: - if (((u64)i << PAGE_CACHE_SHIFT) > i_size_read(inode)) - BUG_ON(1); - page = grab_cache_page(inode->i_mapping, i); - if (!page) { - ret = -ENOMEM; - goto out_unlock; - } - if (!PageUptodate(page)) { - btrfs_readpage(NULL, page); - lock_page(page); - if (!PageUptodate(page)) { - unlock_page(page); - page_cache_release(page); - ret = -EIO; - goto out_unlock; - } - } - wait_on_page_writeback(page); - - page_start = (u64)page->index << PAGE_CACHE_SHIFT; - page_end = page_start + PAGE_CACHE_SIZE - 1; - lock_extent(io_tree, page_start, page_end, GFP_NOFS); - - ordered = btrfs_lookup_ordered_extent(inode, page_start); - if (ordered) { - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - unlock_page(page); - page_cache_release(page); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - goto again; - } - set_page_extent_mapped(page); - - if (i == first_index) - set_extent_bits(io_tree, page_start, page_end, - EXTENT_BOUNDARY, GFP_NOFS); - btrfs_set_extent_delalloc(inode, page_start, page_end); - - set_page_dirty(page); - total_dirty++; - - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - unlock_page(page); - page_cache_release(page); - } - -out_unlock: - kfree(ra); - mutex_unlock(&inode->i_mutex); - balance_dirty_pages_ratelimited_nr(inode->i_mapping, total_dirty); - return ret; -} - -static noinline int relocate_data_extent(struct inode *reloc_inode, - struct btrfs_key *extent_key, - u64 offset) -{ - struct btrfs_root *root = BTRFS_I(reloc_inode)->root; - struct extent_map_tree *em_tree = &BTRFS_I(reloc_inode)->extent_tree; - struct extent_map *em; - u64 start = extent_key->objectid - offset; - u64 end = start + extent_key->offset - 1; - - em = alloc_extent_map(GFP_NOFS); - BUG_ON(!em || IS_ERR(em)); - - em->start = start; - em->len = extent_key->offset; - em->block_len = extent_key->offset; - em->block_start = extent_key->objectid; - em->bdev = root->fs_info->fs_devices->latest_bdev; - set_bit(EXTENT_FLAG_PINNED, &em->flags); - - /* setup extent map to cheat btrfs_readpage */ - lock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS); - while (1) { - int ret; - write_lock(&em_tree->lock); - ret = add_extent_mapping(em_tree, em); - write_unlock(&em_tree->lock); - if (ret != -EEXIST) { - free_extent_map(em); - break; - } - btrfs_drop_extent_cache(reloc_inode, start, end, 0); - } - unlock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS); - - return relocate_inode_pages(reloc_inode, start, extent_key->offset); -} - -struct btrfs_ref_path { - u64 extent_start; - u64 nodes[BTRFS_MAX_LEVEL]; - u64 root_objectid; - u64 root_generation; - u64 owner_objectid; - u32 num_refs; - int lowest_level; - int current_level; - int shared_level; - - struct btrfs_key node_keys[BTRFS_MAX_LEVEL]; - u64 new_nodes[BTRFS_MAX_LEVEL]; -}; - -struct disk_extent { - u64 ram_bytes; - u64 disk_bytenr; - u64 disk_num_bytes; - u64 offset; - u64 num_bytes; - u8 compression; - u8 encryption; - u16 other_encoding; -}; - -static int is_cowonly_root(u64 root_objectid) -{ - if (root_objectid == BTRFS_ROOT_TREE_OBJECTID || - root_objectid == BTRFS_EXTENT_TREE_OBJECTID || - root_objectid == BTRFS_CHUNK_TREE_OBJECTID || - root_objectid == BTRFS_DEV_TREE_OBJECTID || - root_objectid == BTRFS_TREE_LOG_OBJECTID || - root_objectid == BTRFS_CSUM_TREE_OBJECTID) - return 1; - return 0; -} - -static noinline int __next_ref_path(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_ref_path *ref_path, - int first_time) -{ - struct extent_buffer *leaf; - struct btrfs_path *path; - struct btrfs_extent_ref *ref; - struct btrfs_key key; - struct btrfs_key found_key; - u64 bytenr; - u32 nritems; - int level; - int ret = 1; - - path = btrfs_alloc_path(); - if (!path) - return -ENOMEM; - - if (first_time) { - ref_path->lowest_level = -1; - ref_path->current_level = -1; - ref_path->shared_level = -1; - goto walk_up; - } -walk_down: - level = ref_path->current_level - 1; - while (level >= -1) { - u64 parent; - if (level < ref_path->lowest_level) - break; - - if (level >= 0) - bytenr = ref_path->nodes[level]; - else - bytenr = ref_path->extent_start; - BUG_ON(bytenr == 0); - - parent = ref_path->nodes[level + 1]; - ref_path->nodes[level + 1] = 0; - ref_path->current_level = level; - BUG_ON(parent == 0); - - key.objectid = bytenr; - key.offset = parent + 1; - key.type = BTRFS_EXTENT_REF_KEY; - - ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0); - if (ret < 0) - goto out; - BUG_ON(ret == 0); - - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - if (path->slots[0] >= nritems) { - ret = btrfs_next_leaf(extent_root, path); - if (ret < 0) - goto out; - if (ret > 0) - goto next; - leaf = path->nodes[0]; - } - - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid == bytenr && - found_key.type == BTRFS_EXTENT_REF_KEY) { - if (level < ref_path->shared_level) - ref_path->shared_level = level; - goto found; - } -next: - level--; - btrfs_release_path(extent_root, path); - cond_resched(); - } - /* reached lowest level */ - ret = 1; - goto out; -walk_up: - level = ref_path->current_level; - while (level < BTRFS_MAX_LEVEL - 1) { - u64 ref_objectid; - - if (level >= 0) - bytenr = ref_path->nodes[level]; - else - bytenr = ref_path->extent_start; - - BUG_ON(bytenr == 0); - - key.objectid = bytenr; - key.offset = 0; - key.type = BTRFS_EXTENT_REF_KEY; - - ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0); - if (ret < 0) - goto out; - - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - if (path->slots[0] >= nritems) { - ret = btrfs_next_leaf(extent_root, path); - if (ret < 0) - goto out; - if (ret > 0) { - /* the extent was freed by someone */ - if (ref_path->lowest_level == level) - goto out; - btrfs_release_path(extent_root, path); - goto walk_down; - } - leaf = path->nodes[0]; - } - - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid != bytenr || - found_key.type != BTRFS_EXTENT_REF_KEY) { - /* the extent was freed by someone */ - if (ref_path->lowest_level == level) { - ret = 1; - goto out; - } - btrfs_release_path(extent_root, path); - goto walk_down; - } -found: - ref = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_extent_ref); - ref_objectid = btrfs_ref_objectid(leaf, ref); - if (ref_objectid < BTRFS_FIRST_FREE_OBJECTID) { - if (first_time) { - level = (int)ref_objectid; - BUG_ON(level >= BTRFS_MAX_LEVEL); - ref_path->lowest_level = level; - ref_path->current_level = level; - ref_path->nodes[level] = bytenr; - } else { - WARN_ON(ref_objectid != level); - } - } else { - WARN_ON(level != -1); - } - first_time = 0; - - if (ref_path->lowest_level == level) { - ref_path->owner_objectid = ref_objectid; - ref_path->num_refs = btrfs_ref_num_refs(leaf, ref); - } - - /* - * the block is tree root or the block isn't in reference - * counted tree. - */ - if (found_key.objectid == found_key.offset || - is_cowonly_root(btrfs_ref_root(leaf, ref))) { - ref_path->root_objectid = btrfs_ref_root(leaf, ref); - ref_path->root_generation = - btrfs_ref_generation(leaf, ref); - if (level < 0) { - /* special reference from the tree log */ - ref_path->nodes[0] = found_key.offset; - ref_path->current_level = 0; - } - ret = 0; - goto out; - } - - level++; - BUG_ON(ref_path->nodes[level] != 0); - ref_path->nodes[level] = found_key.offset; - ref_path->current_level = level; - - /* - * the reference was created in the running transaction, - * no need to continue walking up. - */ - if (btrfs_ref_generation(leaf, ref) == trans->transid) { - ref_path->root_objectid = btrfs_ref_root(leaf, ref); - ref_path->root_generation = - btrfs_ref_generation(leaf, ref); - ret = 0; - goto out; - } - - btrfs_release_path(extent_root, path); - cond_resched(); - } - /* reached max tree level, but no tree root found. */ - BUG(); -out: - btrfs_free_path(path); - return ret; -} - -static int btrfs_first_ref_path(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_ref_path *ref_path, - u64 extent_start) -{ - memset(ref_path, 0, sizeof(*ref_path)); - ref_path->extent_start = extent_start; - - return __next_ref_path(trans, extent_root, ref_path, 1); -} - -static int btrfs_next_ref_path(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_ref_path *ref_path) -{ - return __next_ref_path(trans, extent_root, ref_path, 0); -} - -static noinline int get_new_locations(struct inode *reloc_inode, - struct btrfs_key *extent_key, - u64 offset, int no_fragment, - struct disk_extent **extents, - int *nr_extents) -{ - struct btrfs_root *root = BTRFS_I(reloc_inode)->root; - struct btrfs_path *path; - struct btrfs_file_extent_item *fi; - struct extent_buffer *leaf; - struct disk_extent *exts = *extents; - struct btrfs_key found_key; - u64 cur_pos; - u64 last_byte; - u32 nritems; - int nr = 0; - int max = *nr_extents; - int ret; - - WARN_ON(!no_fragment && *extents); - if (!exts) { - max = 1; - exts = kmalloc(sizeof(*exts) * max, GFP_NOFS); - if (!exts) - return -ENOMEM; - } - - path = btrfs_alloc_path(); - BUG_ON(!path); - - cur_pos = extent_key->objectid - offset; - last_byte = extent_key->objectid + extent_key->offset; - ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino, - cur_pos, 0); - if (ret < 0) - goto out; - if (ret > 0) { - ret = -ENOENT; - goto out; - } - - while (1) { - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - if (path->slots[0] >= nritems) { - ret = btrfs_next_leaf(root, path); - if (ret < 0) - goto out; - if (ret > 0) - break; - leaf = path->nodes[0]; - } - - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.offset != cur_pos || - found_key.type != BTRFS_EXTENT_DATA_KEY || - found_key.objectid != reloc_inode->i_ino) - break; - - fi = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_file_extent_item); - if (btrfs_file_extent_type(leaf, fi) != - BTRFS_FILE_EXTENT_REG || - btrfs_file_extent_disk_bytenr(leaf, fi) == 0) - break; - - if (nr == max) { - struct disk_extent *old = exts; - max *= 2; - exts = kzalloc(sizeof(*exts) * max, GFP_NOFS); - memcpy(exts, old, sizeof(*exts) * nr); - if (old != *extents) - kfree(old); - } - - exts[nr].disk_bytenr = - btrfs_file_extent_disk_bytenr(leaf, fi); - exts[nr].disk_num_bytes = - btrfs_file_extent_disk_num_bytes(leaf, fi); - exts[nr].offset = btrfs_file_extent_offset(leaf, fi); - exts[nr].num_bytes = btrfs_file_extent_num_bytes(leaf, fi); - exts[nr].ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi); - exts[nr].compression = btrfs_file_extent_compression(leaf, fi); - exts[nr].encryption = btrfs_file_extent_encryption(leaf, fi); - exts[nr].other_encoding = btrfs_file_extent_other_encoding(leaf, - fi); - BUG_ON(exts[nr].offset > 0); - BUG_ON(exts[nr].compression || exts[nr].encryption); - BUG_ON(exts[nr].num_bytes != exts[nr].disk_num_bytes); - - cur_pos += exts[nr].num_bytes; - nr++; - - if (cur_pos + offset >= last_byte) - break; - - if (no_fragment) { - ret = 1; - goto out; - } - path->slots[0]++; - } - - BUG_ON(cur_pos + offset > last_byte); - if (cur_pos + offset < last_byte) { - ret = -ENOENT; - goto out; - } - ret = 0; -out: - btrfs_free_path(path); - if (ret) { - if (exts != *extents) - kfree(exts); - } else { - *extents = exts; - *nr_extents = nr; - } - return ret; -} - -static noinline int replace_one_extent(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *extent_key, - struct btrfs_key *leaf_key, - struct btrfs_ref_path *ref_path, - struct disk_extent *new_extents, - int nr_extents) -{ - struct extent_buffer *leaf; - struct btrfs_file_extent_item *fi; - struct inode *inode = NULL; - struct btrfs_key key; - u64 lock_start = 0; - u64 lock_end = 0; - u64 num_bytes; - u64 ext_offset; - u64 search_end = (u64)-1; - u32 nritems; - int nr_scaned = 0; - int extent_locked = 0; - int extent_type; - int ret; - - memcpy(&key, leaf_key, sizeof(key)); - if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) { - if (key.objectid < ref_path->owner_objectid || - (key.objectid == ref_path->owner_objectid && - key.type < BTRFS_EXTENT_DATA_KEY)) { - key.objectid = ref_path->owner_objectid; - key.type = BTRFS_EXTENT_DATA_KEY; - key.offset = 0; - } - } - - while (1) { - ret = btrfs_search_slot(trans, root, &key, path, 0, 1); - if (ret < 0) - goto out; - - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); -next: - if (extent_locked && ret > 0) { - /* - * the file extent item was modified by someone - * before the extent got locked. - */ - unlock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - extent_locked = 0; - } - - if (path->slots[0] >= nritems) { - if (++nr_scaned > 2) - break; - - BUG_ON(extent_locked); - ret = btrfs_next_leaf(root, path); - if (ret < 0) - goto out; - if (ret > 0) - break; - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - } - - btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - - if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) { - if ((key.objectid > ref_path->owner_objectid) || - (key.objectid == ref_path->owner_objectid && - key.type > BTRFS_EXTENT_DATA_KEY) || - key.offset >= search_end) - break; - } - - if (inode && key.objectid != inode->i_ino) { - BUG_ON(extent_locked); - btrfs_release_path(root, path); - mutex_unlock(&inode->i_mutex); - iput(inode); - inode = NULL; - continue; - } - - if (key.type != BTRFS_EXTENT_DATA_KEY) { - path->slots[0]++; - ret = 1; - goto next; - } - fi = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_file_extent_item); - extent_type = btrfs_file_extent_type(leaf, fi); - if ((extent_type != BTRFS_FILE_EXTENT_REG && - extent_type != BTRFS_FILE_EXTENT_PREALLOC) || - (btrfs_file_extent_disk_bytenr(leaf, fi) != - extent_key->objectid)) { - path->slots[0]++; - ret = 1; - goto next; - } - - num_bytes = btrfs_file_extent_num_bytes(leaf, fi); - ext_offset = btrfs_file_extent_offset(leaf, fi); - - if (search_end == (u64)-1) { - search_end = key.offset - ext_offset + - btrfs_file_extent_ram_bytes(leaf, fi); - } - - if (!extent_locked) { - lock_start = key.offset; - lock_end = lock_start + num_bytes - 1; - } else { - if (lock_start > key.offset || - lock_end + 1 < key.offset + num_bytes) { - unlock_extent(&BTRFS_I(inode)->io_tree, - lock_start, lock_end, GFP_NOFS); - extent_locked = 0; - } - } - - if (!inode) { - btrfs_release_path(root, path); - - inode = btrfs_iget_locked(root->fs_info->sb, - key.objectid, root); - if (inode->i_state & I_NEW) { - BTRFS_I(inode)->root = root; - BTRFS_I(inode)->location.objectid = - key.objectid; - BTRFS_I(inode)->location.type = - BTRFS_INODE_ITEM_KEY; - BTRFS_I(inode)->location.offset = 0; - btrfs_read_locked_inode(inode); - unlock_new_inode(inode); - } - /* - * some code call btrfs_commit_transaction while - * holding the i_mutex, so we can't use mutex_lock - * here. - */ - if (is_bad_inode(inode) || - !mutex_trylock(&inode->i_mutex)) { - iput(inode); - inode = NULL; - key.offset = (u64)-1; - goto skip; - } - } - - if (!extent_locked) { - struct btrfs_ordered_extent *ordered; - - btrfs_release_path(root, path); - - lock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - ordered = btrfs_lookup_first_ordered_extent(inode, - lock_end); - if (ordered && - ordered->file_offset <= lock_end && - ordered->file_offset + ordered->len > lock_start) { - unlock_extent(&BTRFS_I(inode)->io_tree, - lock_start, lock_end, GFP_NOFS); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - key.offset += num_bytes; - goto skip; - } - if (ordered) - btrfs_put_ordered_extent(ordered); - - extent_locked = 1; - continue; - } - - if (nr_extents == 1) { - /* update extent pointer in place */ - btrfs_set_file_extent_disk_bytenr(leaf, fi, - new_extents[0].disk_bytenr); - btrfs_set_file_extent_disk_num_bytes(leaf, fi, - new_extents[0].disk_num_bytes); - btrfs_mark_buffer_dirty(leaf); - - btrfs_drop_extent_cache(inode, key.offset, - key.offset + num_bytes - 1, 0); - - ret = btrfs_inc_extent_ref(trans, root, - new_extents[0].disk_bytenr, - new_extents[0].disk_num_bytes, - leaf->start, - root->root_key.objectid, - trans->transid, - key.objectid); - BUG_ON(ret); - - ret = btrfs_free_extent(trans, root, - extent_key->objectid, - extent_key->offset, - leaf->start, - btrfs_header_owner(leaf), - btrfs_header_generation(leaf), - key.objectid, 0); - BUG_ON(ret); - - btrfs_release_path(root, path); - key.offset += num_bytes; - } else { - BUG_ON(1); -#if 0 - u64 alloc_hint; - u64 extent_len; - int i; - /* - * drop old extent pointer at first, then insert the - * new pointers one bye one - */ - btrfs_release_path(root, path); - ret = btrfs_drop_extents(trans, root, inode, key.offset, - key.offset + num_bytes, - key.offset, &alloc_hint); - BUG_ON(ret); - - for (i = 0; i < nr_extents; i++) { - if (ext_offset >= new_extents[i].num_bytes) { - ext_offset -= new_extents[i].num_bytes; - continue; - } - extent_len = min(new_extents[i].num_bytes - - ext_offset, num_bytes); - - ret = btrfs_insert_empty_item(trans, root, - path, &key, - sizeof(*fi)); - BUG_ON(ret); - - leaf = path->nodes[0]; - fi = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_file_extent_item); - btrfs_set_file_extent_generation(leaf, fi, - trans->transid); - btrfs_set_file_extent_type(leaf, fi, - BTRFS_FILE_EXTENT_REG); - btrfs_set_file_extent_disk_bytenr(leaf, fi, - new_extents[i].disk_bytenr); - btrfs_set_file_extent_disk_num_bytes(leaf, fi, - new_extents[i].disk_num_bytes); - btrfs_set_file_extent_ram_bytes(leaf, fi, - new_extents[i].ram_bytes); - - btrfs_set_file_extent_compression(leaf, fi, - new_extents[i].compression); - btrfs_set_file_extent_encryption(leaf, fi, - new_extents[i].encryption); - btrfs_set_file_extent_other_encoding(leaf, fi, - new_extents[i].other_encoding); - - btrfs_set_file_extent_num_bytes(leaf, fi, - extent_len); - ext_offset += new_extents[i].offset; - btrfs_set_file_extent_offset(leaf, fi, - ext_offset); - btrfs_mark_buffer_dirty(leaf); - - btrfs_drop_extent_cache(inode, key.offset, - key.offset + extent_len - 1, 0); - - ret = btrfs_inc_extent_ref(trans, root, - new_extents[i].disk_bytenr, - new_extents[i].disk_num_bytes, - leaf->start, - root->root_key.objectid, - trans->transid, key.objectid); - BUG_ON(ret); - btrfs_release_path(root, path); - - inode_add_bytes(inode, extent_len); - - ext_offset = 0; - num_bytes -= extent_len; - key.offset += extent_len; - - if (num_bytes == 0) - break; - } - BUG_ON(i >= nr_extents); -#endif - } - - if (extent_locked) { - unlock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - extent_locked = 0; - } -skip: - if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS && - key.offset >= search_end) - break; - - cond_resched(); - } - ret = 0; -out: - btrfs_release_path(root, path); - if (inode) { - mutex_unlock(&inode->i_mutex); - if (extent_locked) { - unlock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - } - iput(inode); - } - return ret; -} - -int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct extent_buffer *buf, u64 orig_start) -{ - int level; - int ret; - - BUG_ON(btrfs_header_generation(buf) != trans->transid); - BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID); - - level = btrfs_header_level(buf); - if (level == 0) { - struct btrfs_leaf_ref *ref; - struct btrfs_leaf_ref *orig_ref; - - orig_ref = btrfs_lookup_leaf_ref(root, orig_start); - if (!orig_ref) - return -ENOENT; - - ref = btrfs_alloc_leaf_ref(root, orig_ref->nritems); - if (!ref) { - btrfs_free_leaf_ref(root, orig_ref); - return -ENOMEM; - } - - ref->nritems = orig_ref->nritems; - memcpy(ref->extents, orig_ref->extents, - sizeof(ref->extents[0]) * ref->nritems); - - btrfs_free_leaf_ref(root, orig_ref); - - ref->root_gen = trans->transid; - ref->bytenr = buf->start; - ref->owner = btrfs_header_owner(buf); - ref->generation = btrfs_header_generation(buf); - - ret = btrfs_add_leaf_ref(root, ref, 0); - WARN_ON(ret); - btrfs_free_leaf_ref(root, ref); - } - return 0; -} - -static noinline int invalidate_extent_cache(struct btrfs_root *root, - struct extent_buffer *leaf, - struct btrfs_block_group_cache *group, - struct btrfs_root *target_root) -{ - struct btrfs_key key; - struct inode *inode = NULL; - struct btrfs_file_extent_item *fi; - struct extent_state *cached_state = NULL; - u64 num_bytes; - u64 skip_objectid = 0; - u32 nritems; - u32 i; - - nritems = btrfs_header_nritems(leaf); - for (i = 0; i < nritems; i++) { - btrfs_item_key_to_cpu(leaf, &key, i); - if (key.objectid == skip_objectid || - key.type != BTRFS_EXTENT_DATA_KEY) - continue; - fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item); - if (btrfs_file_extent_type(leaf, fi) == - BTRFS_FILE_EXTENT_INLINE) - continue; - if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0) - continue; - if (!inode || inode->i_ino != key.objectid) { - iput(inode); - inode = btrfs_ilookup(target_root->fs_info->sb, - key.objectid, target_root, 1); - } - if (!inode) { - skip_objectid = key.objectid; - continue; - } - num_bytes = btrfs_file_extent_num_bytes(leaf, fi); - - lock_extent_bits(&BTRFS_I(inode)->io_tree, key.offset, - key.offset + num_bytes - 1, 0, &cached_state, - GFP_NOFS); - btrfs_drop_extent_cache(inode, key.offset, - key.offset + num_bytes - 1, 1); - unlock_extent_cached(&BTRFS_I(inode)->io_tree, key.offset, - key.offset + num_bytes - 1, &cached_state, - GFP_NOFS); - cond_resched(); - } - iput(inode); - return 0; -} - -static noinline int replace_extents_in_leaf(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct extent_buffer *leaf, - struct btrfs_block_group_cache *group, - struct inode *reloc_inode) -{ - struct btrfs_key key; - struct btrfs_key extent_key; - struct btrfs_file_extent_item *fi; - struct btrfs_leaf_ref *ref; - struct disk_extent *new_extent; - u64 bytenr; - u64 num_bytes; - u32 nritems; - u32 i; - int ext_index; - int nr_extent; - int ret; - - new_extent = kmalloc(sizeof(*new_extent), GFP_NOFS); - BUG_ON(!new_extent); - - ref = btrfs_lookup_leaf_ref(root, leaf->start); - BUG_ON(!ref); - - ext_index = -1; - nritems = btrfs_header_nritems(leaf); - for (i = 0; i < nritems; i++) { - btrfs_item_key_to_cpu(leaf, &key, i); - if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY) - continue; - fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item); - if (btrfs_file_extent_type(leaf, fi) == - BTRFS_FILE_EXTENT_INLINE) - continue; - bytenr = btrfs_file_extent_disk_bytenr(leaf, fi); - num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi); - if (bytenr == 0) - continue; - - ext_index++; - if (bytenr >= group->key.objectid + group->key.offset || - bytenr + num_bytes <= group->key.objectid) - continue; - - extent_key.objectid = bytenr; - extent_key.offset = num_bytes; - extent_key.type = BTRFS_EXTENT_ITEM_KEY; - nr_extent = 1; - ret = get_new_locations(reloc_inode, &extent_key, - group->key.objectid, 1, - &new_extent, &nr_extent); - if (ret > 0) - continue; - BUG_ON(ret < 0); - - BUG_ON(ref->extents[ext_index].bytenr != bytenr); - BUG_ON(ref->extents[ext_index].num_bytes != num_bytes); - ref->extents[ext_index].bytenr = new_extent->disk_bytenr; - ref->extents[ext_index].num_bytes = new_extent->disk_num_bytes; - - btrfs_set_file_extent_disk_bytenr(leaf, fi, - new_extent->disk_bytenr); - btrfs_set_file_extent_disk_num_bytes(leaf, fi, - new_extent->disk_num_bytes); - btrfs_mark_buffer_dirty(leaf); - - ret = btrfs_inc_extent_ref(trans, root, - new_extent->disk_bytenr, - new_extent->disk_num_bytes, - leaf->start, - root->root_key.objectid, - trans->transid, key.objectid); - BUG_ON(ret); - - ret = btrfs_free_extent(trans, root, - bytenr, num_bytes, leaf->start, - btrfs_header_owner(leaf), - btrfs_header_generation(leaf), - key.objectid, 0); - BUG_ON(ret); - cond_resched(); - } - kfree(new_extent); - BUG_ON(ext_index + 1 != ref->nritems); - btrfs_free_leaf_ref(root, ref); - return 0; -} - -int btrfs_free_reloc_root(struct btrfs_trans_handle *trans, - struct btrfs_root *root) -{ - struct btrfs_root *reloc_root; - int ret; - - if (root->reloc_root) { - reloc_root = root->reloc_root; - root->reloc_root = NULL; - list_add(&reloc_root->dead_list, - &root->fs_info->dead_reloc_roots); - - btrfs_set_root_bytenr(&reloc_root->root_item, - reloc_root->node->start); - btrfs_set_root_level(&root->root_item, - btrfs_header_level(reloc_root->node)); - memset(&reloc_root->root_item.drop_progress, 0, - sizeof(struct btrfs_disk_key)); - reloc_root->root_item.drop_level = 0; - - ret = btrfs_update_root(trans, root->fs_info->tree_root, - &reloc_root->root_key, - &reloc_root->root_item); - BUG_ON(ret); - } - return 0; -} - -int btrfs_drop_dead_reloc_roots(struct btrfs_root *root) -{ - struct btrfs_trans_handle *trans; - struct btrfs_root *reloc_root; - struct btrfs_root *prev_root = NULL; - struct list_head dead_roots; - int ret; - unsigned long nr; - - INIT_LIST_HEAD(&dead_roots); - list_splice_init(&root->fs_info->dead_reloc_roots, &dead_roots); - - while (!list_empty(&dead_roots)) { - reloc_root = list_entry(dead_roots.prev, - struct btrfs_root, dead_list); - list_del_init(&reloc_root->dead_list); - - BUG_ON(reloc_root->commit_root != NULL); - while (1) { - trans = btrfs_join_transaction(root, 1); - BUG_ON(IS_ERR(trans)); - - mutex_lock(&root->fs_info->drop_mutex); - ret = btrfs_drop_snapshot(trans, reloc_root); - if (ret != -EAGAIN) - break; - mutex_unlock(&root->fs_info->drop_mutex); - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, root); - BUG_ON(ret); - btrfs_btree_balance_dirty(root, nr); - } - - free_extent_buffer(reloc_root->node); - - ret = btrfs_del_root(trans, root->fs_info->tree_root, - &reloc_root->root_key); - BUG_ON(ret); - mutex_unlock(&root->fs_info->drop_mutex); - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, root); - BUG_ON(ret); - btrfs_btree_balance_dirty(root, nr); - - kfree(prev_root); - prev_root = reloc_root; - } - if (prev_root) { - btrfs_remove_leaf_refs(prev_root, (u64)-1, 0); - kfree(prev_root); - } - return 0; -} - -int btrfs_add_dead_reloc_root(struct btrfs_root *root) -{ - list_add(&root->dead_list, &root->fs_info->dead_reloc_roots); - return 0; -} - -int btrfs_cleanup_reloc_trees(struct btrfs_root *root) -{ - struct btrfs_root *reloc_root; - struct btrfs_trans_handle *trans; - struct btrfs_key location; - int found; - int ret; - - mutex_lock(&root->fs_info->tree_reloc_mutex); - ret = btrfs_find_dead_roots(root, BTRFS_TREE_RELOC_OBJECTID, NULL); - BUG_ON(ret); - found = !list_empty(&root->fs_info->dead_reloc_roots); - mutex_unlock(&root->fs_info->tree_reloc_mutex); - - if (found) { - trans = btrfs_start_transaction(root, 1); - BUG_ON(IS_ERR(trans)); - ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); - } - - location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID; - location.offset = (u64)-1; - location.type = BTRFS_ROOT_ITEM_KEY; - - reloc_root = btrfs_read_fs_root_no_name(root->fs_info, &location); - BUG_ON(!reloc_root); - btrfs_orphan_cleanup(reloc_root); - return 0; -} - -static noinline int init_reloc_tree(struct btrfs_trans_handle *trans, - struct btrfs_root *root) -{ - struct btrfs_root *reloc_root; - struct extent_buffer *eb; - struct btrfs_root_item *root_item; - struct btrfs_key root_key; - int ret; - - BUG_ON(!root->ref_cows); - if (root->reloc_root) - return 0; - - root_item = kmalloc(sizeof(*root_item), GFP_NOFS); - BUG_ON(!root_item); - - ret = btrfs_copy_root(trans, root, root->commit_root, - &eb, BTRFS_TREE_RELOC_OBJECTID); - BUG_ON(ret); - - root_key.objectid = BTRFS_TREE_RELOC_OBJECTID; - root_key.offset = root->root_key.objectid; - root_key.type = BTRFS_ROOT_ITEM_KEY; - - memcpy(root_item, &root->root_item, sizeof(root_item)); - btrfs_set_root_refs(root_item, 0); - btrfs_set_root_bytenr(root_item, eb->start); - btrfs_set_root_level(root_item, btrfs_header_level(eb)); - btrfs_set_root_generation(root_item, trans->transid); - - btrfs_tree_unlock(eb); - free_extent_buffer(eb); - - ret = btrfs_insert_root(trans, root->fs_info->tree_root, - &root_key, root_item); - BUG_ON(ret); - kfree(root_item); - - reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root, - &root_key); - BUG_ON(!reloc_root); - reloc_root->last_trans = trans->transid; - reloc_root->commit_root = NULL; - reloc_root->ref_tree = &root->fs_info->reloc_ref_tree; - - root->reloc_root = reloc_root; - return 0; -} - -/* - * Core function of space balance. - * - * The idea is using reloc trees to relocate tree blocks in reference - * counted roots. There is one reloc tree for each subvol, and all - * reloc trees share same root key objectid. Reloc trees are snapshots - * of the latest committed roots of subvols (root->commit_root). - * - * To relocate a tree block referenced by a subvol, there are two steps. - * COW the block through subvol's reloc tree, then update block pointer - * in the subvol to point to the new block. Since all reloc trees share - * same root key objectid, doing special handing for tree blocks owned - * by them is easy. Once a tree block has been COWed in one reloc tree, - * we can use the resulting new block directly when the same block is - * required to COW again through other reloc trees. By this way, relocated - * tree blocks are shared between reloc trees, so they are also shared - * between subvols. - */ -static noinline int relocate_one_path(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *first_key, - struct btrfs_ref_path *ref_path, - struct btrfs_block_group_cache *group, - struct inode *reloc_inode) -{ - struct btrfs_root *reloc_root; - struct extent_buffer *eb = NULL; - struct btrfs_key *keys; - u64 *nodes; - int level; - int shared_level; - int lowest_level = 0; - int ret; - - if (ref_path->owner_objectid < BTRFS_FIRST_FREE_OBJECTID) - lowest_level = ref_path->owner_objectid; - - if (!root->ref_cows) { - path->lowest_level = lowest_level; - ret = btrfs_search_slot(trans, root, first_key, path, 0, 1); - BUG_ON(ret < 0); - path->lowest_level = 0; - btrfs_release_path(root, path); - return 0; - } - - mutex_lock(&root->fs_info->tree_reloc_mutex); - ret = init_reloc_tree(trans, root); - BUG_ON(ret); - reloc_root = root->reloc_root; - - shared_level = ref_path->shared_level; - ref_path->shared_level = BTRFS_MAX_LEVEL - 1; - - keys = ref_path->node_keys; - nodes = ref_path->new_nodes; - memset(&keys[shared_level + 1], 0, - sizeof(*keys) * (BTRFS_MAX_LEVEL - shared_level - 1)); - memset(&nodes[shared_level + 1], 0, - sizeof(*nodes) * (BTRFS_MAX_LEVEL - shared_level - 1)); - - if (nodes[lowest_level] == 0) { - path->lowest_level = lowest_level; - ret = btrfs_search_slot(trans, reloc_root, first_key, path, - 0, 1); - BUG_ON(ret); - for (level = lowest_level; level < BTRFS_MAX_LEVEL; level++) { - eb = path->nodes[level]; - if (!eb || eb == reloc_root->node) - break; - nodes[level] = eb->start; - if (level == 0) - btrfs_item_key_to_cpu(eb, &keys[level], 0); - else - btrfs_node_key_to_cpu(eb, &keys[level], 0); - } - if (nodes[0] && - ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - eb = path->nodes[0]; - ret = replace_extents_in_leaf(trans, reloc_root, eb, - group, reloc_inode); - BUG_ON(ret); - } - btrfs_release_path(reloc_root, path); - } else { - ret = btrfs_merge_path(trans, reloc_root, keys, nodes, - lowest_level); - BUG_ON(ret); - } - - /* - * replace tree blocks in the fs tree with tree blocks in - * the reloc tree. - */ - ret = btrfs_merge_path(trans, root, keys, nodes, lowest_level); - BUG_ON(ret < 0); - - if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - ret = btrfs_search_slot(trans, reloc_root, first_key, path, - 0, 0); - BUG_ON(ret); - extent_buffer_get(path->nodes[0]); - eb = path->nodes[0]; - btrfs_release_path(reloc_root, path); - ret = invalidate_extent_cache(reloc_root, eb, group, root); - BUG_ON(ret); - free_extent_buffer(eb); - } - - mutex_unlock(&root->fs_info->tree_reloc_mutex); - path->lowest_level = 0; - return 0; -} - -static noinline int relocate_tree_block(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *first_key, - struct btrfs_ref_path *ref_path) -{ - int ret; - - ret = relocate_one_path(trans, root, path, first_key, - ref_path, NULL, NULL); - BUG_ON(ret); - - return 0; -} - -static noinline int del_extent_zero(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_path *path, - struct btrfs_key *extent_key) -{ - int ret; - - ret = btrfs_search_slot(trans, extent_root, extent_key, path, -1, 1); - if (ret) - goto out; - ret = btrfs_del_item(trans, extent_root, path); -out: - btrfs_release_path(extent_root, path); - return ret; -} - -static noinline struct btrfs_root *read_ref_root(struct btrfs_fs_info *fs_info, - struct btrfs_ref_path *ref_path) -{ - struct btrfs_key root_key; - - root_key.objectid = ref_path->root_objectid; - root_key.type = BTRFS_ROOT_ITEM_KEY; - if (is_cowonly_root(ref_path->root_objectid)) - root_key.offset = 0; - else - root_key.offset = (u64)-1; - - return btrfs_read_fs_root_no_name(fs_info, &root_key); -} - -static noinline int relocate_one_extent(struct btrfs_root *extent_root, - struct btrfs_path *path, - struct btrfs_key *extent_key, - struct btrfs_block_group_cache *group, - struct inode *reloc_inode, int pass) -{ - struct btrfs_trans_handle *trans; - struct btrfs_root *found_root; - struct btrfs_ref_path *ref_path = NULL; - struct disk_extent *new_extents = NULL; - int nr_extents = 0; - int loops; - int ret; - int level; - struct btrfs_key first_key; - u64 prev_block = 0; - - - trans = btrfs_start_transaction(extent_root, 1); - BUG_ON(IS_ERR(trans)); - - if (extent_key->objectid == 0) { - ret = del_extent_zero(trans, extent_root, path, extent_key); - goto out; - } - - ref_path = kmalloc(sizeof(*ref_path), GFP_NOFS); - if (!ref_path) { - ret = -ENOMEM; - goto out; - } - - for (loops = 0; ; loops++) { - if (loops == 0) { - ret = btrfs_first_ref_path(trans, extent_root, ref_path, - extent_key->objectid); - } else { - ret = btrfs_next_ref_path(trans, extent_root, ref_path); - } - if (ret < 0) - goto out; - if (ret > 0) - break; - - if (ref_path->root_objectid == BTRFS_TREE_LOG_OBJECTID || - ref_path->root_objectid == BTRFS_TREE_RELOC_OBJECTID) - continue; - - found_root = read_ref_root(extent_root->fs_info, ref_path); - BUG_ON(!found_root); - /* - * for reference counted tree, only process reference paths - * rooted at the latest committed root. - */ - if (found_root->ref_cows && - ref_path->root_generation != found_root->root_key.offset) - continue; - - if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - if (pass == 0) { - /* - * copy data extents to new locations - */ - u64 group_start = group->key.objectid; - ret = relocate_data_extent(reloc_inode, - extent_key, - group_start); - if (ret < 0) - goto out; - break; - } - level = 0; - } else { - level = ref_path->owner_objectid; - } - - if (prev_block != ref_path->nodes[level]) { - struct extent_buffer *eb; - u64 block_start = ref_path->nodes[level]; - u64 block_size = btrfs_level_size(found_root, level); - - eb = read_tree_block(found_root, block_start, - block_size, 0); - btrfs_tree_lock(eb); - BUG_ON(level != btrfs_header_level(eb)); - - if (level == 0) - btrfs_item_key_to_cpu(eb, &first_key, 0); - else - btrfs_node_key_to_cpu(eb, &first_key, 0); - - btrfs_tree_unlock(eb); - free_extent_buffer(eb); - prev_block = block_start; - } - - mutex_lock(&extent_root->fs_info->trans_mutex); - btrfs_record_root_in_trans(found_root); - mutex_unlock(&extent_root->fs_info->trans_mutex); - if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - /* - * try to update data extent references while - * keeping metadata shared between snapshots. - */ - if (pass == 1) { - ret = relocate_one_path(trans, found_root, - path, &first_key, ref_path, - group, reloc_inode); - if (ret < 0) - goto out; - continue; - } - /* - * use fallback method to process the remaining - * references. - */ - if (!new_extents) { - u64 group_start = group->key.objectid; - new_extents = kmalloc(sizeof(*new_extents), - GFP_NOFS); - nr_extents = 1; - ret = get_new_locations(reloc_inode, - extent_key, - group_start, 1, - &new_extents, - &nr_extents); - if (ret) - goto out; - } - ret = replace_one_extent(trans, found_root, - path, extent_key, - &first_key, ref_path, - new_extents, nr_extents); - } else { - ret = relocate_tree_block(trans, found_root, path, - &first_key, ref_path); - } - if (ret < 0) - goto out; - } - ret = 0; -out: - btrfs_end_transaction(trans, extent_root); - kfree(new_extents); - kfree(ref_path); - return ret; -} -#endif - static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) { u64 num_devices; @@ -8050,13 +6531,15 @@ int btrfs_set_block_group_ro(struct btrfs_root *root, alloc_flags = update_block_group_flags(root, cache->flags); if (alloc_flags != cache->flags) - do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, 1); + do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, + CHUNK_ALLOC_FORCE); ret = set_block_group_ro(cache); if (!ret) goto out; alloc_flags = get_alloc_profile(root, cache->space_info->flags); - ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, 1); + ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, + CHUNK_ALLOC_FORCE); if (ret < 0) goto out; ret = set_block_group_ro(cache); @@ -8065,6 +6548,14 @@ out: return ret; } +int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, + struct btrfs_root *root, u64 type) +{ + u64 alloc_flags = get_alloc_profile(root, type); + return do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, + CHUNK_ALLOC_FORCE); +} + /* * helper to account the unused space of all the readonly block group in the * list. takes mirrors into account. @@ -8414,10 +6905,16 @@ int btrfs_read_block_groups(struct btrfs_root *root) ret = -ENOMEM; goto error; } + cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl), + GFP_NOFS); + if (!cache->free_space_ctl) { + kfree(cache); + ret = -ENOMEM; + goto error; + } atomic_set(&cache->count, 1); spin_lock_init(&cache->lock); - spin_lock_init(&cache->tree_lock); cache->fs_info = info; INIT_LIST_HEAD(&cache->list); INIT_LIST_HEAD(&cache->cluster_list); @@ -8425,24 +6922,18 @@ int btrfs_read_block_groups(struct btrfs_root *root) if (need_clear) cache->disk_cache_state = BTRFS_DC_CLEAR; - /* - * we only want to have 32k of ram per block group for keeping - * track of free space, and if we pass 1/2 of that we want to - * start converting things over to using bitmaps - */ - cache->extents_thresh = ((1024 * 32) / 2) / - sizeof(struct btrfs_free_space); - read_extent_buffer(leaf, &cache->item, btrfs_item_ptr_offset(leaf, path->slots[0]), sizeof(cache->item)); memcpy(&cache->key, &found_key, sizeof(found_key)); key.objectid = found_key.objectid + found_key.offset; - btrfs_release_path(root, path); + btrfs_release_path(path); cache->flags = btrfs_block_group_flags(&cache->item); cache->sectorsize = root->sectorsize; + btrfs_init_free_space_ctl(cache); + /* * We need to exclude the super stripes now so that the space * info has super bytes accounted for, otherwise we'll think @@ -8529,6 +7020,12 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, cache = kzalloc(sizeof(*cache), GFP_NOFS); if (!cache) return -ENOMEM; + cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl), + GFP_NOFS); + if (!cache->free_space_ctl) { + kfree(cache); + return -ENOMEM; + } cache->key.objectid = chunk_offset; cache->key.offset = size; @@ -8536,19 +7033,13 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, cache->sectorsize = root->sectorsize; cache->fs_info = root->fs_info; - /* - * we only want to have 32k of ram per block group for keeping track - * of free space, and if we pass 1/2 of that we want to start - * converting things over to using bitmaps - */ - cache->extents_thresh = ((1024 * 32) / 2) / - sizeof(struct btrfs_free_space); atomic_set(&cache->count, 1); spin_lock_init(&cache->lock); - spin_lock_init(&cache->tree_lock); INIT_LIST_HEAD(&cache->list); INIT_LIST_HEAD(&cache->cluster_list); + btrfs_init_free_space_ctl(cache); + btrfs_set_block_group_used(&cache->item, bytes_used); btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid); cache->flags = type; @@ -8603,6 +7094,12 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, BUG_ON(!block_group); BUG_ON(!block_group->ro); + /* + * Free the reserved super bytes from this block group before + * remove it. + */ + free_excluded_extents(root, block_group); + memcpy(&key, &block_group->key, sizeof(key)); if (block_group->flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | @@ -8655,12 +7152,12 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, if (ret < 0) goto out; if (ret > 0) - btrfs_release_path(tree_root, path); + btrfs_release_path(path); if (ret == 0) { ret = btrfs_del_item(trans, tree_root, path); if (ret) goto out; - btrfs_release_path(tree_root, path); + btrfs_release_path(path); } spin_lock(&root->fs_info->block_group_cache_lock); @@ -8706,13 +7203,99 @@ out: return ret; } +int btrfs_init_space_info(struct btrfs_fs_info *fs_info) +{ + struct btrfs_space_info *space_info; + struct btrfs_super_block *disk_super; + u64 features; + u64 flags; + int mixed = 0; + int ret; + + disk_super = &fs_info->super_copy; + if (!btrfs_super_root(disk_super)) + return 1; + + features = btrfs_super_incompat_flags(disk_super); + if (features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) + mixed = 1; + + flags = BTRFS_BLOCK_GROUP_SYSTEM; + ret = update_space_info(fs_info, flags, 0, 0, &space_info); + if (ret) + goto out; + + if (mixed) { + flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA; + ret = update_space_info(fs_info, flags, 0, 0, &space_info); + } else { + flags = BTRFS_BLOCK_GROUP_METADATA; + ret = update_space_info(fs_info, flags, 0, 0, &space_info); + if (ret) + goto out; + + flags = BTRFS_BLOCK_GROUP_DATA; + ret = update_space_info(fs_info, flags, 0, 0, &space_info); + } +out: + return ret; +} + int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) { return unpin_extent_range(root, start, end); } int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr, - u64 num_bytes) + u64 num_bytes, u64 *actual_bytes) { - return btrfs_discard_extent(root, bytenr, num_bytes); + return btrfs_discard_extent(root, bytenr, num_bytes, actual_bytes); +} + +int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_block_group_cache *cache = NULL; + u64 group_trimmed; + u64 start; + u64 end; + u64 trimmed = 0; + int ret = 0; + + cache = btrfs_lookup_block_group(fs_info, range->start); + + while (cache) { + if (cache->key.objectid >= (range->start + range->len)) { + btrfs_put_block_group(cache); + break; + } + + start = max(range->start, cache->key.objectid); + end = min(range->start + range->len, + cache->key.objectid + cache->key.offset); + + if (end - start >= range->minlen) { + if (!block_group_cache_done(cache)) { + ret = cache_block_group(cache, NULL, root, 0); + if (!ret) + wait_block_group_cache_done(cache); + } + ret = btrfs_trim_block_group(cache, + &group_trimmed, + start, + end, + range->minlen); + + trimmed += group_trimmed; + if (ret) { + btrfs_put_block_group(cache); + break; + } + } + + cache = next_block_group(fs_info->tree_root, cache); + } + + range->len = trimmed; + return ret; } diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 5e76a474cb7e..c5d9fbb92bc3 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -10,6 +10,8 @@ #include <linux/swap.h> #include <linux/writeback.h> #include <linux/pagevec.h> +#include <linux/prefetch.h> +#include <linux/cleancache.h> #include "extent_io.h" #include "extent_map.h" #include "compat.h" @@ -101,7 +103,7 @@ void extent_io_exit(void) } void extent_io_tree_init(struct extent_io_tree *tree, - struct address_space *mapping, gfp_t mask) + struct address_space *mapping) { tree->state = RB_ROOT; INIT_RADIX_TREE(&tree->buffer, GFP_ATOMIC); @@ -439,6 +441,15 @@ static int clear_state_bit(struct extent_io_tree *tree, return ret; } +static struct extent_state * +alloc_extent_state_atomic(struct extent_state *prealloc) +{ + if (!prealloc) + prealloc = alloc_extent_state(GFP_ATOMIC); + + return prealloc; +} + /* * clear some bits on a range in the tree. This may require splitting * or inserting elements in the tree, so the gfp mask is used to @@ -529,8 +540,8 @@ hit_next: */ if (state->start < start) { - if (!prealloc) - prealloc = alloc_extent_state(GFP_ATOMIC); + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, start); BUG_ON(err == -EEXIST); prealloc = NULL; @@ -551,8 +562,8 @@ hit_next: * on the first half */ if (state->start <= end && state->end > end) { - if (!prealloc) - prealloc = alloc_extent_state(GFP_ATOMIC); + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, end + 1); BUG_ON(err == -EEXIST); if (wake) @@ -690,6 +701,15 @@ static void cache_state(struct extent_state *state, } } +static void uncache_state(struct extent_state **cached_ptr) +{ + if (cached_ptr && (*cached_ptr)) { + struct extent_state *state = *cached_ptr; + *cached_ptr = NULL; + free_extent_state(state); + } +} + /* * set some bits on a range in the tree. This may require allocations or * sleeping, so the gfp mask is used to indicate what is allowed. @@ -716,8 +736,7 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, again: if (!prealloc && (mask & __GFP_WAIT)) { prealloc = alloc_extent_state(mask); - if (!prealloc) - return -ENOMEM; + BUG_ON(!prealloc); } spin_lock(&tree->lock); @@ -734,6 +753,8 @@ again: */ node = tree_search(tree, start); if (!node) { + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = insert_state(tree, prealloc, start, end, &bits); prealloc = NULL; BUG_ON(err == -EEXIST); @@ -762,20 +783,18 @@ hit_next: if (err) goto out; + next_node = rb_next(node); cache_state(state, cached_state); merge_state(tree, state); if (last_end == (u64)-1) goto out; start = last_end + 1; - if (start < end && prealloc && !need_resched()) { - next_node = rb_next(node); - if (next_node) { - state = rb_entry(next_node, struct extent_state, - rb_node); - if (state->start == start) - goto hit_next; - } + if (next_node && start < end && prealloc && !need_resched()) { + state = rb_entry(next_node, struct extent_state, + rb_node); + if (state->start == start) + goto hit_next; } goto search_again; } @@ -802,6 +821,9 @@ hit_next: err = -EEXIST; goto out; } + + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, start); BUG_ON(err == -EEXIST); prealloc = NULL; @@ -832,14 +854,25 @@ hit_next: this_end = end; else this_end = last_start - 1; + + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); + + /* + * Avoid to free 'prealloc' if it can be merged with + * the later extent. + */ + atomic_inc(&prealloc->refs); err = insert_state(tree, prealloc, start, this_end, &bits); BUG_ON(err == -EEXIST); if (err) { + free_extent_state(prealloc); prealloc = NULL; goto out; } cache_state(prealloc, cached_state); + free_extent_state(prealloc); prealloc = NULL; start = this_end + 1; goto search_again; @@ -856,6 +889,9 @@ hit_next: err = -EEXIST; goto out; } + + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, end + 1); BUG_ON(err == -EEXIST); @@ -932,18 +968,11 @@ int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end, NULL, mask); } -static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask) -{ - return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0, - NULL, mask); -} - int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask) + struct extent_state **cached_state, gfp_t mask) { - return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL, - NULL, mask); + return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, + NULL, cached_state, mask); } static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, @@ -954,11 +983,6 @@ static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, cached_state, mask); } -int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end) -{ - return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK); -} - /* * either insert or lock state struct between start and end use mask to tell * us if waiting is desired. @@ -1012,33 +1036,13 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end, mask); } -int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask) +int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask) { return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, NULL, mask); } /* - * helper function to set pages and extents in the tree dirty - */ -int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end) -{ - unsigned long index = start >> PAGE_CACHE_SHIFT; - unsigned long end_index = end >> PAGE_CACHE_SHIFT; - struct page *page; - - while (index <= end_index) { - page = find_get_page(tree->mapping, index); - BUG_ON(!page); - __set_page_dirty_nobuffers(page); - page_cache_release(page); - index++; - } - return 0; -} - -/* * helper function to set both pages and extents in the tree writeback */ static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end) @@ -1433,12 +1437,13 @@ int extent_clear_unlock_delalloc(struct inode *inode, */ u64 count_range_bits(struct extent_io_tree *tree, u64 *start, u64 search_end, u64 max_bytes, - unsigned long bits) + unsigned long bits, int contig) { struct rb_node *node; struct extent_state *state; u64 cur_start = *start; u64 total_bytes = 0; + u64 last = 0; int found = 0; if (search_end <= cur_start) { @@ -1463,7 +1468,9 @@ u64 count_range_bits(struct extent_io_tree *tree, state = rb_entry(node, struct extent_state, rb_node); if (state->start > search_end) break; - if (state->end >= cur_start && (state->state & bits)) { + if (contig && found && state->start > last + 1) + break; + if (state->end >= cur_start && (state->state & bits) == bits) { total_bytes += min(search_end, state->end) + 1 - max(cur_start, state->start); if (total_bytes >= max_bytes) @@ -1472,6 +1479,9 @@ u64 count_range_bits(struct extent_io_tree *tree, *start = state->start; found = 1; } + last = state->end; + } else if (contig && found) { + break; } node = rb_next(node); if (!node) @@ -1729,6 +1739,9 @@ static void end_bio_extent_readpage(struct bio *bio, int err) do { struct page *page = bvec->bv_page; + struct extent_state *cached = NULL; + struct extent_state *state; + tree = &BTRFS_I(page->mapping->host)->io_tree; start = ((u64)page->index << PAGE_CACHE_SHIFT) + @@ -1743,9 +1756,20 @@ static void end_bio_extent_readpage(struct bio *bio, int err) if (++bvec <= bvec_end) prefetchw(&bvec->bv_page->flags); + spin_lock(&tree->lock); + state = find_first_extent_bit_state(tree, start, EXTENT_LOCKED); + if (state && state->start == start) { + /* + * take a reference on the state, unlock will drop + * the ref + */ + cache_state(state, &cached); + } + spin_unlock(&tree->lock); + if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) { ret = tree->ops->readpage_end_io_hook(page, start, end, - NULL); + state); if (ret) uptodate = 0; } @@ -1758,15 +1782,16 @@ static void end_bio_extent_readpage(struct bio *bio, int err) test_bit(BIO_UPTODATE, &bio->bi_flags); if (err) uptodate = 0; + uncache_state(&cached); continue; } } if (uptodate) { - set_extent_uptodate(tree, start, end, + set_extent_uptodate(tree, start, end, &cached, GFP_ATOMIC); } - unlock_extent(tree, start, end, GFP_ATOMIC); + unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC); if (whole_page) { if (uptodate) { @@ -1790,44 +1815,6 @@ static void end_bio_extent_readpage(struct bio *bio, int err) bio_put(bio); } -/* - * IO done from prepare_write is pretty simple, we just unlock - * the structs in the extent tree when done, and set the uptodate bits - * as appropriate. - */ -static void end_bio_extent_preparewrite(struct bio *bio, int err) -{ - const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); - struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1; - struct extent_io_tree *tree; - u64 start; - u64 end; - - do { - struct page *page = bvec->bv_page; - tree = &BTRFS_I(page->mapping->host)->io_tree; - - start = ((u64)page->index << PAGE_CACHE_SHIFT) + - bvec->bv_offset; - end = start + bvec->bv_len - 1; - - if (--bvec >= bio->bi_io_vec) - prefetchw(&bvec->bv_page->flags); - - if (uptodate) { - set_extent_uptodate(tree, start, end, GFP_ATOMIC); - } else { - ClearPageUptodate(page); - SetPageError(page); - } - - unlock_extent(tree, start, end, GFP_ATOMIC); - - } while (bvec >= bio->bi_io_vec); - - bio_put(bio); -} - struct bio * btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs, gfp_t gfp_flags) @@ -1946,6 +1933,7 @@ void set_page_extent_mapped(struct page *page) static void set_page_extent_head(struct page *page, unsigned long len) { + WARN_ON(!PagePrivate(page)); set_page_private(page, EXTENT_PAGE_PRIVATE_FIRST_PAGE | len << 2); } @@ -1975,7 +1963,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, struct btrfs_ordered_extent *ordered; int ret; int nr = 0; - size_t page_offset = 0; + size_t pg_offset = 0; size_t iosize; size_t disk_io_size; size_t blocksize = inode->i_sb->s_blocksize; @@ -1983,6 +1971,13 @@ static int __extent_read_full_page(struct extent_io_tree *tree, set_page_extent_mapped(page); + if (!PageUptodate(page)) { + if (cleancache_get_page(page) == 0) { + BUG_ON(blocksize != PAGE_SIZE); + goto out; + } + } + end = page_end; while (1) { lock_extent(tree, start, end, GFP_NOFS); @@ -2009,19 +2004,22 @@ static int __extent_read_full_page(struct extent_io_tree *tree, while (cur <= end) { if (cur >= last_byte) { char *userpage; - iosize = PAGE_CACHE_SIZE - page_offset; + struct extent_state *cached = NULL; + + iosize = PAGE_CACHE_SIZE - pg_offset; userpage = kmap_atomic(page, KM_USER0); - memset(userpage + page_offset, 0, iosize); + memset(userpage + pg_offset, 0, iosize); flush_dcache_page(page); kunmap_atomic(userpage, KM_USER0); set_extent_uptodate(tree, cur, cur + iosize - 1, - GFP_NOFS); - unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); + &cached, GFP_NOFS); + unlock_extent_cached(tree, cur, cur + iosize - 1, + &cached, GFP_NOFS); break; } - em = get_extent(inode, page, page_offset, cur, + em = get_extent(inode, page, pg_offset, cur, end - cur + 1, 0); - if (IS_ERR(em) || !em) { + if (IS_ERR_OR_NULL(em)) { SetPageError(page); unlock_extent(tree, cur, end, GFP_NOFS); break; @@ -2056,16 +2054,19 @@ static int __extent_read_full_page(struct extent_io_tree *tree, /* we've found a hole, just zero and go on */ if (block_start == EXTENT_MAP_HOLE) { char *userpage; + struct extent_state *cached = NULL; + userpage = kmap_atomic(page, KM_USER0); - memset(userpage + page_offset, 0, iosize); + memset(userpage + pg_offset, 0, iosize); flush_dcache_page(page); kunmap_atomic(userpage, KM_USER0); set_extent_uptodate(tree, cur, cur + iosize - 1, - GFP_NOFS); - unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); + &cached, GFP_NOFS); + unlock_extent_cached(tree, cur, cur + iosize - 1, + &cached, GFP_NOFS); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; continue; } /* the get_extent function already copied into the page */ @@ -2074,7 +2075,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, check_page_uptodate(tree, page); unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; continue; } /* we have an inline extent but it didn't get marked up @@ -2084,7 +2085,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, SetPageError(page); unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; continue; } @@ -2097,7 +2098,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1; pnr -= page->index; ret = submit_extent_page(READ, tree, page, - sector, disk_io_size, page_offset, + sector, disk_io_size, pg_offset, bdev, bio, pnr, end_bio_extent_readpage, mirror_num, *bio_flags, @@ -2108,8 +2109,9 @@ static int __extent_read_full_page(struct extent_io_tree *tree, if (ret) SetPageError(page); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; } +out: if (!nr) { if (!PageError(page)) SetPageUptodate(page); @@ -2181,10 +2183,12 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, unsigned long nr_written = 0; if (wbc->sync_mode == WB_SYNC_ALL) - write_flags = WRITE_SYNC_PLUG; + write_flags = WRITE_SYNC; else write_flags = WRITE; + trace___extent_writepage(page, inode, wbc); + WARN_ON(!PageLocked(page)); pg_offset = i_size & (PAGE_CACHE_SIZE - 1); if (page->index > end_index || @@ -2301,7 +2305,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, } em = epd->get_extent(inode, page, pg_offset, cur, end - cur + 1, 1); - if (IS_ERR(em) || !em) { + if (IS_ERR_OR_NULL(em)) { SetPageError(page); break; } @@ -2641,7 +2645,7 @@ int extent_readpages(struct extent_io_tree *tree, prefetchw(&page->flags); list_del(&page->lru); if (!add_to_page_cache_lru(page, mapping, - page->index, GFP_KERNEL)) { + page->index, GFP_NOFS)) { __extent_read_full_page(tree, page, get_extent, &bio, 0, &bio_flags); } @@ -2680,125 +2684,6 @@ int extent_invalidatepage(struct extent_io_tree *tree, } /* - * simple commit_write call, set_range_dirty is used to mark both - * the pages and the extent records as dirty - */ -int extent_commit_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to) -{ - loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to; - - set_page_extent_mapped(page); - set_page_dirty(page); - - if (pos > inode->i_size) { - i_size_write(inode, pos); - mark_inode_dirty(inode); - } - return 0; -} - -int extent_prepare_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to, get_extent_t *get_extent) -{ - u64 page_start = (u64)page->index << PAGE_CACHE_SHIFT; - u64 page_end = page_start + PAGE_CACHE_SIZE - 1; - u64 block_start; - u64 orig_block_start; - u64 block_end; - u64 cur_end; - struct extent_map *em; - unsigned blocksize = 1 << inode->i_blkbits; - size_t page_offset = 0; - size_t block_off_start; - size_t block_off_end; - int err = 0; - int iocount = 0; - int ret = 0; - int isnew; - - set_page_extent_mapped(page); - - block_start = (page_start + from) & ~((u64)blocksize - 1); - block_end = (page_start + to - 1) | (blocksize - 1); - orig_block_start = block_start; - - lock_extent(tree, page_start, page_end, GFP_NOFS); - while (block_start <= block_end) { - em = get_extent(inode, page, page_offset, block_start, - block_end - block_start + 1, 1); - if (IS_ERR(em) || !em) - goto err; - - cur_end = min(block_end, extent_map_end(em) - 1); - block_off_start = block_start & (PAGE_CACHE_SIZE - 1); - block_off_end = block_off_start + blocksize; - isnew = clear_extent_new(tree, block_start, cur_end, GFP_NOFS); - - if (!PageUptodate(page) && isnew && - (block_off_end > to || block_off_start < from)) { - void *kaddr; - - kaddr = kmap_atomic(page, KM_USER0); - if (block_off_end > to) - memset(kaddr + to, 0, block_off_end - to); - if (block_off_start < from) - memset(kaddr + block_off_start, 0, - from - block_off_start); - flush_dcache_page(page); - kunmap_atomic(kaddr, KM_USER0); - } - if ((em->block_start != EXTENT_MAP_HOLE && - em->block_start != EXTENT_MAP_INLINE) && - !isnew && !PageUptodate(page) && - (block_off_end > to || block_off_start < from) && - !test_range_bit(tree, block_start, cur_end, - EXTENT_UPTODATE, 1, NULL)) { - u64 sector; - u64 extent_offset = block_start - em->start; - size_t iosize; - sector = (em->block_start + extent_offset) >> 9; - iosize = (cur_end - block_start + blocksize) & - ~((u64)blocksize - 1); - /* - * we've already got the extent locked, but we - * need to split the state such that our end_bio - * handler can clear the lock. - */ - set_extent_bit(tree, block_start, - block_start + iosize - 1, - EXTENT_LOCKED, 0, NULL, NULL, GFP_NOFS); - ret = submit_extent_page(READ, tree, page, - sector, iosize, page_offset, em->bdev, - NULL, 1, - end_bio_extent_preparewrite, 0, - 0, 0); - if (ret && !err) - err = ret; - iocount++; - block_start = block_start + iosize; - } else { - set_extent_uptodate(tree, block_start, cur_end, - GFP_NOFS); - unlock_extent(tree, block_start, cur_end, GFP_NOFS); - block_start = cur_end + 1; - } - page_offset = block_start & (PAGE_CACHE_SIZE - 1); - free_extent_map(em); - } - if (iocount) { - wait_extent_bit(tree, orig_block_start, - block_end, EXTENT_LOCKED); - } - check_page_uptodate(tree, page); -err: - /* FIXME, zero out newly allocated blocks on error */ - return err; -} - -/* * a helper for releasepage, this tests for areas of the page that * are locked or under IO and drops the related state bits if it is safe * to drop the page. @@ -2821,9 +2706,17 @@ int try_release_extent_state(struct extent_map_tree *map, * at this point we can safely clear everything except the * locked bit and the nodatasum bit */ - clear_extent_bit(tree, start, end, + ret = clear_extent_bit(tree, start, end, ~(EXTENT_LOCKED | EXTENT_NODATASUM), 0, 0, NULL, mask); + + /* if clear_extent_bit failed for enomem reasons, + * we can't allow the release to continue. + */ + if (ret < 0) + ret = 0; + else + ret = 1; } return ret; } @@ -2848,7 +2741,7 @@ int try_release_extent_mapping(struct extent_map_tree *map, len = end - start + 1; write_lock(&map->lock); em = lookup_extent_mapping(map, start, len); - if (!em || IS_ERR(em)) { + if (IS_ERR_OR_NULL(em)) { write_unlock(&map->lock); break; } @@ -2876,31 +2769,44 @@ int try_release_extent_mapping(struct extent_map_tree *map, return try_release_extent_state(map, tree, page, mask); } -sector_t extent_bmap(struct address_space *mapping, sector_t iblock, - get_extent_t *get_extent) +/* + * helper function for fiemap, which doesn't want to see any holes. + * This maps until we find something past 'last' + */ +static struct extent_map *get_extent_skip_holes(struct inode *inode, + u64 offset, + u64 last, + get_extent_t *get_extent) { - struct inode *inode = mapping->host; - struct extent_state *cached_state = NULL; - u64 start = iblock << inode->i_blkbits; - sector_t sector = 0; - size_t blksize = (1 << inode->i_blkbits); + u64 sectorsize = BTRFS_I(inode)->root->sectorsize; struct extent_map *em; + u64 len; - lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + blksize - 1, - 0, &cached_state, GFP_NOFS); - em = get_extent(inode, NULL, 0, start, blksize, 0); - unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, - start + blksize - 1, &cached_state, GFP_NOFS); - if (!em || IS_ERR(em)) - return 0; + if (offset >= last) + return NULL; - if (em->block_start > EXTENT_MAP_LAST_BYTE) - goto out; + while(1) { + len = last - offset; + if (len == 0) + break; + len = (len + sectorsize - 1) & ~(sectorsize - 1); + em = get_extent(inode, NULL, 0, offset, len, 0); + if (IS_ERR_OR_NULL(em)) + return em; - sector = (em->block_start + start - em->start) >> inode->i_blkbits; -out: - free_extent_map(em); - return sector; + /* if this isn't a hole return it */ + if (!test_bit(EXTENT_FLAG_VACANCY, &em->flags) && + em->block_start != EXTENT_MAP_HOLE) { + return em; + } + + /* this is a hole, advance to the next extent */ + offset = extent_map_end(em); + free_extent_map(em); + if (offset >= last) + break; + } + return NULL; } int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, @@ -2912,16 +2818,19 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u32 flags = 0; u32 found_type; u64 last; + u64 last_for_get_extent = 0; u64 disko = 0; + u64 isize = i_size_read(inode); struct btrfs_key found_key; struct extent_map *em = NULL; struct extent_state *cached_state = NULL; struct btrfs_path *path; struct btrfs_file_extent_item *item; int end = 0; - u64 em_start = 0, em_len = 0; + u64 em_start = 0; + u64 em_len = 0; + u64 em_end = 0; unsigned long emflags; - int hole = 0; if (len == 0) return -EINVAL; @@ -2931,8 +2840,12 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, return -ENOMEM; path->leave_spinning = 1; + /* + * lookup the last file extent. We're not using i_size here + * because there might be preallocation past i_size + */ ret = btrfs_lookup_file_extent(NULL, BTRFS_I(inode)->root, - path, inode->i_ino, -1, 0); + path, btrfs_ino(inode), -1, 0); if (ret < 0) { btrfs_free_path(path); return ret; @@ -2944,18 +2857,38 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]); found_type = btrfs_key_type(&found_key); - /* No extents, just return */ - if (found_key.objectid != inode->i_ino || + /* No extents, but there might be delalloc bits */ + if (found_key.objectid != btrfs_ino(inode) || found_type != BTRFS_EXTENT_DATA_KEY) { - btrfs_free_path(path); - return 0; + /* have to trust i_size as the end */ + last = (u64)-1; + last_for_get_extent = isize; + } else { + /* + * remember the start of the last extent. There are a + * bunch of different factors that go into the length of the + * extent, so its much less complex to remember where it started + */ + last = found_key.offset; + last_for_get_extent = last + 1; } - last = found_key.offset; btrfs_free_path(path); + /* + * we might have some extents allocated but more delalloc past those + * extents. so, we trust isize unless the start of the last extent is + * beyond isize + */ + if (last < isize) { + last = (u64)-1; + last_for_get_extent = isize; + } + lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0, &cached_state, GFP_NOFS); - em = get_extent(inode, NULL, 0, off, max - off, 0); + + em = get_extent_skip_holes(inode, off, last_for_get_extent, + get_extent); if (!em) goto out; if (IS_ERR(em)) { @@ -2964,22 +2897,38 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, } while (!end) { - hole = 0; - off = em->start + em->len; - if (off >= max) - end = 1; + u64 offset_in_extent; - if (em->block_start == EXTENT_MAP_HOLE) { - hole = 1; - goto next; - } + /* break if the extent we found is outside the range */ + if (em->start >= max || extent_map_end(em) < off) + break; - em_start = em->start; - em_len = em->len; + /* + * get_extent may return an extent that starts before our + * requested range. We have to make sure the ranges + * we return to fiemap always move forward and don't + * overlap, so adjust the offsets here + */ + em_start = max(em->start, off); + /* + * record the offset from the start of the extent + * for adjusting the disk offset below + */ + offset_in_extent = em_start - em->start; + em_end = extent_map_end(em); + em_len = em_end - em_start; + emflags = em->flags; disko = 0; flags = 0; + /* + * bump off for our next call to get_extent + */ + off = extent_map_end(em); + if (off >= max) + end = 1; + if (em->block_start == EXTENT_MAP_LAST_BYTE) { end = 1; flags |= FIEMAP_EXTENT_LAST; @@ -2990,42 +2939,34 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, flags |= (FIEMAP_EXTENT_DELALLOC | FIEMAP_EXTENT_UNKNOWN); } else { - disko = em->block_start; + disko = em->block_start + offset_in_extent; } if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) flags |= FIEMAP_EXTENT_ENCODED; -next: - emflags = em->flags; free_extent_map(em); em = NULL; - if (!end) { - em = get_extent(inode, NULL, 0, off, max - off, 0); - if (!em) - goto out; - if (IS_ERR(em)) { - ret = PTR_ERR(em); - goto out; - } - emflags = em->flags; - } - - if (test_bit(EXTENT_FLAG_VACANCY, &emflags)) { + if ((em_start >= last) || em_len == (u64)-1 || + (last == (u64)-1 && isize <= em_end)) { flags |= FIEMAP_EXTENT_LAST; end = 1; } - if (em_start == last) { + /* now scan forward to see if this is really the last extent. */ + em = get_extent_skip_holes(inode, off, last_for_get_extent, + get_extent); + if (IS_ERR(em)) { + ret = PTR_ERR(em); + goto out; + } + if (!em) { flags |= FIEMAP_EXTENT_LAST; end = 1; } - - if (!hole) { - ret = fiemap_fill_next_extent(fieinfo, em_start, disko, - em_len, flags); - if (ret) - goto out_free; - } + ret = fiemap_fill_next_extent(fieinfo, em_start, disko, + em_len, flags); + if (ret) + goto out_free; } out_free: free_extent_map(em); @@ -3140,8 +3081,7 @@ static inline void btrfs_release_extent_buffer(struct extent_buffer *eb) struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, u64 start, unsigned long len, - struct page *page0, - gfp_t mask) + struct page *page0) { unsigned long num_pages = num_extent_pages(start, len); unsigned long i; @@ -3162,7 +3102,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, } rcu_read_unlock(); - eb = __alloc_extent_buffer(tree, start, len, mask); + eb = __alloc_extent_buffer(tree, start, len, GFP_NOFS); if (!eb) return NULL; @@ -3179,7 +3119,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, i = 0; } for (; i < num_pages; i++, index++) { - p = find_or_create_page(mapping, index, mask | __GFP_HIGHMEM); + p = find_or_create_page(mapping, index, GFP_NOFS | __GFP_HIGHMEM); if (!p) { WARN_ON(1); goto free_eb; @@ -3194,7 +3134,13 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, } if (!PageUptodate(p)) uptodate = 0; - unlock_page(p); + + /* + * see below about how we avoid a nasty race with release page + * and why we unlock later + */ + if (i != 0) + unlock_page(p); } if (uptodate) set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); @@ -3218,9 +3164,26 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, atomic_inc(&eb->refs); spin_unlock(&tree->buffer_lock); radix_tree_preload_end(); + + /* + * there is a race where release page may have + * tried to find this extent buffer in the radix + * but failed. It will tell the VM it is safe to + * reclaim the, and it will clear the page private bit. + * We must make sure to set the page private bit properly + * after the extent buffer is in the radix tree so + * it doesn't get lost + */ + set_page_extent_mapped(eb->first_page); + set_page_extent_head(eb->first_page, eb->len); + if (!page0) + unlock_page(eb->first_page); return eb; free_eb: + if (eb->first_page && !page0) + unlock_page(eb->first_page); + if (!atomic_dec_and_test(&eb->refs)) return exists; btrfs_release_extent_buffer(eb); @@ -3228,8 +3191,7 @@ free_eb: } struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, - u64 start, unsigned long len, - gfp_t mask) + u64 start, unsigned long len) { struct extent_buffer *eb; @@ -3271,10 +3233,11 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree, continue; lock_page(page); + WARN_ON(!PagePrivate(page)); + + set_page_extent_mapped(page); if (i == 0) set_page_extent_head(page, eb->len); - else - set_page_private(page, EXTENT_PAGE_PRIVATE); clear_page_dirty_for_io(page); spin_lock_irq(&page->mapping->tree_lock); @@ -3289,13 +3252,6 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree, return 0; } -int wait_on_extent_buffer_writeback(struct extent_io_tree *tree, - struct extent_buffer *eb) -{ - return wait_on_extent_writeback(tree, eb->start, - eb->start + eb->len - 1); -} - int set_extent_buffer_dirty(struct extent_io_tree *tree, struct extent_buffer *eb) { @@ -3341,7 +3297,7 @@ int set_extent_buffer_uptodate(struct extent_io_tree *tree, num_pages = num_extent_pages(eb->start, eb->len); set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1, - GFP_NOFS); + NULL, GFP_NOFS); for (i = 0; i < num_pages; i++) { page = extent_buffer_page(eb, i); if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) || @@ -3464,6 +3420,13 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, for (i = start_i; i < num_pages; i++) { page = extent_buffer_page(eb, i); + + WARN_ON(!PagePrivate(page)); + + set_page_extent_mapped(page); + if (i == 0) + set_page_extent_head(page, eb->len); + if (inc_all_pages) page_cache_get(page); if (!PageUptodate(page)) { @@ -3569,6 +3532,7 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, "wanted %lu %lu\n", (unsigned long long)eb->start, eb->len, start, min_len); WARN_ON(1); + return -EINVAL; } p = extent_buffer_page(eb, i); @@ -3761,6 +3725,12 @@ static void move_pages(struct page *dst_page, struct page *src_page, kunmap_atomic(dst_kaddr, KM_USER0); } +static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len) +{ + unsigned long distance = (src > dst) ? src - dst : dst - src; + return distance < len; +} + static void copy_pages(struct page *dst_page, struct page *src_page, unsigned long dst_off, unsigned long src_off, unsigned long len) @@ -3768,10 +3738,12 @@ static void copy_pages(struct page *dst_page, struct page *src_page, char *dst_kaddr = kmap_atomic(dst_page, KM_USER0); char *src_kaddr; - if (dst_page != src_page) + if (dst_page != src_page) { src_kaddr = kmap_atomic(src_page, KM_USER1); - else + } else { src_kaddr = dst_kaddr; + BUG_ON(areas_overlap(src_off, dst_off, len)); + } memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len); kunmap_atomic(dst_kaddr, KM_USER0); @@ -3846,7 +3818,7 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, "len %lu len %lu\n", dst_offset, len, dst->len); BUG_ON(1); } - if (dst_offset < src_offset) { + if (!areas_overlap(src_offset, dst_offset, len)) { memcpy_extent_buffer(dst, dst_offset, src_offset, len); return; } diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 7083cfafd061..4e8445a4757c 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -31,6 +31,7 @@ #define EXTENT_BUFFER_UPTODATE 0 #define EXTENT_BUFFER_BLOCKING 1 #define EXTENT_BUFFER_DIRTY 2 +#define EXTENT_BUFFER_CORRUPT 3 /* these are flags for extent_clear_unlock_delalloc */ #define EXTENT_CLEAR_UNLOCK_PAGE 0x1 @@ -152,23 +153,14 @@ static inline int extent_compress_type(unsigned long bio_flags) struct extent_map_tree; -static inline struct extent_state *extent_state_next(struct extent_state *state) -{ - struct rb_node *node; - node = rb_next(&state->rb_node); - if (!node) - return NULL; - return rb_entry(node, struct extent_state, rb_node); -} - typedef struct extent_map *(get_extent_t)(struct inode *inode, struct page *page, - size_t page_offset, + size_t pg_offset, u64 start, u64 len, int create); void extent_io_tree_init(struct extent_io_tree *tree, - struct address_space *mapping, gfp_t mask); + struct address_space *mapping); int try_release_extent_mapping(struct extent_map_tree *map, struct extent_io_tree *tree, struct page *page, gfp_t mask); @@ -191,7 +183,7 @@ void extent_io_exit(void); u64 count_range_bits(struct extent_io_tree *tree, u64 *start, u64 search_end, - u64 max_bytes, unsigned long bits); + u64 max_bytes, unsigned long bits, int contig); void free_extent_state(struct extent_state *state); int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end, @@ -207,21 +199,15 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits, int exclusive_bits, u64 *failed_start, struct extent_state **cached_state, gfp_t mask); int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask); + struct extent_state **cached_state, gfp_t mask); int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); -int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask); -int clear_extent_ordered_metadata(struct extent_io_tree *tree, u64 start, - u64 end, gfp_t mask); int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask); -int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask); int find_first_extent_bit(struct extent_io_tree *tree, u64 start, u64 *start_ret, u64 *end_ret, int bits); struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree, @@ -242,28 +228,17 @@ int extent_readpages(struct extent_io_tree *tree, struct address_space *mapping, struct list_head *pages, unsigned nr_pages, get_extent_t get_extent); -int extent_prepare_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to, get_extent_t *get_extent); -int extent_commit_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to); -sector_t extent_bmap(struct address_space *mapping, sector_t iblock, - get_extent_t *get_extent); int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, __u64 start, __u64 len, get_extent_t *get_extent); -int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end); int set_state_private(struct extent_io_tree *tree, u64 start, u64 private); int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private); void set_page_extent_mapped(struct page *page); struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, u64 start, unsigned long len, - struct page *page0, - gfp_t mask); + struct page *page0); struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, - u64 start, unsigned long len, - gfp_t mask); + u64 start, unsigned long len); void free_extent_buffer(struct extent_buffer *eb); int read_extent_buffer_pages(struct extent_io_tree *tree, struct extent_buffer *eb, u64 start, int wait, @@ -291,16 +266,11 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, unsigned long src_offset, unsigned long len); void memset_extent_buffer(struct extent_buffer *eb, char c, unsigned long start, unsigned long len); -int wait_on_extent_buffer_writeback(struct extent_io_tree *tree, - struct extent_buffer *eb); -int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end); int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits); int clear_extent_buffer_dirty(struct extent_io_tree *tree, struct extent_buffer *eb); int set_extent_buffer_dirty(struct extent_io_tree *tree, struct extent_buffer *eb); -int test_extent_buffer_dirty(struct extent_io_tree *tree, - struct extent_buffer *eb); int set_extent_buffer_uptodate(struct extent_io_tree *tree, struct extent_buffer *eb); int clear_extent_buffer_uptodate(struct extent_io_tree *tree, @@ -318,7 +288,6 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset, unsigned long *map_start, unsigned long *map_len, int km); void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km); -int release_extent_buffer_tail_pages(struct extent_buffer *eb); int extent_range_uptodate(struct extent_io_tree *tree, u64 start, u64 end); int extent_clear_unlock_delalloc(struct inode *inode, diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c index b0e1fce12530..2d0410344ea3 100644 --- a/fs/btrfs/extent_map.c +++ b/fs/btrfs/extent_map.c @@ -28,12 +28,11 @@ void extent_map_exit(void) /** * extent_map_tree_init - initialize extent map tree * @tree: tree to initialize - * @mask: flags for memory allocations during tree operations * * Initialize the extent tree @tree. Should be called for each new inode * or other user of the extent_map interface. */ -void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask) +void extent_map_tree_init(struct extent_map_tree *tree) { tree->map = RB_ROOT; rwlock_init(&tree->lock); @@ -41,18 +40,17 @@ void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask) /** * alloc_extent_map - allocate new extent map structure - * @mask: memory allocation flags * * Allocate a new extent_map structure. The new structure is * returned with a reference count of one and needs to be * freed using free_extent_map() */ -struct extent_map *alloc_extent_map(gfp_t mask) +struct extent_map *alloc_extent_map(void) { struct extent_map *em; - em = kmem_cache_alloc(extent_map_cache, mask); - if (!em || IS_ERR(em)) - return em; + em = kmem_cache_alloc(extent_map_cache, GFP_NOFS); + if (!em) + return NULL; em->in_tree = 0; em->flags = 0; em->compress_type = BTRFS_COMPRESS_NONE; @@ -243,7 +241,7 @@ out: * Insert @em into @tree or perform a simple forward/backward merge with * existing mappings. The extent_map struct passed in will be inserted * into the tree directly, with an additional reference taken, or a - * reference dropped if the merge attempt was successfull. + * reference dropped if the merge attempt was successful. */ int add_extent_mapping(struct extent_map_tree *tree, struct extent_map *em) diff --git a/fs/btrfs/extent_map.h b/fs/btrfs/extent_map.h index 28b44dbd1e35..33a7890b1f40 100644 --- a/fs/btrfs/extent_map.h +++ b/fs/btrfs/extent_map.h @@ -49,14 +49,14 @@ static inline u64 extent_map_block_end(struct extent_map *em) return em->block_start + em->block_len; } -void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask); +void extent_map_tree_init(struct extent_map_tree *tree); struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree, u64 start, u64 len); int add_extent_mapping(struct extent_map_tree *tree, struct extent_map *em); int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em); -struct extent_map *alloc_extent_map(gfp_t mask); +struct extent_map *alloc_extent_map(void); void free_extent_map(struct extent_map *em); int __init extent_map_init(void); void extent_map_exit(void); diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index 4f19a3e1bf32..90d4ee52cd45 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -48,7 +48,8 @@ int btrfs_insert_file_extent(struct btrfs_trans_handle *trans, struct extent_buffer *leaf; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; file_key.objectid = objectid; file_key.offset = pos; btrfs_set_key_type(&file_key, BTRFS_EXTENT_DATA_KEY); @@ -169,6 +170,8 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; if (bio->bi_size > PAGE_CACHE_SIZE * 8) path->reada = 2; @@ -190,7 +193,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, u32 item_size; if (item) - btrfs_release_path(root, path); + btrfs_release_path(path); item = btrfs_lookup_csum(NULL, root->fs_info->csum_root, path, disk_bytenr, 0); if (IS_ERR(item)) { @@ -205,12 +208,13 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, EXTENT_NODATASUM, GFP_NOFS); } else { printk(KERN_INFO "btrfs no csum found " - "for inode %lu start %llu\n", - inode->i_ino, + "for inode %llu start %llu\n", + (unsigned long long) + btrfs_ino(inode), (unsigned long long)offset); } item = NULL; - btrfs_release_path(root, path); + btrfs_release_path(path); goto found; } btrfs_item_key_to_cpu(path->nodes[0], &found_key, @@ -263,7 +267,7 @@ int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode, } int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, - struct list_head *list) + struct list_head *list, int search_commit) { struct btrfs_key key; struct btrfs_path *path; @@ -280,6 +284,12 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, path = btrfs_alloc_path(); BUG_ON(!path); + if (search_commit) { + path->skip_locking = 1; + path->reada = 2; + path->search_commit_root = 1; + } + key.objectid = BTRFS_EXTENT_CSUM_OBJECTID; key.offset = start; key.type = BTRFS_EXTENT_CSUM_KEY; @@ -492,7 +502,6 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, u32 new_size = (bytenr - key->offset) >> blocksize_bits; new_size *= csum_size; ret = btrfs_truncate_item(trans, root, path, new_size, 1); - BUG_ON(ret); } else if (key->offset >= bytenr && csum_end > end_byte && end_byte > key->offset) { /* @@ -505,7 +514,6 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, new_size *= csum_size; ret = btrfs_truncate_item(trans, root, path, new_size, 0); - BUG_ON(ret); key->offset = end_byte; ret = btrfs_set_item_key_safe(trans, root, path, key); @@ -548,10 +556,10 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret > 0) { if (path->slots[0] == 0) - goto out; + break; path->slots[0]--; } else if (ret < 0) { - goto out; + break; } leaf = path->nodes[0]; @@ -576,7 +584,8 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, /* delete the entire item, it is inside our range */ if (key.offset >= bytenr && csum_end <= end_byte) { ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); + if (ret) + goto out; if (key.offset == bytenr) break; } else if (key.offset < bytenr && csum_end > end_byte) { @@ -628,11 +637,12 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, if (key.offset < bytenr) break; } - btrfs_release_path(root, path); + btrfs_release_path(path); } + ret = 0; out: btrfs_free_path(path); - return 0; + return ret; } int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans, @@ -719,7 +729,7 @@ again: * at this point, we know the tree has an item, but it isn't big * enough yet to put our csum in. Grow it */ - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(trans, root, &file_key, path, csum_size, 1); if (ret < 0) @@ -758,12 +768,11 @@ again: goto insert; ret = btrfs_extend_item(trans, root, path, diff); - BUG_ON(ret); goto csum; } insert: - btrfs_release_path(root, path); + btrfs_release_path(path); csum_offset = 0; if (found_next) { u64 tmp = total_bytes + root->sectorsize; @@ -847,7 +856,7 @@ next_sector: } btrfs_mark_buffer_dirty(path->nodes[0]); if (total_bytes < sums->len) { - btrfs_release_path(root, path); + btrfs_release_path(path); cond_resched(); goto again; } diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index c1d3a818731a..c6a22d783c35 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -40,19 +40,276 @@ #include "locking.h" #include "compat.h" +/* + * when auto defrag is enabled we + * queue up these defrag structs to remember which + * inodes need defragging passes + */ +struct inode_defrag { + struct rb_node rb_node; + /* objectid */ + u64 ino; + /* + * transid where the defrag was added, we search for + * extents newer than this + */ + u64 transid; + + /* root objectid */ + u64 root; + + /* last offset we were able to defrag */ + u64 last_offset; + + /* if we've wrapped around back to zero once already */ + int cycled; +}; + +/* pop a record for an inode into the defrag tree. The lock + * must be held already + * + * If you're inserting a record for an older transid than an + * existing record, the transid already in the tree is lowered + * + * If an existing record is found the defrag item you + * pass in is freed + */ +static int __btrfs_add_inode_defrag(struct inode *inode, + struct inode_defrag *defrag) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + struct inode_defrag *entry; + struct rb_node **p; + struct rb_node *parent = NULL; + + p = &root->fs_info->defrag_inodes.rb_node; + while (*p) { + parent = *p; + entry = rb_entry(parent, struct inode_defrag, rb_node); + + if (defrag->ino < entry->ino) + p = &parent->rb_left; + else if (defrag->ino > entry->ino) + p = &parent->rb_right; + else { + /* if we're reinserting an entry for + * an old defrag run, make sure to + * lower the transid of our existing record + */ + if (defrag->transid < entry->transid) + entry->transid = defrag->transid; + if (defrag->last_offset > entry->last_offset) + entry->last_offset = defrag->last_offset; + goto exists; + } + } + BTRFS_I(inode)->in_defrag = 1; + rb_link_node(&defrag->rb_node, parent, p); + rb_insert_color(&defrag->rb_node, &root->fs_info->defrag_inodes); + return 0; + +exists: + kfree(defrag); + return 0; + +} + +/* + * insert a defrag record for this inode if auto defrag is + * enabled + */ +int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, + struct inode *inode) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + struct inode_defrag *defrag; + int ret = 0; + u64 transid; + + if (!btrfs_test_opt(root, AUTO_DEFRAG)) + return 0; + + if (root->fs_info->closing) + return 0; + + if (BTRFS_I(inode)->in_defrag) + return 0; + + if (trans) + transid = trans->transid; + else + transid = BTRFS_I(inode)->root->last_trans; + + defrag = kzalloc(sizeof(*defrag), GFP_NOFS); + if (!defrag) + return -ENOMEM; + + defrag->ino = inode->i_ino; + defrag->transid = transid; + defrag->root = root->root_key.objectid; + + spin_lock(&root->fs_info->defrag_inodes_lock); + if (!BTRFS_I(inode)->in_defrag) + ret = __btrfs_add_inode_defrag(inode, defrag); + spin_unlock(&root->fs_info->defrag_inodes_lock); + return ret; +} + +/* + * must be called with the defrag_inodes lock held + */ +struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, u64 ino, + struct rb_node **next) +{ + struct inode_defrag *entry = NULL; + struct rb_node *p; + struct rb_node *parent = NULL; + + p = info->defrag_inodes.rb_node; + while (p) { + parent = p; + entry = rb_entry(parent, struct inode_defrag, rb_node); + + if (ino < entry->ino) + p = parent->rb_left; + else if (ino > entry->ino) + p = parent->rb_right; + else + return entry; + } + + if (next) { + while (parent && ino > entry->ino) { + parent = rb_next(parent); + entry = rb_entry(parent, struct inode_defrag, rb_node); + } + *next = parent; + } + return NULL; +} + +/* + * run through the list of inodes in the FS that need + * defragging + */ +int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) +{ + struct inode_defrag *defrag; + struct btrfs_root *inode_root; + struct inode *inode; + struct rb_node *n; + struct btrfs_key key; + struct btrfs_ioctl_defrag_range_args range; + u64 first_ino = 0; + int num_defrag; + int defrag_batch = 1024; + + memset(&range, 0, sizeof(range)); + range.len = (u64)-1; + + atomic_inc(&fs_info->defrag_running); + spin_lock(&fs_info->defrag_inodes_lock); + while(1) { + n = NULL; + + /* find an inode to defrag */ + defrag = btrfs_find_defrag_inode(fs_info, first_ino, &n); + if (!defrag) { + if (n) + defrag = rb_entry(n, struct inode_defrag, rb_node); + else if (first_ino) { + first_ino = 0; + continue; + } else { + break; + } + } + + /* remove it from the rbtree */ + first_ino = defrag->ino + 1; + rb_erase(&defrag->rb_node, &fs_info->defrag_inodes); + + if (fs_info->closing) + goto next_free; + + spin_unlock(&fs_info->defrag_inodes_lock); + + /* get the inode */ + key.objectid = defrag->root; + btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY); + key.offset = (u64)-1; + inode_root = btrfs_read_fs_root_no_name(fs_info, &key); + if (IS_ERR(inode_root)) + goto next; + + key.objectid = defrag->ino; + btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); + key.offset = 0; + + inode = btrfs_iget(fs_info->sb, &key, inode_root, NULL); + if (IS_ERR(inode)) + goto next; + + /* do a chunk of defrag */ + BTRFS_I(inode)->in_defrag = 0; + range.start = defrag->last_offset; + num_defrag = btrfs_defrag_file(inode, NULL, &range, defrag->transid, + defrag_batch); + /* + * if we filled the whole defrag batch, there + * must be more work to do. Queue this defrag + * again + */ + if (num_defrag == defrag_batch) { + defrag->last_offset = range.start; + __btrfs_add_inode_defrag(inode, defrag); + /* + * we don't want to kfree defrag, we added it back to + * the rbtree + */ + defrag = NULL; + } else if (defrag->last_offset && !defrag->cycled) { + /* + * we didn't fill our defrag batch, but + * we didn't start at zero. Make sure we loop + * around to the start of the file. + */ + defrag->last_offset = 0; + defrag->cycled = 1; + __btrfs_add_inode_defrag(inode, defrag); + defrag = NULL; + } + + iput(inode); +next: + spin_lock(&fs_info->defrag_inodes_lock); +next_free: + kfree(defrag); + } + spin_unlock(&fs_info->defrag_inodes_lock); + + atomic_dec(&fs_info->defrag_running); + + /* + * during unmount, we use the transaction_wait queue to + * wait for the defragger to stop + */ + wake_up(&fs_info->transaction_wait); + return 0; +} /* simple helper to fault in pages and copy. This should go away * and be replaced with calls into generic code. */ static noinline int btrfs_copy_from_user(loff_t pos, int num_pages, - int write_bytes, + size_t write_bytes, struct page **prepared_pages, struct iov_iter *i) { size_t copied = 0; + size_t total_copied = 0; int pg = 0; int offset = pos & (PAGE_CACHE_SIZE - 1); - int total_copied = 0; while (write_bytes > 0) { size_t count = min_t(size_t, @@ -70,14 +327,26 @@ static noinline int btrfs_copy_from_user(loff_t pos, int num_pages, /* Flush processor's dcache for this page */ flush_dcache_page(page); + + /* + * if we get a partial write, we can end up with + * partially up to date pages. These add + * a lot of complexity, so make sure they don't + * happen by forcing this copy to be retried. + * + * The rest of the btrfs_file_write code will fall + * back to page at a time copies after we return 0. + */ + if (!PageUptodate(page) && copied < count) + copied = 0; + iov_iter_advance(i, copied); write_bytes -= copied; total_copied += copied; /* Return to btrfs_file_aio_write to fault page */ - if (unlikely(copied == 0)) { + if (unlikely(copied == 0)) break; - } if (unlikely(copied < PAGE_CACHE_SIZE - offset)) { offset += copied; @@ -92,12 +361,10 @@ static noinline int btrfs_copy_from_user(loff_t pos, int num_pages, /* * unlocks pages after btrfs_file_write is done with them */ -static noinline void btrfs_drop_pages(struct page **pages, size_t num_pages) +void btrfs_drop_pages(struct page **pages, size_t num_pages) { size_t i; for (i = 0; i < num_pages; i++) { - if (!pages[i]) - break; /* page checked is some magic around finding pages that * have been modified without going through btrfs_set_page_dirty * clear it here @@ -117,17 +384,13 @@ static noinline void btrfs_drop_pages(struct page **pages, size_t num_pages) * this also makes the decision about creating an inline extent vs * doing real data extents, marking pages dirty and delalloc as required. */ -static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct file *file, - struct page **pages, - size_t num_pages, - loff_t pos, - size_t write_bytes) +int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode, + struct page **pages, size_t num_pages, + loff_t pos, size_t write_bytes, + struct extent_state **cached) { int err = 0; int i; - struct inode *inode = fdentry(file)->d_inode; u64 num_bytes; u64 start_pos; u64 end_of_last_block; @@ -140,8 +403,9 @@ static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans, end_of_last_block = start_pos + num_bytes - 1; err = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block, - NULL); - BUG_ON(err); + cached); + if (err) + return err; for (i = 0; i < num_pages; i++) { struct page *p = pages[i]; @@ -149,13 +413,14 @@ static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans, ClearPageChecked(p); set_page_dirty(p); } - if (end_pos > isize) { + + /* + * we've only changed i_size in ram, and we haven't updated + * the disk i_size. There is no need to log the inode + * at this time. + */ + if (end_pos > isize) i_size_write(inode, end_pos); - /* we've only changed i_size in ram, and we haven't updated - * the disk i_size. There is no need to log the inode - * at this time. - */ - } return 0; } @@ -183,9 +448,10 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, } while (1) { if (!split) - split = alloc_extent_map(GFP_NOFS); + split = alloc_extent_map(); if (!split2) - split2 = alloc_extent_map(GFP_NOFS); + split2 = alloc_extent_map(); + BUG_ON(!split || !split2); write_lock(&em_tree->lock); em = lookup_extent_mapping(em_tree, start, len); @@ -289,6 +555,7 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, struct btrfs_path *path; struct btrfs_key key; struct btrfs_key new_key; + u64 ino = btrfs_ino(inode); u64 search_start = start; u64 disk_bytenr = 0; u64 num_bytes = 0; @@ -309,14 +576,14 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, while (1) { recow = 0; - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, ino, search_start, -1); if (ret < 0) break; if (ret > 0 && path->slots[0] > 0 && search_start == start) { leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0] - 1); - if (key.objectid == inode->i_ino && + if (key.objectid == ino && key.type == BTRFS_EXTENT_DATA_KEY) path->slots[0]--; } @@ -337,7 +604,7 @@ next_slot: } btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - if (key.objectid > inode->i_ino || + if (key.objectid > ino || key.type > BTRFS_EXTENT_DATA_KEY || key.offset >= end) break; @@ -367,7 +634,7 @@ next_slot: search_start = max(key.offset, start); if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } @@ -384,7 +651,7 @@ next_slot: ret = btrfs_duplicate_item(trans, root, path, &new_key); if (ret == -EAGAIN) { - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } if (ret < 0) @@ -507,7 +774,7 @@ next_slot: del_nr = 0; del_slot = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } @@ -583,6 +850,7 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans, int del_slot = 0; int recow; int ret; + u64 ino = btrfs_ino(inode); btrfs_drop_extent_cache(inode, start, end - 1, 0); @@ -591,18 +859,19 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans, again: recow = 0; split = start; - key.objectid = inode->i_ino; + key.objectid = ino; key.type = BTRFS_EXTENT_DATA_KEY; key.offset = split; ret = btrfs_search_slot(trans, root, &key, path, -1, 1); + if (ret < 0) + goto out; if (ret > 0 && path->slots[0] > 0) path->slots[0]--; leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - BUG_ON(key.objectid != inode->i_ino || - key.type != BTRFS_EXTENT_DATA_KEY); + BUG_ON(key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY); fi = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item); BUG_ON(btrfs_file_extent_type(leaf, fi) != @@ -619,7 +888,7 @@ again: other_start = 0; other_end = start; if (extent_mergeable(leaf, path->slots[0] - 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { new_key.offset = end; btrfs_set_item_key_safe(trans, root, path, &new_key); @@ -642,7 +911,7 @@ again: other_start = end; other_end = 0; if (extent_mergeable(leaf, path->slots[0] + 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { fi = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item); @@ -670,7 +939,7 @@ again: new_key.offset = split; ret = btrfs_duplicate_item(trans, root, path, &new_key); if (ret == -EAGAIN) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } BUG_ON(ret < 0); @@ -691,7 +960,7 @@ again: ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - inode->i_ino, orig_offset); + ino, orig_offset); BUG_ON(ret); if (split == start) { @@ -707,10 +976,10 @@ again: other_start = end; other_end = 0; if (extent_mergeable(leaf, path->slots[0] + 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } extent_end = other_end; @@ -718,16 +987,16 @@ again: del_nr++; ret = btrfs_free_extent(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - inode->i_ino, orig_offset); + ino, orig_offset); BUG_ON(ret); } other_start = 0; other_end = start; if (extent_mergeable(leaf, path->slots[0] - 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } key.offset = other_start; @@ -735,7 +1004,7 @@ again: del_nr++; ret = btrfs_free_extent(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - inode->i_ino, orig_offset); + ino, orig_offset); BUG_ON(ret); } if (del_nr == 0) { @@ -762,6 +1031,27 @@ out: } /* + * on error we return an unlocked page and the error value + * on success we return a locked page and 0 + */ +static int prepare_uptodate_page(struct page *page, u64 pos) +{ + int ret = 0; + + if ((pos & (PAGE_CACHE_SIZE - 1)) && !PageUptodate(page)) { + ret = btrfs_readpage(NULL, page); + if (ret) + return ret; + lock_page(page); + if (!PageUptodate(page)) { + unlock_page(page); + return -EIO; + } + } + return 0; +} + +/* * this gets pages into the page cache and locks them down, it also properly * waits for data=ordered extents to finish before allowing the pages to be * modified. @@ -776,6 +1066,7 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, unsigned long index = pos >> PAGE_CACHE_SHIFT; struct inode *inode = fdentry(file)->d_inode; int err = 0; + int faili = 0; u64 start_pos; u64 last_pos; @@ -783,25 +1074,33 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, last_pos = ((u64)index + num_pages) << PAGE_CACHE_SHIFT; if (start_pos > inode->i_size) { - err = btrfs_cont_expand(inode, start_pos); + err = btrfs_cont_expand(inode, i_size_read(inode), start_pos); if (err) return err; } - memset(pages, 0, num_pages * sizeof(struct page *)); again: for (i = 0; i < num_pages; i++) { pages[i] = grab_cache_page(inode->i_mapping, index + i); if (!pages[i]) { - int c; - for (c = i - 1; c >= 0; c--) { - unlock_page(pages[c]); - page_cache_release(pages[c]); - } - return -ENOMEM; + faili = i - 1; + err = -ENOMEM; + goto fail; + } + + if (i == 0) + err = prepare_uptodate_page(pages[i], pos); + if (i == num_pages - 1) + err = prepare_uptodate_page(pages[i], + pos + write_bytes); + if (err) { + page_cache_release(pages[i]); + faili = i - 1; + goto fail; } wait_on_page_writeback(pages[i]); } + err = 0; if (start_pos < inode->i_size) { struct btrfs_ordered_extent *ordered; lock_extent_bits(&BTRFS_I(inode)->io_tree, @@ -841,191 +1140,103 @@ again: WARN_ON(!PageLocked(pages[i])); } return 0; +fail: + while (faili >= 0) { + unlock_page(pages[faili]); + page_cache_release(pages[faili]); + faili--; + } + return err; + } -static ssize_t btrfs_file_aio_write(struct kiocb *iocb, - const struct iovec *iov, - unsigned long nr_segs, loff_t pos) +static noinline ssize_t __btrfs_buffered_write(struct file *file, + struct iov_iter *i, + loff_t pos) { - struct file *file = iocb->ki_filp; struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; - struct page *pinned[2]; struct page **pages = NULL; - struct iov_iter i; - loff_t *ppos = &iocb->ki_pos; - loff_t start_pos; - ssize_t num_written = 0; - ssize_t err = 0; - size_t count; - size_t ocount; - int ret = 0; - int nrptrs; unsigned long first_index; unsigned long last_index; - int will_write; - int buffered = 0; - int copied = 0; - int dirty_pages = 0; - - will_write = ((file->f_flags & O_DSYNC) || IS_SYNC(inode) || - (file->f_flags & O_DIRECT)); - - pinned[0] = NULL; - pinned[1] = NULL; - - start_pos = pos; - - vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE); - - mutex_lock(&inode->i_mutex); - - err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ); - if (err) - goto out; - count = ocount; - - current->backing_dev_info = inode->i_mapping->backing_dev_info; - err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode)); - if (err) - goto out; - - if (count == 0) - goto out; - - err = file_remove_suid(file); - if (err) - goto out; - - /* - * If BTRFS flips readonly due to some impossible error - * (fs_info->fs_state now has BTRFS_SUPER_FLAG_ERROR), - * although we have opened a file as writable, we have - * to stop this write operation to ensure FS consistency. - */ - if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { - err = -EROFS; - goto out; - } - - file_update_time(file); - BTRFS_I(inode)->sequence++; - - if (unlikely(file->f_flags & O_DIRECT)) { - num_written = generic_file_direct_write(iocb, iov, &nr_segs, - pos, ppos, count, - ocount); - /* - * the generic O_DIRECT will update in-memory i_size after the - * DIOs are done. But our endio handlers that update the on - * disk i_size never update past the in memory i_size. So we - * need one more update here to catch any additions to the - * file - */ - if (inode->i_size != BTRFS_I(inode)->disk_i_size) { - btrfs_ordered_update_i_size(inode, inode->i_size, NULL); - mark_inode_dirty(inode); - } - - if (num_written < 0) { - ret = num_written; - num_written = 0; - goto out; - } else if (num_written == count) { - /* pick up pos changes done by the generic code */ - pos = *ppos; - goto out; - } - /* - * We are going to do buffered for the rest of the range, so we - * need to make sure to invalidate the buffered pages when we're - * done. - */ - buffered = 1; - pos += num_written; - } + size_t num_written = 0; + int nrptrs; + int ret = 0; - iov_iter_init(&i, iov, nr_segs, count, num_written); - nrptrs = min((iov_iter_count(&i) + PAGE_CACHE_SIZE - 1) / + nrptrs = min((iov_iter_count(i) + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE, PAGE_CACHE_SIZE / (sizeof(struct page *))); pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL); - if (!pages) { - ret = -ENOMEM; - goto out; - } - - /* generic_write_checks can change our pos */ - start_pos = pos; + if (!pages) + return -ENOMEM; first_index = pos >> PAGE_CACHE_SHIFT; - last_index = (pos + iov_iter_count(&i)) >> PAGE_CACHE_SHIFT; + last_index = (pos + iov_iter_count(i)) >> PAGE_CACHE_SHIFT; - /* - * there are lots of better ways to do this, but this code - * makes sure the first and last page in the file range are - * up to date and ready for cow - */ - if ((pos & (PAGE_CACHE_SIZE - 1))) { - pinned[0] = grab_cache_page(inode->i_mapping, first_index); - if (!PageUptodate(pinned[0])) { - ret = btrfs_readpage(NULL, pinned[0]); - BUG_ON(ret); - wait_on_page_locked(pinned[0]); - } else { - unlock_page(pinned[0]); - } - } - if ((pos + iov_iter_count(&i)) & (PAGE_CACHE_SIZE - 1)) { - pinned[1] = grab_cache_page(inode->i_mapping, last_index); - if (!PageUptodate(pinned[1])) { - ret = btrfs_readpage(NULL, pinned[1]); - BUG_ON(ret); - wait_on_page_locked(pinned[1]); - } else { - unlock_page(pinned[1]); - } - } - - while (iov_iter_count(&i) > 0) { + while (iov_iter_count(i) > 0) { size_t offset = pos & (PAGE_CACHE_SIZE - 1); - size_t write_bytes = min(iov_iter_count(&i), + size_t write_bytes = min(iov_iter_count(i), nrptrs * (size_t)PAGE_CACHE_SIZE - offset); size_t num_pages = (write_bytes + offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + size_t dirty_pages; + size_t copied; WARN_ON(num_pages > nrptrs); - memset(pages, 0, sizeof(struct page *) * nrptrs); /* * Fault pages before locking them in prepare_pages * to avoid recursive lock */ - if (unlikely(iov_iter_fault_in_readable(&i, write_bytes))) { + if (unlikely(iov_iter_fault_in_readable(i, write_bytes))) { ret = -EFAULT; - goto out; + break; } ret = btrfs_delalloc_reserve_space(inode, num_pages << PAGE_CACHE_SHIFT); if (ret) - goto out; + break; + /* + * This is going to setup the pages array with the number of + * pages we want, so we don't really need to worry about the + * contents of pages from loop to loop + */ ret = prepare_pages(root, file, pages, num_pages, pos, first_index, last_index, write_bytes); if (ret) { btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT); - goto out; + break; } copied = btrfs_copy_from_user(pos, num_pages, - write_bytes, pages, &i); - dirty_pages = (copied + offset + PAGE_CACHE_SIZE - 1) >> - PAGE_CACHE_SHIFT; + write_bytes, pages, i); + /* + * if we have trouble faulting in the pages, fall + * back to one page at a time + */ + if (copied < write_bytes) + nrptrs = 1; + + if (copied == 0) + dirty_pages = 0; + else + dirty_pages = (copied + offset + + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; + + /* + * If we had a short copy we need to release the excess delaloc + * bytes we reserved. We need to increment outstanding_extents + * because btrfs_delalloc_release_space will decrement it, but + * we still have an outstanding extent for the chunk we actually + * managed to copy. + */ if (num_pages > dirty_pages) { if (copied > 0) atomic_inc( @@ -1036,43 +1247,157 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, } if (copied > 0) { - dirty_and_release_pages(NULL, root, file, pages, - dirty_pages, pos, copied); + ret = btrfs_dirty_pages(root, inode, pages, + dirty_pages, pos, copied, + NULL); + if (ret) { + btrfs_delalloc_release_space(inode, + dirty_pages << PAGE_CACHE_SHIFT); + btrfs_drop_pages(pages, num_pages); + break; + } } btrfs_drop_pages(pages, num_pages); - if (copied > 0) { - if (will_write) { - filemap_fdatawrite_range(inode->i_mapping, pos, - pos + copied - 1); - } else { - balance_dirty_pages_ratelimited_nr( - inode->i_mapping, - dirty_pages); - if (dirty_pages < - (root->leafsize >> PAGE_CACHE_SHIFT) + 1) - btrfs_btree_balance_dirty(root, 1); - btrfs_throttle(root); - } - } + cond_resched(); + + balance_dirty_pages_ratelimited_nr(inode->i_mapping, + dirty_pages); + if (dirty_pages < (root->leafsize >> PAGE_CACHE_SHIFT) + 1) + btrfs_btree_balance_dirty(root, 1); + btrfs_throttle(root); pos += copied; num_written += copied; + } - cond_resched(); + kfree(pages); + + return num_written ? num_written : ret; +} + +static ssize_t __btrfs_direct_write(struct kiocb *iocb, + const struct iovec *iov, + unsigned long nr_segs, loff_t pos, + loff_t *ppos, size_t count, size_t ocount) +{ + struct file *file = iocb->ki_filp; + struct inode *inode = fdentry(file)->d_inode; + struct iov_iter i; + ssize_t written; + ssize_t written_buffered; + loff_t endbyte; + int err; + + written = generic_file_direct_write(iocb, iov, &nr_segs, pos, ppos, + count, ocount); + + /* + * the generic O_DIRECT will update in-memory i_size after the + * DIOs are done. But our endio handlers that update the on + * disk i_size never update past the in memory i_size. So we + * need one more update here to catch any additions to the + * file + */ + if (inode->i_size != BTRFS_I(inode)->disk_i_size) { + btrfs_ordered_update_i_size(inode, inode->i_size, NULL); + mark_inode_dirty(inode); + } + + if (written < 0 || written == count) + return written; + + pos += written; + count -= written; + iov_iter_init(&i, iov, nr_segs, count, written); + written_buffered = __btrfs_buffered_write(file, &i, pos); + if (written_buffered < 0) { + err = written_buffered; + goto out; } + endbyte = pos + written_buffered - 1; + err = filemap_write_and_wait_range(file->f_mapping, pos, endbyte); + if (err) + goto out; + written += written_buffered; + *ppos = pos + written_buffered; + invalidate_mapping_pages(file->f_mapping, pos >> PAGE_CACHE_SHIFT, + endbyte >> PAGE_CACHE_SHIFT); out: - mutex_unlock(&inode->i_mutex); - if (ret) - err = ret; + return written ? written : err; +} - kfree(pages); - if (pinned[0]) - page_cache_release(pinned[0]); - if (pinned[1]) - page_cache_release(pinned[1]); - *ppos = pos; +static ssize_t btrfs_file_aio_write(struct kiocb *iocb, + const struct iovec *iov, + unsigned long nr_segs, loff_t pos) +{ + struct file *file = iocb->ki_filp; + struct inode *inode = fdentry(file)->d_inode; + struct btrfs_root *root = BTRFS_I(inode)->root; + loff_t *ppos = &iocb->ki_pos; + ssize_t num_written = 0; + ssize_t err = 0; + size_t count, ocount; + + vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE); + + mutex_lock(&inode->i_mutex); + + err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + count = ocount; + + current->backing_dev_info = inode->i_mapping->backing_dev_info; + err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode)); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + + if (count == 0) { + mutex_unlock(&inode->i_mutex); + goto out; + } + + err = file_remove_suid(file); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + + /* + * If BTRFS flips readonly due to some impossible error + * (fs_info->fs_state now has BTRFS_SUPER_FLAG_ERROR), + * although we have opened a file as writable, we have + * to stop this write operation to ensure FS consistency. + */ + if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { + mutex_unlock(&inode->i_mutex); + err = -EROFS; + goto out; + } + + file_update_time(file); + BTRFS_I(inode)->sequence++; + + if (unlikely(file->f_flags & O_DIRECT)) { + num_written = __btrfs_direct_write(iocb, iov, nr_segs, + pos, ppos, count, ocount); + } else { + struct iov_iter i; + + iov_iter_init(&i, iov, nr_segs, count, num_written); + + num_written = __btrfs_buffered_write(file, &i, pos); + if (num_written > 0) + *ppos = pos + num_written; + } + + mutex_unlock(&inode->i_mutex); /* * we want to make sure fsync finds this change @@ -1087,43 +1412,12 @@ out: * one running right now. */ BTRFS_I(inode)->last_trans = root->fs_info->generation + 1; - - if (num_written > 0 && will_write) { - struct btrfs_trans_handle *trans; - - err = btrfs_wait_ordered_range(inode, start_pos, num_written); - if (err) + if (num_written > 0 || num_written == -EIOCBQUEUED) { + err = generic_write_sync(file, pos, num_written); + if (err < 0 && num_written > 0) num_written = err; - - if ((file->f_flags & O_DSYNC) || IS_SYNC(inode)) { - trans = btrfs_start_transaction(root, 0); - if (IS_ERR(trans)) { - num_written = PTR_ERR(trans); - goto done; - } - mutex_lock(&inode->i_mutex); - ret = btrfs_log_dentry_safe(trans, root, - file->f_dentry); - mutex_unlock(&inode->i_mutex); - if (ret == 0) { - ret = btrfs_sync_log(trans, root); - if (ret == 0) - btrfs_end_transaction(trans, root); - else - btrfs_commit_transaction(trans, root); - } else if (ret != BTRFS_NO_LOG_SYNC) { - btrfs_commit_transaction(trans, root); - } else { - btrfs_end_transaction(trans, root); - } - } - if (file->f_flags & O_DIRECT && buffered) { - invalidate_mapping_pages(inode->i_mapping, - start_pos >> PAGE_CACHE_SHIFT, - (start_pos + num_written - 1) >> PAGE_CACHE_SHIFT); - } } -done: +out: current->backing_dev_info = NULL; return num_written ? num_written : err; } @@ -1166,6 +1460,7 @@ int btrfs_sync_file(struct file *file, int datasync) int ret = 0; struct btrfs_trans_handle *trans; + trace_btrfs_sync_file(file, datasync); /* we wait first, since the writeback may change the inode */ root->log_batch++; @@ -1293,7 +1588,8 @@ static long btrfs_fallocate(struct file *file, int mode, goto out; if (alloc_start > inode->i_size) { - ret = btrfs_cont_expand(inode, alloc_start); + ret = btrfs_cont_expand(inode, i_size_read(inode), + alloc_start); if (ret) goto out; } @@ -1337,7 +1633,7 @@ static long btrfs_fallocate(struct file *file, int mode, while (1) { em = btrfs_get_extent(inode, NULL, 0, cur_offset, alloc_end - cur_offset, 0); - BUG_ON(IS_ERR(em) || !em); + BUG_ON(IS_ERR_OR_NULL(em)); last_byte = min(extent_map_end(em), alloc_end); last_byte = (last_byte + mask) & ~mask; if (em->block_start == EXTENT_MAP_HOLE || diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index a0390657451b..70d45795d758 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -24,18 +24,18 @@ #include "free-space-cache.h" #include "transaction.h" #include "disk-io.h" +#include "extent_io.h" +#include "inode-map.h" #define BITS_PER_BITMAP (PAGE_CACHE_SIZE * 8) #define MAX_CACHE_BYTES_PER_GIG (32 * 1024) -static void recalculate_thresholds(struct btrfs_block_group_cache - *block_group); -static int link_free_space(struct btrfs_block_group_cache *block_group, +static int link_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info); -struct inode *lookup_free_space_inode(struct btrfs_root *root, - struct btrfs_block_group_cache - *block_group, struct btrfs_path *path) +static struct inode *__lookup_free_space_inode(struct btrfs_root *root, + struct btrfs_path *path, + u64 offset) { struct btrfs_key key; struct btrfs_key location; @@ -45,22 +45,15 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, struct inode *inode = NULL; int ret; - spin_lock(&block_group->lock); - if (block_group->inode) - inode = igrab(block_group->inode); - spin_unlock(&block_group->lock); - if (inode) - return inode; - key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) return ERR_PTR(ret); if (ret > 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return ERR_PTR(-ENOENT); } @@ -69,7 +62,7 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, struct btrfs_free_space_header); btrfs_free_space_key(leaf, header, &disk_key); btrfs_disk_key_to_cpu(&location, &disk_key); - btrfs_release_path(root, path); + btrfs_release_path(path); inode = btrfs_iget(root->fs_info->sb, &location, root, NULL); if (!inode) @@ -81,6 +74,29 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, return ERR_PTR(-ENOENT); } + inode->i_mapping->flags &= ~__GFP_FS; + + return inode; +} + +struct inode *lookup_free_space_inode(struct btrfs_root *root, + struct btrfs_block_group_cache + *block_group, struct btrfs_path *path) +{ + struct inode *inode = NULL; + + spin_lock(&block_group->lock); + if (block_group->inode) + inode = igrab(block_group->inode); + spin_unlock(&block_group->lock); + if (inode) + return inode; + + inode = __lookup_free_space_inode(root, path, + block_group->key.objectid); + if (IS_ERR(inode)) + return inode; + spin_lock(&block_group->lock); if (!root->fs_info->closing) { block_group->inode = igrab(inode); @@ -91,24 +107,18 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, return inode; } -int create_free_space_inode(struct btrfs_root *root, - struct btrfs_trans_handle *trans, - struct btrfs_block_group_cache *block_group, - struct btrfs_path *path) +int __create_free_space_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path, u64 ino, u64 offset) { struct btrfs_key key; struct btrfs_disk_key disk_key; struct btrfs_free_space_header *header; struct btrfs_inode_item *inode_item; struct extent_buffer *leaf; - u64 objectid; int ret; - ret = btrfs_find_free_objectid(trans, root, 0, &objectid); - if (ret < 0) - return ret; - - ret = btrfs_insert_empty_inode(trans, root, path, objectid); + ret = btrfs_insert_empty_inode(trans, root, path, ino); if (ret) return ret; @@ -128,19 +138,18 @@ int create_free_space_inode(struct btrfs_root *root, BTRFS_INODE_PREALLOC | BTRFS_INODE_NODATASUM); btrfs_set_inode_nlink(leaf, inode_item, 1); btrfs_set_inode_transid(leaf, inode_item, trans->transid); - btrfs_set_inode_block_group(leaf, inode_item, - block_group->key.objectid); + btrfs_set_inode_block_group(leaf, inode_item, offset); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(struct btrfs_free_space_header)); if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } leaf = path->nodes[0]; @@ -149,11 +158,27 @@ int create_free_space_inode(struct btrfs_root *root, memset_extent_buffer(leaf, 0, (unsigned long)header, sizeof(*header)); btrfs_set_free_space_key(leaf, header, &disk_key); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } +int create_free_space_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_block_group_cache *block_group, + struct btrfs_path *path) +{ + int ret; + u64 ino; + + ret = btrfs_find_free_objectid(root, &ino); + if (ret < 0) + return ret; + + return __create_free_space_inode(root, trans, path, ino, + block_group->key.objectid); +} + int btrfs_truncate_free_space_cache(struct btrfs_root *root, struct btrfs_trans_handle *trans, struct btrfs_path *path, @@ -184,7 +209,8 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, return ret; } - return btrfs_update_inode(trans, root, inode); + ret = btrfs_update_inode(trans, root, inode); + return ret; } static int readahead_cache(struct inode *inode) @@ -206,15 +232,13 @@ static int readahead_cache(struct inode *inode) return 0; } -int load_free_space_cache(struct btrfs_fs_info *fs_info, - struct btrfs_block_group_cache *block_group) +int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, + struct btrfs_free_space_ctl *ctl, + struct btrfs_path *path, u64 offset) { - struct btrfs_root *root = fs_info->tree_root; - struct inode *inode; struct btrfs_free_space_header *header; struct extent_buffer *leaf; struct page *page; - struct btrfs_path *path; u32 *checksums = NULL, *crc; char *disk_crcs = NULL; struct btrfs_key key; @@ -226,71 +250,43 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, pgoff_t index = 0; unsigned long first_page_offset; int num_checksums; - int ret = 0; - - /* - * If we're unmounting then just return, since this does a search on the - * normal root and not the commit root and we could deadlock. - */ - smp_mb(); - if (fs_info->closing) - return 0; - - /* - * If this block group has been marked to be cleared for one reason or - * another then we can't trust the on disk cache, so just return. - */ - spin_lock(&block_group->lock); - if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { - spin_unlock(&block_group->lock); - return 0; - } - spin_unlock(&block_group->lock); + int ret = 0, ret2; INIT_LIST_HEAD(&bitmaps); - path = btrfs_alloc_path(); - if (!path) - return 0; - - inode = lookup_free_space_inode(root, block_group, path); - if (IS_ERR(inode)) { - btrfs_free_path(path); - return 0; - } - /* Nothing in the space cache, goodbye */ - if (!i_size_read(inode)) { - btrfs_free_path(path); + if (!i_size_read(inode)) goto out; - } key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); - if (ret) { - btrfs_free_path(path); + if (ret < 0) + goto out; + else if (ret > 0) { + btrfs_release_path(path); + ret = 0; goto out; } + ret = -1; + leaf = path->nodes[0]; header = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_free_space_header); num_entries = btrfs_free_space_entries(leaf, header); num_bitmaps = btrfs_free_space_bitmaps(leaf, header); generation = btrfs_free_space_generation(leaf, header); - btrfs_free_path(path); + btrfs_release_path(path); if (BTRFS_I(inode)->generation != generation) { printk(KERN_ERR "btrfs: free space inode generation (%llu) did" - " not match free space cache generation (%llu) for " - "block group %llu\n", + " not match free space cache generation (%llu)\n", (unsigned long long)BTRFS_I(inode)->generation, - (unsigned long long)generation, - (unsigned long long)block_group->key.objectid); - goto free_cache; + (unsigned long long)generation); + goto out; } if (!num_entries) @@ -307,10 +303,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, goto out; ret = readahead_cache(inode); - if (ret) { - ret = 0; + if (ret) goto out; - } while (1) { struct btrfs_free_space_entry *entry; @@ -329,10 +323,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, } page = grab_cache_page(inode->i_mapping, index); - if (!page) { - ret = 0; + if (!page) goto free_cache; - } if (!PageUptodate(page)) { btrfs_readpage(NULL, page); @@ -341,9 +333,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, unlock_page(page); page_cache_release(page); printk(KERN_ERR "btrfs: error reading free " - "space cache: %llu\n", - (unsigned long long) - block_group->key.objectid); + "space cache\n"); goto free_cache; } } @@ -356,13 +346,10 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, gen = addr + (sizeof(u32) * num_checksums); if (*gen != BTRFS_I(inode)->generation) { printk(KERN_ERR "btrfs: space cache generation" - " (%llu) does not match inode (%llu) " - "for block group %llu\n", + " (%llu) does not match inode (%llu)\n", (unsigned long long)*gen, (unsigned long long) - BTRFS_I(inode)->generation, - (unsigned long long) - block_group->key.objectid); + BTRFS_I(inode)->generation); kunmap(page); unlock_page(page); page_cache_release(page); @@ -378,9 +365,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, PAGE_CACHE_SIZE - start_offset); btrfs_csum_final(cur_crc, (char *)&cur_crc); if (cur_crc != *crc) { - printk(KERN_ERR "btrfs: crc mismatch for page %lu in " - "block group %llu\n", index, - (unsigned long long)block_group->key.objectid); + printk(KERN_ERR "btrfs: crc mismatch for page %lu\n", + index); kunmap(page); unlock_page(page); page_cache_release(page); @@ -393,7 +379,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, break; need_loop = 1; - e = kzalloc(sizeof(struct btrfs_free_space), GFP_NOFS); + e = kmem_cache_zalloc(btrfs_free_space_cachep, + GFP_NOFS); if (!e) { kunmap(page); unlock_page(page); @@ -405,31 +392,32 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, e->bytes = le64_to_cpu(entry->bytes); if (!e->bytes) { kunmap(page); - kfree(e); + kmem_cache_free(btrfs_free_space_cachep, e); unlock_page(page); page_cache_release(page); goto free_cache; } if (entry->type == BTRFS_FREE_SPACE_EXTENT) { - spin_lock(&block_group->tree_lock); - ret = link_free_space(block_group, e); - spin_unlock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); + ret = link_free_space(ctl, e); + spin_unlock(&ctl->tree_lock); BUG_ON(ret); } else { e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); if (!e->bitmap) { kunmap(page); - kfree(e); + kmem_cache_free( + btrfs_free_space_cachep, e); unlock_page(page); page_cache_release(page); goto free_cache; } - spin_lock(&block_group->tree_lock); - ret = link_free_space(block_group, e); - block_group->total_bitmaps++; - recalculate_thresholds(block_group); - spin_unlock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); + ret2 = link_free_space(ctl, e); + ctl->total_bitmaps++; + ctl->op->recalc_thresholds(ctl); + spin_unlock(&ctl->tree_lock); list_add_tail(&e->list, &bitmaps); } @@ -469,85 +457,156 @@ next: out: kfree(checksums); kfree(disk_crcs); - iput(inode); return ret; - free_cache: - /* This cache is bogus, make sure it gets cleared */ + __btrfs_remove_free_space_cache(ctl); + goto out; +} + +int load_free_space_cache(struct btrfs_fs_info *fs_info, + struct btrfs_block_group_cache *block_group) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct btrfs_root *root = fs_info->tree_root; + struct inode *inode; + struct btrfs_path *path; + int ret; + bool matched; + u64 used = btrfs_block_group_used(&block_group->item); + + /* + * If we're unmounting then just return, since this does a search on the + * normal root and not the commit root and we could deadlock. + */ + smp_mb(); + if (fs_info->closing) + return 0; + + /* + * If this block group has been marked to be cleared for one reason or + * another then we can't trust the on disk cache, so just return. + */ spin_lock(&block_group->lock); - block_group->disk_cache_state = BTRFS_DC_CLEAR; + if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { + spin_unlock(&block_group->lock); + return 0; + } spin_unlock(&block_group->lock); - btrfs_remove_free_space_cache(block_group); - goto out; + + path = btrfs_alloc_path(); + if (!path) + return 0; + + inode = lookup_free_space_inode(root, block_group, path); + if (IS_ERR(inode)) { + btrfs_free_path(path); + return 0; + } + + ret = __load_free_space_cache(fs_info->tree_root, inode, ctl, + path, block_group->key.objectid); + btrfs_free_path(path); + if (ret <= 0) + goto out; + + spin_lock(&ctl->tree_lock); + matched = (ctl->free_space == (block_group->key.offset - used - + block_group->bytes_super)); + spin_unlock(&ctl->tree_lock); + + if (!matched) { + __btrfs_remove_free_space_cache(ctl); + printk(KERN_ERR "block group %llu has an wrong amount of free " + "space\n", block_group->key.objectid); + ret = -1; + } +out: + if (ret < 0) { + /* This cache is bogus, make sure it gets cleared */ + spin_lock(&block_group->lock); + block_group->disk_cache_state = BTRFS_DC_CLEAR; + spin_unlock(&block_group->lock); + ret = 0; + + printk(KERN_ERR "btrfs: failed to load free space cache " + "for block group %llu\n", block_group->key.objectid); + } + + iput(inode); + return ret; } -int btrfs_write_out_cache(struct btrfs_root *root, - struct btrfs_trans_handle *trans, - struct btrfs_block_group_cache *block_group, - struct btrfs_path *path) +int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, + struct btrfs_free_space_ctl *ctl, + struct btrfs_block_group_cache *block_group, + struct btrfs_trans_handle *trans, + struct btrfs_path *path, u64 offset) { struct btrfs_free_space_header *header; struct extent_buffer *leaf; - struct inode *inode; struct rb_node *node; struct list_head *pos, *n; + struct page **pages; struct page *page; struct extent_state *cached_state = NULL; + struct btrfs_free_cluster *cluster = NULL; + struct extent_io_tree *unpin = NULL; struct list_head bitmap_list; struct btrfs_key key; + u64 start, end, len; u64 bytes = 0; u32 *crc, *checksums; - pgoff_t index = 0, last_index = 0; unsigned long first_page_offset; - int num_checksums; + int index = 0, num_pages = 0; int entries = 0; int bitmaps = 0; - int ret = 0; - - root = root->fs_info->tree_root; + int ret = -1; + bool next_page = false; + bool out_of_space = false; INIT_LIST_HEAD(&bitmap_list); - spin_lock(&block_group->lock); - if (block_group->disk_cache_state < BTRFS_DC_SETUP) { - spin_unlock(&block_group->lock); - return 0; - } - spin_unlock(&block_group->lock); - - inode = lookup_free_space_inode(root, block_group, path); - if (IS_ERR(inode)) - return 0; - - if (!i_size_read(inode)) { - iput(inode); + node = rb_first(&ctl->free_space_offset); + if (!node) return 0; - } - node = rb_first(&block_group->free_space_offset); - if (!node) { - iput(inode); - return 0; - } + if (!i_size_read(inode)) + return -1; - last_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT; + num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; filemap_write_and_wait(inode->i_mapping); btrfs_wait_ordered_range(inode, inode->i_size & ~(root->sectorsize - 1), (u64)-1); /* We need a checksum per page. */ - num_checksums = i_size_read(inode) / PAGE_CACHE_SIZE; - crc = checksums = kzalloc(sizeof(u32) * num_checksums, GFP_NOFS); - if (!crc) { - iput(inode); - return 0; + crc = checksums = kzalloc(sizeof(u32) * num_pages, GFP_NOFS); + if (!crc) + return -1; + + pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS); + if (!pages) { + kfree(crc); + return -1; } /* Since the first page has all of our checksums and our generation we * need to calculate the offset into the page that we can start writing * our entries. */ - first_page_offset = (sizeof(u32) * num_checksums) + sizeof(u64); + first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64); + + /* Get the cluster for this block_group if it exists */ + if (block_group && !list_empty(&block_group->cluster_list)) + cluster = list_entry(block_group->cluster_list.next, + struct btrfs_free_cluster, + block_group_list); + + /* + * We shouldn't have switched the pinned extents yet so this is the + * right one + */ + unpin = root->fs_info->pinned_extents; /* * Lock all pages first so we can lock the extent safely. @@ -557,20 +616,18 @@ int btrfs_write_out_cache(struct btrfs_root *root, * after find_get_page at this point. Just putting this here so people * know and don't freak out. */ - while (index <= last_index) { + while (index < num_pages) { page = grab_cache_page(inode->i_mapping, index); if (!page) { - pgoff_t i = 0; + int i; - while (i < index) { - page = find_get_page(inode->i_mapping, i); - unlock_page(page); - page_cache_release(page); - page_cache_release(page); - i++; + for (i = 0; i < num_pages; i++) { + unlock_page(pages[i]); + page_cache_release(pages[i]); } goto out_free; } + pages[index] = page; index++; } @@ -578,6 +635,13 @@ int btrfs_write_out_cache(struct btrfs_root *root, lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, 0, &cached_state, GFP_NOFS); + /* + * When searching for pinned extents, we need to start at our start + * offset. + */ + if (block_group) + start = block_group->key.objectid; + /* Write out the extent entries */ do { struct btrfs_free_space_entry *entry; @@ -585,18 +649,25 @@ int btrfs_write_out_cache(struct btrfs_root *root, unsigned long offset = 0; unsigned long start_offset = 0; + next_page = false; + if (index == 0) { start_offset = first_page_offset; offset = start_offset; } - page = find_get_page(inode->i_mapping, index); + if (index >= num_pages) { + out_of_space = true; + break; + } + + page = pages[index]; addr = kmap(page); entry = addr + start_offset; memset(addr, 0, PAGE_CACHE_SIZE); - while (1) { + while (node && !next_page) { struct btrfs_free_space *e; e = rb_entry(node, struct btrfs_free_space, offset_index); @@ -612,12 +683,50 @@ int btrfs_write_out_cache(struct btrfs_root *root, entry->type = BTRFS_FREE_SPACE_EXTENT; } node = rb_next(node); - if (!node) - break; + if (!node && cluster) { + node = rb_first(&cluster->root); + cluster = NULL; + } offset += sizeof(struct btrfs_free_space_entry); if (offset + sizeof(struct btrfs_free_space_entry) >= PAGE_CACHE_SIZE) + next_page = true; + entry++; + } + + /* + * We want to add any pinned extents to our free space cache + * so we don't leak the space + */ + while (block_group && !next_page && + (start < block_group->key.objectid + + block_group->key.offset)) { + ret = find_first_extent_bit(unpin, start, &start, &end, + EXTENT_DIRTY); + if (ret) { + ret = 0; break; + } + + /* This pinned extent is out of our range */ + if (start >= block_group->key.objectid + + block_group->key.offset) + break; + + len = block_group->key.objectid + + block_group->key.offset - start; + len = min(len, end + 1 - start); + + entries++; + entry->offset = cpu_to_le64(start); + entry->bytes = cpu_to_le64(len); + entry->type = BTRFS_FREE_SPACE_EXTENT; + + start = end + 1; + offset += sizeof(struct btrfs_free_space_entry); + if (offset + sizeof(struct btrfs_free_space_entry) >= + PAGE_CACHE_SIZE) + next_page = true; entry++; } *crc = ~(u32)0; @@ -630,25 +739,8 @@ int btrfs_write_out_cache(struct btrfs_root *root, bytes += PAGE_CACHE_SIZE; - ClearPageChecked(page); - set_page_extent_mapped(page); - SetPageUptodate(page); - set_page_dirty(page); - - /* - * We need to release our reference we got for grab_cache_page, - * except for the first page which will hold our checksums, we - * do that below. - */ - if (index != 0) { - unlock_page(page); - page_cache_release(page); - } - - page_cache_release(page); - index++; - } while (node); + } while (node || next_page); /* Write out the bitmaps */ list_for_each_safe(pos, n, &bitmap_list) { @@ -656,7 +748,11 @@ int btrfs_write_out_cache(struct btrfs_root *root, struct btrfs_free_space *entry = list_entry(pos, struct btrfs_free_space, list); - page = find_get_page(inode->i_mapping, index); + if (index >= num_pages) { + out_of_space = true; + break; + } + page = pages[index]; addr = kmap(page); memcpy(addr, entry->bitmap, PAGE_CACHE_SIZE); @@ -667,73 +763,67 @@ int btrfs_write_out_cache(struct btrfs_root *root, crc++; bytes += PAGE_CACHE_SIZE; - ClearPageChecked(page); - set_page_extent_mapped(page); - SetPageUptodate(page); - set_page_dirty(page); - unlock_page(page); - page_cache_release(page); - page_cache_release(page); list_del_init(&entry->list); index++; } + if (out_of_space) { + btrfs_drop_pages(pages, num_pages); + unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, + i_size_read(inode) - 1, &cached_state, + GFP_NOFS); + ret = 0; + goto out_free; + } + /* Zero out the rest of the pages just to make sure */ - while (index <= last_index) { + while (index < num_pages) { void *addr; - page = find_get_page(inode->i_mapping, index); - + page = pages[index]; addr = kmap(page); memset(addr, 0, PAGE_CACHE_SIZE); kunmap(page); - ClearPageChecked(page); - set_page_extent_mapped(page); - SetPageUptodate(page); - set_page_dirty(page); - unlock_page(page); - page_cache_release(page); - page_cache_release(page); bytes += PAGE_CACHE_SIZE; index++; } - btrfs_set_extent_delalloc(inode, 0, bytes - 1, &cached_state); - /* Write the checksums and trans id to the first page */ { void *addr; u64 *gen; - page = find_get_page(inode->i_mapping, 0); + page = pages[0]; addr = kmap(page); - memcpy(addr, checksums, sizeof(u32) * num_checksums); - gen = addr + (sizeof(u32) * num_checksums); + memcpy(addr, checksums, sizeof(u32) * num_pages); + gen = addr + (sizeof(u32) * num_pages); *gen = trans->transid; kunmap(page); - ClearPageChecked(page); - set_page_extent_mapped(page); - SetPageUptodate(page); - set_page_dirty(page); - unlock_page(page); - page_cache_release(page); - page_cache_release(page); } - BTRFS_I(inode)->generation = trans->transid; + ret = btrfs_dirty_pages(root, inode, pages, num_pages, 0, + bytes, &cached_state); + btrfs_drop_pages(pages, num_pages); unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, &cached_state, GFP_NOFS); + if (ret) { + ret = 0; + goto out_free; + } + + BTRFS_I(inode)->generation = trans->transid; + filemap_write_and_wait(inode->i_mapping); key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_search_slot(trans, root, &key, path, 1, 1); if (ret < 0) { - ret = 0; + ret = -1; clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS); @@ -746,13 +836,13 @@ int btrfs_write_out_cache(struct btrfs_root *root, path->slots[0]--; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID || - found_key.offset != block_group->key.objectid) { - ret = 0; + found_key.offset != offset) { + ret = -1; clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS); - btrfs_release_path(root, path); + btrfs_release_path(path); goto out_free; } } @@ -762,48 +852,83 @@ int btrfs_write_out_cache(struct btrfs_root *root, btrfs_set_free_space_bitmaps(leaf, header, bitmaps); btrfs_set_free_space_generation(leaf, header, trans->transid); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = 1; out_free: - if (ret == 0) { + if (ret != 1) { invalidate_inode_pages2_range(inode->i_mapping, 0, index); - spin_lock(&block_group->lock); - block_group->disk_cache_state = BTRFS_DC_ERROR; - spin_unlock(&block_group->lock); BTRFS_I(inode)->generation = 0; } kfree(checksums); + kfree(pages); btrfs_update_inode(trans, root, inode); + return ret; +} + +int btrfs_write_out_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_block_group_cache *block_group, + struct btrfs_path *path) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct inode *inode; + int ret = 0; + + root = root->fs_info->tree_root; + + spin_lock(&block_group->lock); + if (block_group->disk_cache_state < BTRFS_DC_SETUP) { + spin_unlock(&block_group->lock); + return 0; + } + spin_unlock(&block_group->lock); + + inode = lookup_free_space_inode(root, block_group, path); + if (IS_ERR(inode)) + return 0; + + ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans, + path, block_group->key.objectid); + if (ret < 0) { + spin_lock(&block_group->lock); + block_group->disk_cache_state = BTRFS_DC_ERROR; + spin_unlock(&block_group->lock); + ret = 0; + + printk(KERN_ERR "btrfs: failed to write free space cace " + "for block group %llu\n", block_group->key.objectid); + } + iput(inode); return ret; } -static inline unsigned long offset_to_bit(u64 bitmap_start, u64 sectorsize, +static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit, u64 offset) { BUG_ON(offset < bitmap_start); offset -= bitmap_start; - return (unsigned long)(div64_u64(offset, sectorsize)); + return (unsigned long)(div_u64(offset, unit)); } -static inline unsigned long bytes_to_bits(u64 bytes, u64 sectorsize) +static inline unsigned long bytes_to_bits(u64 bytes, u32 unit) { - return (unsigned long)(div64_u64(bytes, sectorsize)); + return (unsigned long)(div_u64(bytes, unit)); } -static inline u64 offset_to_bitmap(struct btrfs_block_group_cache *block_group, +static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl, u64 offset) { u64 bitmap_start; u64 bytes_per_bitmap; - bytes_per_bitmap = BITS_PER_BITMAP * block_group->sectorsize; - bitmap_start = offset - block_group->key.objectid; + bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit; + bitmap_start = offset - ctl->start; bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap); bitmap_start *= bytes_per_bitmap; - bitmap_start += block_group->key.objectid; + bitmap_start += ctl->start; return bitmap_start; } @@ -861,10 +986,10 @@ static int tree_insert_offset(struct rb_root *root, u64 offset, * offset. */ static struct btrfs_free_space * -tree_search_offset(struct btrfs_block_group_cache *block_group, +tree_search_offset(struct btrfs_free_space_ctl *ctl, u64 offset, int bitmap_only, int fuzzy) { - struct rb_node *n = block_group->free_space_offset.rb_node; + struct rb_node *n = ctl->free_space_offset.rb_node; struct btrfs_free_space *entry, *prev = NULL; /* find entry that is closest to the 'offset' */ @@ -960,8 +1085,7 @@ tree_search_offset(struct btrfs_block_group_cache *block_group, break; } } - if (entry->offset + BITS_PER_BITMAP * - block_group->sectorsize > offset) + if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset) return entry; } else if (entry->offset + entry->bytes > offset) return entry; @@ -972,7 +1096,7 @@ tree_search_offset(struct btrfs_block_group_cache *block_group, while (1) { if (entry->bitmap) { if (entry->offset + BITS_PER_BITMAP * - block_group->sectorsize > offset) + ctl->unit > offset) break; } else { if (entry->offset + entry->bytes > offset) @@ -988,42 +1112,47 @@ tree_search_offset(struct btrfs_block_group_cache *block_group, } static inline void -__unlink_free_space(struct btrfs_block_group_cache *block_group, +__unlink_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info) { - rb_erase(&info->offset_index, &block_group->free_space_offset); - block_group->free_extents--; + rb_erase(&info->offset_index, &ctl->free_space_offset); + ctl->free_extents--; } -static void unlink_free_space(struct btrfs_block_group_cache *block_group, +static void unlink_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info) { - __unlink_free_space(block_group, info); - block_group->free_space -= info->bytes; + __unlink_free_space(ctl, info); + ctl->free_space -= info->bytes; } -static int link_free_space(struct btrfs_block_group_cache *block_group, +static int link_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info) { int ret = 0; BUG_ON(!info->bitmap && !info->bytes); - ret = tree_insert_offset(&block_group->free_space_offset, info->offset, + ret = tree_insert_offset(&ctl->free_space_offset, info->offset, &info->offset_index, (info->bitmap != NULL)); if (ret) return ret; - block_group->free_space += info->bytes; - block_group->free_extents++; + ctl->free_space += info->bytes; + ctl->free_extents++; return ret; } -static void recalculate_thresholds(struct btrfs_block_group_cache *block_group) +static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) { + struct btrfs_block_group_cache *block_group = ctl->private; u64 max_bytes; u64 bitmap_bytes; u64 extent_bytes; u64 size = block_group->key.offset; + u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize; + int max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg); + + BUG_ON(ctl->total_bitmaps > max_bitmaps); /* * The goal is to keep the total amount of memory used per 1gb of space @@ -1041,10 +1170,10 @@ static void recalculate_thresholds(struct btrfs_block_group_cache *block_group) * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as * we add more bitmaps. */ - bitmap_bytes = (block_group->total_bitmaps + 1) * PAGE_CACHE_SIZE; + bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_CACHE_SIZE; if (bitmap_bytes >= max_bytes) { - block_group->extents_thresh = 0; + ctl->extents_thresh = 0; return; } @@ -1055,47 +1184,43 @@ static void recalculate_thresholds(struct btrfs_block_group_cache *block_group) extent_bytes = max_bytes - bitmap_bytes; extent_bytes = min_t(u64, extent_bytes, div64_u64(max_bytes, 2)); - block_group->extents_thresh = + ctl->extents_thresh = div64_u64(extent_bytes, (sizeof(struct btrfs_free_space))); } -static void bitmap_clear_bits(struct btrfs_block_group_cache *block_group, +static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, u64 offset, u64 bytes) { - unsigned long start, end; - unsigned long i; + unsigned long start, count; - start = offset_to_bit(info->offset, block_group->sectorsize, offset); - end = start + bytes_to_bits(bytes, block_group->sectorsize); - BUG_ON(end > BITS_PER_BITMAP); + start = offset_to_bit(info->offset, ctl->unit, offset); + count = bytes_to_bits(bytes, ctl->unit); + BUG_ON(start + count > BITS_PER_BITMAP); - for (i = start; i < end; i++) - clear_bit(i, info->bitmap); + bitmap_clear(info->bitmap, start, count); info->bytes -= bytes; - block_group->free_space -= bytes; + ctl->free_space -= bytes; } -static void bitmap_set_bits(struct btrfs_block_group_cache *block_group, +static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, u64 offset, u64 bytes) { - unsigned long start, end; - unsigned long i; + unsigned long start, count; - start = offset_to_bit(info->offset, block_group->sectorsize, offset); - end = start + bytes_to_bits(bytes, block_group->sectorsize); - BUG_ON(end > BITS_PER_BITMAP); + start = offset_to_bit(info->offset, ctl->unit, offset); + count = bytes_to_bits(bytes, ctl->unit); + BUG_ON(start + count > BITS_PER_BITMAP); - for (i = start; i < end; i++) - set_bit(i, info->bitmap); + bitmap_set(info->bitmap, start, count); info->bytes += bytes; - block_group->free_space += bytes; + ctl->free_space += bytes; } -static int search_bitmap(struct btrfs_block_group_cache *block_group, +static int search_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *bitmap_info, u64 *offset, u64 *bytes) { @@ -1103,9 +1228,9 @@ static int search_bitmap(struct btrfs_block_group_cache *block_group, unsigned long bits, i; unsigned long next_zero; - i = offset_to_bit(bitmap_info->offset, block_group->sectorsize, + i = offset_to_bit(bitmap_info->offset, ctl->unit, max_t(u64, *offset, bitmap_info->offset)); - bits = bytes_to_bits(*bytes, block_group->sectorsize); + bits = bytes_to_bits(*bytes, ctl->unit); for (i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i); i < BITS_PER_BITMAP; @@ -1120,29 +1245,25 @@ static int search_bitmap(struct btrfs_block_group_cache *block_group, } if (found_bits) { - *offset = (u64)(i * block_group->sectorsize) + - bitmap_info->offset; - *bytes = (u64)(found_bits) * block_group->sectorsize; + *offset = (u64)(i * ctl->unit) + bitmap_info->offset; + *bytes = (u64)(found_bits) * ctl->unit; return 0; } return -1; } -static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache - *block_group, u64 *offset, - u64 *bytes, int debug) +static struct btrfs_free_space * +find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes) { struct btrfs_free_space *entry; struct rb_node *node; int ret; - if (!block_group->free_space_offset.rb_node) + if (!ctl->free_space_offset.rb_node) return NULL; - entry = tree_search_offset(block_group, - offset_to_bitmap(block_group, *offset), - 0, 1); + entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1); if (!entry) return NULL; @@ -1152,7 +1273,7 @@ static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache continue; if (entry->bitmap) { - ret = search_bitmap(block_group, entry, offset, bytes); + ret = search_bitmap(ctl, entry, offset, bytes); if (!ret) return entry; continue; @@ -1166,33 +1287,28 @@ static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache return NULL; } -static void add_new_bitmap(struct btrfs_block_group_cache *block_group, +static void add_new_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, u64 offset) { - u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize; - int max_bitmaps = (int)div64_u64(block_group->key.offset + - bytes_per_bg - 1, bytes_per_bg); - BUG_ON(block_group->total_bitmaps >= max_bitmaps); - - info->offset = offset_to_bitmap(block_group, offset); + info->offset = offset_to_bitmap(ctl, offset); info->bytes = 0; - link_free_space(block_group, info); - block_group->total_bitmaps++; + link_free_space(ctl, info); + ctl->total_bitmaps++; - recalculate_thresholds(block_group); + ctl->op->recalc_thresholds(ctl); } -static void free_bitmap(struct btrfs_block_group_cache *block_group, +static void free_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *bitmap_info) { - unlink_free_space(block_group, bitmap_info); + unlink_free_space(ctl, bitmap_info); kfree(bitmap_info->bitmap); - kfree(bitmap_info); - block_group->total_bitmaps--; - recalculate_thresholds(block_group); + kmem_cache_free(btrfs_free_space_cachep, bitmap_info); + ctl->total_bitmaps--; + ctl->op->recalc_thresholds(ctl); } -static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_group, +static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *bitmap_info, u64 *offset, u64 *bytes) { @@ -1201,8 +1317,7 @@ static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_gro int ret; again: - end = bitmap_info->offset + - (u64)(BITS_PER_BITMAP * block_group->sectorsize) - 1; + end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1; /* * XXX - this can go away after a few releases. @@ -1217,24 +1332,22 @@ again: search_start = *offset; search_bytes = *bytes; search_bytes = min(search_bytes, end - search_start + 1); - ret = search_bitmap(block_group, bitmap_info, &search_start, - &search_bytes); + ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes); BUG_ON(ret < 0 || search_start != *offset); if (*offset > bitmap_info->offset && *offset + *bytes > end) { - bitmap_clear_bits(block_group, bitmap_info, *offset, - end - *offset + 1); + bitmap_clear_bits(ctl, bitmap_info, *offset, end - *offset + 1); *bytes -= end - *offset + 1; *offset = end + 1; } else if (*offset >= bitmap_info->offset && *offset + *bytes <= end) { - bitmap_clear_bits(block_group, bitmap_info, *offset, *bytes); + bitmap_clear_bits(ctl, bitmap_info, *offset, *bytes); *bytes = 0; } if (*bytes) { struct rb_node *next = rb_next(&bitmap_info->offset_index); if (!bitmap_info->bytes) - free_bitmap(block_group, bitmap_info); + free_bitmap(ctl, bitmap_info); /* * no entry after this bitmap, but we still have bytes to @@ -1261,33 +1374,42 @@ again: */ search_start = *offset; search_bytes = *bytes; - ret = search_bitmap(block_group, bitmap_info, &search_start, + ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes); if (ret < 0 || search_start != *offset) return -EAGAIN; goto again; } else if (!bitmap_info->bytes) - free_bitmap(block_group, bitmap_info); + free_bitmap(ctl, bitmap_info); return 0; } -static int insert_into_bitmap(struct btrfs_block_group_cache *block_group, - struct btrfs_free_space *info) +static bool use_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) { - struct btrfs_free_space *bitmap_info; - int added = 0; - u64 bytes, offset, end; - int ret; + struct btrfs_block_group_cache *block_group = ctl->private; /* * If we are below the extents threshold then we can add this as an * extent, and don't have to deal with the bitmap */ - if (block_group->free_extents < block_group->extents_thresh && - info->bytes > block_group->sectorsize * 4) - return 0; + if (ctl->free_extents < ctl->extents_thresh) { + /* + * If this block group has some small extents we don't want to + * use up all of our free slots in the cache with them, we want + * to reserve them to larger extents, however if we have plent + * of cache left then go ahead an dadd them, no sense in adding + * the overhead of a bitmap if we don't have to. + */ + if (info->bytes <= block_group->sectorsize * 4) { + if (ctl->free_extents * 2 <= ctl->extents_thresh) + return false; + } else { + return false; + } + } /* * some block groups are so tiny they can't be enveloped by a bitmap, so @@ -1295,31 +1417,42 @@ static int insert_into_bitmap(struct btrfs_block_group_cache *block_group, */ if (BITS_PER_BITMAP * block_group->sectorsize > block_group->key.offset) - return 0; + return false; + + return true; +} + +static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) +{ + struct btrfs_free_space *bitmap_info; + int added = 0; + u64 bytes, offset, end; + int ret; bytes = info->bytes; offset = info->offset; + if (!ctl->op->use_bitmap(ctl, info)) + return 0; + again: - bitmap_info = tree_search_offset(block_group, - offset_to_bitmap(block_group, offset), + bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 1, 0); if (!bitmap_info) { BUG_ON(added); goto new_bitmap; } - end = bitmap_info->offset + - (u64)(BITS_PER_BITMAP * block_group->sectorsize); + end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit); if (offset >= bitmap_info->offset && offset + bytes > end) { - bitmap_set_bits(block_group, bitmap_info, offset, - end - offset); + bitmap_set_bits(ctl, bitmap_info, offset, end - offset); bytes -= end - offset; offset = end; added = 0; } else if (offset >= bitmap_info->offset && offset + bytes <= end) { - bitmap_set_bits(block_group, bitmap_info, offset, bytes); + bitmap_set_bits(ctl, bitmap_info, offset, bytes); bytes = 0; } else { BUG(); @@ -1333,19 +1466,19 @@ again: new_bitmap: if (info && info->bitmap) { - add_new_bitmap(block_group, info, offset); + add_new_bitmap(ctl, info, offset); added = 1; info = NULL; goto again; } else { - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); /* no pre-allocated info, allocate a new one */ if (!info) { - info = kzalloc(sizeof(struct btrfs_free_space), - GFP_NOFS); + info = kmem_cache_zalloc(btrfs_free_space_cachep, + GFP_NOFS); if (!info) { - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); ret = -ENOMEM; goto out; } @@ -1353,7 +1486,7 @@ new_bitmap: /* allocate the bitmap */ info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); if (!info->bitmap) { ret = -ENOMEM; goto out; @@ -1365,13 +1498,13 @@ out: if (info) { if (info->bitmap) kfree(info->bitmap); - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); } return ret; } -bool try_merge_free_space(struct btrfs_block_group_cache *block_group, +static bool try_merge_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, bool update_stat) { struct btrfs_free_space *left_info; @@ -1385,54 +1518,54 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group, * are adding, if there is remove that struct and add a new one to * cover the entire range */ - right_info = tree_search_offset(block_group, offset + bytes, 0, 0); + right_info = tree_search_offset(ctl, offset + bytes, 0, 0); if (right_info && rb_prev(&right_info->offset_index)) left_info = rb_entry(rb_prev(&right_info->offset_index), struct btrfs_free_space, offset_index); else - left_info = tree_search_offset(block_group, offset - 1, 0, 0); + left_info = tree_search_offset(ctl, offset - 1, 0, 0); if (right_info && !right_info->bitmap) { if (update_stat) - unlink_free_space(block_group, right_info); + unlink_free_space(ctl, right_info); else - __unlink_free_space(block_group, right_info); + __unlink_free_space(ctl, right_info); info->bytes += right_info->bytes; - kfree(right_info); + kmem_cache_free(btrfs_free_space_cachep, right_info); merged = true; } if (left_info && !left_info->bitmap && left_info->offset + left_info->bytes == offset) { if (update_stat) - unlink_free_space(block_group, left_info); + unlink_free_space(ctl, left_info); else - __unlink_free_space(block_group, left_info); + __unlink_free_space(ctl, left_info); info->offset = left_info->offset; info->bytes += left_info->bytes; - kfree(left_info); + kmem_cache_free(btrfs_free_space_cachep, left_info); merged = true; } return merged; } -int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, - u64 offset, u64 bytes) +int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl, + u64 offset, u64 bytes) { struct btrfs_free_space *info; int ret = 0; - info = kzalloc(sizeof(struct btrfs_free_space), GFP_NOFS); + info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS); if (!info) return -ENOMEM; info->offset = offset; info->bytes = bytes; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); - if (try_merge_free_space(block_group, info, true)) + if (try_merge_free_space(ctl, info, true)) goto link; /* @@ -1440,7 +1573,7 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, * extent then we know we're going to have to allocate a new extent, so * before we do that see if we need to drop this into a bitmap */ - ret = insert_into_bitmap(block_group, info); + ret = insert_into_bitmap(ctl, info); if (ret < 0) { goto out; } else if (ret) { @@ -1448,11 +1581,11 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, goto out; } link: - ret = link_free_space(block_group, info); + ret = link_free_space(ctl, info); if (ret) - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); out: - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); if (ret) { printk(KERN_CRIT "btrfs: unable to add free space :%d\n", ret); @@ -1465,21 +1598,21 @@ out: int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group, u64 offset, u64 bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *info; struct btrfs_free_space *next_info = NULL; int ret = 0; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); again: - info = tree_search_offset(block_group, offset, 0, 0); + info = tree_search_offset(ctl, offset, 0, 0); if (!info) { /* * oops didn't find an extent that matched the space we wanted * to remove, look for a bitmap instead */ - info = tree_search_offset(block_group, - offset_to_bitmap(block_group, offset), + info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 1, 0); if (!info) { WARN_ON(1); @@ -1494,8 +1627,8 @@ again: offset_index); if (next_info->bitmap) - end = next_info->offset + BITS_PER_BITMAP * - block_group->sectorsize - 1; + end = next_info->offset + + BITS_PER_BITMAP * ctl->unit - 1; else end = next_info->offset + next_info->bytes; @@ -1515,20 +1648,20 @@ again: } if (info->bytes == bytes) { - unlink_free_space(block_group, info); + unlink_free_space(ctl, info); if (info->bitmap) { kfree(info->bitmap); - block_group->total_bitmaps--; + ctl->total_bitmaps--; } - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); goto out_lock; } if (!info->bitmap && info->offset == offset) { - unlink_free_space(block_group, info); + unlink_free_space(ctl, info); info->offset += bytes; info->bytes -= bytes; - link_free_space(block_group, info); + link_free_space(ctl, info); goto out_lock; } @@ -1542,13 +1675,13 @@ again: * first unlink the old info and then * insert it again after the hole we're creating */ - unlink_free_space(block_group, info); + unlink_free_space(ctl, info); if (offset + bytes < info->offset + info->bytes) { u64 old_end = info->offset + info->bytes; info->offset = offset + bytes; info->bytes = old_end - info->offset; - ret = link_free_space(block_group, info); + ret = link_free_space(ctl, info); WARN_ON(ret); if (ret) goto out_lock; @@ -1556,9 +1689,9 @@ again: /* the hole we're creating ends at the end * of the info struct, just free the info */ - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); } - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); /* step two, insert a new info struct to cover * anything before the hole @@ -1569,12 +1702,12 @@ again: goto out; } - ret = remove_from_bitmap(block_group, info, &offset, &bytes); + ret = remove_from_bitmap(ctl, info, &offset, &bytes); if (ret == -EAGAIN) goto again; BUG_ON(ret); out_lock: - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); out: return ret; } @@ -1582,11 +1715,12 @@ out: void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, u64 bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *info; struct rb_node *n; int count = 0; - for (n = rb_first(&block_group->free_space_offset); n; n = rb_next(n)) { + for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) { info = rb_entry(n, struct btrfs_free_space, offset_index); if (info->bytes >= bytes) count++; @@ -1601,19 +1735,28 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, "\n", count); } -u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group) +static struct btrfs_free_space_op free_space_op = { + .recalc_thresholds = recalculate_thresholds, + .use_bitmap = use_bitmap, +}; + +void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group) { - struct btrfs_free_space *info; - struct rb_node *n; - u64 ret = 0; + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; - for (n = rb_first(&block_group->free_space_offset); n; - n = rb_next(n)) { - info = rb_entry(n, struct btrfs_free_space, offset_index); - ret += info->bytes; - } + spin_lock_init(&ctl->tree_lock); + ctl->unit = block_group->sectorsize; + ctl->start = block_group->key.objectid; + ctl->private = block_group; + ctl->op = &free_space_op; - return ret; + /* + * we only want to have 32k of ram per block group for keeping + * track of free space, and if we pass 1/2 of that we want to + * start converting things over to using bitmaps + */ + ctl->extents_thresh = ((1024 * 32) / 2) / + sizeof(struct btrfs_free_space); } /* @@ -1627,32 +1770,31 @@ __btrfs_return_cluster_to_free_space( struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry; struct rb_node *node; - bool bitmap; spin_lock(&cluster->lock); if (cluster->block_group != block_group) goto out; - bitmap = cluster->points_to_bitmap; cluster->block_group = NULL; cluster->window_start = 0; list_del_init(&cluster->block_group_list); - cluster->points_to_bitmap = false; - - if (bitmap) - goto out; node = rb_first(&cluster->root); while (node) { + bool bitmap; + entry = rb_entry(node, struct btrfs_free_space, offset_index); node = rb_next(&entry->offset_index); rb_erase(&entry->offset_index, &cluster->root); - BUG_ON(entry->bitmap); - try_merge_free_space(block_group, entry, false); - tree_insert_offset(&block_group->free_space_offset, - entry->offset, &entry->offset_index, 0); + + bitmap = (entry->bitmap != NULL); + if (!bitmap) + try_merge_free_space(ctl, entry, false); + tree_insert_offset(&ctl->free_space_offset, + entry->offset, &entry->offset_index, bitmap); } cluster->root = RB_ROOT; @@ -1662,14 +1804,38 @@ out: return 0; } -void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) +void __btrfs_remove_free_space_cache_locked(struct btrfs_free_space_ctl *ctl) { struct btrfs_free_space *info; struct rb_node *node; + + while ((node = rb_last(&ctl->free_space_offset)) != NULL) { + info = rb_entry(node, struct btrfs_free_space, offset_index); + unlink_free_space(ctl, info); + kfree(info->bitmap); + kmem_cache_free(btrfs_free_space_cachep, info); + if (need_resched()) { + spin_unlock(&ctl->tree_lock); + cond_resched(); + spin_lock(&ctl->tree_lock); + } + } +} + +void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl) +{ + spin_lock(&ctl->tree_lock); + __btrfs_remove_free_space_cache_locked(ctl); + spin_unlock(&ctl->tree_lock); +} + +void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_cluster *cluster; struct list_head *head; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); while ((head = block_group->cluster_list.next) != &block_group->cluster_list) { cluster = list_entry(head, struct btrfs_free_cluster, @@ -1678,57 +1844,46 @@ void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) WARN_ON(cluster->block_group != block_group); __btrfs_return_cluster_to_free_space(block_group, cluster); if (need_resched()) { - spin_unlock(&block_group->tree_lock); - cond_resched(); - spin_lock(&block_group->tree_lock); - } - } - - while ((node = rb_last(&block_group->free_space_offset)) != NULL) { - info = rb_entry(node, struct btrfs_free_space, offset_index); - unlink_free_space(block_group, info); - if (info->bitmap) - kfree(info->bitmap); - kfree(info); - if (need_resched()) { - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); cond_resched(); - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); } } + __btrfs_remove_free_space_cache_locked(ctl); + spin_unlock(&ctl->tree_lock); - spin_unlock(&block_group->tree_lock); } u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group, u64 offset, u64 bytes, u64 empty_size) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry = NULL; u64 bytes_search = bytes + empty_size; u64 ret = 0; - spin_lock(&block_group->tree_lock); - entry = find_free_space(block_group, &offset, &bytes_search, 0); + spin_lock(&ctl->tree_lock); + entry = find_free_space(ctl, &offset, &bytes_search); if (!entry) goto out; ret = offset; if (entry->bitmap) { - bitmap_clear_bits(block_group, entry, offset, bytes); + bitmap_clear_bits(ctl, entry, offset, bytes); if (!entry->bytes) - free_bitmap(block_group, entry); + free_bitmap(ctl, entry); } else { - unlink_free_space(block_group, entry); + unlink_free_space(ctl, entry); entry->offset += bytes; entry->bytes -= bytes; if (!entry->bytes) - kfree(entry); + kmem_cache_free(btrfs_free_space_cachep, entry); else - link_free_space(block_group, entry); + link_free_space(ctl, entry); } out: - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); return ret; } @@ -1745,6 +1900,7 @@ int btrfs_return_cluster_to_free_space( struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster) { + struct btrfs_free_space_ctl *ctl; int ret; /* first, get a safe pointer to the block group */ @@ -1763,10 +1919,12 @@ int btrfs_return_cluster_to_free_space( atomic_inc(&block_group->count); spin_unlock(&cluster->lock); + ctl = block_group->free_space_ctl; + /* now return any extents the cluster had on it */ - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); ret = __btrfs_return_cluster_to_free_space(block_group, cluster); - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); /* finally drop our ref */ btrfs_put_block_group(block_group); @@ -1775,50 +1933,24 @@ int btrfs_return_cluster_to_free_space( static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, + struct btrfs_free_space *entry, u64 bytes, u64 min_start) { - struct btrfs_free_space *entry; + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; int err; u64 search_start = cluster->window_start; u64 search_bytes = bytes; u64 ret = 0; - spin_lock(&block_group->tree_lock); - spin_lock(&cluster->lock); - - if (!cluster->points_to_bitmap) - goto out; - - if (cluster->block_group != block_group) - goto out; - - /* - * search_start is the beginning of the bitmap, but at some point it may - * be a good idea to point to the actual start of the free area in the - * bitmap, so do the offset_to_bitmap trick anyway, and set bitmap_only - * to 1 to make sure we get the bitmap entry - */ - entry = tree_search_offset(block_group, - offset_to_bitmap(block_group, search_start), - 1, 0); - if (!entry || !entry->bitmap) - goto out; - search_start = min_start; search_bytes = bytes; - err = search_bitmap(block_group, entry, &search_start, - &search_bytes); + err = search_bitmap(ctl, entry, &search_start, &search_bytes); if (err) - goto out; + return 0; ret = search_start; - bitmap_clear_bits(block_group, entry, ret, bytes); - if (entry->bytes == 0) - free_bitmap(block_group, entry); -out: - spin_unlock(&cluster->lock); - spin_unlock(&block_group->tree_lock); + bitmap_clear_bits(ctl, entry, ret, bytes); return ret; } @@ -1832,14 +1964,11 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, u64 bytes, u64 min_start) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry = NULL; struct rb_node *node; u64 ret = 0; - if (cluster->points_to_bitmap) - return btrfs_alloc_from_bitmap(block_group, cluster, bytes, - min_start); - spin_lock(&cluster->lock); if (bytes > cluster->max_size) goto out; @@ -1852,11 +1981,9 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, goto out; entry = rb_entry(node, struct btrfs_free_space, offset_index); - while(1) { - if (entry->bytes < bytes || entry->offset < min_start) { - struct rb_node *node; - + if (entry->bytes < bytes || + (!entry->bitmap && entry->offset < min_start)) { node = rb_next(&entry->offset_index); if (!node) break; @@ -1864,10 +1991,26 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, offset_index); continue; } - ret = entry->offset; - entry->offset += bytes; - entry->bytes -= bytes; + if (entry->bitmap) { + ret = btrfs_alloc_from_bitmap(block_group, + cluster, entry, bytes, + min_start); + if (ret == 0) { + node = rb_next(&entry->offset_index); + if (!node) + break; + entry = rb_entry(node, struct btrfs_free_space, + offset_index); + continue; + } + } else { + + ret = entry->offset; + + entry->offset += bytes; + entry->bytes -= bytes; + } if (entry->bytes == 0) rb_erase(&entry->offset_index, &cluster->root); @@ -1879,15 +2022,20 @@ out: if (!ret) return 0; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); - block_group->free_space -= bytes; + ctl->free_space -= bytes; if (entry->bytes == 0) { - block_group->free_extents--; - kfree(entry); + ctl->free_extents--; + if (entry->bitmap) { + kfree(entry->bitmap); + ctl->total_bitmaps--; + ctl->op->recalc_thresholds(ctl); + } + kmem_cache_free(btrfs_free_space_cachep, entry); } - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); return ret; } @@ -1897,6 +2045,7 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, u64 offset, u64 bytes, u64 min_bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; unsigned long next_zero; unsigned long i; unsigned long search_bits; @@ -1904,12 +2053,13 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, unsigned long found_bits; unsigned long start = 0; unsigned long total_found = 0; + int ret; bool found = false; i = offset_to_bit(entry->offset, block_group->sectorsize, max_t(u64, offset, entry->offset)); - search_bits = bytes_to_bits(min_bytes, block_group->sectorsize); - total_bits = bytes_to_bits(bytes, block_group->sectorsize); + search_bits = bytes_to_bits(bytes, block_group->sectorsize); + total_bits = bytes_to_bits(min_bytes, block_group->sectorsize); again: found_bits = 0; @@ -1926,7 +2076,7 @@ again: } if (!found_bits) - return -1; + return -ENOSPC; if (!found) { start = i; @@ -1950,192 +2100,212 @@ again: cluster->window_start = start * block_group->sectorsize + entry->offset; - cluster->points_to_bitmap = true; + rb_erase(&entry->offset_index, &ctl->free_space_offset); + ret = tree_insert_offset(&cluster->root, entry->offset, + &entry->offset_index, 1); + BUG_ON(ret); return 0; } /* - * here we try to find a cluster of blocks in a block group. The goal - * is to find at least bytes free and up to empty_size + bytes free. - * We might not find them all in one contiguous area. - * - * returns zero and sets up cluster if things worked out, otherwise - * it returns -enospc + * This searches the block group for just extents to fill the cluster with. */ -int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_group_cache *block_group, - struct btrfs_free_cluster *cluster, - u64 offset, u64 bytes, u64 empty_size) +static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, + struct btrfs_free_cluster *cluster, + u64 offset, u64 bytes, u64 min_bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct btrfs_free_space *first = NULL; struct btrfs_free_space *entry = NULL; + struct btrfs_free_space *prev = NULL; + struct btrfs_free_space *last; struct rb_node *node; - struct btrfs_free_space *next; - struct btrfs_free_space *last = NULL; - u64 min_bytes; u64 window_start; u64 window_free; - u64 max_extent = 0; - bool found_bitmap = false; - int ret; - - /* for metadata, allow allocates with more holes */ - if (btrfs_test_opt(root, SSD_SPREAD)) { - min_bytes = bytes + empty_size; - } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) { - /* - * we want to do larger allocations when we are - * flushing out the delayed refs, it helps prevent - * making more work as we go along. - */ - if (trans->transaction->delayed_refs.flushing) - min_bytes = max(bytes, (bytes + empty_size) >> 1); - else - min_bytes = max(bytes, (bytes + empty_size) >> 4); - } else - min_bytes = max(bytes, (bytes + empty_size) >> 2); + u64 max_extent; + u64 max_gap = 128 * 1024; - spin_lock(&block_group->tree_lock); - spin_lock(&cluster->lock); - - /* someone already found a cluster, hooray */ - if (cluster->block_group) { - ret = 0; - goto out; - } -again: - entry = tree_search_offset(block_group, offset, found_bitmap, 1); - if (!entry) { - ret = -ENOSPC; - goto out; - } + entry = tree_search_offset(ctl, offset, 0, 1); + if (!entry) + return -ENOSPC; /* - * If found_bitmap is true, we exhausted our search for extent entries, - * and we just want to search all of the bitmaps that we can find, and - * ignore any extent entries we find. + * We don't want bitmaps, so just move along until we find a normal + * extent entry. */ - while (entry->bitmap || found_bitmap || - (!entry->bitmap && entry->bytes < min_bytes)) { - struct rb_node *node = rb_next(&entry->offset_index); - - if (entry->bitmap && entry->bytes > bytes + empty_size) { - ret = btrfs_bitmap_cluster(block_group, entry, cluster, - offset, bytes + empty_size, - min_bytes); - if (!ret) - goto got_it; - } - - if (!node) { - ret = -ENOSPC; - goto out; - } + while (entry->bitmap) { + node = rb_next(&entry->offset_index); + if (!node) + return -ENOSPC; entry = rb_entry(node, struct btrfs_free_space, offset_index); } - /* - * We already searched all the extent entries from the passed in offset - * to the end and didn't find enough space for the cluster, and we also - * didn't find any bitmaps that met our criteria, just go ahead and exit - */ - if (found_bitmap) { - ret = -ENOSPC; - goto out; - } - - cluster->points_to_bitmap = false; window_start = entry->offset; window_free = entry->bytes; - last = entry; max_extent = entry->bytes; + first = entry; + last = entry; + prev = entry; - while (1) { - /* out window is just right, lets fill it */ - if (window_free >= bytes + empty_size) - break; - - node = rb_next(&last->offset_index); - if (!node) { - if (found_bitmap) - goto again; - ret = -ENOSPC; - goto out; - } - next = rb_entry(node, struct btrfs_free_space, offset_index); + while (window_free <= min_bytes) { + node = rb_next(&entry->offset_index); + if (!node) + return -ENOSPC; + entry = rb_entry(node, struct btrfs_free_space, offset_index); - /* - * we found a bitmap, so if this search doesn't result in a - * cluster, we know to go and search again for the bitmaps and - * start looking for space there - */ - if (next->bitmap) { - if (!found_bitmap) - offset = next->offset; - found_bitmap = true; - last = next; + if (entry->bitmap) continue; - } - /* * we haven't filled the empty size and the window is * very large. reset and try again */ - if (next->offset - (last->offset + last->bytes) > 128 * 1024 || - next->offset - window_start > (bytes + empty_size) * 2) { - entry = next; + if (entry->offset - (prev->offset + prev->bytes) > max_gap || + entry->offset - window_start > (min_bytes * 2)) { + first = entry; window_start = entry->offset; window_free = entry->bytes; last = entry; max_extent = entry->bytes; } else { - last = next; - window_free += next->bytes; + last = entry; + window_free += entry->bytes; if (entry->bytes > max_extent) max_extent = entry->bytes; } + prev = entry; } - cluster->window_start = entry->offset; + cluster->window_start = first->offset; + + node = &first->offset_index; /* * now we've found our entries, pull them out of the free space * cache and put them into the cluster rbtree - * - * The cluster includes an rbtree, but only uses the offset index - * of each free space cache entry. */ - while (1) { + do { + int ret; + + entry = rb_entry(node, struct btrfs_free_space, offset_index); node = rb_next(&entry->offset_index); - if (entry->bitmap && node) { - entry = rb_entry(node, struct btrfs_free_space, - offset_index); + if (entry->bitmap) continue; - } else if (entry->bitmap && !node) { - break; - } - rb_erase(&entry->offset_index, &block_group->free_space_offset); + rb_erase(&entry->offset_index, &ctl->free_space_offset); ret = tree_insert_offset(&cluster->root, entry->offset, &entry->offset_index, 0); BUG_ON(ret); + } while (node && entry != last); - if (!node || entry == last) - break; + cluster->max_size = max_extent; + + return 0; +} + +/* + * This specifically looks for bitmaps that may work in the cluster, we assume + * that we have already failed to find extents that will work. + */ +static int setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, + struct btrfs_free_cluster *cluster, + u64 offset, u64 bytes, u64 min_bytes) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct btrfs_free_space *entry; + struct rb_node *node; + int ret = -ENOSPC; + + if (ctl->total_bitmaps == 0) + return -ENOSPC; + entry = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 0, 1); + if (!entry) + return -ENOSPC; + + node = &entry->offset_index; + do { entry = rb_entry(node, struct btrfs_free_space, offset_index); + node = rb_next(&entry->offset_index); + if (!entry->bitmap) + continue; + if (entry->bytes < min_bytes) + continue; + ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset, + bytes, min_bytes); + } while (ret && node); + + return ret; +} + +/* + * here we try to find a cluster of blocks in a block group. The goal + * is to find at least bytes free and up to empty_size + bytes free. + * We might not find them all in one contiguous area. + * + * returns zero and sets up cluster if things worked out, otherwise + * it returns -enospc + */ +int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_block_group_cache *block_group, + struct btrfs_free_cluster *cluster, + u64 offset, u64 bytes, u64 empty_size) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + u64 min_bytes; + int ret; + + /* for metadata, allow allocates with more holes */ + if (btrfs_test_opt(root, SSD_SPREAD)) { + min_bytes = bytes + empty_size; + } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) { + /* + * we want to do larger allocations when we are + * flushing out the delayed refs, it helps prevent + * making more work as we go along. + */ + if (trans->transaction->delayed_refs.flushing) + min_bytes = max(bytes, (bytes + empty_size) >> 1); + else + min_bytes = max(bytes, (bytes + empty_size) >> 4); + } else + min_bytes = max(bytes, (bytes + empty_size) >> 2); + + spin_lock(&ctl->tree_lock); + + /* + * If we know we don't have enough space to make a cluster don't even + * bother doing all the work to try and find one. + */ + if (ctl->free_space < min_bytes) { + spin_unlock(&ctl->tree_lock); + return -ENOSPC; } - cluster->max_size = max_extent; -got_it: - ret = 0; - atomic_inc(&block_group->count); - list_add_tail(&cluster->block_group_list, &block_group->cluster_list); - cluster->block_group = block_group; + spin_lock(&cluster->lock); + + /* someone already found a cluster, hooray */ + if (cluster->block_group) { + ret = 0; + goto out; + } + + ret = setup_cluster_no_bitmap(block_group, cluster, offset, bytes, + min_bytes); + if (ret) + ret = setup_cluster_bitmap(block_group, cluster, offset, + bytes, min_bytes); + + if (!ret) { + atomic_inc(&block_group->count); + list_add_tail(&cluster->block_group_list, + &block_group->cluster_list); + cluster->block_group = block_group; + } out: spin_unlock(&cluster->lock); - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); return ret; } @@ -2149,8 +2319,239 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster) spin_lock_init(&cluster->refill_lock); cluster->root = RB_ROOT; cluster->max_size = 0; - cluster->points_to_bitmap = false; INIT_LIST_HEAD(&cluster->block_group_list); cluster->block_group = NULL; } +int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, + u64 *trimmed, u64 start, u64 end, u64 minlen) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct btrfs_free_space *entry = NULL; + struct btrfs_fs_info *fs_info = block_group->fs_info; + u64 bytes = 0; + u64 actually_trimmed; + int ret = 0; + + *trimmed = 0; + + while (start < end) { + spin_lock(&ctl->tree_lock); + + if (ctl->free_space < minlen) { + spin_unlock(&ctl->tree_lock); + break; + } + + entry = tree_search_offset(ctl, start, 0, 1); + if (!entry) + entry = tree_search_offset(ctl, + offset_to_bitmap(ctl, start), + 1, 1); + + if (!entry || entry->offset >= end) { + spin_unlock(&ctl->tree_lock); + break; + } + + if (entry->bitmap) { + ret = search_bitmap(ctl, entry, &start, &bytes); + if (!ret) { + if (start >= end) { + spin_unlock(&ctl->tree_lock); + break; + } + bytes = min(bytes, end - start); + bitmap_clear_bits(ctl, entry, start, bytes); + if (entry->bytes == 0) + free_bitmap(ctl, entry); + } else { + start = entry->offset + BITS_PER_BITMAP * + block_group->sectorsize; + spin_unlock(&ctl->tree_lock); + ret = 0; + continue; + } + } else { + start = entry->offset; + bytes = min(entry->bytes, end - start); + unlink_free_space(ctl, entry); + kmem_cache_free(btrfs_free_space_cachep, entry); + } + + spin_unlock(&ctl->tree_lock); + + if (bytes >= minlen) { + int update_ret; + update_ret = btrfs_update_reserved_bytes(block_group, + bytes, 1, 1); + + ret = btrfs_error_discard_extent(fs_info->extent_root, + start, + bytes, + &actually_trimmed); + + btrfs_add_free_space(block_group, start, bytes); + if (!update_ret) + btrfs_update_reserved_bytes(block_group, + bytes, 0, 1); + + if (ret) + break; + *trimmed += actually_trimmed; + } + start += bytes; + bytes = 0; + + if (fatal_signal_pending(current)) { + ret = -ERESTARTSYS; + break; + } + + cond_resched(); + } + + return ret; +} + +/* + * Find the left-most item in the cache tree, and then return the + * smallest inode number in the item. + * + * Note: the returned inode number may not be the smallest one in + * the tree, if the left-most item is a bitmap. + */ +u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root) +{ + struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl; + struct btrfs_free_space *entry = NULL; + u64 ino = 0; + + spin_lock(&ctl->tree_lock); + + if (RB_EMPTY_ROOT(&ctl->free_space_offset)) + goto out; + + entry = rb_entry(rb_first(&ctl->free_space_offset), + struct btrfs_free_space, offset_index); + + if (!entry->bitmap) { + ino = entry->offset; + + unlink_free_space(ctl, entry); + entry->offset++; + entry->bytes--; + if (!entry->bytes) + kmem_cache_free(btrfs_free_space_cachep, entry); + else + link_free_space(ctl, entry); + } else { + u64 offset = 0; + u64 count = 1; + int ret; + + ret = search_bitmap(ctl, entry, &offset, &count); + BUG_ON(ret); + + ino = offset; + bitmap_clear_bits(ctl, entry, offset, 1); + if (entry->bytes == 0) + free_bitmap(ctl, entry); + } +out: + spin_unlock(&ctl->tree_lock); + + return ino; +} + +struct inode *lookup_free_ino_inode(struct btrfs_root *root, + struct btrfs_path *path) +{ + struct inode *inode = NULL; + + spin_lock(&root->cache_lock); + if (root->cache_inode) + inode = igrab(root->cache_inode); + spin_unlock(&root->cache_lock); + if (inode) + return inode; + + inode = __lookup_free_space_inode(root, path, 0); + if (IS_ERR(inode)) + return inode; + + spin_lock(&root->cache_lock); + if (!root->fs_info->closing) + root->cache_inode = igrab(inode); + spin_unlock(&root->cache_lock); + + return inode; +} + +int create_free_ino_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path) +{ + return __create_free_space_inode(root, trans, path, + BTRFS_FREE_INO_OBJECTID, 0); +} + +int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_path *path; + struct inode *inode; + int ret = 0; + u64 root_gen = btrfs_root_generation(&root->root_item); + + /* + * If we're unmounting then just return, since this does a search on the + * normal root and not the commit root and we could deadlock. + */ + smp_mb(); + if (fs_info->closing) + return 0; + + path = btrfs_alloc_path(); + if (!path) + return 0; + + inode = lookup_free_ino_inode(root, path); + if (IS_ERR(inode)) + goto out; + + if (root_gen != BTRFS_I(inode)->generation) + goto out_put; + + ret = __load_free_space_cache(root, inode, ctl, path, 0); + + if (ret < 0) + printk(KERN_ERR "btrfs: failed to load free ino cache for " + "root %llu\n", root->root_key.objectid); +out_put: + iput(inode); +out: + btrfs_free_path(path); + return ret; +} + +int btrfs_write_out_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct inode *inode; + int ret; + + inode = lookup_free_ino_inode(root, path); + if (IS_ERR(inode)) + return 0; + + ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0); + if (ret < 0) + printk(KERN_ERR "btrfs: failed to write free ino cache " + "for root %llu\n", root->root_key.objectid); + + iput(inode); + return ret; +} diff --git a/fs/btrfs/free-space-cache.h b/fs/btrfs/free-space-cache.h index e49ca5c321b5..8f2613f779ed 100644 --- a/fs/btrfs/free-space-cache.h +++ b/fs/btrfs/free-space-cache.h @@ -27,6 +27,25 @@ struct btrfs_free_space { struct list_head list; }; +struct btrfs_free_space_ctl { + spinlock_t tree_lock; + struct rb_root free_space_offset; + u64 free_space; + int extents_thresh; + int free_extents; + int total_bitmaps; + int unit; + u64 start; + struct btrfs_free_space_op *op; + void *private; +}; + +struct btrfs_free_space_op { + void (*recalc_thresholds)(struct btrfs_free_space_ctl *ctl); + bool (*use_bitmap)(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info); +}; + struct inode *lookup_free_space_inode(struct btrfs_root *root, struct btrfs_block_group_cache *block_group, struct btrfs_path *path); @@ -45,17 +64,38 @@ int btrfs_write_out_cache(struct btrfs_root *root, struct btrfs_trans_handle *trans, struct btrfs_block_group_cache *block_group, struct btrfs_path *path); -int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, - u64 bytenr, u64 size); + +struct inode *lookup_free_ino_inode(struct btrfs_root *root, + struct btrfs_path *path); +int create_free_ino_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path); +int load_free_ino_cache(struct btrfs_fs_info *fs_info, + struct btrfs_root *root); +int btrfs_write_out_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path); + +void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group); +int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl, + u64 bytenr, u64 size); +static inline int +btrfs_add_free_space(struct btrfs_block_group_cache *block_group, + u64 bytenr, u64 size) +{ + return __btrfs_add_free_space(block_group->free_space_ctl, + bytenr, size); +} int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group, u64 bytenr, u64 size); +void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl); void btrfs_remove_free_space_cache(struct btrfs_block_group_cache - *block_group); + *block_group); u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group, u64 offset, u64 bytes, u64 empty_size); +u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root); void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, u64 bytes); -u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group); int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_block_group_cache *block_group, @@ -68,4 +108,6 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, int btrfs_return_cluster_to_free_space( struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster); +int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, + u64 *trimmed, u64 start, u64 end, u64 minlen); #endif diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c index 64f1150bb48d..baa74f3db691 100644 --- a/fs/btrfs/inode-item.c +++ b/fs/btrfs/inode-item.c @@ -130,7 +130,6 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans, item_size - (ptr + sub_item_len - item_start)); ret = btrfs_truncate_item(trans, root, path, item_size - sub_item_len, 1); - BUG_ON(ret); out: btrfs_free_path(path); return ret; @@ -167,7 +166,6 @@ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans, old_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]); ret = btrfs_extend_item(trans, root, path, ins_len); - BUG_ON(ret); ref = btrfs_item_ptr(path->nodes[0], path->slots[0], struct btrfs_inode_ref); ref = (struct btrfs_inode_ref *)((unsigned long)ref + old_size); diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c index c56eb5909172..3262cd17a12f 100644 --- a/fs/btrfs/inode-map.c +++ b/fs/btrfs/inode-map.c @@ -16,11 +16,446 @@ * Boston, MA 021110-1307, USA. */ +#include <linux/delay.h> +#include <linux/kthread.h> +#include <linux/pagemap.h> + #include "ctree.h" #include "disk-io.h" +#include "free-space-cache.h" +#include "inode-map.h" #include "transaction.h" -int btrfs_find_highest_inode(struct btrfs_root *root, u64 *objectid) +static int caching_kthread(void *data) +{ + struct btrfs_root *root = data; + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_key key; + struct btrfs_path *path; + struct extent_buffer *leaf; + u64 last = (u64)-1; + int slot; + int ret; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + /* Since the commit root is read-only, we can safely skip locking. */ + path->skip_locking = 1; + path->search_commit_root = 1; + path->reada = 2; + + key.objectid = BTRFS_FIRST_FREE_OBJECTID; + key.offset = 0; + key.type = BTRFS_INODE_ITEM_KEY; +again: + /* need to make sure the commit_root doesn't disappear */ + mutex_lock(&root->fs_commit_mutex); + + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + + while (1) { + smp_mb(); + if (fs_info->closing) + goto out; + + leaf = path->nodes[0]; + slot = path->slots[0]; + if (slot >= btrfs_header_nritems(leaf)) { + ret = btrfs_next_leaf(root, path); + if (ret < 0) + goto out; + else if (ret > 0) + break; + + if (need_resched() || + btrfs_transaction_in_commit(fs_info)) { + leaf = path->nodes[0]; + + if (btrfs_header_nritems(leaf) == 0) { + WARN_ON(1); + break; + } + + /* + * Save the key so we can advances forward + * in the next search. + */ + btrfs_item_key_to_cpu(leaf, &key, 0); + btrfs_release_path(path); + root->cache_progress = last; + mutex_unlock(&root->fs_commit_mutex); + schedule_timeout(1); + goto again; + } else + continue; + } + + btrfs_item_key_to_cpu(leaf, &key, slot); + + if (key.type != BTRFS_INODE_ITEM_KEY) + goto next; + + if (key.objectid >= root->highest_objectid) + break; + + if (last != (u64)-1 && last + 1 != key.objectid) { + __btrfs_add_free_space(ctl, last + 1, + key.objectid - last - 1); + wake_up(&root->cache_wait); + } + + last = key.objectid; +next: + path->slots[0]++; + } + + if (last < root->highest_objectid - 1) { + __btrfs_add_free_space(ctl, last + 1, + root->highest_objectid - last - 1); + } + + spin_lock(&root->cache_lock); + root->cached = BTRFS_CACHE_FINISHED; + spin_unlock(&root->cache_lock); + + root->cache_progress = (u64)-1; + btrfs_unpin_free_ino(root); +out: + wake_up(&root->cache_wait); + mutex_unlock(&root->fs_commit_mutex); + + btrfs_free_path(path); + + return ret; +} + +static void start_caching(struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct task_struct *tsk; + int ret; + u64 objectid; + + spin_lock(&root->cache_lock); + if (root->cached != BTRFS_CACHE_NO) { + spin_unlock(&root->cache_lock); + return; + } + + root->cached = BTRFS_CACHE_STARTED; + spin_unlock(&root->cache_lock); + + ret = load_free_ino_cache(root->fs_info, root); + if (ret == 1) { + spin_lock(&root->cache_lock); + root->cached = BTRFS_CACHE_FINISHED; + spin_unlock(&root->cache_lock); + return; + } + + /* + * It can be quite time-consuming to fill the cache by searching + * through the extent tree, and this can keep ino allocation path + * waiting. Therefore at start we quickly find out the highest + * inode number and we know we can use inode numbers which fall in + * [highest_ino + 1, BTRFS_LAST_FREE_OBJECTID]. + */ + ret = btrfs_find_free_objectid(root, &objectid); + if (!ret && objectid <= BTRFS_LAST_FREE_OBJECTID) { + __btrfs_add_free_space(ctl, objectid, + BTRFS_LAST_FREE_OBJECTID - objectid + 1); + } + + tsk = kthread_run(caching_kthread, root, "btrfs-ino-cache-%llu\n", + root->root_key.objectid); + BUG_ON(IS_ERR(tsk)); +} + +int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid) +{ +again: + *objectid = btrfs_find_ino_for_alloc(root); + + if (*objectid != 0) + return 0; + + start_caching(root); + + wait_event(root->cache_wait, + root->cached == BTRFS_CACHE_FINISHED || + root->free_ino_ctl->free_space > 0); + + if (root->cached == BTRFS_CACHE_FINISHED && + root->free_ino_ctl->free_space == 0) + return -ENOSPC; + else + goto again; +} + +void btrfs_return_ino(struct btrfs_root *root, u64 objectid) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_free_space_ctl *pinned = root->free_ino_pinned; +again: + if (root->cached == BTRFS_CACHE_FINISHED) { + __btrfs_add_free_space(ctl, objectid, 1); + } else { + /* + * If we are in the process of caching free ino chunks, + * to avoid adding the same inode number to the free_ino + * tree twice due to cross transaction, we'll leave it + * in the pinned tree until a transaction is committed + * or the caching work is done. + */ + + mutex_lock(&root->fs_commit_mutex); + spin_lock(&root->cache_lock); + if (root->cached == BTRFS_CACHE_FINISHED) { + spin_unlock(&root->cache_lock); + mutex_unlock(&root->fs_commit_mutex); + goto again; + } + spin_unlock(&root->cache_lock); + + start_caching(root); + + if (objectid <= root->cache_progress || + objectid > root->highest_objectid) + __btrfs_add_free_space(ctl, objectid, 1); + else + __btrfs_add_free_space(pinned, objectid, 1); + + mutex_unlock(&root->fs_commit_mutex); + } +} + +/* + * When a transaction is committed, we'll move those inode numbers which + * are smaller than root->cache_progress from pinned tree to free_ino tree, + * and others will just be dropped, because the commit root we were + * searching has changed. + * + * Must be called with root->fs_commit_mutex held + */ +void btrfs_unpin_free_ino(struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct rb_root *rbroot = &root->free_ino_pinned->free_space_offset; + struct btrfs_free_space *info; + struct rb_node *n; + u64 count; + + while (1) { + n = rb_first(rbroot); + if (!n) + break; + + info = rb_entry(n, struct btrfs_free_space, offset_index); + BUG_ON(info->bitmap); + + if (info->offset > root->cache_progress) + goto free; + else if (info->offset + info->bytes > root->cache_progress) + count = root->cache_progress - info->offset + 1; + else + count = info->bytes; + + __btrfs_add_free_space(ctl, info->offset, count); +free: + rb_erase(&info->offset_index, rbroot); + kfree(info); + } +} + +#define INIT_THRESHOLD (((1024 * 32) / 2) / sizeof(struct btrfs_free_space)) +#define INODES_PER_BITMAP (PAGE_CACHE_SIZE * 8) + +/* + * The goal is to keep the memory used by the free_ino tree won't + * exceed the memory if we use bitmaps only. + */ +static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) +{ + struct btrfs_free_space *info; + struct rb_node *n; + int max_ino; + int max_bitmaps; + + n = rb_last(&ctl->free_space_offset); + if (!n) { + ctl->extents_thresh = INIT_THRESHOLD; + return; + } + info = rb_entry(n, struct btrfs_free_space, offset_index); + + /* + * Find the maximum inode number in the filesystem. Note we + * ignore the fact that this can be a bitmap, because we are + * not doing precise calculation. + */ + max_ino = info->bytes - 1; + + max_bitmaps = ALIGN(max_ino, INODES_PER_BITMAP) / INODES_PER_BITMAP; + if (max_bitmaps <= ctl->total_bitmaps) { + ctl->extents_thresh = 0; + return; + } + + ctl->extents_thresh = (max_bitmaps - ctl->total_bitmaps) * + PAGE_CACHE_SIZE / sizeof(*info); +} + +/* + * We don't fall back to bitmap, if we are below the extents threshold + * or this chunk of inode numbers is a big one. + */ +static bool use_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) +{ + if (ctl->free_extents < ctl->extents_thresh || + info->bytes > INODES_PER_BITMAP / 10) + return false; + + return true; +} + +static struct btrfs_free_space_op free_ino_op = { + .recalc_thresholds = recalculate_thresholds, + .use_bitmap = use_bitmap, +}; + +static void pinned_recalc_thresholds(struct btrfs_free_space_ctl *ctl) +{ +} + +static bool pinned_use_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) +{ + /* + * We always use extents for two reasons: + * + * - The pinned tree is only used during the process of caching + * work. + * - Make code simpler. See btrfs_unpin_free_ino(). + */ + return false; +} + +static struct btrfs_free_space_op pinned_free_ino_op = { + .recalc_thresholds = pinned_recalc_thresholds, + .use_bitmap = pinned_use_bitmap, +}; + +void btrfs_init_free_ino_ctl(struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_free_space_ctl *pinned = root->free_ino_pinned; + + spin_lock_init(&ctl->tree_lock); + ctl->unit = 1; + ctl->start = 0; + ctl->private = NULL; + ctl->op = &free_ino_op; + + /* + * Initially we allow to use 16K of ram to cache chunks of + * inode numbers before we resort to bitmaps. This is somewhat + * arbitrary, but it will be adjusted in runtime. + */ + ctl->extents_thresh = INIT_THRESHOLD; + + spin_lock_init(&pinned->tree_lock); + pinned->unit = 1; + pinned->start = 0; + pinned->private = NULL; + pinned->extents_thresh = 0; + pinned->op = &pinned_free_ino_op; +} + +int btrfs_save_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_path *path; + struct inode *inode; + u64 alloc_hint = 0; + int ret; + int prealloc; + bool retry = false; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; +again: + inode = lookup_free_ino_inode(root, path); + if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { + ret = PTR_ERR(inode); + goto out; + } + + if (IS_ERR(inode)) { + BUG_ON(retry); + retry = true; + + ret = create_free_ino_inode(root, trans, path); + if (ret) + goto out; + goto again; + } + + BTRFS_I(inode)->generation = 0; + ret = btrfs_update_inode(trans, root, inode); + WARN_ON(ret); + + if (i_size_read(inode) > 0) { + ret = btrfs_truncate_free_space_cache(root, trans, path, inode); + if (ret) + goto out_put; + } + + spin_lock(&root->cache_lock); + if (root->cached != BTRFS_CACHE_FINISHED) { + ret = -1; + spin_unlock(&root->cache_lock); + goto out_put; + } + spin_unlock(&root->cache_lock); + + spin_lock(&ctl->tree_lock); + prealloc = sizeof(struct btrfs_free_space) * ctl->free_extents; + prealloc = ALIGN(prealloc, PAGE_CACHE_SIZE); + prealloc += ctl->total_bitmaps * PAGE_CACHE_SIZE; + spin_unlock(&ctl->tree_lock); + + /* Just to make sure we have enough space */ + prealloc += 8 * PAGE_CACHE_SIZE; + + ret = btrfs_check_data_free_space(inode, prealloc); + if (ret) + goto out_put; + + ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc, + prealloc, prealloc, &alloc_hint); + if (ret) + goto out_put; + btrfs_free_reserved_data_space(inode, prealloc); + +out_put: + iput(inode); +out: + if (ret == 0) + ret = btrfs_write_out_ino_cache(root, trans, path); + + btrfs_free_path(path); + return ret; +} + +static int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid) { struct btrfs_path *path; int ret; @@ -30,7 +465,8 @@ int btrfs_find_highest_inode(struct btrfs_root *root, u64 *objectid) int slot; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; search_key.objectid = BTRFS_LAST_FREE_OBJECTID; search_key.type = -1; @@ -54,15 +490,14 @@ error: return ret; } -int btrfs_find_free_objectid(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - u64 dirid, u64 *objectid) +int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid) { int ret; mutex_lock(&root->objectid_mutex); if (unlikely(root->highest_objectid < BTRFS_FIRST_FREE_OBJECTID)) { - ret = btrfs_find_highest_inode(root, &root->highest_objectid); + ret = btrfs_find_highest_objectid(root, + &root->highest_objectid); if (ret) goto out; } diff --git a/fs/btrfs/inode-map.h b/fs/btrfs/inode-map.h new file mode 100644 index 000000000000..ddb347bfee23 --- /dev/null +++ b/fs/btrfs/inode-map.h @@ -0,0 +1,13 @@ +#ifndef __BTRFS_INODE_MAP +#define __BTRFS_INODE_MAP + +void btrfs_init_free_ino_ctl(struct btrfs_root *root); +void btrfs_unpin_free_ino(struct btrfs_root *root); +void btrfs_return_ino(struct btrfs_root *root, u64 objectid); +int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid); +int btrfs_save_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans); + +int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid); + +#endif diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index bcc461a9695f..39a9d5750efd 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -37,6 +37,7 @@ #include <linux/posix_acl.h> #include <linux/falloc.h> #include <linux/slab.h> +#include <linux/ratelimit.h> #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -50,6 +51,8 @@ #include "tree-log.h" #include "compression.h" #include "locking.h" +#include "free-space-cache.h" +#include "inode-map.h" struct btrfs_iget_args { u64 ino; @@ -70,6 +73,7 @@ static struct kmem_cache *btrfs_inode_cachep; struct kmem_cache *btrfs_trans_handle_cachep; struct kmem_cache *btrfs_transaction_cachep; struct kmem_cache *btrfs_path_cachep; +struct kmem_cache *btrfs_free_space_cachep; #define S_SHIFT 12 static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { @@ -82,7 +86,8 @@ static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { [S_IFLNK >> S_SHIFT] = BTRFS_FT_SYMLINK, }; -static void btrfs_truncate(struct inode *inode); +static int btrfs_setsize(struct inode *inode, loff_t newsize); +static int btrfs_truncate(struct inode *inode); static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end); static noinline int cow_file_range(struct inode *inode, struct page *locked_page, @@ -90,13 +95,14 @@ static noinline int cow_file_range(struct inode *inode, unsigned long *nr_written, int unlock); static int btrfs_init_inode_security(struct btrfs_trans_handle *trans, - struct inode *inode, struct inode *dir) + struct inode *inode, struct inode *dir, + const struct qstr *qstr) { int err; err = btrfs_init_acl(trans, inode, dir); if (!err) - err = btrfs_xattr_security_init(trans, inode, dir); + err = btrfs_xattr_security_init(trans, inode, dir, qstr); return err; } @@ -108,6 +114,7 @@ static int btrfs_init_inode_security(struct btrfs_trans_handle *trans, static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode, u64 start, size_t size, size_t compressed_size, + int compress_type, struct page **compressed_pages) { struct btrfs_key key; @@ -122,12 +129,9 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, size_t cur_size = size; size_t datasize; unsigned long offset; - int compress_type = BTRFS_COMPRESS_NONE; - if (compressed_size && compressed_pages) { - compress_type = root->fs_info->compress_type; + if (compressed_size && compressed_pages) cur_size = compressed_size; - } path = btrfs_alloc_path(); if (!path) @@ -136,7 +140,7 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, path->leave_spinning = 1; btrfs_set_trans_block_group(trans, inode); - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); key.offset = start; btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY); datasize = btrfs_file_extent_calc_inline_size(cur_size); @@ -217,7 +221,7 @@ fail: static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode, u64 start, u64 end, - size_t compressed_size, + size_t compressed_size, int compress_type, struct page **compressed_pages) { u64 isize = i_size_read(inode); @@ -250,7 +254,7 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans, inline_len = min_t(u64, isize, actual_end); ret = insert_inline_extent(trans, root, inode, start, inline_len, compressed_size, - compressed_pages); + compress_type, compressed_pages); BUG_ON(ret); btrfs_delalloc_release_metadata(inode, end + 1 - start); btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0); @@ -287,6 +291,7 @@ static noinline int add_async_extent(struct async_cow *cow, struct async_extent *async_extent; async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS); + BUG_ON(!async_extent); async_extent->start = start; async_extent->ram_size = ram_size; async_extent->compressed_size = compressed_size; @@ -337,6 +342,10 @@ static noinline int compress_file_range(struct inode *inode, int will_compress; int compress_type = root->fs_info->compress_type; + /* if this is a small write inside eof, kick off a defragbot */ + if (end <= BTRFS_I(inode)->disk_i_size && (end - start + 1) < 16 * 1024) + btrfs_add_inode_defrag(NULL, inode); + actual_end = min_t(u64, isize, end + 1); again: will_compress = 0; @@ -381,9 +390,11 @@ again: */ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) && (btrfs_test_opt(root, COMPRESS) || - (BTRFS_I(inode)->force_compress))) { + (BTRFS_I(inode)->force_compress) || + (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) { WARN_ON(pages); pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); + BUG_ON(!pages); if (BTRFS_I(inode)->force_compress) compress_type = BTRFS_I(inode)->force_compress; @@ -426,12 +437,13 @@ again: * to make an uncompressed inline extent. */ ret = cow_file_range_inline(trans, root, inode, - start, end, 0, NULL); + start, end, 0, 0, NULL); } else { /* try making a compressed inline extent */ ret = cow_file_range_inline(trans, root, inode, start, end, - total_compressed, pages); + total_compressed, + compress_type, pages); } if (ret == 0) { /* @@ -643,7 +655,8 @@ retry: async_extent->start + async_extent->ram_size - 1, 0); - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); + BUG_ON(!em); em->start = async_extent->start; em->len = async_extent->ram_size; em->orig_start = em->start; @@ -738,6 +751,15 @@ static u64 get_extent_allocation_hint(struct inode *inode, u64 start, return alloc_hint; } +static inline bool is_free_space_inode(struct btrfs_root *root, + struct inode *inode) +{ + if (root == root->fs_info->tree_root || + BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) + return true; + return false; +} + /* * when extent_io.c finds a delayed allocation range in the file, * the call backs end up in this code. The basic idea is to @@ -770,7 +792,7 @@ static noinline int cow_file_range(struct inode *inode, struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; int ret = 0; - BUG_ON(root == root->fs_info->tree_root); + BUG_ON(is_free_space_inode(root, inode)); trans = btrfs_join_transaction(root, 1); BUG_ON(IS_ERR(trans)); btrfs_set_trans_block_group(trans, inode); @@ -781,10 +803,14 @@ static noinline int cow_file_range(struct inode *inode, disk_num_bytes = num_bytes; ret = 0; + /* if this is a small write inside eof, kick off defrag */ + if (end <= BTRFS_I(inode)->disk_i_size && num_bytes < 64 * 1024) + btrfs_add_inode_defrag(trans, inode); + if (start == 0) { /* lets try to make an inline extent */ ret = cow_file_range_inline(trans, root, inode, - start, end, 0, NULL); + start, end, 0, 0, NULL); if (ret == 0) { extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree, @@ -819,7 +845,8 @@ static noinline int cow_file_range(struct inode *inode, (u64)-1, &ins, 1); BUG_ON(ret); - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); + BUG_ON(!em); em->start = start; em->orig_start = em->start; ram_size = ins.offset; @@ -946,6 +973,7 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page, 1, 0, NULL, GFP_NOFS); while (start < end) { async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS); + BUG_ON(!async_cow); async_cow->inode = inode; async_cow->root = root; async_cow->locked_page = locked_page; @@ -999,7 +1027,7 @@ static noinline int csum_exist_in_range(struct btrfs_root *root, LIST_HEAD(list); ret = btrfs_lookup_csums_range(root->fs_info->csum_root, bytenr, - bytenr + num_bytes - 1, &list); + bytenr + num_bytes - 1, &list, 0); if (ret == 0 && list_empty(&list)) return 0; @@ -1040,29 +1068,31 @@ static noinline int run_delalloc_nocow(struct inode *inode, int type; int nocow; int check_prev = 1; - bool nolock = false; + bool nolock; + u64 ino = btrfs_ino(inode); path = btrfs_alloc_path(); BUG_ON(!path); - if (root == root->fs_info->tree_root) { - nolock = true; + + nolock = is_free_space_inode(root, inode); + + if (nolock) trans = btrfs_join_transaction_nolock(root, 1); - } else { + else trans = btrfs_join_transaction(root, 1); - } BUG_ON(IS_ERR(trans)); cow_start = (u64)-1; cur_offset = start; while (1) { - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, ino, cur_offset, 0); BUG_ON(ret < 0); if (ret > 0 && path->slots[0] > 0 && check_prev) { leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1); - if (found_key.objectid == inode->i_ino && + if (found_key.objectid == ino && found_key.type == BTRFS_EXTENT_DATA_KEY) path->slots[0]--; } @@ -1083,7 +1113,7 @@ next_slot: num_bytes = 0; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid > inode->i_ino || + if (found_key.objectid > ino || found_key.type > BTRFS_EXTENT_DATA_KEY || found_key.offset > end) break; @@ -1118,7 +1148,7 @@ next_slot: goto out_check; if (btrfs_extent_readonly(root, disk_bytenr)) goto out_check; - if (btrfs_cross_ref_exist(trans, root, inode->i_ino, + if (btrfs_cross_ref_exist(trans, root, ino, found_key.offset - extent_offset, disk_bytenr)) goto out_check; @@ -1155,7 +1185,7 @@ out_check: goto next_slot; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (cow_start != (u64)-1) { ret = cow_file_range(inode, locked_page, cow_start, found_key.offset - 1, page_started, @@ -1168,7 +1198,8 @@ out_check: struct extent_map *em; struct extent_map_tree *em_tree; em_tree = &BTRFS_I(inode)->extent_tree; - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); + BUG_ON(!em); em->start = cur_offset; em->orig_start = em->start; em->len = num_bytes; @@ -1212,7 +1243,7 @@ out_check: if (cur_offset > end) break; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (cur_offset <= end && cow_start == (u64)-1) cow_start = cur_offset; @@ -1250,7 +1281,8 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page, ret = run_delalloc_nocow(inode, locked_page, start, end, page_started, 0, nr_written); else if (!btrfs_test_opt(root, COMPRESS) && - !(BTRFS_I(inode)->force_compress)) + !(BTRFS_I(inode)->force_compress) && + !(BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS)) ret = cow_file_range(inode, locked_page, start, end, page_started, nr_written, 1); else @@ -1299,14 +1331,13 @@ static int btrfs_set_bit_hook(struct inode *inode, /* * set_bit and clear bit hooks normally require _irqsave/restore - * but in this case, we are only testeing for the DELALLOC + * but in this case, we are only testing for the DELALLOC * bit, which is only set or cleared with irqs on */ if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - int do_list = (root->root_key.objectid != - BTRFS_ROOT_TREE_OBJECTID); + bool do_list = !is_free_space_inode(root, inode); if (*bits & EXTENT_FIRST_DELALLOC) *bits &= ~EXTENT_FIRST_DELALLOC; @@ -1333,14 +1364,13 @@ static int btrfs_clear_bit_hook(struct inode *inode, { /* * set_bit and clear bit hooks normally require _irqsave/restore - * but in this case, we are only testeing for the DELALLOC + * but in this case, we are only testing for the DELALLOC * bit, which is only set or cleared with irqs on */ if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - int do_list = (root->root_key.objectid != - BTRFS_ROOT_TREE_OBJECTID); + bool do_list = !is_free_space_inode(root, inode); if (*bits & EXTENT_FIRST_DELALLOC) *bits &= ~EXTENT_FIRST_DELALLOC; @@ -1447,7 +1477,7 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; - if (root == root->fs_info->tree_root) + if (is_free_space_inode(root, inode)) ret = btrfs_bio_wq_end_io(root->fs_info, bio, 2); else ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); @@ -1457,8 +1487,11 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, if (bio_flags & EXTENT_BIO_COMPRESSED) { return btrfs_submit_compressed_read(inode, bio, mirror_num, bio_flags); - } else if (!skip_sum) - btrfs_lookup_bio_sums(root, inode, bio, NULL); + } else if (!skip_sum) { + ret = btrfs_lookup_bio_sums(root, inode, bio, NULL); + if (ret) + return ret; + } goto mapit; } else if (!skip_sum) { /* csum items have already been cloned */ @@ -1630,7 +1663,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, &hint, 0); BUG_ON(ret); - ins.objectid = inode->i_ino; + ins.objectid = btrfs_ino(inode); ins.offset = file_pos; ins.type = BTRFS_EXTENT_DATA_KEY; ret = btrfs_insert_empty_item(trans, root, path, &ins, sizeof(*fi)); @@ -1661,7 +1694,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, ins.type = BTRFS_EXTENT_ITEM_KEY; ret = btrfs_alloc_reserved_file_extent(trans, root, root->root_key.objectid, - inode->i_ino, file_pos, &ins); + btrfs_ino(inode), file_pos, &ins); BUG_ON(ret); btrfs_free_path(path); @@ -1687,7 +1720,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) struct extent_state *cached_state = NULL; int compress_type = 0; int ret; - bool nolock = false; + bool nolock; ret = btrfs_dec_test_ordered_pending(inode, &ordered_extent, start, end - start + 1); @@ -1695,7 +1728,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) return 0; BUG_ON(!ordered_extent); - nolock = (root == root->fs_info->tree_root); + nolock = is_free_space_inode(root, inode); if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) { BUG_ON(!list_empty(&ordered_extent->list)); @@ -1757,9 +1790,12 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) add_pending_csums(trans, inode, ordered_extent->file_offset, &ordered_extent->list); - btrfs_ordered_update_i_size(inode, 0, ordered_extent); - ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent); + if (!ret) { + ret = btrfs_update_inode(trans, root, inode); + BUG_ON(ret); + } + ret = 0; out: if (nolock) { if (trans) @@ -1781,6 +1817,8 @@ out: static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, struct extent_state *state, int uptodate) { + trace_btrfs_writepage_end_io_hook(page, start, end, uptodate); + ClearPagePrivate2(page); return btrfs_finish_ordered_io(page->mapping->host, start, end); } @@ -1836,7 +1874,7 @@ static int btrfs_io_failed_hook(struct bio *failed_bio, } read_unlock(&em_tree->lock); - if (!em || IS_ERR(em)) { + if (IS_ERR_OR_NULL(em)) { kfree(failrec); return -EIO; } @@ -1891,10 +1929,10 @@ static int btrfs_io_failed_hook(struct bio *failed_bio, else rw = READ; - BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, + ret = BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, failrec->last_mirror, failrec->bio_flags, 0); - return 0; + return ret; } /* @@ -1910,7 +1948,7 @@ static int btrfs_clean_io_failures(struct inode *inode, u64 start) private = 0; if (count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private, - (u64)-1, 1, EXTENT_DIRTY)) { + (u64)-1, 1, EXTENT_DIRTY, 0)) { ret = get_state_private(&BTRFS_I(inode)->io_failure_tree, start, &private_failure); if (ret == 0) { @@ -1985,12 +2023,11 @@ good: return 0; zeroit: - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u " - "private %llu\n", page->mapping->host->i_ino, + printk_ratelimited(KERN_INFO "btrfs csum failed ino %llu off %llu csum %u " + "private %llu\n", + (unsigned long long)btrfs_ino(page->mapping->host), (unsigned long long)start, csum, (unsigned long long)private); - } memset(kaddr + offset, 1, end - start + 1); flush_dcache_page(page); kunmap_atomic(kaddr, KM_USER0); @@ -2206,8 +2243,6 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) insert = 1; #endif insert = 1; - } else { - WARN_ON(!BTRFS_I(inode)->orphan_meta_reserved); } if (!BTRFS_I(inode)->orphan_meta_reserved) { @@ -2227,7 +2262,7 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) /* insert an orphan item to track this unlinked/truncated file */ if (insert >= 1) { - ret = btrfs_insert_orphan_item(trans, root, inode->i_ino); + ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); BUG_ON(ret); } @@ -2264,7 +2299,7 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode) spin_unlock(&root->orphan_lock); if (trans && delete_item) { - ret = btrfs_del_orphan_item(trans, root, inode->i_ino); + ret = btrfs_del_orphan_item(trans, root, btrfs_ino(inode)); BUG_ON(ret); } @@ -2278,7 +2313,7 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode) * this cleans up any orphans that may be left on the list from the last use * of this root. */ -void btrfs_orphan_cleanup(struct btrfs_root *root) +int btrfs_orphan_cleanup(struct btrfs_root *root) { struct btrfs_path *path; struct extent_buffer *leaf; @@ -2288,10 +2323,13 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) int ret = 0, nr_unlink = 0, nr_truncate = 0; if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) - return; + return 0; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + ret = -ENOMEM; + goto out; + } path->reada = -1; key.objectid = BTRFS_ORPHAN_OBJECTID; @@ -2300,18 +2338,16 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) while (1) { ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); - if (ret < 0) { - printk(KERN_ERR "Error searching slot for orphan: %d" - "\n", ret); - break; - } + if (ret < 0) + goto out; /* * if ret == 0 means we found what we were searching for, which - * is weird, but possible, so only screw with path if we didnt + * is weird, but possible, so only screw with path if we didn't * find the key and see if we have stuff that matches */ if (ret > 0) { + ret = 0; if (path->slots[0] == 0) break; path->slots[0]--; @@ -2328,7 +2364,7 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) break; /* release the path since we're done with it */ - btrfs_release_path(root, path); + btrfs_release_path(path); /* * this is where we are basically btrfs_lookup, without the @@ -2339,7 +2375,10 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) found_key.type = BTRFS_INODE_ITEM_KEY; found_key.offset = 0; inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL); - BUG_ON(IS_ERR(inode)); + if (IS_ERR(inode)) { + ret = PTR_ERR(inode); + goto out; + } /* * add this inode to the orphan list so btrfs_orphan_del does @@ -2357,7 +2396,10 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) */ if (is_bad_inode(inode)) { trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out; + } btrfs_orphan_del(trans, inode); btrfs_end_transaction(trans, root); iput(inode); @@ -2366,17 +2408,22 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) /* if we have links, this was a truncate, lets do that */ if (inode->i_nlink) { + if (!S_ISREG(inode->i_mode)) { + WARN_ON(1); + iput(inode); + continue; + } nr_truncate++; - btrfs_truncate(inode); + ret = btrfs_truncate(inode); } else { nr_unlink++; } /* this will do delete_inode and everything for us */ iput(inode); + if (ret) + goto out; } - btrfs_free_path(path); - root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; if (root->orphan_block_rsv) @@ -2385,14 +2432,20 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) if (root->orphan_block_rsv || root->orphan_item_inserted) { trans = btrfs_join_transaction(root, 1); - BUG_ON(IS_ERR(trans)); - btrfs_end_transaction(trans, root); + if (!IS_ERR(trans)) + btrfs_end_transaction(trans, root); } if (nr_unlink) printk(KERN_INFO "btrfs: unlinked %d orphans\n", nr_unlink); if (nr_truncate) printk(KERN_INFO "btrfs: truncated %d orphans\n", nr_truncate); + +out: + if (ret) + printk(KERN_CRIT "btrfs: could not do orphan cleanup %d\n", ret); + btrfs_free_path(path); + return ret; } /* @@ -2508,7 +2561,8 @@ static void btrfs_read_locked_inode(struct inode *inode) * try to precache a NULL acl entry for files that don't have * any xattrs or acls */ - maybe_acls = acls_after_inode_item(leaf, path->slots[0], inode->i_ino); + maybe_acls = acls_after_inode_item(leaf, path->slots[0], + btrfs_ino(inode)); if (!maybe_acls) cache_no_acl(inode); @@ -2559,6 +2613,13 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, struct btrfs_inode_item *item, struct inode *inode) { + if (!leaf->map_token) + map_private_extent_buffer(leaf, (unsigned long)item, + sizeof(struct btrfs_inode_item), + &leaf->map_token, &leaf->kaddr, + &leaf->map_start, &leaf->map_len, + KM_USER1); + btrfs_set_inode_uid(leaf, item, inode->i_uid); btrfs_set_inode_gid(leaf, item, inode->i_gid); btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size); @@ -2587,6 +2648,11 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, btrfs_set_inode_rdev(leaf, item, inode->i_rdev); btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags); btrfs_set_inode_block_group(leaf, item, BTRFS_I(inode)->block_group); + + if (leaf->map_token) { + unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); + leaf->map_token = NULL; + } } /* @@ -2600,11 +2666,26 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, struct extent_buffer *leaf; int ret; + /* + * If root is tree root, it means this inode is used to + * store free space information. And these inodes are updated + * when committing the transaction, so they needn't delaye to + * be updated, or deadlock will occured. + */ + if (!is_free_space_inode(root, inode)) { + ret = btrfs_delayed_update_inode(trans, root, inode); + if (!ret) + btrfs_set_inode_last_trans(trans, inode); + return ret; + } + path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; + path->leave_spinning = 1; - ret = btrfs_lookup_inode(trans, root, path, - &BTRFS_I(inode)->location, 1); + ret = btrfs_lookup_inode(trans, root, path, &BTRFS_I(inode)->location, + 1); if (ret) { if (ret > 0) ret = -ENOENT; @@ -2614,7 +2695,7 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, btrfs_unlock_up_safe(path, 1); leaf = path->nodes[0]; inode_item = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_inode_item); + struct btrfs_inode_item); fill_inode_item(trans, leaf, inode_item, inode); btrfs_mark_buffer_dirty(leaf); @@ -2625,16 +2706,15 @@ failed: return ret; } - /* * unlink helper that gets used here in inode.c and in the tree logging * recovery code. It remove a link in a directory with a given name, and * also drops the back refs in the inode to the directory */ -int btrfs_unlink_inode(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct inode *dir, struct inode *inode, - const char *name, int name_len) +static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *dir, struct inode *inode, + const char *name, int name_len) { struct btrfs_path *path; int ret = 0; @@ -2642,6 +2722,8 @@ int btrfs_unlink_inode(struct btrfs_trans_handle *trans, struct btrfs_dir_item *di; struct btrfs_key key; u64 index; + u64 ino = btrfs_ino(inode); + u64 dir_ino = btrfs_ino(dir); path = btrfs_alloc_path(); if (!path) { @@ -2650,7 +2732,7 @@ int btrfs_unlink_inode(struct btrfs_trans_handle *trans, } path->leave_spinning = 1; - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, root, path, dir_ino, name, name_len, -1); if (IS_ERR(di)) { ret = PTR_ERR(di); @@ -2665,33 +2747,23 @@ int btrfs_unlink_inode(struct btrfs_trans_handle *trans, ret = btrfs_delete_one_dir_name(trans, root, path, di); if (ret) goto err; - btrfs_release_path(root, path); + btrfs_release_path(path); - ret = btrfs_del_inode_ref(trans, root, name, name_len, - inode->i_ino, - dir->i_ino, &index); + ret = btrfs_del_inode_ref(trans, root, name, name_len, ino, + dir_ino, &index); if (ret) { printk(KERN_INFO "btrfs failed to delete reference to %.*s, " - "inode %lu parent %lu\n", name_len, name, - inode->i_ino, dir->i_ino); + "inode %llu parent %llu\n", name_len, name, + (unsigned long long)ino, (unsigned long long)dir_ino); goto err; } - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, - index, name, name_len, -1); - if (IS_ERR(di)) { - ret = PTR_ERR(di); - goto err; - } - if (!di) { - ret = -ENOENT; + ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); + if (ret) goto err; - } - ret = btrfs_delete_one_dir_name(trans, root, path, di); - btrfs_release_path(root, path); ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len, - inode, dir->i_ino); + inode, dir_ino); BUG_ON(ret != 0 && ret != -ENOENT); ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len, @@ -2706,12 +2778,25 @@ err: btrfs_i_size_write(dir, dir->i_size - name_len * 2); inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME; btrfs_update_inode(trans, root, dir); - btrfs_drop_nlink(inode); - ret = btrfs_update_inode(trans, root, inode); out: return ret; } +int btrfs_unlink_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *dir, struct inode *inode, + const char *name, int name_len) +{ + int ret; + ret = __btrfs_unlink_inode(trans, root, dir, inode, name, name_len); + if (!ret) { + btrfs_drop_nlink(inode); + ret = btrfs_update_inode(trans, root, inode); + } + return ret; +} + + /* helper to check if there is any shared block in the path */ static int check_path_shared(struct btrfs_root *root, struct btrfs_path *path) @@ -2756,12 +2841,14 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, int check_link = 1; int err = -ENOSPC; int ret; + u64 ino = btrfs_ino(inode); + u64 dir_ino = btrfs_ino(dir); trans = btrfs_start_transaction(root, 10); if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC) return trans; - if (inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) + if (ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) return ERR_PTR(-ENOSPC); /* check if there is someone else holds reference */ @@ -2802,7 +2889,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, } else { check_link = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_lookup_inode(trans, root, path, &BTRFS_I(inode)->location, 0); @@ -2816,11 +2903,11 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, } else { check_link = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret == 0 && S_ISREG(inode->i_mode)) { ret = btrfs_lookup_file_extent(trans, root, path, - inode->i_ino, (u64)-1, 0); + ino, (u64)-1, 0); if (ret < 0) { err = ret; goto out; @@ -2828,7 +2915,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, BUG_ON(ret == 0); if (check_path_shared(root, path)) goto out; - btrfs_release_path(root, path); + btrfs_release_path(path); } if (!check_link) { @@ -2836,7 +2923,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, goto out; } - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, root, path, dir_ino, dentry->d_name.name, dentry->d_name.len, 0); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2849,11 +2936,11 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, err = 0; goto out; } - btrfs_release_path(root, path); + btrfs_release_path(path); ref = btrfs_lookup_inode_ref(trans, root, path, dentry->d_name.name, dentry->d_name.len, - inode->i_ino, dir->i_ino, 0); + ino, dir_ino, 0); if (IS_ERR(ref)) { err = PTR_ERR(ref); goto out; @@ -2862,9 +2949,17 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, if (check_path_shared(root, path)) goto out; index = btrfs_inode_ref_index(path->nodes[0], ref); - btrfs_release_path(root, path); + btrfs_release_path(path); - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, index, + /* + * This is a commit root search, if we can lookup inode item and other + * relative items in the commit root, it means the transaction of + * dir/file creation has been committed, and the dir index item that we + * delay to insert has also been inserted into the commit root. So + * we needn't worry about the delayed insertion of the dir index item + * here. + */ + di = btrfs_lookup_dir_index_item(trans, root, path, dir_ino, index, dentry->d_name.name, dentry->d_name.len, 0); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2939,54 +3034,47 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, struct btrfs_key key; u64 index; int ret; + u64 dir_ino = btrfs_ino(dir); path = btrfs_alloc_path(); if (!path) return -ENOMEM; - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, root, path, dir_ino, name, name_len, -1); - BUG_ON(!di || IS_ERR(di)); + BUG_ON(IS_ERR_OR_NULL(di)); leaf = path->nodes[0]; btrfs_dir_item_key_to_cpu(leaf, di, &key); WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); ret = btrfs_delete_one_dir_name(trans, root, path, di); BUG_ON(ret); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_del_root_ref(trans, root->fs_info->tree_root, objectid, root->root_key.objectid, - dir->i_ino, &index, name, name_len); + dir_ino, &index, name, name_len); if (ret < 0) { BUG_ON(ret != -ENOENT); - di = btrfs_search_dir_index_item(root, path, dir->i_ino, + di = btrfs_search_dir_index_item(root, path, dir_ino, name, name_len); - BUG_ON(!di || IS_ERR(di)); + BUG_ON(IS_ERR_OR_NULL(di)); leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(root, path); + btrfs_release_path(path); index = key.offset; } + btrfs_release_path(path); - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, - index, name, name_len, -1); - BUG_ON(!di || IS_ERR(di)); - - leaf = path->nodes[0]; - btrfs_dir_item_key_to_cpu(leaf, di, &key); - WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); - ret = btrfs_delete_one_dir_name(trans, root, path, di); + ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); BUG_ON(ret); - btrfs_release_path(root, path); btrfs_i_size_write(dir, dir->i_size - name_len * 2); dir->i_mtime = dir->i_ctime = CURRENT_TIME; ret = btrfs_update_inode(trans, root, dir); BUG_ON(ret); - btrfs_free_path(path); return 0; } @@ -2999,7 +3087,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) unsigned long nr = 0; if (inode->i_size > BTRFS_EMPTY_DIR_SIZE || - inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) + btrfs_ino(inode) == BTRFS_FIRST_FREE_OBJECTID) return -ENOTEMPTY; trans = __unlink_start_trans(dir, dentry); @@ -3008,7 +3096,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) btrfs_set_trans_block_group(trans, dir); - if (unlikely(inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { + if (unlikely(btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { err = btrfs_unlink_subvol(trans, root, dir, BTRFS_I(inode)->location.objectid, dentry->d_name.name, @@ -3033,178 +3121,6 @@ out: return err; } -#if 0 -/* - * when truncating bytes in a file, it is possible to avoid reading - * the leaves that contain only checksum items. This can be the - * majority of the IO required to delete a large file, but it must - * be done carefully. - * - * The keys in the level just above the leaves are checked to make sure - * the lowest key in a given leaf is a csum key, and starts at an offset - * after the new size. - * - * Then the key for the next leaf is checked to make sure it also has - * a checksum item for the same file. If it does, we know our target leaf - * contains only checksum items, and it can be safely freed without reading - * it. - * - * This is just an optimization targeted at large files. It may do - * nothing. It will return 0 unless things went badly. - */ -static noinline int drop_csum_leaves(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct inode *inode, u64 new_size) -{ - struct btrfs_key key; - int ret; - int nritems; - struct btrfs_key found_key; - struct btrfs_key other_key; - struct btrfs_leaf_ref *ref; - u64 leaf_gen; - u64 leaf_start; - - path->lowest_level = 1; - key.objectid = inode->i_ino; - key.type = BTRFS_CSUM_ITEM_KEY; - key.offset = new_size; -again: - ret = btrfs_search_slot(trans, root, &key, path, -1, 1); - if (ret < 0) - goto out; - - if (path->nodes[1] == NULL) { - ret = 0; - goto out; - } - ret = 0; - btrfs_node_key_to_cpu(path->nodes[1], &found_key, path->slots[1]); - nritems = btrfs_header_nritems(path->nodes[1]); - - if (!nritems) - goto out; - - if (path->slots[1] >= nritems) - goto next_node; - - /* did we find a key greater than anything we want to delete? */ - if (found_key.objectid > inode->i_ino || - (found_key.objectid == inode->i_ino && found_key.type > key.type)) - goto out; - - /* we check the next key in the node to make sure the leave contains - * only checksum items. This comparison doesn't work if our - * leaf is the last one in the node - */ - if (path->slots[1] + 1 >= nritems) { -next_node: - /* search forward from the last key in the node, this - * will bring us into the next node in the tree - */ - btrfs_node_key_to_cpu(path->nodes[1], &found_key, nritems - 1); - - /* unlikely, but we inc below, so check to be safe */ - if (found_key.offset == (u64)-1) - goto out; - - /* search_forward needs a path with locks held, do the - * search again for the original key. It is possible - * this will race with a balance and return a path that - * we could modify, but this drop is just an optimization - * and is allowed to miss some leaves. - */ - btrfs_release_path(root, path); - found_key.offset++; - - /* setup a max key for search_forward */ - other_key.offset = (u64)-1; - other_key.type = key.type; - other_key.objectid = key.objectid; - - path->keep_locks = 1; - ret = btrfs_search_forward(root, &found_key, &other_key, - path, 0, 0); - path->keep_locks = 0; - if (ret || found_key.objectid != key.objectid || - found_key.type != key.type) { - ret = 0; - goto out; - } - - key.offset = found_key.offset; - btrfs_release_path(root, path); - cond_resched(); - goto again; - } - - /* we know there's one more slot after us in the tree, - * read that key so we can verify it is also a checksum item - */ - btrfs_node_key_to_cpu(path->nodes[1], &other_key, path->slots[1] + 1); - - if (found_key.objectid < inode->i_ino) - goto next_key; - - if (found_key.type != key.type || found_key.offset < new_size) - goto next_key; - - /* - * if the key for the next leaf isn't a csum key from this objectid, - * we can't be sure there aren't good items inside this leaf. - * Bail out - */ - if (other_key.objectid != inode->i_ino || other_key.type != key.type) - goto out; - - leaf_start = btrfs_node_blockptr(path->nodes[1], path->slots[1]); - leaf_gen = btrfs_node_ptr_generation(path->nodes[1], path->slots[1]); - /* - * it is safe to delete this leaf, it contains only - * csum items from this inode at an offset >= new_size - */ - ret = btrfs_del_leaf(trans, root, path, leaf_start); - BUG_ON(ret); - - if (root->ref_cows && leaf_gen < trans->transid) { - ref = btrfs_alloc_leaf_ref(root, 0); - if (ref) { - ref->root_gen = root->root_key.offset; - ref->bytenr = leaf_start; - ref->owner = 0; - ref->generation = leaf_gen; - ref->nritems = 0; - - btrfs_sort_leaf_ref(ref); - - ret = btrfs_add_leaf_ref(root, ref, 0); - WARN_ON(ret); - btrfs_free_leaf_ref(root, ref); - } else { - WARN_ON(1); - } - } -next_key: - btrfs_release_path(root, path); - - if (other_key.objectid == inode->i_ino && - other_key.type == key.type && other_key.offset > key.offset) { - key.offset = other_key.offset; - cond_resched(); - goto again; - } - ret = 0; -out: - /* fixup any changes we've made to the path */ - path->lowest_level = 0; - path->keep_locks = 0; - btrfs_release_path(root, path); - return ret; -} - -#endif - /* * this can truncate away extent items, csum items and directory items. * It starts at a high offset and removes keys until it can't find @@ -3240,17 +3156,27 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, int encoding; int ret; int err = 0; + u64 ino = btrfs_ino(inode); BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY); if (root->ref_cows || root == root->fs_info->tree_root) btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0); + /* + * This function is also used to drop the items in the log tree before + * we relog the inode, so if root != BTRFS_I(inode)->root, it means + * it is used to drop the loged items. So we shouldn't kill the delayed + * items. + */ + if (min_type == 0 && root == BTRFS_I(inode)->root) + btrfs_kill_delayed_inode_items(inode); + path = btrfs_alloc_path(); BUG_ON(!path); path->reada = -1; - key.objectid = inode->i_ino; + key.objectid = ino; key.offset = (u64)-1; key.type = (u8)-1; @@ -3278,7 +3204,7 @@ search_again: found_type = btrfs_key_type(&found_key); encoding = 0; - if (found_key.objectid != inode->i_ino) + if (found_key.objectid != ino) break; if (found_type < min_type) @@ -3368,7 +3294,6 @@ search_again: btrfs_file_extent_calc_inline_size(size); ret = btrfs_truncate_item(trans, root, path, size, 1); - BUG_ON(ret); } else if (root->ref_cows) { inode_sub_bytes(inode, item_end + 1 - found_key.offset); @@ -3397,7 +3322,7 @@ delete: ret = btrfs_free_extent(trans, root, extent_start, extent_num_bytes, 0, btrfs_header_owner(leaf), - inode->i_ino, extent_offset); + ino, extent_offset); BUG_ON(ret); } @@ -3406,7 +3331,9 @@ delete: if (path->slots[0] == 0 || path->slots[0] != pending_del_slot) { - if (root->ref_cows) { + if (root->ref_cows && + BTRFS_I(inode)->location.objectid != + BTRFS_FREE_INO_OBJECTID) { err = -EAGAIN; goto out; } @@ -3417,7 +3344,7 @@ delete: BUG_ON(ret); pending_del_nr = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); goto search_again; } else { path->slots[0]--; @@ -3533,7 +3460,13 @@ out: return ret; } -int btrfs_cont_expand(struct inode *inode, loff_t size) +/* + * This function puts in dummy file extents for the area we're creating a hole + * for. So if we are truncating this file to a larger size we need to insert + * these file extents so that btrfs_get_extent will return a EXTENT_MAP_HOLE for + * the range between oldsize and size + */ +int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -3541,7 +3474,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) struct extent_map *em = NULL; struct extent_state *cached_state = NULL; u64 mask = root->sectorsize - 1; - u64 hole_start = (inode->i_size + mask) & ~mask; + u64 hole_start = (oldsize + mask) & ~mask; u64 block_end = (size + mask) & ~mask; u64 last_byte; u64 cur_offset; @@ -3569,7 +3502,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) while (1) { em = btrfs_get_extent(inode, NULL, 0, cur_offset, block_end - cur_offset, 0); - BUG_ON(IS_ERR(em) || !em); + BUG_ON(IS_ERR_OR_NULL(em)); last_byte = min(extent_map_end(em), block_end); last_byte = (last_byte + mask) & ~mask; if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) { @@ -3586,13 +3519,15 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) err = btrfs_drop_extents(trans, inode, cur_offset, cur_offset + hole_size, &hint_byte, 1); - BUG_ON(err); + if (err) + break; err = btrfs_insert_file_extent(trans, root, - inode->i_ino, cur_offset, 0, + btrfs_ino(inode), cur_offset, 0, 0, hole_size, 0, hole_size, 0, 0, 0); - BUG_ON(err); + if (err) + break; btrfs_drop_extent_cache(inode, hole_start, last_byte - 1, 0); @@ -3612,81 +3547,41 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) return err; } -static int btrfs_setattr_size(struct inode *inode, struct iattr *attr) +static int btrfs_setsize(struct inode *inode, loff_t newsize) { - struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_trans_handle *trans; - unsigned long nr; + loff_t oldsize = i_size_read(inode); int ret; - if (attr->ia_size == inode->i_size) + if (newsize == oldsize) return 0; - if (attr->ia_size > inode->i_size) { - unsigned long limit; - limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur; - if (attr->ia_size > inode->i_sb->s_maxbytes) - return -EFBIG; - if (limit != RLIM_INFINITY && attr->ia_size > limit) { - send_sig(SIGXFSZ, current, 0); - return -EFBIG; - } - } - - trans = btrfs_start_transaction(root, 5); - if (IS_ERR(trans)) - return PTR_ERR(trans); - - btrfs_set_trans_block_group(trans, inode); - - ret = btrfs_orphan_add(trans, inode); - BUG_ON(ret); - - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); - - if (attr->ia_size > inode->i_size) { - ret = btrfs_cont_expand(inode, attr->ia_size); + if (newsize > oldsize) { + i_size_write(inode, newsize); + btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); + truncate_pagecache(inode, oldsize, newsize); + ret = btrfs_cont_expand(inode, oldsize, newsize); if (ret) { - btrfs_truncate(inode); + btrfs_setsize(inode, oldsize); return ret; } - i_size_write(inode, attr->ia_size); - btrfs_ordered_update_i_size(inode, inode->i_size, NULL); + mark_inode_dirty(inode); + } else { - trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); - btrfs_set_trans_block_group(trans, inode); - trans->block_rsv = root->orphan_block_rsv; - BUG_ON(!trans->block_rsv); + /* + * We're truncating a file that used to have good data down to + * zero. Make sure it gets into the ordered flush list so that + * any new writes get down to disk quickly. + */ + if (newsize == 0) + BTRFS_I(inode)->ordered_data_close = 1; - ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); - if (inode->i_nlink > 0) { - ret = btrfs_orphan_del(trans, inode); - BUG_ON(ret); - } - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); - return 0; + /* we don't support swapfiles, so vmtruncate shouldn't fail */ + truncate_setsize(inode, newsize); + ret = btrfs_truncate(inode); } - /* - * We're truncating a file that used to have good data down to - * zero. Make sure it gets into the ordered flush list so that - * any new writes get down to disk quickly. - */ - if (attr->ia_size == 0) - BTRFS_I(inode)->ordered_data_close = 1; - - /* we don't support swapfiles, so vmtruncate shouldn't fail */ - ret = vmtruncate(inode, attr->ia_size); - BUG_ON(ret); - - return 0; + return ret; } static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) @@ -3703,7 +3598,7 @@ static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) return err; if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) { - err = btrfs_setattr_size(inode, attr); + err = btrfs_setsize(inode, attr->ia_size); if (err) return err; } @@ -3726,9 +3621,11 @@ void btrfs_evict_inode(struct inode *inode) unsigned long nr; int ret; + trace_btrfs_inode_evict(inode); + truncate_inode_pages(&inode->i_data, 0); if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 || - root == root->fs_info->tree_root)) + is_free_space_inode(root, inode))) goto no_delete; if (is_bad_inode(inode)) { @@ -3781,6 +3678,10 @@ void btrfs_evict_inode(struct inode *inode) BUG_ON(ret); } + if (!(root == root->fs_info->tree_root || + root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) + btrfs_return_ino(root, btrfs_ino(inode)); + nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); @@ -3806,12 +3707,12 @@ static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry, path = btrfs_alloc_path(); BUG_ON(!path); - di = btrfs_lookup_dir_item(NULL, root, path, dir->i_ino, name, + di = btrfs_lookup_dir_item(NULL, root, path, btrfs_ino(dir), name, namelen, 0); if (IS_ERR(di)) ret = PTR_ERR(di); - if (!di || IS_ERR(di)) + if (IS_ERR_OR_NULL(di)) goto out_err; btrfs_dir_item_key_to_cpu(path->nodes[0], di, location); @@ -3859,7 +3760,7 @@ static int fixup_tree_root_location(struct btrfs_root *root, leaf = path->nodes[0]; ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); - if (btrfs_root_ref_dirid(leaf, ref) != dir->i_ino || + if (btrfs_root_ref_dirid(leaf, ref) != btrfs_ino(dir) || btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len) goto out; @@ -3869,7 +3770,7 @@ static int fixup_tree_root_location(struct btrfs_root *root, if (ret) goto out; - btrfs_release_path(root->fs_info->tree_root, path); + btrfs_release_path(path); new_root = btrfs_read_fs_root_no_name(root->fs_info, location); if (IS_ERR(new_root)) { @@ -3898,6 +3799,7 @@ static void inode_tree_add(struct inode *inode) struct btrfs_inode *entry; struct rb_node **p; struct rb_node *parent; + u64 ino = btrfs_ino(inode); again: p = &root->inode_tree.rb_node; parent = NULL; @@ -3910,9 +3812,9 @@ again: parent = *p; entry = rb_entry(parent, struct btrfs_inode, rb_node); - if (inode->i_ino < entry->vfs_inode.i_ino) + if (ino < btrfs_ino(&entry->vfs_inode)) p = &parent->rb_left; - else if (inode->i_ino > entry->vfs_inode.i_ino) + else if (ino > btrfs_ino(&entry->vfs_inode)) p = &parent->rb_right; else { WARN_ON(!(entry->vfs_inode.i_state & @@ -3976,9 +3878,9 @@ again: prev = node; entry = rb_entry(node, struct btrfs_inode, rb_node); - if (objectid < entry->vfs_inode.i_ino) + if (objectid < btrfs_ino(&entry->vfs_inode)) node = node->rb_left; - else if (objectid > entry->vfs_inode.i_ino) + else if (objectid > btrfs_ino(&entry->vfs_inode)) node = node->rb_right; else break; @@ -3986,7 +3888,7 @@ again: if (!node) { while (prev) { entry = rb_entry(prev, struct btrfs_inode, rb_node); - if (objectid <= entry->vfs_inode.i_ino) { + if (objectid <= btrfs_ino(&entry->vfs_inode)) { node = prev; break; } @@ -3995,7 +3897,7 @@ again: } while (node) { entry = rb_entry(node, struct btrfs_inode, rb_node); - objectid = entry->vfs_inode.i_ino + 1; + objectid = btrfs_ino(&entry->vfs_inode) + 1; inode = igrab(&entry->vfs_inode); if (inode) { spin_unlock(&root->inode_lock); @@ -4033,7 +3935,7 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p) static int btrfs_find_actor(struct inode *inode, void *opaque) { struct btrfs_iget_args *args = opaque; - return args->ino == inode->i_ino && + return args->ino == btrfs_ino(inode) && args->root == BTRFS_I(inode)->root; } @@ -4068,7 +3970,6 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, BTRFS_I(inode)->root = root; memcpy(&BTRFS_I(inode)->location, location, sizeof(*location)); btrfs_read_locked_inode(inode); - inode_tree_add(inode); unlock_new_inode(inode); if (new) @@ -4143,8 +4044,10 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry) if (!IS_ERR(inode) && root != sub_root) { down_read(&root->fs_info->cleanup_work_sem); if (!(inode->i_sb->s_flags & MS_RDONLY)) - btrfs_orphan_cleanup(sub_root); + ret = btrfs_orphan_cleanup(sub_root); up_read(&root->fs_info->cleanup_work_sem); + if (ret) + inode = ERR_PTR(ret); } return inode; @@ -4177,7 +4080,7 @@ static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry, return d_splice_alias(inode, dentry); } -static unsigned char btrfs_filetype_table[] = { +unsigned char btrfs_filetype_table[] = { DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK }; @@ -4191,11 +4094,11 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, struct btrfs_key key; struct btrfs_key found_key; struct btrfs_path *path; + struct list_head ins_list; + struct list_head del_list; int ret; - u32 nritems; struct extent_buffer *leaf; int slot; - int advance; unsigned char d_type; int over = 0; u32 di_cur; @@ -4205,6 +4108,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, char tmp_name[32]; char *name_ptr; int name_len; + int is_curr = 0; /* filp->f_pos points to the current index? */ /* FIXME, use a real flag for deciding about the key type */ if (root->fs_info->tree_root == root) @@ -4212,9 +4116,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, /* special case for "." */ if (filp->f_pos == 0) { - over = filldir(dirent, ".", 1, - 1, inode->i_ino, - DT_DIR); + over = filldir(dirent, ".", 1, 1, btrfs_ino(inode), DT_DIR); if (over) return 0; filp->f_pos = 1; @@ -4229,36 +4131,36 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, filp->f_pos = 2; } path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; path->reada = 2; + if (key_type == BTRFS_DIR_INDEX_KEY) { + INIT_LIST_HEAD(&ins_list); + INIT_LIST_HEAD(&del_list); + btrfs_get_delayed_items(inode, &ins_list, &del_list); + } + btrfs_set_key_type(&key, key_type); key.offset = filp->f_pos; - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) goto err; - advance = 0; while (1) { leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); slot = path->slots[0]; - if (advance || slot >= nritems) { - if (slot >= nritems - 1) { - ret = btrfs_next_leaf(root, path); - if (ret) - break; - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - slot = path->slots[0]; - } else { - slot++; - path->slots[0]++; - } + if (slot >= btrfs_header_nritems(leaf)) { + ret = btrfs_next_leaf(root, path); + if (ret < 0) + goto err; + else if (ret > 0) + break; + continue; } - advance = 1; item = btrfs_item_nr(leaf, slot); btrfs_item_key_to_cpu(leaf, &found_key, slot); @@ -4267,9 +4169,14 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, if (btrfs_key_type(&found_key) != key_type) break; if (found_key.offset < filp->f_pos) - continue; + goto next; + if (key_type == BTRFS_DIR_INDEX_KEY && + btrfs_should_delete_dir_index(&del_list, + found_key.offset)) + goto next; filp->f_pos = found_key.offset; + is_curr = 1; di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item); di_cur = 0; @@ -4278,6 +4185,9 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, while (di_cur < di_total) { struct btrfs_key location; + if (verify_dir_item(root, leaf, di)) + break; + name_len = btrfs_dir_name_len(leaf, di); if (name_len <= sizeof(tmp_name)) { name_ptr = tmp_name; @@ -4317,6 +4227,17 @@ skip: di_cur += di_len; di = (struct btrfs_dir_item *)((char *)di + di_len); } +next: + path->slots[0]++; + } + + if (key_type == BTRFS_DIR_INDEX_KEY) { + if (is_curr) + filp->f_pos++; + ret = btrfs_readdir_delayed_dir_index(filp, dirent, filldir, + &ins_list); + if (ret) + goto nopos; } /* Reached end of directory/root. Bump pos past the last item. */ @@ -4331,6 +4252,8 @@ skip: nopos: ret = 0; err: + if (key_type == BTRFS_DIR_INDEX_KEY) + btrfs_put_delayed_items(&ins_list, &del_list); btrfs_free_path(path); return ret; } @@ -4346,7 +4269,8 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) return 0; smp_mb(); - nolock = (root->fs_info->closing && root == root->fs_info->tree_root); + if (root->fs_info->closing && is_free_space_inode(root, inode)) + nolock = true; if (wbc->sync_mode == WB_SYNC_ALL) { if (nolock) @@ -4370,7 +4294,7 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) * FIXME, needs more benchmarking...there are no reasons other than performance * to keep or drop this code. */ -void btrfs_dirty_inode(struct inode *inode) +void btrfs_dirty_inode(struct inode *inode, int flags) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; @@ -4389,25 +4313,25 @@ void btrfs_dirty_inode(struct inode *inode) btrfs_end_transaction(trans, root); trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { - if (printk_ratelimit()) { - printk(KERN_ERR "btrfs: fail to " - "dirty inode %lu error %ld\n", - inode->i_ino, PTR_ERR(trans)); - } + printk_ratelimited(KERN_ERR "btrfs: fail to " + "dirty inode %llu error %ld\n", + (unsigned long long)btrfs_ino(inode), + PTR_ERR(trans)); return; } btrfs_set_trans_block_group(trans, inode); ret = btrfs_update_inode(trans, root, inode); if (ret) { - if (printk_ratelimit()) { - printk(KERN_ERR "btrfs: fail to " - "dirty inode %lu error %d\n", - inode->i_ino, ret); - } + printk_ratelimited(KERN_ERR "btrfs: fail to " + "dirty inode %llu error %d\n", + (unsigned long long)btrfs_ino(inode), + ret); } } btrfs_end_transaction(trans, root); + if (BTRFS_I(inode)->delayed_node) + btrfs_balance_delayed_items(root); } /* @@ -4423,7 +4347,7 @@ static int btrfs_set_inode_index_count(struct inode *inode) struct extent_buffer *leaf; int ret; - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY); key.offset = (u64)-1; @@ -4455,7 +4379,7 @@ static int btrfs_set_inode_index_count(struct inode *inode) leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid != inode->i_ino || + if (found_key.objectid != btrfs_ino(inode) || btrfs_key_type(&found_key) != BTRFS_DIR_INDEX_KEY) { BTRFS_I(inode)->index_cnt = 2; goto out; @@ -4476,9 +4400,12 @@ int btrfs_set_inode_index(struct inode *dir, u64 *index) int ret = 0; if (BTRFS_I(dir)->index_cnt == (u64)-1) { - ret = btrfs_set_inode_index_count(dir); - if (ret) - return ret; + ret = btrfs_inode_delayed_dir_index_count(dir); + if (ret) { + ret = btrfs_set_inode_index_count(dir); + if (ret) + return ret; + } } *index = BTRFS_I(dir)->index_cnt; @@ -4509,12 +4436,23 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, BUG_ON(!path); inode = new_inode(root->fs_info->sb); - if (!inode) + if (!inode) { + btrfs_free_path(path); return ERR_PTR(-ENOMEM); + } + + /* + * we have to initialize this early, so we can reclaim the inode + * number if we fail afterwards in this function. + */ + inode->i_ino = objectid; if (dir) { + trace_btrfs_inode_request(dir); + ret = btrfs_set_inode_index(dir, index); if (ret) { + btrfs_free_path(path); iput(inode); return ERR_PTR(ret); } @@ -4554,7 +4492,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, goto fail; inode_init_owner(inode, dir, mode); - inode->i_ino = objectid; inode_set_bytes(inode, 0); inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0], @@ -4581,12 +4518,16 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, if ((mode & S_IFREG)) { if (btrfs_test_opt(root, NODATASUM)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; - if (btrfs_test_opt(root, NODATACOW)) + if (btrfs_test_opt(root, NODATACOW) || + (BTRFS_I(dir)->flags & BTRFS_INODE_NODATACOW)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW; } insert_inode_hash(inode); inode_tree_add(inode); + + trace_btrfs_inode_new(inode); + return inode; fail: if (dir) @@ -4614,29 +4555,29 @@ int btrfs_add_link(struct btrfs_trans_handle *trans, int ret = 0; struct btrfs_key key; struct btrfs_root *root = BTRFS_I(parent_inode)->root; + u64 ino = btrfs_ino(inode); + u64 parent_ino = btrfs_ino(parent_inode); - if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { memcpy(&key, &BTRFS_I(inode)->root->root_key, sizeof(key)); } else { - key.objectid = inode->i_ino; + key.objectid = ino; btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); key.offset = 0; } - if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, key.objectid, root->root_key.objectid, - parent_inode->i_ino, - index, name, name_len); + parent_ino, index, name, name_len); } else if (add_backref) { - ret = btrfs_insert_inode_ref(trans, root, - name, name_len, inode->i_ino, - parent_inode->i_ino, index); + ret = btrfs_insert_inode_ref(trans, root, name, name_len, ino, + parent_ino, index); } if (ret == 0) { ret = btrfs_insert_dir_item(trans, root, name, name_len, - parent_inode->i_ino, &key, + parent_inode, &key, btrfs_inode_type(inode), index); BUG_ON(ret); @@ -4679,10 +4620,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, if (!new_valid_dev(rdev)) return -EINVAL; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; - /* * 2 for inode item and ref * 2 for dir items @@ -4694,14 +4631,19 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, btrfs_set_trans_block_group(trans, dir); + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_unlock; + inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, + dentry->d_name.len, btrfs_ino(dir), objectid, BTRFS_I(dir)->block_group, mode, &index); - err = PTR_ERR(inode); - if (IS_ERR(inode)) + if (IS_ERR(inode)) { + err = PTR_ERR(inode); goto out_unlock; + } - err = btrfs_init_inode_security(trans, inode, dir); + err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); if (err) { drop_inode = 1; goto out_unlock; @@ -4741,9 +4683,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, u64 objectid; u64 index = 0; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; /* * 2 for inode item and ref * 2 for dir items @@ -4755,14 +4694,19 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, btrfs_set_trans_block_group(trans, dir); + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_unlock; + inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, + dentry->d_name.len, btrfs_ino(dir), objectid, BTRFS_I(dir)->block_group, mode, &index); - err = PTR_ERR(inode); - if (IS_ERR(inode)) + if (IS_ERR(inode)) { + err = PTR_ERR(inode); goto out_unlock; + } - err = btrfs_init_inode_security(trans, inode, dir); + err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); if (err) { drop_inode = 1; goto out_unlock; @@ -4803,30 +4747,31 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, int err; int drop_inode = 0; - if (inode->i_nlink == 0) - return -ENOENT; - /* do not allow sys_link's with other subvols of the same device */ if (root->objectid != BTRFS_I(inode)->root->objectid) - return -EPERM; + return -EXDEV; - btrfs_inc_nlink(inode); - inode->i_ctime = CURRENT_TIME; + if (inode->i_nlink == ~0U) + return -EMLINK; err = btrfs_set_inode_index(dir, &index); if (err) goto fail; /* - * 1 item for inode ref + * 2 items for inode and inode ref * 2 items for dir items + * 1 item for parent inode */ - trans = btrfs_start_transaction(root, 3); + trans = btrfs_start_transaction(root, 5); if (IS_ERR(trans)) { err = PTR_ERR(trans); goto fail; } + btrfs_inc_nlink(inode); + inode->i_ctime = CURRENT_TIME; + btrfs_set_trans_block_group(trans, dir); ihold(inode); @@ -4865,10 +4810,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) u64 index = 0; unsigned long nr = 1; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; - /* * 2 items for inode and ref * 2 items for dir items @@ -4879,8 +4820,12 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) return PTR_ERR(trans); btrfs_set_trans_block_group(trans, dir); + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_fail; + inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, + dentry->d_name.len, btrfs_ino(dir), objectid, BTRFS_I(dir)->block_group, S_IFDIR | mode, &index); if (IS_ERR(inode)) { @@ -4890,7 +4835,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) drop_on_err = 1; - err = btrfs_init_inode_security(trans, inode, dir); + err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); if (err) goto out_fail; @@ -4964,6 +4909,8 @@ static noinline int uncompress_inline(struct btrfs_path *path, inline_size = btrfs_file_extent_inline_item_len(leaf, btrfs_item_nr(leaf, path->slots[0])); tmp = kmalloc(inline_size, GFP_NOFS); + if (!tmp) + return -ENOMEM; ptr = btrfs_file_extent_inline_start(item); read_extent_buffer(leaf, tmp, ptr, inline_size); @@ -5001,7 +4948,7 @@ struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page, u64 bytenr; u64 extent_start = 0; u64 extent_end = 0; - u64 objectid = inode->i_ino; + u64 objectid = btrfs_ino(inode); u32 found_type; struct btrfs_path *path = NULL; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -5029,7 +4976,7 @@ again: else goto out; } - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) { err = -ENOMEM; goto out; @@ -5183,7 +5130,7 @@ again: kunmap(page); free_extent_map(em); em = NULL; - btrfs_release_path(root, path); + btrfs_release_path(path); trans = btrfs_join_transaction(root, 1); if (IS_ERR(trans)) return ERR_CAST(trans); @@ -5196,7 +5143,7 @@ again: btrfs_mark_buffer_dirty(leaf); } set_extent_uptodate(io_tree, em->start, - extent_map_end(em) - 1, GFP_NOFS); + extent_map_end(em) - 1, NULL, GFP_NOFS); goto insert; } else { printk(KERN_ERR "btrfs unknown found_type %d\n", found_type); @@ -5209,7 +5156,7 @@ not_found_em: em->block_start = EXTENT_MAP_HOLE; set_bit(EXTENT_FLAG_VACANCY, &em->flags); insert: - btrfs_release_path(root, path); + btrfs_release_path(path); if (em->start > start || extent_map_end(em) <= start) { printk(KERN_ERR "Btrfs: bad extent! em: [%llu %llu] passed " "[%llu %llu]\n", (unsigned long long)em->start, @@ -5263,6 +5210,9 @@ insert: } write_unlock(&em_tree->lock); out: + + trace_btrfs_get_extent(root, em); + if (path) btrfs_free_path(path); if (trans) { @@ -5277,23 +5227,161 @@ out: return em; } +struct extent_map *btrfs_get_extent_fiemap(struct inode *inode, struct page *page, + size_t pg_offset, u64 start, u64 len, + int create) +{ + struct extent_map *em; + struct extent_map *hole_em = NULL; + u64 range_start = start; + u64 end; + u64 found; + u64 found_end; + int err = 0; + + em = btrfs_get_extent(inode, page, pg_offset, start, len, create); + if (IS_ERR(em)) + return em; + if (em) { + /* + * if our em maps to a hole, there might + * actually be delalloc bytes behind it + */ + if (em->block_start != EXTENT_MAP_HOLE) + return em; + else + hole_em = em; + } + + /* check to see if we've wrapped (len == -1 or similar) */ + end = start + len; + if (end < start) + end = (u64)-1; + else + end -= 1; + + em = NULL; + + /* ok, we didn't find anything, lets look for delalloc */ + found = count_range_bits(&BTRFS_I(inode)->io_tree, &range_start, + end, len, EXTENT_DELALLOC, 1); + found_end = range_start + found; + if (found_end < range_start) + found_end = (u64)-1; + + /* + * we didn't find anything useful, return + * the original results from get_extent() + */ + if (range_start > end || found_end <= start) { + em = hole_em; + hole_em = NULL; + goto out; + } + + /* adjust the range_start to make sure it doesn't + * go backwards from the start they passed in + */ + range_start = max(start,range_start); + found = found_end - range_start; + + if (found > 0) { + u64 hole_start = start; + u64 hole_len = len; + + em = alloc_extent_map(); + if (!em) { + err = -ENOMEM; + goto out; + } + /* + * when btrfs_get_extent can't find anything it + * returns one huge hole + * + * make sure what it found really fits our range, and + * adjust to make sure it is based on the start from + * the caller + */ + if (hole_em) { + u64 calc_end = extent_map_end(hole_em); + + if (calc_end <= start || (hole_em->start > end)) { + free_extent_map(hole_em); + hole_em = NULL; + } else { + hole_start = max(hole_em->start, start); + hole_len = calc_end - hole_start; + } + } + em->bdev = NULL; + if (hole_em && range_start > hole_start) { + /* our hole starts before our delalloc, so we + * have to return just the parts of the hole + * that go until the delalloc starts + */ + em->len = min(hole_len, + range_start - hole_start); + em->start = hole_start; + em->orig_start = hole_start; + /* + * don't adjust block start at all, + * it is fixed at EXTENT_MAP_HOLE + */ + em->block_start = hole_em->block_start; + em->block_len = hole_len; + } else { + em->start = range_start; + em->len = found; + em->orig_start = range_start; + em->block_start = EXTENT_MAP_DELALLOC; + em->block_len = found; + } + } else if (hole_em) { + return hole_em; + } +out: + + free_extent_map(hole_em); + if (err) { + free_extent_map(em); + return ERR_PTR(err); + } + return em; +} + static struct extent_map *btrfs_new_extent_direct(struct inode *inode, + struct extent_map *em, u64 start, u64 len) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; - struct extent_map *em; struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; struct btrfs_key ins; u64 alloc_hint; int ret; + bool insert = false; - btrfs_drop_extent_cache(inode, start, start + len - 1, 0); + /* + * Ok if the extent map we looked up is a hole and is for the exact + * range we want, there is no reason to allocate a new one, however if + * it is not right then we need to free this one and drop the cache for + * our range. + */ + if (em->block_start != EXTENT_MAP_HOLE || em->start != start || + em->len != len) { + free_extent_map(em); + em = NULL; + insert = true; + btrfs_drop_extent_cache(inode, start, start + len - 1, 0); + } trans = btrfs_join_transaction(root, 0); if (IS_ERR(trans)) return ERR_CAST(trans); + if (start <= BTRFS_I(inode)->disk_i_size && len < 64 * 1024) + btrfs_add_inode_defrag(trans, inode); + trans->block_rsv = &root->fs_info->delalloc_block_rsv; alloc_hint = get_extent_allocation_hint(inode, start, len); @@ -5304,10 +5392,12 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, goto out; } - em = alloc_extent_map(GFP_NOFS); if (!em) { - em = ERR_PTR(-ENOMEM); - goto out; + em = alloc_extent_map(); + if (!em) { + em = ERR_PTR(-ENOMEM); + goto out; + } } em->start = start; @@ -5317,9 +5407,15 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, em->block_start = ins.objectid; em->block_len = ins.offset; em->bdev = root->fs_info->fs_devices->latest_bdev; + + /* + * We need to do this because if we're using the original em we searched + * for, we could have EXTENT_FLAG_VACANCY set, and we don't want that. + */ + em->flags = 0; set_bit(EXTENT_FLAG_PINNED, &em->flags); - while (1) { + while (insert) { write_lock(&em_tree->lock); ret = add_extent_mapping(em_tree, em); write_unlock(&em_tree->lock); @@ -5363,7 +5459,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans, if (!path) return -ENOMEM; - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode), offset, 0); if (ret < 0) goto out; @@ -5380,7 +5476,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans, ret = 0; leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, slot); - if (key.objectid != inode->i_ino || + if (key.objectid != btrfs_ino(inode) || key.type != BTRFS_EXTENT_DATA_KEY) { /* not our file or wrong item type, must cow */ goto out; @@ -5414,7 +5510,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans, * look for other files referencing this extent, if we * find any we must cow */ - if (btrfs_cross_ref_exist(trans, root, inode->i_ino, + if (btrfs_cross_ref_exist(trans, root, btrfs_ino(inode), key.offset - backref_offset, disk_bytenr)) goto out; @@ -5537,8 +5633,7 @@ must_cow: * it above */ len = bh_result->b_size; - free_extent_map(em); - em = btrfs_new_extent_direct(inode, start, len); + em = btrfs_new_extent_direct(inode, em, start, len); if (IS_ERR(em)) return PTR_ERR(em); len = min(len, em->len - (start - em->start)); @@ -5605,9 +5700,10 @@ static void btrfs_endio_direct_read(struct bio *bio, int err) flush_dcache_page(bvec->bv_page); if (csum != *private) { - printk(KERN_ERR "btrfs csum failed ino %lu off" + printk(KERN_ERR "btrfs csum failed ino %llu off" " %llu csum %u private %u\n", - inode->i_ino, (unsigned long long)start, + (unsigned long long)btrfs_ino(inode), + (unsigned long long)start, csum, *private); err = -EIO; } @@ -5624,6 +5720,10 @@ static void btrfs_endio_direct_read(struct bio *bio, int err) kfree(dip->csums); kfree(dip); + + /* If we had a csum failure make sure to clear the uptodate flag */ + if (err) + clear_bit(BIO_UPTODATE, &bio->bi_flags); dio_end_io(bio, err); } @@ -5697,8 +5797,10 @@ again: } add_pending_csums(trans, inode, ordered->file_offset, &ordered->list); - btrfs_ordered_update_i_size(inode, 0, ordered); - btrfs_update_inode(trans, root, inode); + ret = btrfs_ordered_update_i_size(inode, 0, ordered); + if (!ret) + btrfs_update_inode(trans, root, inode); + ret = 0; out_unlock: unlock_extent_cached(&BTRFS_I(inode)->io_tree, ordered->file_offset, ordered->file_offset + ordered->len - 1, @@ -5725,6 +5827,10 @@ out_done: kfree(dip->csums); kfree(dip); + + /* If we had an error make sure to clear the uptodate flag */ + if (err) + clear_bit(BIO_UPTODATE, &bio->bi_flags); dio_end_io(bio, err); } @@ -5744,9 +5850,9 @@ static void btrfs_end_dio_bio(struct bio *bio, int err) struct btrfs_dio_private *dip = bio->bi_private; if (err) { - printk(KERN_ERR "btrfs direct IO failed ino %lu rw %lu " + printk(KERN_ERR "btrfs direct IO failed ino %llu rw %lu " "sector %#Lx len %u err no %d\n", - dip->inode->i_ino, bio->bi_rw, + (unsigned long long)btrfs_ino(dip->inode), bio->bi_rw, (unsigned long long)bio->bi_sector, bio->bi_size, err); dip->errors = 1; @@ -5780,7 +5886,7 @@ static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev, static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, int rw, u64 file_offset, int skip_sum, - u32 *csums) + u32 *csums, int async_submit) { int write = rw & REQ_WRITE; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -5791,18 +5897,33 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, if (ret) goto err; - if (write && !skip_sum) { + if (skip_sum) + goto map; + + if (write && async_submit) { ret = btrfs_wq_submit_bio(root->fs_info, inode, rw, bio, 0, 0, file_offset, __btrfs_submit_bio_start_direct_io, __btrfs_submit_bio_done); goto err; - } else if (!skip_sum) - btrfs_lookup_bio_sums_dio(root, inode, bio, + } else if (write) { + /* + * If we aren't doing async submit, calculate the csum of the + * bio now. + */ + ret = btrfs_csum_one_bio(root, inode, bio, file_offset, 1); + if (ret) + goto err; + } else if (!skip_sum) { + ret = btrfs_lookup_bio_sums_dio(root, inode, bio, file_offset, csums); + if (ret) + goto err; + } - ret = btrfs_map_bio(root, rw, bio, 0, 1); +map: + ret = btrfs_map_bio(root, rw, bio, 0, async_submit); err: bio_put(bio); return ret; @@ -5824,22 +5945,30 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, int nr_pages = 0; u32 *csums = dip->csums; int ret = 0; - - bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); - if (!bio) - return -ENOMEM; - bio->bi_private = dip; - bio->bi_end_io = btrfs_end_dio_bio; - atomic_inc(&dip->pending_bios); + int async_submit = 0; + int write = rw & REQ_WRITE; map_length = orig_bio->bi_size; ret = btrfs_map_block(map_tree, READ, start_sector << 9, &map_length, NULL, 0); if (ret) { - bio_put(bio); + bio_put(orig_bio); return -EIO; } + if (map_length >= orig_bio->bi_size) { + bio = orig_bio; + goto submit; + } + + async_submit = 1; + bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); + if (!bio) + return -ENOMEM; + bio->bi_private = dip; + bio->bi_end_io = btrfs_end_dio_bio; + atomic_inc(&dip->pending_bios); + while (bvec <= (orig_bio->bi_io_vec + orig_bio->bi_vcnt - 1)) { if (unlikely(map_length < submit_len + bvec->bv_len || bio_add_page(bio, bvec->bv_page, bvec->bv_len, @@ -5853,14 +5982,15 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, atomic_inc(&dip->pending_bios); ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, - csums); + csums, async_submit); if (ret) { bio_put(bio); atomic_dec(&dip->pending_bios); goto out_err; } - if (!skip_sum) + /* Write's use the ordered csums */ + if (!write && !skip_sum) csums = csums + nr_pages; start_sector += submit_len >> 9; file_offset += submit_len; @@ -5889,8 +6019,9 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, } } +submit: ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, - csums); + csums, async_submit); if (!ret) return 0; @@ -5928,9 +6059,11 @@ static void btrfs_submit_direct(int rw, struct bio *bio, struct inode *inode, } dip->csums = NULL; - if (!skip_sum) { + /* Write's use the ordered csum stuff, so we don't need dip->csums */ + if (!write && !skip_sum) { dip->csums = kmalloc(sizeof(u32) * bio->bi_vcnt, GFP_NOFS); if (!dip->csums) { + kfree(dip); ret = -ENOMEM; goto free_ordered; } @@ -5983,6 +6116,7 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io unsigned long nr_segs) { int seg; + int i; size_t size; unsigned long addr; unsigned blocksize_mask = root->sectorsize - 1; @@ -5997,8 +6131,22 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io addr = (unsigned long)iov[seg].iov_base; size = iov[seg].iov_len; end += size; - if ((addr & blocksize_mask) || (size & blocksize_mask)) + if ((addr & blocksize_mask) || (size & blocksize_mask)) goto out; + + /* If this is a write we don't need to check anymore */ + if (rw & WRITE) + continue; + + /* + * Check to make sure we don't have duplicate iov_base's in this + * iovec, if so return EINVAL, otherwise we'll get csum errors + * when reading back. + */ + for (i = seg + 1; i < nr_segs; i++) { + if (iov[seg].iov_base == iov[i].iov_base) + goto out; + } } retval = 0; out: @@ -6099,7 +6247,7 @@ out: static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, __u64 start, __u64 len) { - return extent_fiemap(inode, fieinfo, start, len, btrfs_get_extent); + return extent_fiemap(inode, fieinfo, start, len, btrfs_get_extent_fiemap); } int btrfs_readpage(struct file *file, struct page *page) @@ -6349,28 +6497,42 @@ out: return ret; } -static void btrfs_truncate(struct inode *inode) +static int btrfs_truncate(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; int ret; + int err = 0; struct btrfs_trans_handle *trans; unsigned long nr; u64 mask = root->sectorsize - 1; - if (!S_ISREG(inode->i_mode)) { - WARN_ON(1); - return; - } - ret = btrfs_truncate_page(inode->i_mapping, inode->i_size); if (ret) - return; + return ret; btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1); btrfs_ordered_update_i_size(inode, inode->i_size, NULL); + trans = btrfs_start_transaction(root, 5); + if (IS_ERR(trans)) + return PTR_ERR(trans); + + btrfs_set_trans_block_group(trans, inode); + + ret = btrfs_orphan_add(trans, inode); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; + } + + nr = trans->blocks_used; + btrfs_end_transaction(trans, root); + btrfs_btree_balance_dirty(root, nr); + + /* Now start a transaction for the truncate */ trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; @@ -6397,29 +6559,38 @@ static void btrfs_truncate(struct inode *inode) while (1) { if (!trans) { trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; } ret = btrfs_block_rsv_check(trans, root, root->orphan_block_rsv, 0, 5); - if (ret) { - BUG_ON(ret != -EAGAIN); + if (ret == -EAGAIN) { ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); + if (ret) + return ret; trans = NULL; continue; + } else if (ret) { + err = ret; + break; } ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, BTRFS_EXTENT_DATA_KEY); - if (ret != -EAGAIN) + if (ret != -EAGAIN) { + err = ret; break; + } ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + if (ret) { + err = ret; + break; + } nr = trans->blocks_used; btrfs_end_transaction(trans, root); @@ -6429,16 +6600,27 @@ static void btrfs_truncate(struct inode *inode) if (ret == 0 && inode->i_nlink > 0) { ret = btrfs_orphan_del(trans, inode); - BUG_ON(ret); + if (ret) + err = ret; + } else if (ret && inode->i_nlink > 0) { + /* + * Failed to do the truncate, remove us from the in memory + * orphan list. + */ + ret = btrfs_orphan_del(NULL, inode); } ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + if (ret && !err) + err = ret; nr = trans->blocks_used; ret = btrfs_end_transaction_throttle(trans, root); - BUG_ON(ret); + if (ret && !err) + err = ret; btrfs_btree_balance_dirty(root, nr); + + return err; } /* @@ -6505,19 +6687,21 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) ei->index_cnt = (u64)-1; ei->last_unlink_trans = 0; - spin_lock_init(&ei->accounting_lock); atomic_set(&ei->outstanding_extents, 0); - ei->reserved_extents = 0; + atomic_set(&ei->reserved_extents, 0); ei->ordered_data_close = 0; ei->orphan_meta_reserved = 0; ei->dummy_inode = 0; + ei->in_defrag = 0; ei->force_compress = BTRFS_COMPRESS_NONE; + ei->delayed_node = NULL; + inode = &ei->vfs_inode; - extent_map_tree_init(&ei->extent_tree, GFP_NOFS); - extent_io_tree_init(&ei->io_tree, &inode->i_data, GFP_NOFS); - extent_io_tree_init(&ei->io_failure_tree, &inode->i_data, GFP_NOFS); + extent_map_tree_init(&ei->extent_tree); + extent_io_tree_init(&ei->io_tree, &inode->i_data); + extent_io_tree_init(&ei->io_failure_tree, &inode->i_data); mutex_init(&ei->log_mutex); btrfs_ordered_inode_tree_init(&ei->ordered_tree); INIT_LIST_HEAD(&ei->i_orphan); @@ -6543,7 +6727,7 @@ void btrfs_destroy_inode(struct inode *inode) WARN_ON(!list_empty(&inode->i_dentry)); WARN_ON(inode->i_data.nrpages); WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents)); - WARN_ON(BTRFS_I(inode)->reserved_extents); + WARN_ON(atomic_read(&BTRFS_I(inode)->reserved_extents)); /* * This can happen where we create an inode, but somebody else also @@ -6581,8 +6765,8 @@ void btrfs_destroy_inode(struct inode *inode) spin_lock(&root->orphan_lock); if (!list_empty(&BTRFS_I(inode)->i_orphan)) { - printk(KERN_INFO "BTRFS: inode %lu still on the orphan list\n", - inode->i_ino); + printk(KERN_INFO "BTRFS: inode %llu still on the orphan list\n", + (unsigned long long)btrfs_ino(inode)); list_del_init(&BTRFS_I(inode)->i_orphan); } spin_unlock(&root->orphan_lock); @@ -6604,6 +6788,7 @@ void btrfs_destroy_inode(struct inode *inode) inode_tree_del(inode); btrfs_drop_extent_cache(inode, 0, (u64)-1, 0); free: + btrfs_remove_delayed_node(inode); call_rcu(&inode->i_rcu, btrfs_i_callback); } @@ -6612,7 +6797,7 @@ int btrfs_drop_inode(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; if (btrfs_root_refs(&root->root_item) == 0 && - root != root->fs_info->tree_root) + !is_free_space_inode(root, inode)) return 1; else return generic_drop_inode(inode); @@ -6635,6 +6820,8 @@ void btrfs_destroy_cachep(void) kmem_cache_destroy(btrfs_transaction_cachep); if (btrfs_path_cachep) kmem_cache_destroy(btrfs_path_cachep); + if (btrfs_free_space_cachep) + kmem_cache_destroy(btrfs_free_space_cachep); } int btrfs_init_cachep(void) @@ -6663,6 +6850,12 @@ int btrfs_init_cachep(void) if (!btrfs_path_cachep) goto fail; + btrfs_free_space_cachep = kmem_cache_create("btrfs_free_space_cache", + sizeof(struct btrfs_free_space), 0, + SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + if (!btrfs_free_space_cachep) + goto fail; + return 0; fail: btrfs_destroy_cachep(); @@ -6681,6 +6874,26 @@ static int btrfs_getattr(struct vfsmount *mnt, return 0; } +/* + * If a file is moved, it will inherit the cow and compression flags of the new + * directory. + */ +static void fixup_inode_flags(struct inode *dir, struct inode *inode) +{ + struct btrfs_inode *b_dir = BTRFS_I(dir); + struct btrfs_inode *b_inode = BTRFS_I(inode); + + if (b_dir->flags & BTRFS_INODE_NODATACOW) + b_inode->flags |= BTRFS_INODE_NODATACOW; + else + b_inode->flags &= ~BTRFS_INODE_NODATACOW; + + if (b_dir->flags & BTRFS_INODE_COMPRESS) + b_inode->flags |= BTRFS_INODE_COMPRESS; + else + b_inode->flags &= ~BTRFS_INODE_COMPRESS; +} + static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry) { @@ -6693,16 +6906,17 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, u64 index = 0; u64 root_objectid; int ret; + u64 old_ino = btrfs_ino(old_inode); - if (new_dir->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) + if (btrfs_ino(new_dir) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) return -EPERM; /* we only allow rename subvolume link between subvolumes */ - if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) + if (old_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) return -EXDEV; - if (old_inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID || - (new_inode && new_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) + if (old_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID || + (new_inode && btrfs_ino(new_inode) == BTRFS_FIRST_FREE_OBJECTID)) return -ENOTEMPTY; if (S_ISDIR(old_inode->i_mode) && new_inode && @@ -6718,7 +6932,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, filemap_flush(old_inode->i_mapping); /* close the racy window with snapshot create/destroy ioctl */ - if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) down_read(&root->fs_info->subvol_sem); /* * We want to reserve the absolute worst case amount of items. So if @@ -6729,8 +6943,10 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, * should cover the worst case number of items we'll modify. */ trans = btrfs_start_transaction(root, 20); - if (IS_ERR(trans)) - return PTR_ERR(trans); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out_notrans; + } btrfs_set_trans_block_group(trans, new_dir); @@ -6741,15 +6957,15 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (ret) goto out_fail; - if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { /* force full log commit if subvolume involved. */ root->fs_info->last_trans_log_full_commit = trans->transid; } else { ret = btrfs_insert_inode_ref(trans, dest, new_dentry->d_name.name, new_dentry->d_name.len, - old_inode->i_ino, - new_dir->i_ino, index); + old_ino, + btrfs_ino(new_dir), index); if (ret) goto out_fail; /* @@ -6765,10 +6981,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, * make sure the inode gets flushed if it is replacing * something. */ - if (new_inode && new_inode->i_size && - old_inode && S_ISREG(old_inode->i_mode)) { + if (new_inode && new_inode->i_size && S_ISREG(old_inode->i_mode)) btrfs_add_ordered_operation(trans, root, old_inode); - } old_dir->i_ctime = old_dir->i_mtime = ctime; new_dir->i_ctime = new_dir->i_mtime = ctime; @@ -6777,23 +6991,24 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (old_dentry->d_parent != new_dentry->d_parent) btrfs_record_unlink_dir(trans, old_dir, old_inode, 1); - if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { root_objectid = BTRFS_I(old_inode)->root->root_key.objectid; ret = btrfs_unlink_subvol(trans, root, old_dir, root_objectid, old_dentry->d_name.name, old_dentry->d_name.len); } else { - btrfs_inc_nlink(old_dentry->d_inode); - ret = btrfs_unlink_inode(trans, root, old_dir, - old_dentry->d_inode, - old_dentry->d_name.name, - old_dentry->d_name.len); + ret = __btrfs_unlink_inode(trans, root, old_dir, + old_dentry->d_inode, + old_dentry->d_name.name, + old_dentry->d_name.len); + if (!ret) + ret = btrfs_update_inode(trans, root, old_inode); } BUG_ON(ret); if (new_inode) { new_inode->i_ctime = CURRENT_TIME; - if (unlikely(new_inode->i_ino == + if (unlikely(btrfs_ino(new_inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { root_objectid = BTRFS_I(new_inode)->location.objectid; ret = btrfs_unlink_subvol(trans, dest, new_dir, @@ -6814,12 +7029,14 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, } } + fixup_inode_flags(new_dir, old_inode); + ret = btrfs_add_link(trans, new_dir, old_inode, new_dentry->d_name.name, new_dentry->d_name.len, 0, index); BUG_ON(ret); - if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { struct dentry *parent = dget_parent(new_dentry); btrfs_log_new_name(trans, old_inode, old_dir, parent); dput(parent); @@ -6827,8 +7044,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, } out_fail: btrfs_end_transaction_throttle(trans, root); - - if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) +out_notrans: + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) up_read(&root->fs_info->subvol_sem); return ret; @@ -6882,58 +7099,6 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput) return 0; } -int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput, - int sync) -{ - struct btrfs_inode *binode; - struct inode *inode = NULL; - - spin_lock(&root->fs_info->delalloc_lock); - while (!list_empty(&root->fs_info->delalloc_inodes)) { - binode = list_entry(root->fs_info->delalloc_inodes.next, - struct btrfs_inode, delalloc_inodes); - inode = igrab(&binode->vfs_inode); - if (inode) { - list_move_tail(&binode->delalloc_inodes, - &root->fs_info->delalloc_inodes); - break; - } - - list_del_init(&binode->delalloc_inodes); - cond_resched_lock(&root->fs_info->delalloc_lock); - } - spin_unlock(&root->fs_info->delalloc_lock); - - if (inode) { - if (sync) { - filemap_write_and_wait(inode->i_mapping); - /* - * We have to do this because compression doesn't - * actually set PG_writeback until it submits the pages - * for IO, which happens in an async thread, so we could - * race and not actually wait for any writeback pages - * because they've not been submitted yet. Technically - * this could still be the case for the ordered stuff - * since the async thread may not have started to do its - * work yet. If this becomes the case then we need to - * figure out a way to make sure that in writepage we - * wait for any async pages to be submitted before - * returning so that fdatawait does what its supposed to - * do. - */ - btrfs_wait_ordered_range(inode, 0, (u64)-1); - } else { - filemap_flush(inode->i_mapping); - } - if (delay_iput) - btrfs_add_delayed_iput(inode); - else - iput(inode); - return 1; - } - return 0; -} - static int btrfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname) { @@ -6957,9 +7122,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root)) return -ENAMETOOLONG; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; /* * 2 items for inode item and ref * 2 items for dir items @@ -6971,15 +7133,20 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, btrfs_set_trans_block_group(trans, dir); + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_unlock; + inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, + dentry->d_name.len, btrfs_ino(dir), objectid, BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO, &index); - err = PTR_ERR(inode); - if (IS_ERR(inode)) + if (IS_ERR(inode)) { + err = PTR_ERR(inode); goto out_unlock; + } - err = btrfs_init_inode_security(trans, inode, dir); + err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); if (err) { drop_inode = 1; goto out_unlock; @@ -7003,7 +7170,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, path = btrfs_alloc_path(); BUG_ON(!path); - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); key.offset = 0; btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY); datasize = btrfs_file_extent_calc_inline_size(name_len); @@ -7011,6 +7178,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, datasize); if (err) { drop_inode = 1; + btrfs_free_path(path); goto out_unlock; } leaf = path->nodes[0]; @@ -7215,7 +7383,6 @@ static const struct address_space_operations btrfs_aops = { .writepage = btrfs_writepage, .writepages = btrfs_writepages, .readpages = btrfs_readpages, - .sync_page = block_sync_page, .direct_IO = btrfs_direct_IO, .invalidatepage = btrfs_invalidatepage, .releasepage = btrfs_releasepage, @@ -7231,7 +7398,6 @@ static const struct address_space_operations btrfs_symlink_aops = { }; static const struct inode_operations btrfs_file_inode_operations = { - .truncate = btrfs_truncate, .getattr = btrfs_getattr, .setattr = btrfs_setattr, .setxattr = btrfs_setxattr, diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 02d224e8c83f..85e818ce00c5 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -40,6 +40,7 @@ #include <linux/xattr.h> #include <linux/vmalloc.h> #include <linux/slab.h> +#include <linux/blkdev.h> #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -49,6 +50,7 @@ #include "print-tree.h" #include "volumes.h" #include "locking.h" +#include "inode-map.h" /* Mask out flags that are inappropriate for the given type of inode. */ static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags) @@ -80,6 +82,13 @@ static unsigned int btrfs_flags_to_ioctl(unsigned int flags) iflags |= FS_NOATIME_FL; if (flags & BTRFS_INODE_DIRSYNC) iflags |= FS_DIRSYNC_FL; + if (flags & BTRFS_INODE_NODATACOW) + iflags |= FS_NOCOW_FL; + + if ((flags & BTRFS_INODE_COMPRESS) && !(flags & BTRFS_INODE_NOCOMPRESS)) + iflags |= FS_COMPR_FL; + else if (flags & BTRFS_INODE_NOCOMPRESS) + iflags |= FS_NOCOMP_FL; return iflags; } @@ -138,6 +147,21 @@ static int btrfs_ioctl_getflags(struct file *file, void __user *arg) return 0; } +static int check_flags(unsigned int flags) +{ + if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \ + FS_NOATIME_FL | FS_NODUMP_FL | \ + FS_SYNC_FL | FS_DIRSYNC_FL | \ + FS_NOCOMP_FL | FS_COMPR_FL | + FS_NOCOW_FL)) + return -EOPNOTSUPP; + + if ((flags & FS_NOCOMP_FL) && (flags & FS_COMPR_FL)) + return -EINVAL; + + return 0; +} + static int btrfs_ioctl_setflags(struct file *file, void __user *arg) { struct inode *inode = file->f_path.dentry->d_inode; @@ -153,12 +177,11 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) if (copy_from_user(&flags, arg, sizeof(flags))) return -EFAULT; - if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \ - FS_NOATIME_FL | FS_NODUMP_FL | \ - FS_SYNC_FL | FS_DIRSYNC_FL)) - return -EOPNOTSUPP; + ret = check_flags(flags); + if (ret) + return ret; - if (!is_owner_or_cap(inode)) + if (!inode_owner_or_capable(inode)) return -EACCES; mutex_lock(&inode->i_mutex); @@ -200,7 +223,25 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) ip->flags |= BTRFS_INODE_DIRSYNC; else ip->flags &= ~BTRFS_INODE_DIRSYNC; + if (flags & FS_NOCOW_FL) + ip->flags |= BTRFS_INODE_NODATACOW; + else + ip->flags &= ~BTRFS_INODE_NODATACOW; + /* + * The COMPRESS flag can only be changed by users, while the NOCOMPRESS + * flag may be changed automatically if compression code won't make + * things smaller. + */ + if (flags & FS_NOCOMP_FL) { + ip->flags &= ~BTRFS_INODE_COMPRESS; + ip->flags |= BTRFS_INODE_NOCOMPRESS; + } else if (flags & FS_COMPR_FL) { + ip->flags |= BTRFS_INODE_COMPRESS; + ip->flags &= ~BTRFS_INODE_NOCOMPRESS; + } else { + ip->flags &= ~(BTRFS_INODE_COMPRESS | BTRFS_INODE_NOCOMPRESS); + } trans = btrfs_join_transaction(root, 1); BUG_ON(IS_ERR(trans)); @@ -213,9 +254,11 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) btrfs_end_transaction(trans, root); mnt_drop_write(file->f_path.mnt); + + ret = 0; out_unlock: mutex_unlock(&inode->i_mutex); - return 0; + return ret; } static int btrfs_ioctl_getversion(struct file *file, int __user *arg) @@ -225,6 +268,50 @@ static int btrfs_ioctl_getversion(struct file *file, int __user *arg) return put_user(inode->i_generation, arg); } +static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) +{ + struct btrfs_root *root = fdentry(file)->d_sb->s_fs_info; + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_device *device; + struct request_queue *q; + struct fstrim_range range; + u64 minlen = ULLONG_MAX; + u64 num_devices = 0; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + rcu_read_lock(); + list_for_each_entry_rcu(device, &fs_info->fs_devices->devices, + dev_list) { + if (!device->bdev) + continue; + q = bdev_get_queue(device->bdev); + if (blk_queue_discard(q)) { + num_devices++; + minlen = min((u64)q->limits.discard_granularity, + minlen); + } + } + rcu_read_unlock(); + if (!num_devices) + return -EOPNOTSUPP; + + if (copy_from_user(&range, arg, sizeof(range))) + return -EFAULT; + + range.minlen = max(range.minlen, minlen); + ret = btrfs_trim_fs(root, &range); + if (ret < 0) + return ret; + + if (copy_to_user(arg, &range, sizeof(range))) + return -EFAULT; + + return 0; +} + static noinline int create_subvol(struct btrfs_root *root, struct dentry *dentry, char *name, int namelen, @@ -244,8 +331,7 @@ static noinline int create_subvol(struct btrfs_root *root, u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID; u64 index = 0; - ret = btrfs_find_free_objectid(NULL, root->fs_info->tree_root, - 0, &objectid); + ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid); if (ret) { dput(parent); return ret; @@ -294,6 +380,10 @@ static noinline int create_subvol(struct btrfs_root *root, inode_item->nbytes = cpu_to_le64(root->leafsize); inode_item->mode = cpu_to_le32(S_IFDIR | 0755); + root_item.flags = 0; + root_item.byte_limit = 0; + inode_item->flags = cpu_to_le64(BTRFS_INODE_ROOT_ITEM_INIT); + btrfs_set_root_bytenr(&root_item, leaf->start); btrfs_set_root_generation(&root_item, trans->transid); btrfs_set_root_level(&root_item, 0); @@ -333,7 +423,7 @@ static noinline int create_subvol(struct btrfs_root *root, BUG_ON(ret); ret = btrfs_insert_dir_item(trans, root, - name, namelen, dir->i_ino, &key, + name, namelen, dir, &key, BTRFS_FT_DIR, index); if (ret) goto fail; @@ -344,7 +434,7 @@ static noinline int create_subvol(struct btrfs_root *root, ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, objectid, root->root_key.objectid, - dir->i_ino, index, name, namelen); + btrfs_ino(dir), index, name, namelen); BUG_ON(ret); @@ -409,7 +499,9 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, if (ret) goto fail; - btrfs_orphan_cleanup(pending_snapshot->snap); + ret = btrfs_orphan_cleanup(pending_snapshot->snap); + if (ret) + goto fail; parent = dget_parent(dentry); inode = btrfs_lookup_dentry(parent->d_inode, dentry); @@ -564,6 +656,106 @@ out_unlock: return error; } +/* + * When we're defragging a range, we don't want to kick it off again + * if it is really just waiting for delalloc to send it down. + * If we find a nice big extent or delalloc range for the bytes in the + * file you want to defrag, we return 0 to let you know to skip this + * part of the file + */ +static int check_defrag_in_cache(struct inode *inode, u64 offset, int thresh) +{ + struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; + struct extent_map *em = NULL; + struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; + u64 end; + + read_lock(&em_tree->lock); + em = lookup_extent_mapping(em_tree, offset, PAGE_CACHE_SIZE); + read_unlock(&em_tree->lock); + + if (em) { + end = extent_map_end(em); + free_extent_map(em); + if (end - offset > thresh) + return 0; + } + /* if we already have a nice delalloc here, just stop */ + thresh /= 2; + end = count_range_bits(io_tree, &offset, offset + thresh, + thresh, EXTENT_DELALLOC, 1); + if (end >= thresh) + return 0; + return 1; +} + +/* + * helper function to walk through a file and find extents + * newer than a specific transid, and smaller than thresh. + * + * This is used by the defragging code to find new and small + * extents + */ +static int find_new_extents(struct btrfs_root *root, + struct inode *inode, u64 newer_than, + u64 *off, int thresh) +{ + struct btrfs_path *path; + struct btrfs_key min_key; + struct btrfs_key max_key; + struct extent_buffer *leaf; + struct btrfs_file_extent_item *extent; + int type; + int ret; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + min_key.objectid = inode->i_ino; + min_key.type = BTRFS_EXTENT_DATA_KEY; + min_key.offset = *off; + + max_key.objectid = inode->i_ino; + max_key.type = (u8)-1; + max_key.offset = (u64)-1; + + path->keep_locks = 1; + + while(1) { + ret = btrfs_search_forward(root, &min_key, &max_key, + path, 0, newer_than); + if (ret != 0) + goto none; + if (min_key.objectid != inode->i_ino) + goto none; + if (min_key.type != BTRFS_EXTENT_DATA_KEY) + goto none; + + leaf = path->nodes[0]; + extent = btrfs_item_ptr(leaf, path->slots[0], + struct btrfs_file_extent_item); + + type = btrfs_file_extent_type(leaf, extent); + if (type == BTRFS_FILE_EXTENT_REG && + btrfs_file_extent_num_bytes(leaf, extent) < thresh && + check_defrag_in_cache(inode, min_key.offset, thresh)) { + *off = min_key.offset; + btrfs_free_path(path); + return 0; + } + + if (min_key.offset == (u64)-1) + goto none; + + min_key.offset++; + btrfs_release_path(path); + } +none: + btrfs_free_path(path); + return -ENOENT; +} + static int should_defrag_range(struct inode *inode, u64 start, u64 len, int thresh, u64 *last_len, u64 *skip, u64 *defrag_end) @@ -573,10 +765,6 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; int ret = 1; - - if (thresh == 0) - thresh = 256 * 1024; - /* * make sure that once we start defragging and extent, we keep on * defragging it @@ -635,27 +823,176 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, return ret; } -static int btrfs_defrag_file(struct file *file, - struct btrfs_ioctl_defrag_range_args *range) +/* + * it doesn't do much good to defrag one or two pages + * at a time. This pulls in a nice chunk of pages + * to COW and defrag. + * + * It also makes sure the delalloc code has enough + * dirty data to avoid making new small extents as part + * of the defrag + * + * It's a good idea to start RA on this range + * before calling this. + */ +static int cluster_pages_for_defrag(struct inode *inode, + struct page **pages, + unsigned long start_index, + int num_pages) { - struct inode *inode = fdentry(file)->d_inode; - struct btrfs_root *root = BTRFS_I(inode)->root; - struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; + unsigned long file_end; + u64 isize = i_size_read(inode); + u64 page_start; + u64 page_end; + int ret; + int i; + int i_done; struct btrfs_ordered_extent *ordered; - struct page *page; + struct extent_state *cached_state = NULL; + + if (isize == 0) + return 0; + file_end = (isize - 1) >> PAGE_CACHE_SHIFT; + + ret = btrfs_delalloc_reserve_space(inode, + num_pages << PAGE_CACHE_SHIFT); + if (ret) + return ret; +again: + ret = 0; + i_done = 0; + + /* step one, lock all the pages */ + for (i = 0; i < num_pages; i++) { + struct page *page; + page = grab_cache_page(inode->i_mapping, + start_index + i); + if (!page) + break; + + if (!PageUptodate(page)) { + btrfs_readpage(NULL, page); + lock_page(page); + if (!PageUptodate(page)) { + unlock_page(page); + page_cache_release(page); + ret = -EIO; + break; + } + } + isize = i_size_read(inode); + file_end = (isize - 1) >> PAGE_CACHE_SHIFT; + if (!isize || page->index > file_end || + page->mapping != inode->i_mapping) { + /* whoops, we blew past eof, skip this page */ + unlock_page(page); + page_cache_release(page); + break; + } + pages[i] = page; + i_done++; + } + if (!i_done || ret) + goto out; + + if (!(inode->i_sb->s_flags & MS_ACTIVE)) + goto out; + + /* + * so now we have a nice long stream of locked + * and up to date pages, lets wait on them + */ + for (i = 0; i < i_done; i++) + wait_on_page_writeback(pages[i]); + + page_start = page_offset(pages[0]); + page_end = page_offset(pages[i_done - 1]) + PAGE_CACHE_SIZE; + + lock_extent_bits(&BTRFS_I(inode)->io_tree, + page_start, page_end - 1, 0, &cached_state, + GFP_NOFS); + ordered = btrfs_lookup_first_ordered_extent(inode, page_end - 1); + if (ordered && + ordered->file_offset + ordered->len > page_start && + ordered->file_offset < page_end) { + btrfs_put_ordered_extent(ordered); + unlock_extent_cached(&BTRFS_I(inode)->io_tree, + page_start, page_end - 1, + &cached_state, GFP_NOFS); + for (i = 0; i < i_done; i++) { + unlock_page(pages[i]); + page_cache_release(pages[i]); + } + btrfs_wait_ordered_range(inode, page_start, + page_end - page_start); + goto again; + } + if (ordered) + btrfs_put_ordered_extent(ordered); + + clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, + page_end - 1, EXTENT_DIRTY | EXTENT_DELALLOC | + EXTENT_DO_ACCOUNTING, 0, 0, &cached_state, + GFP_NOFS); + + if (i_done != num_pages) { + atomic_inc(&BTRFS_I(inode)->outstanding_extents); + btrfs_delalloc_release_space(inode, + (num_pages - i_done) << PAGE_CACHE_SHIFT); + } + + + btrfs_set_extent_delalloc(inode, page_start, page_end - 1, + &cached_state); + + unlock_extent_cached(&BTRFS_I(inode)->io_tree, + page_start, page_end - 1, &cached_state, + GFP_NOFS); + + for (i = 0; i < i_done; i++) { + clear_page_dirty_for_io(pages[i]); + ClearPageChecked(pages[i]); + set_page_extent_mapped(pages[i]); + set_page_dirty(pages[i]); + unlock_page(pages[i]); + page_cache_release(pages[i]); + } + return i_done; +out: + for (i = 0; i < i_done; i++) { + unlock_page(pages[i]); + page_cache_release(pages[i]); + } + btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT); + return ret; + +} + +int btrfs_defrag_file(struct inode *inode, struct file *file, + struct btrfs_ioctl_defrag_range_args *range, + u64 newer_than, unsigned long max_to_defrag) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_super_block *disk_super; + struct file_ra_state *ra = NULL; unsigned long last_index; - unsigned long ra_pages = root->fs_info->bdi.ra_pages; - unsigned long total_read = 0; u64 features; - u64 page_start; - u64 page_end; u64 last_len = 0; u64 skip = 0; u64 defrag_end = 0; + u64 newer_off = range->start; + int newer_left = 0; unsigned long i; int ret; + int defrag_count = 0; int compress_type = BTRFS_COMPRESS_ZLIB; + int extent_thresh = range->extent_thresh; + int newer_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT; + u64 new_align = ~((u64)128 * 1024 - 1); + struct page **pages = NULL; + + if (extent_thresh == 0) + extent_thresh = 256 * 1024; if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) { if (range->compress_type > BTRFS_COMPRESS_TYPES) @@ -667,6 +1004,27 @@ static int btrfs_defrag_file(struct file *file, if (inode->i_size == 0) return 0; + /* + * if we were not given a file, allocate a readahead + * context + */ + if (!file) { + ra = kzalloc(sizeof(*ra), GFP_NOFS); + if (!ra) + return -ENOMEM; + file_ra_state_init(ra, inode->i_mapping); + } else { + ra = &file->f_ra; + } + + pages = kmalloc(sizeof(struct page *) * newer_cluster, + GFP_NOFS); + if (!pages) { + ret = -ENOMEM; + goto out_ra; + } + + /* find the last page to defrag */ if (range->start + range->len > range->start) { last_index = min_t(u64, inode->i_size - 1, range->start + range->len - 1) >> PAGE_CACHE_SHIFT; @@ -674,11 +1032,37 @@ static int btrfs_defrag_file(struct file *file, last_index = (inode->i_size - 1) >> PAGE_CACHE_SHIFT; } - i = range->start >> PAGE_CACHE_SHIFT; - while (i <= last_index) { - if (!should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT, + if (newer_than) { + ret = find_new_extents(root, inode, newer_than, + &newer_off, 64 * 1024); + if (!ret) { + range->start = newer_off; + /* + * we always align our defrag to help keep + * the extents in the file evenly spaced + */ + i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; + newer_left = newer_cluster; + } else + goto out_ra; + } else { + i = range->start >> PAGE_CACHE_SHIFT; + } + if (!max_to_defrag) + max_to_defrag = last_index - 1; + + while (i <= last_index && defrag_count < max_to_defrag) { + /* + * make sure we stop running if someone unmounts + * the FS + */ + if (!(inode->i_sb->s_flags & MS_ACTIVE)) + break; + + if (!newer_than && + !should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT, PAGE_CACHE_SIZE, - range->extent_thresh, + extent_thresh, &last_len, &skip, &defrag_end)) { unsigned long next; @@ -690,92 +1074,39 @@ static int btrfs_defrag_file(struct file *file, i = max(i + 1, next); continue; } - - if (total_read % ra_pages == 0) { - btrfs_force_ra(inode->i_mapping, &file->f_ra, file, i, - min(last_index, i + ra_pages - 1)); - } - total_read++; - mutex_lock(&inode->i_mutex); if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) BTRFS_I(inode)->force_compress = compress_type; - ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); - if (ret) - goto err_unlock; -again: - if (inode->i_size == 0 || - i > ((inode->i_size - 1) >> PAGE_CACHE_SHIFT)) { - ret = 0; - goto err_reservations; - } - - page = grab_cache_page(inode->i_mapping, i); - if (!page) { - ret = -ENOMEM; - goto err_reservations; - } - - if (!PageUptodate(page)) { - btrfs_readpage(NULL, page); - lock_page(page); - if (!PageUptodate(page)) { - unlock_page(page); - page_cache_release(page); - ret = -EIO; - goto err_reservations; - } - } + btrfs_force_ra(inode->i_mapping, ra, file, i, newer_cluster); - if (page->mapping != inode->i_mapping) { - unlock_page(page); - page_cache_release(page); - goto again; - } - - wait_on_page_writeback(page); + ret = cluster_pages_for_defrag(inode, pages, i, newer_cluster); + if (ret < 0) + goto out_ra; - if (PageDirty(page)) { - btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); - goto loop_unlock; - } + defrag_count += ret; + balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret); + i += ret; - page_start = (u64)page->index << PAGE_CACHE_SHIFT; - page_end = page_start + PAGE_CACHE_SIZE - 1; - lock_extent(io_tree, page_start, page_end, GFP_NOFS); + if (newer_than) { + if (newer_off == (u64)-1) + break; - ordered = btrfs_lookup_ordered_extent(inode, page_start); - if (ordered) { - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - unlock_page(page); - page_cache_release(page); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - goto again; + newer_off = max(newer_off + 1, + (u64)i << PAGE_CACHE_SHIFT); + + ret = find_new_extents(root, inode, + newer_than, &newer_off, + 64 * 1024); + if (!ret) { + range->start = newer_off; + i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; + newer_left = newer_cluster; + } else { + break; + } + } else { + i++; } - set_page_extent_mapped(page); - - /* - * this makes sure page_mkwrite is called on the - * page if it is dirtied again later - */ - clear_page_dirty_for_io(page); - clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start, - page_end, EXTENT_DIRTY | EXTENT_DELALLOC | - EXTENT_DO_ACCOUNTING, GFP_NOFS); - - btrfs_set_extent_delalloc(inode, page_start, page_end, NULL); - ClearPageChecked(page); - set_page_dirty(page); - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - -loop_unlock: - unlock_page(page); - page_cache_release(page); - mutex_unlock(&inode->i_mutex); - - balance_dirty_pages_ratelimited_nr(inode->i_mapping, 1); - i++; } if ((range->flags & BTRFS_DEFRAG_RANGE_START_IO)) @@ -807,12 +1138,14 @@ loop_unlock: btrfs_set_super_incompat_flags(disk_super, features); } - return 0; + if (!file) + kfree(ra); + return defrag_count; -err_reservations: - btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); -err_unlock: - mutex_unlock(&inode->i_mutex); +out_ra: + if (!file) + kfree(ra); + kfree(pages); return ret; } @@ -1038,7 +1371,7 @@ static noinline int btrfs_ioctl_subvol_getflags(struct file *file, int ret = 0; u64 flags = 0; - if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) return -EINVAL; down_read(&root->fs_info->subvol_sem); @@ -1065,18 +1398,21 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file, if (root->fs_info->sb->s_flags & MS_RDONLY) return -EROFS; - if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) return -EINVAL; if (copy_from_user(&flags, arg, sizeof(flags))) return -EFAULT; - if (flags & ~BTRFS_SUBVOL_CREATE_ASYNC) + if (flags & BTRFS_SUBVOL_CREATE_ASYNC) return -EINVAL; if (flags & ~BTRFS_SUBVOL_RDONLY) return -EOPNOTSUPP; + if (!inode_owner_or_capable(inode)) + return -EACCES; + down_write(&root->fs_info->subvol_sem); /* nothing to do */ @@ -1097,7 +1433,7 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file, goto out_reset; } - ret = btrfs_update_root(trans, root, + ret = btrfs_update_root(trans, root->fs_info->tree_root, &root->root_key, &root->root_item); btrfs_commit_transaction(trans, root); @@ -1185,7 +1521,6 @@ static noinline int copy_to_sk(struct btrfs_root *root, int nritems; int i; int slot; - int found = 0; int ret = 0; leaf = path->nodes[0]; @@ -1232,7 +1567,7 @@ static noinline int copy_to_sk(struct btrfs_root *root, item_off, item_len); *sk_offset += item_len; } - found++; + (*num_found)++; if (*num_found >= sk->nr_items) break; @@ -1251,7 +1586,6 @@ advance_key: } else ret = 1; overflow: - *num_found += found; return ret; } @@ -1308,7 +1642,7 @@ static noinline int search_ioctl(struct inode *inode, } ret = copy_to_sk(root, path, &key, sk, args->buf, &sk_offset, &num_found); - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret || num_found >= sk->nr_items) break; @@ -1415,7 +1749,7 @@ static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info, if (key.offset == BTRFS_FIRST_FREE_OBJECTID) break; - btrfs_release_path(root, path); + btrfs_release_path(path); key.objectid = key.offset; key.offset = (u64)-1; dirid = key.objectid; @@ -1545,7 +1879,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, goto out_dput; } - if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) { err = -EINVAL; goto out_dput; } @@ -1663,7 +1997,10 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp) /* the rest are all set to zero by kzalloc */ range->len = (u64)-1; } - ret = btrfs_defrag_file(file, range); + ret = btrfs_defrag_file(fdentry(file)->d_inode, file, + range, 0, 0); + if (ret > 0) + ret = 0; kfree(range); break; default: @@ -1715,6 +2052,75 @@ static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg) return ret; } +static long btrfs_ioctl_fs_info(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_fs_info_args fi_args; + struct btrfs_device *device; + struct btrfs_device *next; + struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + fi_args.num_devices = fs_devices->num_devices; + fi_args.max_id = 0; + memcpy(&fi_args.fsid, root->fs_info->fsid, sizeof(fi_args.fsid)); + + mutex_lock(&fs_devices->device_list_mutex); + list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) { + if (device->devid > fi_args.max_id) + fi_args.max_id = device->devid; + } + mutex_unlock(&fs_devices->device_list_mutex); + + if (copy_to_user(arg, &fi_args, sizeof(fi_args))) + return -EFAULT; + + return 0; +} + +static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_dev_info_args *di_args; + struct btrfs_device *dev; + struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; + int ret = 0; + char *s_uuid = NULL; + char empty_uuid[BTRFS_UUID_SIZE] = {0}; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + di_args = memdup_user(arg, sizeof(*di_args)); + if (IS_ERR(di_args)) + return PTR_ERR(di_args); + + if (memcmp(empty_uuid, di_args->uuid, BTRFS_UUID_SIZE) != 0) + s_uuid = di_args->uuid; + + mutex_lock(&fs_devices->device_list_mutex); + dev = btrfs_find_device(root, di_args->devid, s_uuid, NULL); + mutex_unlock(&fs_devices->device_list_mutex); + + if (!dev) { + ret = -ENODEV; + goto out; + } + + di_args->devid = dev->devid; + di_args->bytes_used = dev->bytes_used; + di_args->total_bytes = dev->total_bytes; + memcpy(di_args->uuid, dev->uuid, sizeof(di_args->uuid)); + strncpy(di_args->path, dev->name, sizeof(di_args->path)); + +out: + if (ret == 0 && copy_to_user(arg, di_args, sizeof(*di_args))) + ret = -EFAULT; + + kfree(di_args); + return ret; +} + static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, u64 off, u64 olen, u64 destoff) { @@ -1831,7 +2237,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, } /* clone data */ - key.objectid = src->i_ino; + key.objectid = btrfs_ino(src); key.type = BTRFS_EXTENT_DATA_KEY; key.offset = 0; @@ -1858,7 +2264,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, btrfs_item_key_to_cpu(leaf, &key, slot); if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY || - key.objectid != src->i_ino) + key.objectid != btrfs_ino(src)) break; if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) { @@ -1894,14 +2300,14 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, datal = btrfs_file_extent_ram_bytes(leaf, extent); } - btrfs_release_path(root, path); + btrfs_release_path(path); if (key.offset + datal <= off || key.offset >= off+len) goto next; memcpy(&new_key, &key, sizeof(new_key)); - new_key.objectid = inode->i_ino; + new_key.objectid = btrfs_ino(inode); if (off <= key.offset) new_key.offset = key.offset + destoff - off; else @@ -1955,7 +2361,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, ret = btrfs_inc_extent_ref(trans, root, disko, diskl, 0, root->root_key.objectid, - inode->i_ino, + btrfs_ino(inode), new_key.offset - datao); BUG_ON(ret); } @@ -2004,7 +2410,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, } btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); inode->i_mtime = inode->i_ctime = CURRENT_TIME; @@ -2025,12 +2431,12 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, btrfs_end_transaction(trans, root); } next: - btrfs_release_path(root, path); + btrfs_release_path(path); key.offset++; } ret = 0; out: - btrfs_release_path(root, path); + btrfs_release_path(path); unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); out_unlock: mutex_unlock(&src->i_mutex); @@ -2199,7 +2605,7 @@ long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg) struct btrfs_ioctl_space_info space; struct btrfs_ioctl_space_info *dest; struct btrfs_ioctl_space_info *dest_orig; - struct btrfs_ioctl_space_info *user_dest; + struct btrfs_ioctl_space_info __user *user_dest; struct btrfs_space_info *info; u64 types[] = {BTRFS_BLOCK_GROUP_DATA, BTRFS_BLOCK_GROUP_SYSTEM, @@ -2208,7 +2614,7 @@ long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg) int num_types = 4; int alloc_size; int ret = 0; - int slot_count = 0; + u64 slot_count = 0; int i, c; if (copy_from_user(&space_args, @@ -2247,7 +2653,7 @@ long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg) goto out; } - slot_count = min_t(int, space_args.space_slots, slot_count); + slot_count = min_t(u64, space_args.space_slots, slot_count); alloc_size = sizeof(*dest) * slot_count; @@ -2267,6 +2673,9 @@ long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg) for (i = 0; i < num_types; i++) { struct btrfs_space_info *tmp; + if (!slot_count) + break; + info = NULL; rcu_read_lock(); list_for_each_entry_rcu(tmp, &root->fs_info->space_info, @@ -2288,7 +2697,10 @@ long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg) memcpy(dest, &space, sizeof(space)); dest++; space_args.total_spaces++; + slot_count--; } + if (!slot_count) + break; } up_read(&info->groups_sem); } @@ -2339,12 +2751,17 @@ static noinline long btrfs_ioctl_start_sync(struct file *file, void __user *argp struct btrfs_root *root = BTRFS_I(file->f_dentry->d_inode)->root; struct btrfs_trans_handle *trans; u64 transid; + int ret; trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) return PTR_ERR(trans); transid = trans->transid; - btrfs_commit_transaction_async(trans, root, 0); + ret = btrfs_commit_transaction_async(trans, root, 0); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; + } if (argp) if (copy_to_user(argp, &transid, sizeof(transid))) @@ -2366,6 +2783,58 @@ static noinline long btrfs_ioctl_wait_sync(struct file *file, void __user *argp) return btrfs_wait_for_commit(root, transid); } +static long btrfs_ioctl_scrub(struct btrfs_root *root, void __user *arg) +{ + int ret; + struct btrfs_ioctl_scrub_args *sa; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + ret = btrfs_scrub_dev(root, sa->devid, sa->start, sa->end, + &sa->progress, sa->flags & BTRFS_SCRUB_READONLY); + + if (copy_to_user(arg, sa, sizeof(*sa))) + ret = -EFAULT; + + kfree(sa); + return ret; +} + +static long btrfs_ioctl_scrub_cancel(struct btrfs_root *root, void __user *arg) +{ + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + return btrfs_scrub_cancel(root); +} + +static long btrfs_ioctl_scrub_progress(struct btrfs_root *root, + void __user *arg) +{ + struct btrfs_ioctl_scrub_args *sa; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + ret = btrfs_scrub_progress(root, sa->devid, &sa->progress); + + if (copy_to_user(arg, sa, sizeof(*sa))) + ret = -EFAULT; + + kfree(sa); + return ret; +} + long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { @@ -2379,6 +2848,8 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_setflags(file, argp); case FS_IOC_GETVERSION: return btrfs_ioctl_getversion(file, argp); + case FITRIM: + return btrfs_ioctl_fitrim(file, argp); case BTRFS_IOC_SNAP_CREATE: return btrfs_ioctl_snap_create(file, argp, 0); case BTRFS_IOC_SNAP_CREATE_V2: @@ -2403,6 +2874,10 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_add_dev(root, argp); case BTRFS_IOC_RM_DEV: return btrfs_ioctl_rm_dev(root, argp); + case BTRFS_IOC_FS_INFO: + return btrfs_ioctl_fs_info(root, argp); + case BTRFS_IOC_DEV_INFO: + return btrfs_ioctl_dev_info(root, argp); case BTRFS_IOC_BALANCE: return btrfs_balance(root->fs_info->dev_root); case BTRFS_IOC_CLONE: @@ -2426,6 +2901,12 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_start_sync(file, argp); case BTRFS_IOC_WAIT_SYNC: return btrfs_ioctl_wait_sync(file, argp); + case BTRFS_IOC_SCRUB: + return btrfs_ioctl_scrub(root, argp); + case BTRFS_IOC_SCRUB_CANCEL: + return btrfs_ioctl_scrub_cancel(root, argp); + case BTRFS_IOC_SCRUB_PROGRESS: + return btrfs_ioctl_scrub_progress(root, argp); } return -ENOTTY; diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h index 8fb382167b13..ad1ea789fcb4 100644 --- a/fs/btrfs/ioctl.h +++ b/fs/btrfs/ioctl.h @@ -32,6 +32,8 @@ struct btrfs_ioctl_vol_args { #define BTRFS_SUBVOL_CREATE_ASYNC (1ULL << 0) #define BTRFS_SUBVOL_RDONLY (1ULL << 1) +#define BTRFS_FSID_SIZE 16 +#define BTRFS_UUID_SIZE 16 #define BTRFS_SUBVOL_NAME_MAX 4039 struct btrfs_ioctl_vol_args_v2 { @@ -42,6 +44,71 @@ struct btrfs_ioctl_vol_args_v2 { char name[BTRFS_SUBVOL_NAME_MAX + 1]; }; +/* + * structure to report errors and progress to userspace, either as a + * result of a finished scrub, a canceled scrub or a progress inquiry + */ +struct btrfs_scrub_progress { + __u64 data_extents_scrubbed; /* # of data extents scrubbed */ + __u64 tree_extents_scrubbed; /* # of tree extents scrubbed */ + __u64 data_bytes_scrubbed; /* # of data bytes scrubbed */ + __u64 tree_bytes_scrubbed; /* # of tree bytes scrubbed */ + __u64 read_errors; /* # of read errors encountered (EIO) */ + __u64 csum_errors; /* # of failed csum checks */ + __u64 verify_errors; /* # of occurences, where the metadata + * of a tree block did not match the + * expected values, like generation or + * logical */ + __u64 no_csum; /* # of 4k data block for which no csum + * is present, probably the result of + * data written with nodatasum */ + __u64 csum_discards; /* # of csum for which no data was found + * in the extent tree. */ + __u64 super_errors; /* # of bad super blocks encountered */ + __u64 malloc_errors; /* # of internal kmalloc errors. These + * will likely cause an incomplete + * scrub */ + __u64 uncorrectable_errors; /* # of errors where either no intact + * copy was found or the writeback + * failed */ + __u64 corrected_errors; /* # of errors corrected */ + __u64 last_physical; /* last physical address scrubbed. In + * case a scrub was aborted, this can + * be used to restart the scrub */ + __u64 unverified_errors; /* # of occurences where a read for a + * full (64k) bio failed, but the re- + * check succeeded for each 4k piece. + * Intermittent error. */ +}; + +#define BTRFS_SCRUB_READONLY 1 +struct btrfs_ioctl_scrub_args { + __u64 devid; /* in */ + __u64 start; /* in */ + __u64 end; /* in */ + __u64 flags; /* in */ + struct btrfs_scrub_progress progress; /* out */ + /* pad to 1k */ + __u64 unused[(1024-32-sizeof(struct btrfs_scrub_progress))/8]; +}; + +#define BTRFS_DEVICE_PATH_NAME_MAX 1024 +struct btrfs_ioctl_dev_info_args { + __u64 devid; /* in/out */ + __u8 uuid[BTRFS_UUID_SIZE]; /* in/out */ + __u64 bytes_used; /* out */ + __u64 total_bytes; /* out */ + __u64 unused[379]; /* pad to 4k */ + __u8 path[BTRFS_DEVICE_PATH_NAME_MAX]; /* out */ +}; + +struct btrfs_ioctl_fs_info_args { + __u64 max_id; /* out */ + __u64 num_devices; /* out */ + __u8 fsid[BTRFS_FSID_SIZE]; /* out */ + __u64 reserved[124]; /* pad to 1k */ +}; + #define BTRFS_INO_LOOKUP_PATH_MAX 4080 struct btrfs_ioctl_ino_lookup_args { __u64 treeid; @@ -114,37 +181,6 @@ struct btrfs_ioctl_clone_range_args { #define BTRFS_DEFRAG_RANGE_COMPRESS 1 #define BTRFS_DEFRAG_RANGE_START_IO 2 -struct btrfs_ioctl_defrag_range_args { - /* start of the defrag operation */ - __u64 start; - - /* number of bytes to defrag, use (u64)-1 to say all */ - __u64 len; - - /* - * flags for the operation, which can include turning - * on compression for this one defrag - */ - __u64 flags; - - /* - * any extent bigger than this will be considered - * already defragged. Use 0 to take the kernel default - * Use 1 to say every single extent must be rewritten - */ - __u32 extent_thresh; - - /* - * which compression method to use if turning on compression - * for this defrag operation. If unspecified, zlib will - * be used - */ - __u32 compress_type; - - /* spare for later */ - __u32 unused[4]; -}; - struct btrfs_ioctl_space_info { __u64 flags; __u64 total_bytes; @@ -203,4 +239,13 @@ struct btrfs_ioctl_space_args { struct btrfs_ioctl_vol_args_v2) #define BTRFS_IOC_SUBVOL_GETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 25, __u64) #define BTRFS_IOC_SUBVOL_SETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 26, __u64) +#define BTRFS_IOC_SCRUB _IOWR(BTRFS_IOCTL_MAGIC, 27, \ + struct btrfs_ioctl_scrub_args) +#define BTRFS_IOC_SCRUB_CANCEL _IO(BTRFS_IOCTL_MAGIC, 28) +#define BTRFS_IOC_SCRUB_PROGRESS _IOWR(BTRFS_IOCTL_MAGIC, 29, \ + struct btrfs_ioctl_scrub_args) +#define BTRFS_IOC_DEV_INFO _IOWR(BTRFS_IOCTL_MAGIC, 30, \ + struct btrfs_ioctl_dev_info_args) +#define BTRFS_IOC_FS_INFO _IOR(BTRFS_IOCTL_MAGIC, 31, \ + struct btrfs_ioctl_fs_info_args) #endif diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c index 6151f2ea38bb..66fa43dc3f0f 100644 --- a/fs/btrfs/locking.c +++ b/fs/btrfs/locking.c @@ -185,31 +185,6 @@ sleep: return 0; } -/* - * Very quick trylock, this does not spin or schedule. It returns - * 1 with the spinlock held if it was able to take the lock, or it - * returns zero if it was unable to take the lock. - * - * After this call, scheduling is not safe without first calling - * btrfs_set_lock_blocking() - */ -int btrfs_try_tree_lock(struct extent_buffer *eb) -{ - if (spin_trylock(&eb->lock)) { - if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) { - /* - * we've got the spinlock, but the real owner is - * blocking. Drop the spinlock and return failure - */ - spin_unlock(&eb->lock); - return 0; - } - return 1; - } - /* someone else has the spinlock giveup */ - return 0; -} - int btrfs_tree_unlock(struct extent_buffer *eb) { /* diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h index 6c4ce457168c..5c33a560a2f1 100644 --- a/fs/btrfs/locking.h +++ b/fs/btrfs/locking.h @@ -21,8 +21,6 @@ int btrfs_tree_lock(struct extent_buffer *eb); int btrfs_tree_unlock(struct extent_buffer *eb); - -int btrfs_try_tree_lock(struct extent_buffer *eb); int btrfs_try_spin_lock(struct extent_buffer *eb); void btrfs_set_lock_blocking(struct extent_buffer *eb); diff --git a/fs/btrfs/lzo.c b/fs/btrfs/lzo.c index cc9b450399df..a178f5ebea78 100644 --- a/fs/btrfs/lzo.c +++ b/fs/btrfs/lzo.c @@ -280,6 +280,7 @@ static int lzo_decompress_biovec(struct list_head *ws, unsigned long tot_out; unsigned long tot_len; char *buf; + bool may_late_unmap, need_unmap; data_in = kmap(pages_in[0]); tot_len = read_compress_length(data_in); @@ -300,11 +301,13 @@ static int lzo_decompress_biovec(struct list_head *ws, tot_in += in_len; working_bytes = in_len; + may_late_unmap = need_unmap = false; /* fast path: avoid using the working buffer */ if (in_page_bytes_left >= in_len) { buf = data_in + in_offset; bytes = in_len; + may_late_unmap = true; goto cont; } @@ -329,14 +332,17 @@ cont: if (working_bytes == 0 && tot_in >= tot_len) break; - kunmap(pages_in[page_in_index]); - page_in_index++; - if (page_in_index >= total_pages_in) { + if (page_in_index + 1 >= total_pages_in) { ret = -1; - data_in = NULL; goto done; } - data_in = kmap(pages_in[page_in_index]); + + if (may_late_unmap) + need_unmap = true; + else + kunmap(pages_in[page_in_index]); + + data_in = kmap(pages_in[++page_in_index]); in_page_bytes_left = PAGE_CACHE_SIZE; in_offset = 0; @@ -346,6 +352,8 @@ cont: out_len = lzo1x_worst_compress(PAGE_CACHE_SIZE); ret = lzo1x_decompress_safe(buf, in_len, workspace->buf, &out_len); + if (need_unmap) + kunmap(pages_in[page_in_index - 1]); if (ret != LZO_E_OK) { printk(KERN_WARNING "btrfs decompress failed\n"); ret = -1; @@ -363,8 +371,7 @@ cont: break; } done: - if (data_in) - kunmap(pages_in[page_in_index]); + kunmap(pages_in[page_in_index]); return ret; } diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index 083a55477375..a1c940425307 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -202,6 +202,8 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset, INIT_LIST_HEAD(&entry->list); INIT_LIST_HEAD(&entry->root_extent_list); + trace_btrfs_ordered_extent_add(inode, entry); + spin_lock(&tree->lock); node = tree_insert(&tree->tree, file_offset, &entry->rb_node); @@ -387,6 +389,8 @@ int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry) struct list_head *cur; struct btrfs_ordered_sum *sum; + trace_btrfs_ordered_extent_put(entry->inode, entry); + if (atomic_dec_and_test(&entry->refs)) { while (!list_empty(&entry->list)) { cur = entry->list.next; @@ -420,6 +424,8 @@ static int __btrfs_remove_ordered_extent(struct inode *inode, spin_lock(&root->fs_info->ordered_extent_lock); list_del_init(&entry->root_extent_list); + trace_btrfs_ordered_extent_remove(inode, entry); + /* * we have no more ordered extents for this inode and * no dirty pages. We can safely remove it from the @@ -585,6 +591,8 @@ void btrfs_start_ordered_extent(struct inode *inode, u64 start = entry->file_offset; u64 end = start + entry->len - 1; + trace_btrfs_ordered_extent_start(inode, entry); + /* * pages in the range can be dirty, clean or writeback. We * start IO on any dirty ones so the wait doesn't stall waiting diff --git a/fs/btrfs/ref-cache.c b/fs/btrfs/ref-cache.c index a97314cf6bd6..82d569cb6267 100644 --- a/fs/btrfs/ref-cache.c +++ b/fs/btrfs/ref-cache.c @@ -23,56 +23,6 @@ #include "ref-cache.h" #include "transaction.h" -/* - * leaf refs are used to cache the information about which extents - * a given leaf has references on. This allows us to process that leaf - * in btrfs_drop_snapshot without needing to read it back from disk. - */ - -/* - * kmalloc a leaf reference struct and update the counters for the - * total ref cache size - */ -struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root, - int nr_extents) -{ - struct btrfs_leaf_ref *ref; - size_t size = btrfs_leaf_ref_size(nr_extents); - - ref = kmalloc(size, GFP_NOFS); - if (ref) { - spin_lock(&root->fs_info->ref_cache_lock); - root->fs_info->total_ref_cache_size += size; - spin_unlock(&root->fs_info->ref_cache_lock); - - memset(ref, 0, sizeof(*ref)); - atomic_set(&ref->usage, 1); - INIT_LIST_HEAD(&ref->list); - } - return ref; -} - -/* - * free a leaf reference struct and update the counters for the - * total ref cache size - */ -void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref) -{ - if (!ref) - return; - WARN_ON(atomic_read(&ref->usage) == 0); - if (atomic_dec_and_test(&ref->usage)) { - size_t size = btrfs_leaf_ref_size(ref->nritems); - - BUG_ON(ref->in_tree); - kfree(ref); - - spin_lock(&root->fs_info->ref_cache_lock); - root->fs_info->total_ref_cache_size -= size; - spin_unlock(&root->fs_info->ref_cache_lock); - } -} - static struct rb_node *tree_insert(struct rb_root *root, u64 bytenr, struct rb_node *node) { @@ -116,117 +66,3 @@ static struct rb_node *tree_search(struct rb_root *root, u64 bytenr) } return NULL; } - -int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen, - int shared) -{ - struct btrfs_leaf_ref *ref = NULL; - struct btrfs_leaf_ref_tree *tree = root->ref_tree; - - if (shared) - tree = &root->fs_info->shared_ref_tree; - if (!tree) - return 0; - - spin_lock(&tree->lock); - while (!list_empty(&tree->list)) { - ref = list_entry(tree->list.next, struct btrfs_leaf_ref, list); - BUG_ON(ref->tree != tree); - if (ref->root_gen > max_root_gen) - break; - if (!xchg(&ref->in_tree, 0)) { - cond_resched_lock(&tree->lock); - continue; - } - - rb_erase(&ref->rb_node, &tree->root); - list_del_init(&ref->list); - - spin_unlock(&tree->lock); - btrfs_free_leaf_ref(root, ref); - cond_resched(); - spin_lock(&tree->lock); - } - spin_unlock(&tree->lock); - return 0; -} - -/* - * find the leaf ref for a given extent. This returns the ref struct with - * a usage reference incremented - */ -struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root, - u64 bytenr) -{ - struct rb_node *rb; - struct btrfs_leaf_ref *ref = NULL; - struct btrfs_leaf_ref_tree *tree = root->ref_tree; -again: - if (tree) { - spin_lock(&tree->lock); - rb = tree_search(&tree->root, bytenr); - if (rb) - ref = rb_entry(rb, struct btrfs_leaf_ref, rb_node); - if (ref) - atomic_inc(&ref->usage); - spin_unlock(&tree->lock); - if (ref) - return ref; - } - if (tree != &root->fs_info->shared_ref_tree) { - tree = &root->fs_info->shared_ref_tree; - goto again; - } - return NULL; -} - -/* - * add a fully filled in leaf ref struct - * remove all the refs older than a given root generation - */ -int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref, - int shared) -{ - int ret = 0; - struct rb_node *rb; - struct btrfs_leaf_ref_tree *tree = root->ref_tree; - - if (shared) - tree = &root->fs_info->shared_ref_tree; - - spin_lock(&tree->lock); - rb = tree_insert(&tree->root, ref->bytenr, &ref->rb_node); - if (rb) { - ret = -EEXIST; - } else { - atomic_inc(&ref->usage); - ref->tree = tree; - ref->in_tree = 1; - list_add_tail(&ref->list, &tree->list); - } - spin_unlock(&tree->lock); - return ret; -} - -/* - * remove a single leaf ref from the tree. This drops the ref held by the tree - * only - */ -int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref) -{ - struct btrfs_leaf_ref_tree *tree; - - if (!xchg(&ref->in_tree, 0)) - return 0; - - tree = ref->tree; - spin_lock(&tree->lock); - - rb_erase(&ref->rb_node, &tree->root); - list_del_init(&ref->list); - - spin_unlock(&tree->lock); - - btrfs_free_leaf_ref(root, ref); - return 0; -} diff --git a/fs/btrfs/ref-cache.h b/fs/btrfs/ref-cache.h index e2a55cb2072b..24f7001f6387 100644 --- a/fs/btrfs/ref-cache.h +++ b/fs/btrfs/ref-cache.h @@ -49,28 +49,4 @@ static inline size_t btrfs_leaf_ref_size(int nr_extents) return sizeof(struct btrfs_leaf_ref) + sizeof(struct btrfs_extent_info) * nr_extents; } - -static inline void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree) -{ - tree->root = RB_ROOT; - INIT_LIST_HEAD(&tree->list); - spin_lock_init(&tree->lock); -} - -static inline int btrfs_leaf_ref_tree_empty(struct btrfs_leaf_ref_tree *tree) -{ - return RB_EMPTY_ROOT(&tree->root); -} - -void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree); -struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root, - int nr_extents); -void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref); -struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root, - u64 bytenr); -int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref, - int shared); -int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen, - int shared); -int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref); #endif diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 1f5556acb530..ca38eca70af0 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -30,6 +30,7 @@ #include "btrfs_inode.h" #include "async-thread.h" #include "free-space-cache.h" +#include "inode-map.h" /* * backref_node, mapping_node and tree_block start with this @@ -507,6 +508,7 @@ static int update_backref_cache(struct btrfs_trans_handle *trans, return 1; } + static int should_ignore_root(struct btrfs_root *root) { struct btrfs_root *reloc_root; @@ -529,7 +531,6 @@ static int should_ignore_root(struct btrfs_root *root) */ return 1; } - /* * find reloc tree by address of tree root */ @@ -709,7 +710,7 @@ again: WARN_ON(cur->checked); if (!list_empty(&cur->upper)) { /* - * the backref was added previously when processsing + * the backref was added previously when processing * backref of type BTRFS_TREE_BLOCK_REF_KEY */ BUG_ON(!list_is_singular(&cur->upper)); @@ -961,7 +962,7 @@ again: lower = upper; upper = NULL; } - btrfs_release_path(root, path2); + btrfs_release_path(path2); next: if (ptr < end) { ptr += btrfs_extent_inline_ref_size(key.type); @@ -974,7 +975,7 @@ next: if (ptr >= end) path1->slots[0]++; } - btrfs_release_path(rc->extent_root, path1); + btrfs_release_path(path1); cur->checked = 1; WARN_ON(exist); @@ -1157,6 +1158,7 @@ static int clone_backref_node(struct btrfs_trans_handle *trans, new_node->bytenr = dest->node->start; new_node->level = node->level; new_node->lowest = node->lowest; + new_node->checked = 1; new_node->root = dest; if (!node->lowest) { @@ -1408,9 +1410,9 @@ again: prev = node; entry = rb_entry(node, struct btrfs_inode, rb_node); - if (objectid < entry->vfs_inode.i_ino) + if (objectid < btrfs_ino(&entry->vfs_inode)) node = node->rb_left; - else if (objectid > entry->vfs_inode.i_ino) + else if (objectid > btrfs_ino(&entry->vfs_inode)) node = node->rb_right; else break; @@ -1418,7 +1420,7 @@ again: if (!node) { while (prev) { entry = rb_entry(prev, struct btrfs_inode, rb_node); - if (objectid <= entry->vfs_inode.i_ino) { + if (objectid <= btrfs_ino(&entry->vfs_inode)) { node = prev; break; } @@ -1433,7 +1435,7 @@ again: return inode; } - objectid = entry->vfs_inode.i_ino + 1; + objectid = btrfs_ino(&entry->vfs_inode) + 1; if (cond_resched_lock(&root->inode_lock)) goto again; @@ -1469,7 +1471,7 @@ static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr, return -ENOMEM; bytenr -= BTRFS_I(reloc_inode)->index_cnt; - ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino, + ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(reloc_inode), bytenr, 0); if (ret < 0) goto out; @@ -1557,11 +1559,11 @@ int replace_file_extents(struct btrfs_trans_handle *trans, if (first) { inode = find_next_inode(root, key.objectid); first = 0; - } else if (inode && inode->i_ino < key.objectid) { + } else if (inode && btrfs_ino(inode) < key.objectid) { btrfs_add_delayed_iput(inode); inode = find_next_inode(root, key.objectid); } - if (inode && inode->i_ino == key.objectid) { + if (inode && btrfs_ino(inode) == key.objectid) { end = key.offset + btrfs_file_extent_num_bytes(leaf, fi); WARN_ON(!IS_ALIGNED(key.offset, @@ -1723,6 +1725,7 @@ again: eb = read_tree_block(dest, old_bytenr, blocksize, old_ptr_gen); + BUG_ON(!eb); btrfs_tree_lock(eb); if (cow) { ret = btrfs_cow_block(trans, dest, eb, parent, @@ -1747,7 +1750,7 @@ again: btrfs_node_key_to_cpu(path->nodes[level], &key, path->slots[level]); - btrfs_release_path(src, path); + btrfs_release_path(path); path->lowest_level = level; ret = btrfs_search_slot(trans, src, &key, path, 0, 1); @@ -1891,6 +1894,7 @@ static int invalidate_extent_cache(struct btrfs_root *root, struct inode *inode = NULL; u64 objectid; u64 start, end; + u64 ino; objectid = min_key->objectid; while (1) { @@ -1903,17 +1907,18 @@ static int invalidate_extent_cache(struct btrfs_root *root, inode = find_next_inode(root, objectid); if (!inode) break; + ino = btrfs_ino(inode); - if (inode->i_ino > max_key->objectid) { + if (ino > max_key->objectid) { iput(inode); break; } - objectid = inode->i_ino + 1; + objectid = ino + 1; if (!S_ISREG(inode->i_mode)) continue; - if (unlikely(min_key->objectid == inode->i_ino)) { + if (unlikely(min_key->objectid == ino)) { if (min_key->type > BTRFS_EXTENT_DATA_KEY) continue; if (min_key->type < BTRFS_EXTENT_DATA_KEY) @@ -1926,7 +1931,7 @@ static int invalidate_extent_cache(struct btrfs_root *root, start = 0; } - if (unlikely(max_key->objectid == inode->i_ino)) { + if (unlikely(max_key->objectid == ino)) { if (max_key->type < BTRFS_EXTENT_DATA_KEY) continue; if (max_key->type > BTRFS_EXTENT_DATA_KEY) { @@ -2344,7 +2349,7 @@ struct btrfs_root *select_one_root(struct btrfs_trans_handle *trans, root = next->root; BUG_ON(!root); - /* no other choice for non-refernce counted tree */ + /* no other choice for non-references counted tree */ if (!root->ref_cows) return root; @@ -2494,7 +2499,7 @@ static int do_relocation(struct btrfs_trans_handle *trans, path->locks[upper->level] = 0; slot = path->slots[upper->level]; - btrfs_release_path(NULL, path); + btrfs_release_path(path); } else { ret = btrfs_bin_search(upper->eb, key, upper->level, &slot); @@ -2512,6 +2517,10 @@ static int do_relocation(struct btrfs_trans_handle *trans, blocksize = btrfs_level_size(root, node->level); generation = btrfs_node_ptr_generation(upper->eb, slot); eb = read_tree_block(root, bytenr, blocksize, generation); + if (!eb) { + err = -EIO; + goto next; + } btrfs_tree_lock(eb); btrfs_set_lock_blocking(eb); @@ -2669,6 +2678,7 @@ static int get_tree_block_key(struct reloc_control *rc, BUG_ON(block->key_ready); eb = read_tree_block(rc->extent_root, block->bytenr, block->key.objectid, block->key.offset); + BUG_ON(!eb); WARN_ON(btrfs_header_level(eb) != block->level); if (block->level == 0) btrfs_item_key_to_cpu(eb, &block->key, 0); @@ -2730,7 +2740,7 @@ static int relocate_tree_block(struct btrfs_trans_handle *trans, } else { path->lowest_level = node->level; ret = btrfs_search_slot(trans, root, key, path, 0, 1); - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret > 0) ret = 0; } @@ -2863,7 +2873,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end, struct extent_map *em; int ret = 0; - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) return -ENOMEM; @@ -3112,7 +3122,7 @@ static int add_tree_block(struct reloc_control *rc, #endif } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); BUG_ON(level == -1); @@ -3213,7 +3223,7 @@ static int delete_block_group_cache(struct btrfs_fs_info *fs_info, key.offset = 0; inode = btrfs_iget(fs_info->sb, &key, root, NULL); - if (!inode || IS_ERR(inode) || is_bad_inode(inode)) { + if (IS_ERR_OR_NULL(inode) || is_bad_inode(inode)) { if (inode && !IS_ERR(inode)) iput(inode); return -ENOENT; @@ -3498,7 +3508,7 @@ int add_data_references(struct reloc_control *rc, } path->slots[0]++; } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); if (err) free_block_list(blocks); return err; @@ -3561,7 +3571,7 @@ next: EXTENT_DIRTY); if (ret == 0 && start <= key.objectid) { - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); rc->search_start = end + 1; } else { rc->search_start = key.objectid + key.offset; @@ -3569,7 +3579,7 @@ next: return 0; } } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); return ret; } @@ -3653,6 +3663,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc) u32 item_size; int ret; int err = 0; + int progress = 0; path = btrfs_alloc_path(); if (!path) @@ -3665,9 +3676,10 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc) } while (1) { + progress++; trans = btrfs_start_transaction(rc->extent_root, 0); BUG_ON(IS_ERR(trans)); - +restart: if (update_backref_cache(trans, &rc->backref_cache)) { btrfs_end_transaction(trans, rc->extent_root); continue; @@ -3704,7 +3716,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc) flags = BTRFS_EXTENT_FLAG_DATA; if (path_change) { - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); path->search_commit_root = 1; path->skip_locking = 1; @@ -3727,7 +3739,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc) (flags & BTRFS_EXTENT_FLAG_DATA)) { ret = add_data_references(rc, &key, path, &blocks); } else { - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); ret = 0; } if (ret < 0) { @@ -3780,8 +3792,17 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc) } } } + if (trans && progress && err == -ENOSPC) { + ret = btrfs_force_chunk_alloc(trans, rc->extent_root, + rc->block_group->flags); + if (ret == 0) { + err = 0; + progress = 0; + goto restart; + } + } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY, GFP_NOFS); @@ -3849,7 +3870,7 @@ static int __insert_orphan_inode(struct btrfs_trans_handle *trans, btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); out: btrfs_free_path(path); return ret; @@ -3879,7 +3900,7 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info, if (IS_ERR(trans)) return ERR_CAST(trans); - err = btrfs_find_free_objectid(trans, root, objectid, &objectid); + err = btrfs_find_free_objectid(root, &objectid); if (err) goto out; @@ -3917,7 +3938,7 @@ static struct reloc_control *alloc_reloc_control(void) INIT_LIST_HEAD(&rc->reloc_roots); backref_cache_init(&rc->backref_cache); mapping_tree_init(&rc->reloc_root_tree); - extent_io_tree_init(&rc->processed_blocks, NULL, GFP_NOFS); + extent_io_tree_init(&rc->processed_blocks, NULL); return rc; } @@ -4091,7 +4112,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) } leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(root->fs_info->tree_root, path); + btrfs_release_path(path); if (key.objectid != BTRFS_TREE_RELOC_OBJECTID || key.type != BTRFS_ROOT_ITEM_KEY) @@ -4123,7 +4144,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) key.offset--; } - btrfs_release_path(root->fs_info->tree_root, path); + btrfs_release_path(path); if (list_empty(&reloc_roots)) goto out; @@ -4197,7 +4218,7 @@ out: if (IS_ERR(fs_root)) err = PTR_ERR(fs_root); else - btrfs_orphan_cleanup(fs_root); + err = btrfs_orphan_cleanup(fs_root); } return err; } @@ -4224,7 +4245,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len) disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt; ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr, - disk_bytenr + len - 1, &list); + disk_bytenr + len - 1, &list, 0); while (!list_empty(&list)) { sums = list_entry(list.next, struct btrfs_ordered_sum, list); diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index 6a1086e83ffc..ebe45443de06 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -22,53 +22,6 @@ #include "print-tree.h" /* - * search forward for a root, starting with objectid 'search_start' - * if a root key is found, the objectid we find is filled into 'found_objectid' - * and 0 is returned. < 0 is returned on error, 1 if there is nothing - * left in the tree. - */ -int btrfs_search_root(struct btrfs_root *root, u64 search_start, - u64 *found_objectid) -{ - struct btrfs_path *path; - struct btrfs_key search_key; - int ret; - - root = root->fs_info->tree_root; - search_key.objectid = search_start; - search_key.type = (u8)-1; - search_key.offset = (u64)-1; - - path = btrfs_alloc_path(); - BUG_ON(!path); -again: - ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0); - if (ret < 0) - goto out; - if (ret == 0) { - ret = 1; - goto out; - } - if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) { - ret = btrfs_next_leaf(root, path); - if (ret) - goto out; - } - btrfs_item_key_to_cpu(path->nodes[0], &search_key, path->slots[0]); - if (search_key.type != BTRFS_ROOT_ITEM_KEY) { - search_key.offset++; - btrfs_release_path(root, path); - goto again; - } - ret = 0; - *found_objectid = search_key.objectid; - -out: - btrfs_free_path(path); - return ret; -} - -/* * lookup the root with the highest offset for a given objectid. The key we do * find is copied into 'key'. If we find something return 0, otherwise 1, < 0 * on error. @@ -88,7 +41,8 @@ int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, search_key.offset = (u64)-1; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0); if (ret < 0) goto out; @@ -229,7 +183,7 @@ again: memcpy(&found_key, &key, sizeof(key)); key.offset++; - btrfs_release_path(root, path); + btrfs_release_path(path); dead_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root, &found_key); @@ -291,7 +245,7 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root) } btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(tree_root, path); + btrfs_release_path(path); if (key.objectid != BTRFS_ORPHAN_OBJECTID || key.type != BTRFS_ORPHAN_ITEM_KEY) @@ -332,7 +286,8 @@ int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *leaf; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; ret = btrfs_search_slot(trans, root, key, path, -1, 1); if (ret < 0) goto out; @@ -383,18 +338,22 @@ again: *sequence = btrfs_root_ref_sequence(leaf, ref); ret = btrfs_del_item(trans, tree_root, path); - BUG_ON(ret); + if (ret) { + err = ret; + goto out; + } } else err = -ENOENT; if (key.type == BTRFS_ROOT_BACKREF_KEY) { - btrfs_release_path(tree_root, path); + btrfs_release_path(path); key.objectid = ref_id; key.type = BTRFS_ROOT_REF_KEY; key.offset = root_id; goto again; } +out: btrfs_free_path(path); return err; } @@ -461,7 +420,7 @@ again: btrfs_mark_buffer_dirty(leaf); if (key.type == BTRFS_ROOT_BACKREF_KEY) { - btrfs_release_path(tree_root, path); + btrfs_release_path(path); key.objectid = ref_id; key.type = BTRFS_ROOT_REF_KEY; key.offset = root_id; @@ -471,3 +430,21 @@ again: btrfs_free_path(path); return 0; } + +/* + * Old btrfs forgets to init root_item->flags and root_item->byte_limit + * for subvolumes. To work around this problem, we steal a bit from + * root_item->inode_item->flags, and use it to indicate if those fields + * have been properly initialized. + */ +void btrfs_check_and_init_root_item(struct btrfs_root_item *root_item) +{ + u64 inode_flags = le64_to_cpu(root_item->inode.flags); + + if (!(inode_flags & BTRFS_INODE_ROOT_ITEM_INIT)) { + inode_flags |= BTRFS_INODE_ROOT_ITEM_INIT; + root_item->inode.flags = cpu_to_le64(inode_flags); + root_item->flags = 0; + root_item->byte_limit = 0; + } +} diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c new file mode 100644 index 000000000000..6dfed0c27ac3 --- /dev/null +++ b/fs/btrfs/scrub.c @@ -0,0 +1,1369 @@ +/* + * Copyright (C) 2011 STRATO. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/sched.h> +#include <linux/pagemap.h> +#include <linux/writeback.h> +#include <linux/blkdev.h> +#include <linux/rbtree.h> +#include <linux/slab.h> +#include <linux/workqueue.h> +#include "ctree.h" +#include "volumes.h" +#include "disk-io.h" +#include "ordered-data.h" + +/* + * This is only the first step towards a full-features scrub. It reads all + * extent and super block and verifies the checksums. In case a bad checksum + * is found or the extent cannot be read, good data will be written back if + * any can be found. + * + * Future enhancements: + * - To enhance the performance, better read-ahead strategies for the + * extent-tree can be employed. + * - In case an unrepairable extent is encountered, track which files are + * affected and report them + * - In case of a read error on files with nodatasum, map the file and read + * the extent to trigger a writeback of the good copy + * - track and record media errors, throw out bad devices + * - add a mode to also read unallocated space + * - make the prefetch cancellable + */ + +struct scrub_bio; +struct scrub_page; +struct scrub_dev; +static void scrub_bio_end_io(struct bio *bio, int err); +static void scrub_checksum(struct btrfs_work *work); +static int scrub_checksum_data(struct scrub_dev *sdev, + struct scrub_page *spag, void *buffer); +static int scrub_checksum_tree_block(struct scrub_dev *sdev, + struct scrub_page *spag, u64 logical, + void *buffer); +static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer); +static int scrub_fixup_check(struct scrub_bio *sbio, int ix); +static void scrub_fixup_end_io(struct bio *bio, int err); +static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector, + struct page *page); +static void scrub_fixup(struct scrub_bio *sbio, int ix); + +#define SCRUB_PAGES_PER_BIO 16 /* 64k per bio */ +#define SCRUB_BIOS_PER_DEV 16 /* 1 MB per device in flight */ + +struct scrub_page { + u64 flags; /* extent flags */ + u64 generation; + u64 mirror_num; + int have_csum; + u8 csum[BTRFS_CSUM_SIZE]; +}; + +struct scrub_bio { + int index; + struct scrub_dev *sdev; + struct bio *bio; + int err; + u64 logical; + u64 physical; + struct scrub_page spag[SCRUB_PAGES_PER_BIO]; + u64 count; + int next_free; + struct btrfs_work work; +}; + +struct scrub_dev { + struct scrub_bio *bios[SCRUB_BIOS_PER_DEV]; + struct btrfs_device *dev; + int first_free; + int curr; + atomic_t in_flight; + spinlock_t list_lock; + wait_queue_head_t list_wait; + u16 csum_size; + struct list_head csum_list; + atomic_t cancel_req; + int readonly; + /* + * statistics + */ + struct btrfs_scrub_progress stat; + spinlock_t stat_lock; +}; + +static void scrub_free_csums(struct scrub_dev *sdev) +{ + while (!list_empty(&sdev->csum_list)) { + struct btrfs_ordered_sum *sum; + sum = list_first_entry(&sdev->csum_list, + struct btrfs_ordered_sum, list); + list_del(&sum->list); + kfree(sum); + } +} + +static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev) +{ + int i; + int j; + struct page *last_page; + + if (!sdev) + return; + + for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) { + struct scrub_bio *sbio = sdev->bios[i]; + struct bio *bio; + + if (!sbio) + break; + + bio = sbio->bio; + if (bio) { + last_page = NULL; + for (j = 0; j < bio->bi_vcnt; ++j) { + if (bio->bi_io_vec[j].bv_page == last_page) + continue; + last_page = bio->bi_io_vec[j].bv_page; + __free_page(last_page); + } + bio_put(bio); + } + kfree(sbio); + } + + scrub_free_csums(sdev); + kfree(sdev); +} + +static noinline_for_stack +struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev) +{ + struct scrub_dev *sdev; + int i; + int j; + int ret; + struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; + + sdev = kzalloc(sizeof(*sdev), GFP_NOFS); + if (!sdev) + goto nomem; + sdev->dev = dev; + for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) { + struct bio *bio; + struct scrub_bio *sbio; + + sbio = kzalloc(sizeof(*sbio), GFP_NOFS); + if (!sbio) + goto nomem; + sdev->bios[i] = sbio; + + bio = bio_kmalloc(GFP_NOFS, SCRUB_PAGES_PER_BIO); + if (!bio) + goto nomem; + + sbio->index = i; + sbio->sdev = sdev; + sbio->bio = bio; + sbio->count = 0; + sbio->work.func = scrub_checksum; + bio->bi_private = sdev->bios[i]; + bio->bi_end_io = scrub_bio_end_io; + bio->bi_sector = 0; + bio->bi_bdev = dev->bdev; + bio->bi_size = 0; + + for (j = 0; j < SCRUB_PAGES_PER_BIO; ++j) { + struct page *page; + page = alloc_page(GFP_NOFS); + if (!page) + goto nomem; + + ret = bio_add_page(bio, page, PAGE_SIZE, 0); + if (!ret) + goto nomem; + } + WARN_ON(bio->bi_vcnt != SCRUB_PAGES_PER_BIO); + + if (i != SCRUB_BIOS_PER_DEV-1) + sdev->bios[i]->next_free = i + 1; + else + sdev->bios[i]->next_free = -1; + } + sdev->first_free = 0; + sdev->curr = -1; + atomic_set(&sdev->in_flight, 0); + atomic_set(&sdev->cancel_req, 0); + sdev->csum_size = btrfs_super_csum_size(&fs_info->super_copy); + INIT_LIST_HEAD(&sdev->csum_list); + + spin_lock_init(&sdev->list_lock); + spin_lock_init(&sdev->stat_lock); + init_waitqueue_head(&sdev->list_wait); + return sdev; + +nomem: + scrub_free_dev(sdev); + return ERR_PTR(-ENOMEM); +} + +/* + * scrub_recheck_error gets called when either verification of the page + * failed or the bio failed to read, e.g. with EIO. In the latter case, + * recheck_error gets called for every page in the bio, even though only + * one may be bad + */ +static void scrub_recheck_error(struct scrub_bio *sbio, int ix) +{ + if (sbio->err) { + if (scrub_fixup_io(READ, sbio->sdev->dev->bdev, + (sbio->physical + ix * PAGE_SIZE) >> 9, + sbio->bio->bi_io_vec[ix].bv_page) == 0) { + if (scrub_fixup_check(sbio, ix) == 0) + return; + } + } + + scrub_fixup(sbio, ix); +} + +static int scrub_fixup_check(struct scrub_bio *sbio, int ix) +{ + int ret = 1; + struct page *page; + void *buffer; + u64 flags = sbio->spag[ix].flags; + + page = sbio->bio->bi_io_vec[ix].bv_page; + buffer = kmap_atomic(page, KM_USER0); + if (flags & BTRFS_EXTENT_FLAG_DATA) { + ret = scrub_checksum_data(sbio->sdev, + sbio->spag + ix, buffer); + } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + ret = scrub_checksum_tree_block(sbio->sdev, + sbio->spag + ix, + sbio->logical + ix * PAGE_SIZE, + buffer); + } else { + WARN_ON(1); + } + kunmap_atomic(buffer, KM_USER0); + + return ret; +} + +static void scrub_fixup_end_io(struct bio *bio, int err) +{ + complete((struct completion *)bio->bi_private); +} + +static void scrub_fixup(struct scrub_bio *sbio, int ix) +{ + struct scrub_dev *sdev = sbio->sdev; + struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; + struct btrfs_multi_bio *multi = NULL; + u64 logical = sbio->logical + ix * PAGE_SIZE; + u64 length; + int i; + int ret; + DECLARE_COMPLETION_ONSTACK(complete); + + if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) && + (sbio->spag[ix].have_csum == 0)) { + /* + * nodatasum, don't try to fix anything + * FIXME: we can do better, open the inode and trigger a + * writeback + */ + goto uncorrectable; + } + + length = PAGE_SIZE; + ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length, + &multi, 0); + if (ret || !multi || length < PAGE_SIZE) { + printk(KERN_ERR + "scrub_fixup: btrfs_map_block failed us for %llu\n", + (unsigned long long)logical); + WARN_ON(1); + return; + } + + if (multi->num_stripes == 1) + /* there aren't any replicas */ + goto uncorrectable; + + /* + * first find a good copy + */ + for (i = 0; i < multi->num_stripes; ++i) { + if (i == sbio->spag[ix].mirror_num) + continue; + + if (scrub_fixup_io(READ, multi->stripes[i].dev->bdev, + multi->stripes[i].physical >> 9, + sbio->bio->bi_io_vec[ix].bv_page)) { + /* I/O-error, this is not a good copy */ + continue; + } + + if (scrub_fixup_check(sbio, ix) == 0) + break; + } + if (i == multi->num_stripes) + goto uncorrectable; + + if (!sdev->readonly) { + /* + * bi_io_vec[ix].bv_page now contains good data, write it back + */ + if (scrub_fixup_io(WRITE, sdev->dev->bdev, + (sbio->physical + ix * PAGE_SIZE) >> 9, + sbio->bio->bi_io_vec[ix].bv_page)) { + /* I/O-error, writeback failed, give up */ + goto uncorrectable; + } + } + + kfree(multi); + spin_lock(&sdev->stat_lock); + ++sdev->stat.corrected_errors; + spin_unlock(&sdev->stat_lock); + + if (printk_ratelimit()) + printk(KERN_ERR "btrfs: fixed up at %llu\n", + (unsigned long long)logical); + return; + +uncorrectable: + kfree(multi); + spin_lock(&sdev->stat_lock); + ++sdev->stat.uncorrectable_errors; + spin_unlock(&sdev->stat_lock); + + if (printk_ratelimit()) + printk(KERN_ERR "btrfs: unable to fixup at %llu\n", + (unsigned long long)logical); +} + +static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector, + struct page *page) +{ + struct bio *bio = NULL; + int ret; + DECLARE_COMPLETION_ONSTACK(complete); + + /* we are going to wait on this IO */ + rw |= REQ_SYNC; + + bio = bio_alloc(GFP_NOFS, 1); + bio->bi_bdev = bdev; + bio->bi_sector = sector; + bio_add_page(bio, page, PAGE_SIZE, 0); + bio->bi_end_io = scrub_fixup_end_io; + bio->bi_private = &complete; + submit_bio(rw, bio); + + wait_for_completion(&complete); + + ret = !test_bit(BIO_UPTODATE, &bio->bi_flags); + bio_put(bio); + return ret; +} + +static void scrub_bio_end_io(struct bio *bio, int err) +{ + struct scrub_bio *sbio = bio->bi_private; + struct scrub_dev *sdev = sbio->sdev; + struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + + sbio->err = err; + + btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work); +} + +static void scrub_checksum(struct btrfs_work *work) +{ + struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); + struct scrub_dev *sdev = sbio->sdev; + struct page *page; + void *buffer; + int i; + u64 flags; + u64 logical; + int ret; + + if (sbio->err) { + for (i = 0; i < sbio->count; ++i) + scrub_recheck_error(sbio, i); + + sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1); + sbio->bio->bi_flags |= 1 << BIO_UPTODATE; + sbio->bio->bi_phys_segments = 0; + sbio->bio->bi_idx = 0; + + for (i = 0; i < sbio->count; i++) { + struct bio_vec *bi; + bi = &sbio->bio->bi_io_vec[i]; + bi->bv_offset = 0; + bi->bv_len = PAGE_SIZE; + } + + spin_lock(&sdev->stat_lock); + ++sdev->stat.read_errors; + spin_unlock(&sdev->stat_lock); + goto out; + } + for (i = 0; i < sbio->count; ++i) { + page = sbio->bio->bi_io_vec[i].bv_page; + buffer = kmap_atomic(page, KM_USER0); + flags = sbio->spag[i].flags; + logical = sbio->logical + i * PAGE_SIZE; + ret = 0; + if (flags & BTRFS_EXTENT_FLAG_DATA) { + ret = scrub_checksum_data(sdev, sbio->spag + i, buffer); + } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + ret = scrub_checksum_tree_block(sdev, sbio->spag + i, + logical, buffer); + } else if (flags & BTRFS_EXTENT_FLAG_SUPER) { + BUG_ON(i); + (void)scrub_checksum_super(sbio, buffer); + } else { + WARN_ON(1); + } + kunmap_atomic(buffer, KM_USER0); + if (ret) + scrub_recheck_error(sbio, i); + } + +out: + spin_lock(&sdev->list_lock); + sbio->next_free = sdev->first_free; + sdev->first_free = sbio->index; + spin_unlock(&sdev->list_lock); + atomic_dec(&sdev->in_flight); + wake_up(&sdev->list_wait); +} + +static int scrub_checksum_data(struct scrub_dev *sdev, + struct scrub_page *spag, void *buffer) +{ + u8 csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + int fail = 0; + struct btrfs_root *root = sdev->dev->dev_root; + + if (!spag->have_csum) + return 0; + + crc = btrfs_csum_data(root, buffer, crc, PAGE_SIZE); + btrfs_csum_final(crc, csum); + if (memcmp(csum, spag->csum, sdev->csum_size)) + fail = 1; + + spin_lock(&sdev->stat_lock); + ++sdev->stat.data_extents_scrubbed; + sdev->stat.data_bytes_scrubbed += PAGE_SIZE; + if (fail) + ++sdev->stat.csum_errors; + spin_unlock(&sdev->stat_lock); + + return fail; +} + +static int scrub_checksum_tree_block(struct scrub_dev *sdev, + struct scrub_page *spag, u64 logical, + void *buffer) +{ + struct btrfs_header *h; + struct btrfs_root *root = sdev->dev->dev_root; + struct btrfs_fs_info *fs_info = root->fs_info; + u8 csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + int fail = 0; + int crc_fail = 0; + + /* + * we don't use the getter functions here, as we + * a) don't have an extent buffer and + * b) the page is already kmapped + */ + h = (struct btrfs_header *)buffer; + + if (logical != le64_to_cpu(h->bytenr)) + ++fail; + + if (spag->generation != le64_to_cpu(h->generation)) + ++fail; + + if (memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE)) + ++fail; + + if (memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid, + BTRFS_UUID_SIZE)) + ++fail; + + crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc, + PAGE_SIZE - BTRFS_CSUM_SIZE); + btrfs_csum_final(crc, csum); + if (memcmp(csum, h->csum, sdev->csum_size)) + ++crc_fail; + + spin_lock(&sdev->stat_lock); + ++sdev->stat.tree_extents_scrubbed; + sdev->stat.tree_bytes_scrubbed += PAGE_SIZE; + if (crc_fail) + ++sdev->stat.csum_errors; + if (fail) + ++sdev->stat.verify_errors; + spin_unlock(&sdev->stat_lock); + + return fail || crc_fail; +} + +static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer) +{ + struct btrfs_super_block *s; + u64 logical; + struct scrub_dev *sdev = sbio->sdev; + struct btrfs_root *root = sdev->dev->dev_root; + struct btrfs_fs_info *fs_info = root->fs_info; + u8 csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + int fail = 0; + + s = (struct btrfs_super_block *)buffer; + logical = sbio->logical; + + if (logical != le64_to_cpu(s->bytenr)) + ++fail; + + if (sbio->spag[0].generation != le64_to_cpu(s->generation)) + ++fail; + + if (memcmp(s->fsid, fs_info->fsid, BTRFS_UUID_SIZE)) + ++fail; + + crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc, + PAGE_SIZE - BTRFS_CSUM_SIZE); + btrfs_csum_final(crc, csum); + if (memcmp(csum, s->csum, sbio->sdev->csum_size)) + ++fail; + + if (fail) { + /* + * if we find an error in a super block, we just report it. + * They will get written with the next transaction commit + * anyway + */ + spin_lock(&sdev->stat_lock); + ++sdev->stat.super_errors; + spin_unlock(&sdev->stat_lock); + } + + return fail; +} + +static int scrub_submit(struct scrub_dev *sdev) +{ + struct scrub_bio *sbio; + + if (sdev->curr == -1) + return 0; + + sbio = sdev->bios[sdev->curr]; + + sbio->bio->bi_sector = sbio->physical >> 9; + sbio->bio->bi_size = sbio->count * PAGE_SIZE; + sbio->bio->bi_next = NULL; + sbio->bio->bi_flags |= 1 << BIO_UPTODATE; + sbio->bio->bi_comp_cpu = -1; + sbio->bio->bi_bdev = sdev->dev->bdev; + sbio->err = 0; + sdev->curr = -1; + atomic_inc(&sdev->in_flight); + + submit_bio(0, sbio->bio); + + return 0; +} + +static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len, + u64 physical, u64 flags, u64 gen, u64 mirror_num, + u8 *csum, int force) +{ + struct scrub_bio *sbio; + +again: + /* + * grab a fresh bio or wait for one to become available + */ + while (sdev->curr == -1) { + spin_lock(&sdev->list_lock); + sdev->curr = sdev->first_free; + if (sdev->curr != -1) { + sdev->first_free = sdev->bios[sdev->curr]->next_free; + sdev->bios[sdev->curr]->next_free = -1; + sdev->bios[sdev->curr]->count = 0; + spin_unlock(&sdev->list_lock); + } else { + spin_unlock(&sdev->list_lock); + wait_event(sdev->list_wait, sdev->first_free != -1); + } + } + sbio = sdev->bios[sdev->curr]; + if (sbio->count == 0) { + sbio->physical = physical; + sbio->logical = logical; + } else if (sbio->physical + sbio->count * PAGE_SIZE != physical || + sbio->logical + sbio->count * PAGE_SIZE != logical) { + scrub_submit(sdev); + goto again; + } + sbio->spag[sbio->count].flags = flags; + sbio->spag[sbio->count].generation = gen; + sbio->spag[sbio->count].have_csum = 0; + sbio->spag[sbio->count].mirror_num = mirror_num; + if (csum) { + sbio->spag[sbio->count].have_csum = 1; + memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size); + } + ++sbio->count; + if (sbio->count == SCRUB_PAGES_PER_BIO || force) + scrub_submit(sdev); + + return 0; +} + +static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len, + u8 *csum) +{ + struct btrfs_ordered_sum *sum = NULL; + int ret = 0; + unsigned long i; + unsigned long num_sectors; + u32 sectorsize = sdev->dev->dev_root->sectorsize; + + while (!list_empty(&sdev->csum_list)) { + sum = list_first_entry(&sdev->csum_list, + struct btrfs_ordered_sum, list); + if (sum->bytenr > logical) + return 0; + if (sum->bytenr + sum->len > logical) + break; + + ++sdev->stat.csum_discards; + list_del(&sum->list); + kfree(sum); + sum = NULL; + } + if (!sum) + return 0; + + num_sectors = sum->len / sectorsize; + for (i = 0; i < num_sectors; ++i) { + if (sum->sums[i].bytenr == logical) { + memcpy(csum, &sum->sums[i].sum, sdev->csum_size); + ret = 1; + break; + } + } + if (ret && i == num_sectors - 1) { + list_del(&sum->list); + kfree(sum); + } + return ret; +} + +/* scrub extent tries to collect up to 64 kB for each bio */ +static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len, + u64 physical, u64 flags, u64 gen, u64 mirror_num) +{ + int ret; + u8 csum[BTRFS_CSUM_SIZE]; + + while (len) { + u64 l = min_t(u64, len, PAGE_SIZE); + int have_csum = 0; + + if (flags & BTRFS_EXTENT_FLAG_DATA) { + /* push csums to sbio */ + have_csum = scrub_find_csum(sdev, logical, l, csum); + if (have_csum == 0) + ++sdev->stat.no_csum; + } + ret = scrub_page(sdev, logical, l, physical, flags, gen, + mirror_num, have_csum ? csum : NULL, 0); + if (ret) + return ret; + len -= l; + logical += l; + physical += l; + } + return 0; +} + +static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev, + struct map_lookup *map, int num, u64 base, u64 length) +{ + struct btrfs_path *path; + struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + struct btrfs_root *root = fs_info->extent_root; + struct btrfs_root *csum_root = fs_info->csum_root; + struct btrfs_extent_item *extent; + u64 flags; + int ret; + int slot; + int i; + u64 nstripes; + int start_stripe; + struct extent_buffer *l; + struct btrfs_key key; + u64 physical; + u64 logical; + u64 generation; + u64 mirror_num; + + u64 increment = map->stripe_len; + u64 offset; + + nstripes = length; + offset = 0; + do_div(nstripes, map->stripe_len); + if (map->type & BTRFS_BLOCK_GROUP_RAID0) { + offset = map->stripe_len * num; + increment = map->stripe_len * map->num_stripes; + mirror_num = 0; + } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { + int factor = map->num_stripes / map->sub_stripes; + offset = map->stripe_len * (num / map->sub_stripes); + increment = map->stripe_len * factor; + mirror_num = num % map->sub_stripes; + } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { + increment = map->stripe_len; + mirror_num = num % map->num_stripes; + } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { + increment = map->stripe_len; + mirror_num = num % map->num_stripes; + } else { + increment = map->stripe_len; + mirror_num = 0; + } + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + path->reada = 2; + path->search_commit_root = 1; + path->skip_locking = 1; + + /* + * find all extents for each stripe and just read them to get + * them into the page cache + * FIXME: we can do better. build a more intelligent prefetching + */ + logical = base + offset; + physical = map->stripes[num].physical; + ret = 0; + for (i = 0; i < nstripes; ++i) { + key.objectid = logical; + key.type = BTRFS_EXTENT_ITEM_KEY; + key.offset = (u64)0; + + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + + l = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(l, &key, slot); + if (key.objectid != logical) { + ret = btrfs_previous_item(root, path, 0, + BTRFS_EXTENT_ITEM_KEY); + if (ret < 0) + goto out; + } + + while (1) { + l = path->nodes[0]; + slot = path->slots[0]; + if (slot >= btrfs_header_nritems(l)) { + ret = btrfs_next_leaf(root, path); + if (ret == 0) + continue; + if (ret < 0) + goto out; + + break; + } + btrfs_item_key_to_cpu(l, &key, slot); + + if (key.objectid >= logical + map->stripe_len) + break; + + path->slots[0]++; + } + btrfs_release_path(path); + logical += increment; + physical += map->stripe_len; + cond_resched(); + } + + /* + * collect all data csums for the stripe to avoid seeking during + * the scrub. This might currently (crc32) end up to be about 1MB + */ + start_stripe = 0; +again: + logical = base + offset + start_stripe * increment; + for (i = start_stripe; i < nstripes; ++i) { + ret = btrfs_lookup_csums_range(csum_root, logical, + logical + map->stripe_len - 1, + &sdev->csum_list, 1); + if (ret) + goto out; + + logical += increment; + cond_resched(); + } + /* + * now find all extents for each stripe and scrub them + */ + logical = base + offset + start_stripe * increment; + physical = map->stripes[num].physical + start_stripe * map->stripe_len; + ret = 0; + for (i = start_stripe; i < nstripes; ++i) { + /* + * canceled? + */ + if (atomic_read(&fs_info->scrub_cancel_req) || + atomic_read(&sdev->cancel_req)) { + ret = -ECANCELED; + goto out; + } + /* + * check to see if we have to pause + */ + if (atomic_read(&fs_info->scrub_pause_req)) { + /* push queued extents */ + scrub_submit(sdev); + wait_event(sdev->list_wait, + atomic_read(&sdev->in_flight) == 0); + atomic_inc(&fs_info->scrubs_paused); + wake_up(&fs_info->scrub_pause_wait); + mutex_lock(&fs_info->scrub_lock); + while (atomic_read(&fs_info->scrub_pause_req)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrub_pause_req) == 0); + mutex_lock(&fs_info->scrub_lock); + } + atomic_dec(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + wake_up(&fs_info->scrub_pause_wait); + scrub_free_csums(sdev); + start_stripe = i; + goto again; + } + + key.objectid = logical; + key.type = BTRFS_EXTENT_ITEM_KEY; + key.offset = (u64)0; + + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + + l = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(l, &key, slot); + if (key.objectid != logical) { + ret = btrfs_previous_item(root, path, 0, + BTRFS_EXTENT_ITEM_KEY); + if (ret < 0) + goto out; + } + + while (1) { + l = path->nodes[0]; + slot = path->slots[0]; + if (slot >= btrfs_header_nritems(l)) { + ret = btrfs_next_leaf(root, path); + if (ret == 0) + continue; + if (ret < 0) + goto out; + + break; + } + btrfs_item_key_to_cpu(l, &key, slot); + + if (key.objectid + key.offset <= logical) + goto next; + + if (key.objectid >= logical + map->stripe_len) + break; + + if (btrfs_key_type(&key) != BTRFS_EXTENT_ITEM_KEY) + goto next; + + extent = btrfs_item_ptr(l, slot, + struct btrfs_extent_item); + flags = btrfs_extent_flags(l, extent); + generation = btrfs_extent_generation(l, extent); + + if (key.objectid < logical && + (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) { + printk(KERN_ERR + "btrfs scrub: tree block %llu spanning " + "stripes, ignored. logical=%llu\n", + (unsigned long long)key.objectid, + (unsigned long long)logical); + goto next; + } + + /* + * trim extent to this stripe + */ + if (key.objectid < logical) { + key.offset -= logical - key.objectid; + key.objectid = logical; + } + if (key.objectid + key.offset > + logical + map->stripe_len) { + key.offset = logical + map->stripe_len - + key.objectid; + } + + ret = scrub_extent(sdev, key.objectid, key.offset, + key.objectid - logical + physical, + flags, generation, mirror_num); + if (ret) + goto out; + +next: + path->slots[0]++; + } + btrfs_release_path(path); + logical += increment; + physical += map->stripe_len; + spin_lock(&sdev->stat_lock); + sdev->stat.last_physical = physical; + spin_unlock(&sdev->stat_lock); + } + /* push queued extents */ + scrub_submit(sdev); + +out: + btrfs_free_path(path); + return ret < 0 ? ret : 0; +} + +static noinline_for_stack int scrub_chunk(struct scrub_dev *sdev, + u64 chunk_tree, u64 chunk_objectid, u64 chunk_offset, u64 length) +{ + struct btrfs_mapping_tree *map_tree = + &sdev->dev->dev_root->fs_info->mapping_tree; + struct map_lookup *map; + struct extent_map *em; + int i; + int ret = -EINVAL; + + read_lock(&map_tree->map_tree.lock); + em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1); + read_unlock(&map_tree->map_tree.lock); + + if (!em) + return -EINVAL; + + map = (struct map_lookup *)em->bdev; + if (em->start != chunk_offset) + goto out; + + if (em->len < length) + goto out; + + for (i = 0; i < map->num_stripes; ++i) { + if (map->stripes[i].dev == sdev->dev) { + ret = scrub_stripe(sdev, map, i, chunk_offset, length); + if (ret) + goto out; + } + } +out: + free_extent_map(em); + + return ret; +} + +static noinline_for_stack +int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end) +{ + struct btrfs_dev_extent *dev_extent = NULL; + struct btrfs_path *path; + struct btrfs_root *root = sdev->dev->dev_root; + struct btrfs_fs_info *fs_info = root->fs_info; + u64 length; + u64 chunk_tree; + u64 chunk_objectid; + u64 chunk_offset; + int ret; + int slot; + struct extent_buffer *l; + struct btrfs_key key; + struct btrfs_key found_key; + struct btrfs_block_group_cache *cache; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + path->reada = 2; + path->search_commit_root = 1; + path->skip_locking = 1; + + key.objectid = sdev->dev->devid; + key.offset = 0ull; + key.type = BTRFS_DEV_EXTENT_KEY; + + + while (1) { + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + ret = 0; + + l = path->nodes[0]; + slot = path->slots[0]; + + btrfs_item_key_to_cpu(l, &found_key, slot); + + if (found_key.objectid != sdev->dev->devid) + break; + + if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY) + break; + + if (found_key.offset >= end) + break; + + if (found_key.offset < key.offset) + break; + + dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent); + length = btrfs_dev_extent_length(l, dev_extent); + + if (found_key.offset + length <= start) { + key.offset = found_key.offset + length; + btrfs_release_path(path); + continue; + } + + chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent); + chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent); + chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent); + + /* + * get a reference on the corresponding block group to prevent + * the chunk from going away while we scrub it + */ + cache = btrfs_lookup_block_group(fs_info, chunk_offset); + if (!cache) { + ret = -ENOENT; + goto out; + } + ret = scrub_chunk(sdev, chunk_tree, chunk_objectid, + chunk_offset, length); + btrfs_put_block_group(cache); + if (ret) + break; + + key.offset = found_key.offset + length; + btrfs_release_path(path); + } + +out: + btrfs_free_path(path); + return ret; +} + +static noinline_for_stack int scrub_supers(struct scrub_dev *sdev) +{ + int i; + u64 bytenr; + u64 gen; + int ret; + struct btrfs_device *device = sdev->dev; + struct btrfs_root *root = device->dev_root; + + gen = root->fs_info->last_trans_committed; + + for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { + bytenr = btrfs_sb_offset(i); + if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes) + break; + + ret = scrub_page(sdev, bytenr, PAGE_SIZE, bytenr, + BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1); + if (ret) + return ret; + } + wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0); + + return 0; +} + +/* + * get a reference count on fs_info->scrub_workers. start worker if necessary + */ +static noinline_for_stack int scrub_workers_get(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + if (fs_info->scrub_workers_refcnt == 0) + btrfs_start_workers(&fs_info->scrub_workers, 1); + ++fs_info->scrub_workers_refcnt; + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} + +static noinline_for_stack void scrub_workers_put(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + if (--fs_info->scrub_workers_refcnt == 0) + btrfs_stop_workers(&fs_info->scrub_workers); + WARN_ON(fs_info->scrub_workers_refcnt < 0); + mutex_unlock(&fs_info->scrub_lock); +} + + +int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, + struct btrfs_scrub_progress *progress, int readonly) +{ + struct scrub_dev *sdev; + struct btrfs_fs_info *fs_info = root->fs_info; + int ret; + struct btrfs_device *dev; + + if (root->fs_info->closing) + return -EINVAL; + + /* + * check some assumptions + */ + if (root->sectorsize != PAGE_SIZE || + root->sectorsize != root->leafsize || + root->sectorsize != root->nodesize) { + printk(KERN_ERR "btrfs_scrub: size assumptions fail\n"); + return -EINVAL; + } + + ret = scrub_workers_get(root); + if (ret) + return ret; + + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + dev = btrfs_find_device(root, devid, NULL, NULL); + if (!dev || dev->missing) { + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return -ENODEV; + } + mutex_lock(&fs_info->scrub_lock); + + if (!dev->in_fs_metadata) { + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return -ENODEV; + } + + if (dev->scrub_device) { + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return -EINPROGRESS; + } + sdev = scrub_setup_dev(dev); + if (IS_ERR(sdev)) { + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return PTR_ERR(sdev); + } + sdev->readonly = readonly; + dev->scrub_device = sdev; + + atomic_inc(&fs_info->scrubs_running); + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + + down_read(&fs_info->scrub_super_lock); + ret = scrub_supers(sdev); + up_read(&fs_info->scrub_super_lock); + + if (!ret) + ret = scrub_enumerate_chunks(sdev, start, end); + + wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0); + + atomic_dec(&fs_info->scrubs_running); + wake_up(&fs_info->scrub_pause_wait); + + if (progress) + memcpy(progress, &sdev->stat, sizeof(*progress)); + + mutex_lock(&fs_info->scrub_lock); + dev->scrub_device = NULL; + mutex_unlock(&fs_info->scrub_lock); + + scrub_free_dev(sdev); + scrub_workers_put(root); + + return ret; +} + +int btrfs_scrub_pause(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + atomic_inc(&fs_info->scrub_pause_req); + while (atomic_read(&fs_info->scrubs_paused) != + atomic_read(&fs_info->scrubs_running)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrubs_paused) == + atomic_read(&fs_info->scrubs_running)); + mutex_lock(&fs_info->scrub_lock); + } + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} + +int btrfs_scrub_continue(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + atomic_dec(&fs_info->scrub_pause_req); + wake_up(&fs_info->scrub_pause_wait); + return 0; +} + +int btrfs_scrub_pause_super(struct btrfs_root *root) +{ + down_write(&root->fs_info->scrub_super_lock); + return 0; +} + +int btrfs_scrub_continue_super(struct btrfs_root *root) +{ + up_write(&root->fs_info->scrub_super_lock); + return 0; +} + +int btrfs_scrub_cancel(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + if (!atomic_read(&fs_info->scrubs_running)) { + mutex_unlock(&fs_info->scrub_lock); + return -ENOTCONN; + } + + atomic_inc(&fs_info->scrub_cancel_req); + while (atomic_read(&fs_info->scrubs_running)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrubs_running) == 0); + mutex_lock(&fs_info->scrub_lock); + } + atomic_dec(&fs_info->scrub_cancel_req); + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} + +int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct scrub_dev *sdev; + + mutex_lock(&fs_info->scrub_lock); + sdev = dev->scrub_device; + if (!sdev) { + mutex_unlock(&fs_info->scrub_lock); + return -ENOTCONN; + } + atomic_inc(&sdev->cancel_req); + while (dev->scrub_device) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + dev->scrub_device == NULL); + mutex_lock(&fs_info->scrub_lock); + } + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} +int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_device *dev; + int ret; + + /* + * we have to hold the device_list_mutex here so the device + * does not go away in cancel_dev. FIXME: find a better solution + */ + mutex_lock(&fs_info->fs_devices->device_list_mutex); + dev = btrfs_find_device(root, devid, NULL, NULL); + if (!dev) { + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + return -ENODEV; + } + ret = btrfs_scrub_cancel_dev(root, dev); + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + + return ret; +} + +int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, + struct btrfs_scrub_progress *progress) +{ + struct btrfs_device *dev; + struct scrub_dev *sdev = NULL; + + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + dev = btrfs_find_device(root, devid, NULL, NULL); + if (dev) + sdev = dev->scrub_device; + if (sdev) + memcpy(progress, &sdev->stat, sizeof(*progress)); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + + return dev ? (sdev ? 0 : -ENOTCONN) : -ENODEV; +} diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index a004008f7d28..9b2e7e5bc3ef 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -39,7 +39,9 @@ #include <linux/miscdevice.h> #include <linux/magic.h> #include <linux/slab.h> +#include <linux/cleancache.h> #include "compat.h" +#include "delayed-inode.h" #include "ctree.h" #include "disk-io.h" #include "transaction.h" @@ -52,6 +54,9 @@ #include "export.h" #include "compression.h" +#define CREATE_TRACE_POINTS +#include <trace/events/btrfs.h> + static const struct super_operations btrfs_super_ops; static const char *btrfs_decode_error(struct btrfs_fs_info *fs_info, int errno, @@ -155,7 +160,8 @@ enum { Opt_nossd, Opt_ssd_spread, Opt_thread_pool, Opt_noacl, Opt_compress, Opt_compress_type, Opt_compress_force, Opt_compress_force_type, Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard, - Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed, Opt_err, + Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed, + Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, Opt_err, }; static match_table_t tokens = { @@ -184,6 +190,9 @@ static match_table_t tokens = { {Opt_space_cache, "space_cache"}, {Opt_clear_cache, "clear_cache"}, {Opt_user_subvol_rm_allowed, "user_subvol_rm_allowed"}, + {Opt_enospc_debug, "enospc_debug"}, + {Opt_subvolrootid, "subvolrootid=%d"}, + {Opt_defrag, "autodefrag"}, {Opt_err, NULL}, }; @@ -227,6 +236,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) break; case Opt_subvol: case Opt_subvolid: + case Opt_subvolrootid: case Opt_device: /* * These are parsed by btrfs_parse_early_options @@ -358,6 +368,13 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) case Opt_user_subvol_rm_allowed: btrfs_set_opt(info->mount_opt, USER_SUBVOL_RM_ALLOWED); break; + case Opt_enospc_debug: + btrfs_set_opt(info->mount_opt, ENOSPC_DEBUG); + break; + case Opt_defrag: + printk(KERN_INFO "btrfs: enabling auto defrag"); + btrfs_set_opt(info->mount_opt, AUTO_DEFRAG); + break; case Opt_err: printk(KERN_INFO "btrfs: unrecognized mount option " "'%s'\n", p); @@ -380,7 +397,7 @@ out: */ static int btrfs_parse_early_options(const char *options, fmode_t flags, void *holder, char **subvol_name, u64 *subvol_objectid, - struct btrfs_fs_devices **fs_devices) + u64 *subvol_rootid, struct btrfs_fs_devices **fs_devices) { substring_t args[MAX_OPT_ARGS]; char *opts, *orig, *p; @@ -421,6 +438,18 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags, *subvol_objectid = intarg; } break; + case Opt_subvolrootid: + intarg = 0; + error = match_int(&args[0], &intarg); + if (!error) { + /* we want the original fs_tree */ + if (!intarg) + *subvol_rootid = + BTRFS_FS_TREE_OBJECTID; + else + *subvol_rootid = intarg; + } + break; case Opt_device: error = btrfs_scan_one_device(match_strdup(&args[0]), flags, holder, fs_devices); @@ -484,8 +513,10 @@ static struct dentry *get_default_root(struct super_block *sb, */ dir_id = btrfs_super_root_dir(&root->fs_info->super_copy); di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0); - if (IS_ERR(di)) + if (IS_ERR(di)) { + btrfs_free_path(path); return ERR_CAST(di); + } if (!di) { /* * Ok the default dir item isn't there. This is weird since @@ -602,6 +633,7 @@ static int btrfs_fill_super(struct super_block *sb, sb->s_root = root_dentry; save_mount_options(sb, data); + cleancache_init_fs(sb); return 0; fail_close: @@ -615,6 +647,8 @@ int btrfs_sync_fs(struct super_block *sb, int wait) struct btrfs_root *root = btrfs_sb(sb); int ret; + trace_btrfs_sync_fs(wait); + if (!wait) { filemap_flush(root->fs_info->btree_inode->i_mapping); return 0; @@ -634,6 +668,7 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) { struct btrfs_root *root = btrfs_sb(vfs->mnt_sb); struct btrfs_fs_info *info = root->fs_info; + char *compress_type; if (btrfs_test_opt(root, DEGRADED)) seq_puts(seq, ",degraded"); @@ -652,8 +687,16 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) if (info->thread_pool_size != min_t(unsigned long, num_online_cpus() + 2, 8)) seq_printf(seq, ",thread_pool=%d", info->thread_pool_size); - if (btrfs_test_opt(root, COMPRESS)) - seq_puts(seq, ",compress"); + if (btrfs_test_opt(root, COMPRESS)) { + if (info->compress_type == BTRFS_COMPRESS_ZLIB) + compress_type = "zlib"; + else + compress_type = "lzo"; + if (btrfs_test_opt(root, FORCE_COMPRESS)) + seq_printf(seq, ",compress-force=%s", compress_type); + else + seq_printf(seq, ",compress=%s", compress_type); + } if (btrfs_test_opt(root, NOSSD)) seq_puts(seq, ",nossd"); if (btrfs_test_opt(root, SSD_SPREAD)) @@ -668,6 +711,12 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) seq_puts(seq, ",discard"); if (!(root->fs_info->sb->s_flags & MS_POSIXACL)) seq_puts(seq, ",noacl"); + if (btrfs_test_opt(root, SPACE_CACHE)) + seq_puts(seq, ",space_cache"); + if (btrfs_test_opt(root, CLEAR_CACHE)) + seq_puts(seq, ",clear_cache"); + if (btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED)) + seq_puts(seq, ",user_subvol_rm_allowed"); return 0; } @@ -700,7 +749,7 @@ static int btrfs_set_super(struct super_block *s, void *data) * for multiple device setup. Make sure to keep it in sync. */ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, - const char *dev_name, void *data) + const char *device_name, void *data) { struct block_device *bdev = NULL; struct super_block *s; @@ -711,6 +760,7 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, fmode_t mode = FMODE_READ; char *subvol_name = NULL; u64 subvol_objectid = 0; + u64 subvol_rootid = 0; int error = 0; if (!(flags & MS_RDONLY)) @@ -718,11 +768,11 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, error = btrfs_parse_early_options(data, mode, fs_type, &subvol_name, &subvol_objectid, - &fs_devices); + &subvol_rootid, &fs_devices); if (error) return ERR_PTR(error); - error = btrfs_scan_one_device(dev_name, mode, fs_type, &fs_devices); + error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices); if (error) goto error_free_subvol_name; @@ -782,15 +832,17 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, s->s_flags |= MS_ACTIVE; } - root = get_default_root(s, subvol_objectid); - if (IS_ERR(root)) { - error = PTR_ERR(root); - deactivate_locked_super(s); - goto error_free_subvol_name; - } /* if they gave us a subvolume name bind mount into that */ if (strcmp(subvol_name, ".")) { struct dentry *new_root; + + root = get_default_root(s, subvol_rootid); + if (IS_ERR(root)) { + error = PTR_ERR(root); + deactivate_locked_super(s); + goto error_free_subvol_name; + } + mutex_lock(&root->d_inode->i_mutex); new_root = lookup_one_len(subvol_name, root, strlen(subvol_name)); @@ -811,6 +863,13 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, } dput(root); root = new_root; + } else { + root = get_default_root(s, subvol_objectid); + if (IS_ERR(root)) { + error = PTR_ERR(root); + deactivate_locked_super(s); + goto error_free_subvol_name; + } } kfree(subvol_name); @@ -864,6 +923,32 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) return 0; } +/* Used to sort the devices by max_avail(descending sort) */ +static int btrfs_cmp_device_free_bytes(const void *dev_info1, + const void *dev_info2) +{ + if (((struct btrfs_device_info *)dev_info1)->max_avail > + ((struct btrfs_device_info *)dev_info2)->max_avail) + return -1; + else if (((struct btrfs_device_info *)dev_info1)->max_avail < + ((struct btrfs_device_info *)dev_info2)->max_avail) + return 1; + else + return 0; +} + +/* + * sort the devices by max_avail, in which max free extent size of each device + * is stored.(Descending Sort) + */ +static inline void btrfs_descending_sort_devices( + struct btrfs_device_info *devices, + size_t nr_devices) +{ + sort(devices, nr_devices, sizeof(struct btrfs_device_info), + btrfs_cmp_device_free_bytes, NULL); +} + /* * The helper to calc the free space on the devices that can be used to store * file data. @@ -1157,10 +1242,14 @@ static int __init init_btrfs_fs(void) if (err) goto free_extent_io; - err = btrfs_interface_init(); + err = btrfs_delayed_inode_init(); if (err) goto free_extent_map; + err = btrfs_interface_init(); + if (err) + goto free_delayed_inode; + err = register_filesystem(&btrfs_fs_type); if (err) goto unregister_ioctl; @@ -1170,6 +1259,8 @@ static int __init init_btrfs_fs(void) unregister_ioctl: btrfs_interface_exit(); +free_delayed_inode: + btrfs_delayed_inode_exit(); free_extent_map: extent_map_exit(); free_extent_io: @@ -1186,6 +1277,7 @@ free_sysfs: static void __exit exit_btrfs_fs(void) { btrfs_destroy_cachep(); + btrfs_delayed_inode_exit(); extent_map_exit(); extent_io_exit(); btrfs_interface_exit(); diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c index 4ce16ef702a3..c3c223ae6691 100644 --- a/fs/btrfs/sysfs.c +++ b/fs/btrfs/sysfs.c @@ -174,86 +174,9 @@ static const struct sysfs_ops btrfs_root_attr_ops = { .store = btrfs_root_attr_store, }; -static struct kobj_type btrfs_root_ktype = { - .default_attrs = btrfs_root_attrs, - .sysfs_ops = &btrfs_root_attr_ops, - .release = btrfs_root_release, -}; - -static struct kobj_type btrfs_super_ktype = { - .default_attrs = btrfs_super_attrs, - .sysfs_ops = &btrfs_super_attr_ops, - .release = btrfs_super_release, -}; - /* /sys/fs/btrfs/ entry */ static struct kset *btrfs_kset; -int btrfs_sysfs_add_super(struct btrfs_fs_info *fs) -{ - int error; - char *name; - char c; - int len = strlen(fs->sb->s_id) + 1; - int i; - - name = kmalloc(len, GFP_NOFS); - if (!name) { - error = -ENOMEM; - goto fail; - } - - for (i = 0; i < len; i++) { - c = fs->sb->s_id[i]; - if (c == '/' || c == '\\') - c = '!'; - name[i] = c; - } - name[len] = '\0'; - - fs->super_kobj.kset = btrfs_kset; - error = kobject_init_and_add(&fs->super_kobj, &btrfs_super_ktype, - NULL, "%s", name); - kfree(name); - if (error) - goto fail; - - return 0; - -fail: - printk(KERN_ERR "btrfs: sysfs creation for super failed\n"); - return error; -} - -int btrfs_sysfs_add_root(struct btrfs_root *root) -{ - int error; - - error = kobject_init_and_add(&root->root_kobj, &btrfs_root_ktype, - &root->fs_info->super_kobj, - "%s", root->name); - if (error) - goto fail; - - return 0; - -fail: - printk(KERN_ERR "btrfs: sysfs creation for root failed\n"); - return error; -} - -void btrfs_sysfs_del_root(struct btrfs_root *root) -{ - kobject_put(&root->root_kobj); - wait_for_completion(&root->kobj_unregister); -} - -void btrfs_sysfs_del_super(struct btrfs_fs_info *fs) -{ - kobject_put(&fs->super_kobj); - wait_for_completion(&fs->kobj_unregister); -} - int btrfs_init_sysfs(void) { btrfs_kset = kset_create_and_add("btrfs", NULL, fs_kobj); diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 3d73c8d93bbb..dc80f7156923 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -27,15 +27,14 @@ #include "transaction.h" #include "locking.h" #include "tree-log.h" +#include "inode-map.h" #define BTRFS_ROOT_TRANS_TAG 0 static noinline void put_transaction(struct btrfs_transaction *transaction) { - WARN_ON(transaction->use_count == 0); - transaction->use_count--; - if (transaction->use_count == 0) { - list_del_init(&transaction->list); + WARN_ON(atomic_read(&transaction->use_count) == 0); + if (atomic_dec_and_test(&transaction->use_count)) { memset(transaction, 0, sizeof(*transaction)); kmem_cache_free(btrfs_transaction_cachep, transaction); } @@ -57,16 +56,17 @@ static noinline int join_transaction(struct btrfs_root *root) if (!cur_trans) { cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS); - BUG_ON(!cur_trans); + if (!cur_trans) + return -ENOMEM; root->fs_info->generation++; - cur_trans->num_writers = 1; + atomic_set(&cur_trans->num_writers, 1); cur_trans->num_joined = 0; cur_trans->transid = root->fs_info->generation; init_waitqueue_head(&cur_trans->writer_wait); init_waitqueue_head(&cur_trans->commit_wait); cur_trans->in_commit = 0; cur_trans->blocked = 0; - cur_trans->use_count = 1; + atomic_set(&cur_trans->use_count, 1); cur_trans->commit_done = 0; cur_trans->start_time = get_seconds(); @@ -81,13 +81,12 @@ static noinline int join_transaction(struct btrfs_root *root) INIT_LIST_HEAD(&cur_trans->pending_snapshots); list_add_tail(&cur_trans->list, &root->fs_info->trans_list); extent_io_tree_init(&cur_trans->dirty_pages, - root->fs_info->btree_inode->i_mapping, - GFP_NOFS); + root->fs_info->btree_inode->i_mapping); spin_lock(&root->fs_info->new_trans_lock); root->fs_info->running_transaction = cur_trans; spin_unlock(&root->fs_info->new_trans_lock); } else { - cur_trans->num_writers++; + atomic_inc(&cur_trans->num_writers); cur_trans->num_joined++; } @@ -144,7 +143,7 @@ static void wait_current_trans(struct btrfs_root *root) cur_trans = root->fs_info->running_transaction; if (cur_trans && cur_trans->blocked) { DEFINE_WAIT(wait); - cur_trans->use_count++; + atomic_inc(&cur_trans->use_count); while (1) { prepare_to_wait(&root->fs_info->transaction_wait, &wait, TASK_UNINTERRUPTIBLE); @@ -180,6 +179,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root, { struct btrfs_trans_handle *h; struct btrfs_transaction *cur_trans; + int retries = 0; int ret; if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) @@ -195,10 +195,15 @@ again: wait_current_trans(root); ret = join_transaction(root); - BUG_ON(ret); + if (ret < 0) { + kmem_cache_free(btrfs_trans_handle_cachep, h); + if (type != TRANS_JOIN_NOLOCK) + mutex_unlock(&root->fs_info->trans_mutex); + return ERR_PTR(ret); + } cur_trans = root->fs_info->running_transaction; - cur_trans->use_count++; + atomic_inc(&cur_trans->use_count); if (type != TRANS_JOIN_NOLOCK) mutex_unlock(&root->fs_info->trans_mutex); @@ -218,10 +223,18 @@ again: if (num_items > 0) { ret = btrfs_trans_reserve_metadata(h, root, num_items); - if (ret == -EAGAIN) { + if (ret == -EAGAIN && !retries) { + retries++; btrfs_commit_transaction(h, root); goto again; + } else if (ret == -EAGAIN) { + /* + * We have already retried and got EAGAIN, so really we + * don't have space, so set ret to -ENOSPC. + */ + ret = -ENOSPC; } + if (ret < 0) { btrfs_end_transaction(h, root); return ERR_PTR(ret); @@ -321,7 +334,7 @@ int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid) goto out_unlock; /* nothing committing|committed */ } - cur_trans->use_count++; + atomic_inc(&cur_trans->use_count); mutex_unlock(&root->fs_info->trans_mutex); wait_for_commit(root, cur_trans); @@ -334,49 +347,6 @@ out_unlock: return ret; } -#if 0 -/* - * rate limit against the drop_snapshot code. This helps to slow down new - * operations if the drop_snapshot code isn't able to keep up. - */ -static void throttle_on_drops(struct btrfs_root *root) -{ - struct btrfs_fs_info *info = root->fs_info; - int harder_count = 0; - -harder: - if (atomic_read(&info->throttles)) { - DEFINE_WAIT(wait); - int thr; - thr = atomic_read(&info->throttle_gen); - - do { - prepare_to_wait(&info->transaction_throttle, - &wait, TASK_UNINTERRUPTIBLE); - if (!atomic_read(&info->throttles)) { - finish_wait(&info->transaction_throttle, &wait); - break; - } - schedule(); - finish_wait(&info->transaction_throttle, &wait); - } while (thr == atomic_read(&info->throttle_gen)); - harder_count++; - - if (root->fs_info->total_ref_cache_size > 1 * 1024 * 1024 && - harder_count < 2) - goto harder; - - if (root->fs_info->total_ref_cache_size > 5 * 1024 * 1024 && - harder_count < 10) - goto harder; - - if (root->fs_info->total_ref_cache_size > 10 * 1024 * 1024 && - harder_count < 20) - goto harder; - } -} -#endif - void btrfs_throttle(struct btrfs_root *root) { mutex_lock(&root->fs_info->trans_mutex); @@ -451,18 +421,14 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, wake_up_process(info->transaction_kthread); } - if (lock) - mutex_lock(&info->trans_mutex); WARN_ON(cur_trans != info->running_transaction); - WARN_ON(cur_trans->num_writers < 1); - cur_trans->num_writers--; + WARN_ON(atomic_read(&cur_trans->num_writers) < 1); + atomic_dec(&cur_trans->num_writers); smp_mb(); if (waitqueue_active(&cur_trans->writer_wait)) wake_up(&cur_trans->writer_wait); put_transaction(cur_trans); - if (lock) - mutex_unlock(&info->trans_mutex); if (current->journal_info == trans) current->journal_info = NULL; @@ -478,19 +444,40 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, int btrfs_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root) { - return __btrfs_end_transaction(trans, root, 0, 1); + int ret; + + ret = __btrfs_end_transaction(trans, root, 0, 1); + if (ret) + return ret; + return 0; } int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans, struct btrfs_root *root) { - return __btrfs_end_transaction(trans, root, 1, 1); + int ret; + + ret = __btrfs_end_transaction(trans, root, 1, 1); + if (ret) + return ret; + return 0; } int btrfs_end_transaction_nolock(struct btrfs_trans_handle *trans, struct btrfs_root *root) { - return __btrfs_end_transaction(trans, root, 0, 0); + int ret; + + ret = __btrfs_end_transaction(trans, root, 0, 0); + if (ret) + return ret; + return 0; +} + +int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans, + struct btrfs_root *root) +{ + return __btrfs_end_transaction(trans, root, 1, 1); } /* @@ -751,8 +738,14 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, btrfs_update_reloc_root(trans, root); btrfs_orphan_commit_root(trans, root); + btrfs_save_ino_cache(root, trans); + if (root->commit_root != root->node) { + mutex_lock(&root->fs_commit_mutex); switch_commit_root(root); + btrfs_unpin_free_ino(root); + mutex_unlock(&root->fs_commit_mutex); + btrfs_set_root_node(&root->root_item, root->node); } @@ -800,97 +793,6 @@ int btrfs_defrag_root(struct btrfs_root *root, int cacheonly) return ret; } -#if 0 -/* - * when dropping snapshots, we generate a ton of delayed refs, and it makes - * sense not to join the transaction while it is trying to flush the current - * queue of delayed refs out. - * - * This is used by the drop snapshot code only - */ -static noinline int wait_transaction_pre_flush(struct btrfs_fs_info *info) -{ - DEFINE_WAIT(wait); - - mutex_lock(&info->trans_mutex); - while (info->running_transaction && - info->running_transaction->delayed_refs.flushing) { - prepare_to_wait(&info->transaction_wait, &wait, - TASK_UNINTERRUPTIBLE); - mutex_unlock(&info->trans_mutex); - - schedule(); - - mutex_lock(&info->trans_mutex); - finish_wait(&info->transaction_wait, &wait); - } - mutex_unlock(&info->trans_mutex); - return 0; -} - -/* - * Given a list of roots that need to be deleted, call btrfs_drop_snapshot on - * all of them - */ -int btrfs_drop_dead_root(struct btrfs_root *root) -{ - struct btrfs_trans_handle *trans; - struct btrfs_root *tree_root = root->fs_info->tree_root; - unsigned long nr; - int ret; - - while (1) { - /* - * we don't want to jump in and create a bunch of - * delayed refs if the transaction is starting to close - */ - wait_transaction_pre_flush(tree_root->fs_info); - trans = btrfs_start_transaction(tree_root, 1); - - /* - * we've joined a transaction, make sure it isn't - * closing right now - */ - if (trans->transaction->delayed_refs.flushing) { - btrfs_end_transaction(trans, tree_root); - continue; - } - - ret = btrfs_drop_snapshot(trans, root); - if (ret != -EAGAIN) - break; - - ret = btrfs_update_root(trans, tree_root, - &root->root_key, - &root->root_item); - if (ret) - break; - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, tree_root); - BUG_ON(ret); - - btrfs_btree_balance_dirty(tree_root, nr); - cond_resched(); - } - BUG_ON(ret); - - ret = btrfs_del_root(trans, tree_root, &root->root_key); - BUG_ON(ret); - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, tree_root); - BUG_ON(ret); - - free_extent_buffer(root->node); - free_extent_buffer(root->commit_root); - kfree(root); - - btrfs_btree_balance_dirty(tree_root, nr); - return ret; -} -#endif - /* * new snapshots need to be created at a very specific time in the * transaction commit. This does the actual creation @@ -921,7 +823,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, goto fail; } - ret = btrfs_find_free_objectid(trans, tree_root, 0, &objectid); + ret = btrfs_find_free_objectid(tree_root, &objectid); if (ret) { pending->error = ret; goto fail; @@ -958,7 +860,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, BUG_ON(ret); ret = btrfs_insert_dir_item(trans, parent_root, dentry->d_name.name, dentry->d_name.len, - parent_inode->i_ino, &key, + parent_inode, &key, BTRFS_FT_DIR, index); BUG_ON(ret); @@ -970,6 +872,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, record_root_in_trans(trans, root); btrfs_set_root_last_snapshot(&root->root_item, trans->transid); memcpy(new_root_item, &root->root_item, sizeof(*new_root_item)); + btrfs_check_and_init_root_item(new_root_item); root_flags = btrfs_root_flags(new_root_item); if (pending->readonly) @@ -999,7 +902,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, */ ret = btrfs_add_root_ref(trans, tree_root, objectid, parent_root->root_key.objectid, - parent_inode->i_ino, index, + btrfs_ino(parent_inode), index, dentry->d_name.name, dentry->d_name.len); BUG_ON(ret); dput(parent); @@ -1027,6 +930,14 @@ static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans, int ret; list_for_each_entry(pending, head, list) { + /* + * We must deal with the delayed items before creating + * snapshots, or we will create a snapthot with inconsistent + * information. + */ + ret = btrfs_run_delayed_items(trans, fs_info->fs_root); + BUG_ON(ret); + ret = create_pending_snapshot(trans, fs_info, pending); BUG_ON(ret); } @@ -1156,7 +1067,8 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, struct btrfs_transaction *cur_trans; ac = kmalloc(sizeof(*ac), GFP_NOFS); - BUG_ON(!ac); + if (!ac) + return -ENOMEM; INIT_DELAYED_WORK(&ac->work, do_async_commit); ac->root = root; @@ -1170,7 +1082,7 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, /* take transaction reference */ mutex_lock(&root->fs_info->trans_mutex); cur_trans = trans->transaction; - cur_trans->use_count++; + atomic_inc(&cur_trans->use_count); mutex_unlock(&root->fs_info->trans_mutex); btrfs_end_transaction(trans, root); @@ -1229,7 +1141,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, mutex_lock(&root->fs_info->trans_mutex); if (cur_trans->in_commit) { - cur_trans->use_count++; + atomic_inc(&cur_trans->use_count); mutex_unlock(&root->fs_info->trans_mutex); btrfs_end_transaction(trans, root); @@ -1251,7 +1163,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, prev_trans = list_entry(cur_trans->list.prev, struct btrfs_transaction, list); if (!prev_trans->commit_done) { - prev_trans->use_count++; + atomic_inc(&prev_trans->use_count); mutex_unlock(&root->fs_info->trans_mutex); wait_for_commit(root, prev_trans); @@ -1279,6 +1191,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, BUG_ON(ret); } + ret = btrfs_run_delayed_items(trans, root); + BUG_ON(ret); + /* * rename don't use btrfs_join_transaction, so, once we * set the transaction to blocked above, we aren't going @@ -1292,24 +1207,28 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, TASK_UNINTERRUPTIBLE); smp_mb(); - if (cur_trans->num_writers > 1) + if (atomic_read(&cur_trans->num_writers) > 1) schedule_timeout(MAX_SCHEDULE_TIMEOUT); else if (should_grow) schedule_timeout(1); mutex_lock(&root->fs_info->trans_mutex); finish_wait(&cur_trans->writer_wait, &wait); - } while (cur_trans->num_writers > 1 || + } while (atomic_read(&cur_trans->num_writers) > 1 || (should_grow && cur_trans->num_joined != joined)); ret = create_pending_snapshots(trans, root->fs_info); BUG_ON(ret); + ret = btrfs_run_delayed_items(trans, root); + BUG_ON(ret); + ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); BUG_ON(ret); WARN_ON(cur_trans != trans->transaction); + btrfs_scrub_pause(root); /* btrfs_commit_tree_roots is responsible for getting the * various roots consistent with each other. Every pointer * in the tree of tree roots has to point to the most up to date @@ -1386,11 +1305,16 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, wake_up(&cur_trans->commit_wait); + list_del_init(&cur_trans->list); put_transaction(cur_trans); put_transaction(cur_trans); + trace_btrfs_transaction_commit(root); + mutex_unlock(&root->fs_info->trans_mutex); + btrfs_scrub_continue(root); + if (current->journal_info == trans) current->journal_info = NULL; @@ -1418,6 +1342,8 @@ int btrfs_clean_old_snapshots(struct btrfs_root *root) root = list_entry(list.next, struct btrfs_root, root_list); list_del(&root->root_list); + btrfs_kill_all_delayed_nodes(root); + if (btrfs_header_backref_rev(root->node) < BTRFS_MIXED_BACKREF_REV) btrfs_drop_snapshot(root, NULL, 0); diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h index 229a594cacd5..804c88639e5d 100644 --- a/fs/btrfs/transaction.h +++ b/fs/btrfs/transaction.h @@ -27,11 +27,11 @@ struct btrfs_transaction { * total writers in this transaction, it must be zero before the * transaction can end */ - unsigned long num_writers; + atomic_t num_writers; unsigned long num_joined; int in_commit; - int use_count; + atomic_t use_count; int commit_done; int blocked; struct list_head list; @@ -101,11 +101,8 @@ struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r, int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid); int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root); -int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans, - struct btrfs_root *root); int btrfs_add_dead_root(struct btrfs_root *root); -int btrfs_drop_dead_root(struct btrfs_root *root); int btrfs_defrag_root(struct btrfs_root *root, int cacheonly); int btrfs_clean_old_snapshots(struct btrfs_root *root); int btrfs_commit_transaction(struct btrfs_trans_handle *trans, @@ -115,6 +112,8 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, int wait_for_unblock); int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans, struct btrfs_root *root); +int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans, + struct btrfs_root *root); int btrfs_should_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root); void btrfs_throttle(struct btrfs_root *root); diff --git a/fs/btrfs/tree-defrag.c b/fs/btrfs/tree-defrag.c index 992ab425599d..3b580ee8ab1d 100644 --- a/fs/btrfs/tree-defrag.c +++ b/fs/btrfs/tree-defrag.c @@ -97,7 +97,7 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans, ret = 0; goto out; } - btrfs_release_path(root, path); + btrfs_release_path(path); wret = btrfs_search_slot(trans, root, &key, path, 0, 1); if (wret < 0) { diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index a4bbb854dfd2..592396c6dc47 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -333,13 +333,13 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans, goto insert; if (item_size == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } dst_copy = kmalloc(item_size, GFP_NOFS); src_copy = kmalloc(item_size, GFP_NOFS); if (!dst_copy || !src_copy) { - btrfs_release_path(root, path); + btrfs_release_path(path); kfree(dst_copy); kfree(src_copy); return -ENOMEM; @@ -361,13 +361,13 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans, * sync */ if (ret == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } } insert: - btrfs_release_path(root, path); + btrfs_release_path(path); /* try to insert the key into the destination tree */ ret = btrfs_insert_empty_item(trans, root, path, key, item_size); @@ -382,7 +382,6 @@ insert: } else if (found_size < item_size) { ret = btrfs_extend_item(trans, root, path, item_size - found_size); - BUG_ON(ret); } } else if (ret) { return ret; @@ -438,7 +437,7 @@ insert: } no_copy: btrfs_mark_buffer_dirty(path->nodes[0]); - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } @@ -519,7 +518,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, * file. This must be done before the btrfs_drop_extents run * so we don't try to drop this extent. */ - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode), start, 0); if (ret == 0 && @@ -544,11 +543,11 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, * we don't have to do anything */ if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto out; } } - btrfs_release_path(root, path); + btrfs_release_path(path); saved_nbytes = inode_get_bytes(inode); /* drop any overlapping extents */ @@ -590,6 +589,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, ins.objectid, ins.offset, 0, root->root_key.objectid, key->objectid, offset); + BUG_ON(ret); } else { /* * insert the extent pointer in the extent @@ -600,7 +600,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, key->objectid, offset, &ins); BUG_ON(ret); } - btrfs_release_path(root, path); + btrfs_release_path(path); if (btrfs_file_extent_compression(eb, item)) { csum_start = ins.objectid; @@ -614,7 +614,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, ret = btrfs_lookup_csums_range(root->log_root, csum_start, csum_end - 1, - &ordered_sums); + &ordered_sums, 0); BUG_ON(ret); while (!list_empty(&ordered_sums)) { struct btrfs_ordered_sum *sums; @@ -629,7 +629,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, kfree(sums); } } else { - btrfs_release_path(root, path); + btrfs_release_path(path); } } else if (found_type == BTRFS_FILE_EXTENT_INLINE) { /* inline extents are easy, we just overwrite them */ @@ -675,10 +675,13 @@ static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans, return -ENOMEM; read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len); - btrfs_release_path(root, path); + btrfs_release_path(path); inode = read_one_inode(root, location.objectid); - BUG_ON(!inode); + if (!inode) { + kfree(name); + return -EIO; + } ret = link_to_fixup_dir(trans, root, path, location.objectid); BUG_ON(ret); @@ -713,7 +716,7 @@ static noinline int inode_in_dir(struct btrfs_root *root, goto out; } else goto out; - btrfs_release_path(root, path); + btrfs_release_path(path); di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0); if (di && !IS_ERR(di)) { @@ -724,7 +727,7 @@ static noinline int inode_in_dir(struct btrfs_root *root, goto out; match = 1; out: - btrfs_release_path(root, path); + btrfs_release_path(path); return match; } @@ -799,12 +802,12 @@ static noinline int add_inode_ref(struct btrfs_trans_handle *trans, struct inode *dir; int ret; struct btrfs_inode_ref *ref; - struct btrfs_dir_item *di; struct inode *inode; char *name; int namelen; unsigned long ref_ptr; unsigned long ref_end; + int search_done = 0; /* * it is possible that we didn't log all the parent directories @@ -817,7 +820,10 @@ static noinline int add_inode_ref(struct btrfs_trans_handle *trans, return -ENOENT; inode = read_one_inode(root, key->objectid); - BUG_ON(!inode); + if (!inode) { + iput(dir); + return -EIO; + } ref_ptr = btrfs_item_ptr_offset(eb, slot); ref_end = ref_ptr + btrfs_item_size_nr(eb, slot); @@ -832,7 +838,7 @@ again: read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen); /* if we already have a perfect match, we're done */ - if (inode_in_dir(root, path, dir->i_ino, inode->i_ino, + if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode), btrfs_inode_ref_index(eb, ref), name, namelen)) { goto out; @@ -845,7 +851,10 @@ again: * existing back reference, and we don't want to create * dangling pointers in the directory. */ -conflict_again: + + if (search_done) + goto insert; + ret = btrfs_search_slot(NULL, root, key, path, 0, 0); if (ret == 0) { char *victim_name; @@ -881,42 +890,26 @@ conflict_again: if (!backref_in_log(log, key, victim_name, victim_name_len)) { btrfs_inc_nlink(inode); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_unlink_inode(trans, root, dir, inode, victim_name, victim_name_len); - kfree(victim_name); - btrfs_release_path(root, path); - goto conflict_again; } kfree(victim_name); ptr = (unsigned long)(victim_ref + 1) + victim_name_len; } BUG_ON(ret); - } - btrfs_release_path(root, path); - - /* look for a conflicting sequence number */ - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, - btrfs_inode_ref_index(eb, ref), - name, namelen, 0); - if (di && !IS_ERR(di)) { - ret = drop_one_dir_item(trans, root, path, dir, di); - BUG_ON(ret); - } - btrfs_release_path(root, path); - - /* look for a conflicting name */ - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, - name, namelen, 0); - if (di && !IS_ERR(di)) { - ret = drop_one_dir_item(trans, root, path, dir, di); - BUG_ON(ret); + /* + * NOTE: we have searched root tree and checked the + * coresponding ref, it does not need to check again. + */ + search_done = 1; } - btrfs_release_path(root, path); + btrfs_release_path(path); +insert: /* insert our name */ ret = btrfs_add_link(trans, dir, inode, name, namelen, 0, btrfs_inode_ref_index(eb, ref)); @@ -935,7 +928,7 @@ out: BUG_ON(ret); out_nowrite: - btrfs_release_path(root, path); + btrfs_release_path(path); iput(dir); iput(inode); return 0; @@ -973,8 +966,9 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, unsigned long ptr; unsigned long ptr_end; int name_len; + u64 ino = btrfs_ino(inode); - key.objectid = inode->i_ino; + key.objectid = ino; key.type = BTRFS_INODE_REF_KEY; key.offset = (u64)-1; @@ -993,7 +987,7 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, } btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); - if (key.objectid != inode->i_ino || + if (key.objectid != ino || key.type != BTRFS_INODE_REF_KEY) break; ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]); @@ -1012,9 +1006,9 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, if (key.offset == 0) break; key.offset--; - btrfs_release_path(root, path); + btrfs_release_path(path); } - btrfs_release_path(root, path); + btrfs_release_path(path); if (nlink != inode->i_nlink) { inode->i_nlink = nlink; btrfs_update_inode(trans, root, inode); @@ -1024,10 +1018,10 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, if (inode->i_nlink == 0) { if (S_ISDIR(inode->i_mode)) { ret = replay_dir_deletes(trans, root, NULL, path, - inode->i_ino, 1); + ino, 1); BUG_ON(ret); } - ret = insert_orphan_item(trans, root, inode->i_ino); + ret = insert_orphan_item(trans, root, ino); BUG_ON(ret); } btrfs_free_path(path); @@ -1063,11 +1057,13 @@ static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans, break; ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); + if (ret) + goto out; - btrfs_release_path(root, path); + btrfs_release_path(path); inode = read_one_inode(root, key.offset); - BUG_ON(!inode); + if (!inode) + return -EIO; ret = fixup_inode_link_count(trans, root, inode); BUG_ON(ret); @@ -1081,8 +1077,10 @@ static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans, */ key.offset = (u64)-1; } - btrfs_release_path(root, path); - return 0; + ret = 0; +out: + btrfs_release_path(path); + return ret; } @@ -1101,7 +1099,8 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans, struct inode *inode; inode = read_one_inode(root, objectid); - BUG_ON(!inode); + if (!inode) + return -EIO; key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID; btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY); @@ -1109,7 +1108,7 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans, ret = btrfs_insert_empty_item(trans, root, path, &key, 0); - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret == 0) { btrfs_inc_nlink(inode); btrfs_update_inode(trans, root, inode); @@ -1188,7 +1187,8 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, int ret; dir = read_one_inode(root, key->objectid); - BUG_ON(!dir); + if (!dir) + return -EIO; name_len = btrfs_dir_name_len(eb, di); name = kmalloc(name_len, GFP_NOFS); @@ -1205,7 +1205,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, exists = 1; else exists = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); if (key->type == BTRFS_DIR_ITEM_KEY) { dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid, @@ -1218,7 +1218,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, } else { BUG(); } - if (!dst_di || IS_ERR(dst_di)) { + if (IS_ERR_OR_NULL(dst_di)) { /* we need a sequence number to insert, so we only * do inserts for the BTRFS_DIR_INDEX_KEY types */ @@ -1249,13 +1249,13 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, if (key->type == BTRFS_DIR_INDEX_KEY) goto insert; out: - btrfs_release_path(root, path); + btrfs_release_path(path); kfree(name); iput(dir); return 0; insert: - btrfs_release_path(root, path); + btrfs_release_path(path); ret = insert_one_name(trans, root, path, key->objectid, key->offset, name, name_len, log_type, &log_key); @@ -1286,6 +1286,8 @@ static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans, ptr_end = ptr + item_size; while (ptr < ptr_end) { di = (struct btrfs_dir_item *)ptr; + if (verify_dir_item(root, eb, di)) + return -EIO; name_len = btrfs_dir_name_len(eb, di); ret = replay_one_name(trans, root, path, eb, di, key); BUG_ON(ret); @@ -1374,7 +1376,7 @@ next: *end_ret = found_end; ret = 0; out: - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } @@ -1412,6 +1414,11 @@ again: ptr_end = ptr + item_size; while (ptr < ptr_end) { di = (struct btrfs_dir_item *)ptr; + if (verify_dir_item(root, eb, di)) { + ret = -EIO; + goto out; + } + name_len = btrfs_dir_name_len(eb, di); name = kmalloc(name_len, GFP_NOFS); if (!name) { @@ -1432,12 +1439,15 @@ again: dir_key->offset, name, name_len, 0); } - if (!log_di || IS_ERR(log_di)) { + if (IS_ERR_OR_NULL(log_di)) { btrfs_dir_item_key_to_cpu(eb, di, &location); - btrfs_release_path(root, path); - btrfs_release_path(log, log_path); + btrfs_release_path(path); + btrfs_release_path(log_path); inode = read_one_inode(root, location.objectid); - BUG_ON(!inode); + if (!inode) { + kfree(name); + return -EIO; + } ret = link_to_fixup_dir(trans, root, path, location.objectid); @@ -1459,7 +1469,7 @@ again: ret = 0; goto out; } - btrfs_release_path(log, log_path); + btrfs_release_path(log_path); kfree(name); ptr = (unsigned long)(di + 1); @@ -1467,8 +1477,8 @@ again: } ret = 0; out: - btrfs_release_path(root, path); - btrfs_release_path(log, log_path); + btrfs_release_path(path); + btrfs_release_path(log_path); return ret; } @@ -1556,7 +1566,7 @@ again: break; dir_key.offset = found_key.offset + 1; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (range_end == (u64)-1) break; range_start = range_end + 1; @@ -1567,11 +1577,11 @@ next_type: if (key_type == BTRFS_DIR_LOG_ITEM_KEY) { key_type = BTRFS_DIR_LOG_INDEX_KEY; dir_key.type = BTRFS_DIR_INDEX_KEY; - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } out: - btrfs_release_path(root, path); + btrfs_release_path(path); btrfs_free_path(log_path); iput(dir); return ret; @@ -1821,7 +1831,8 @@ static int walk_log_tree(struct btrfs_trans_handle *trans, int orig_level; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; level = btrfs_header_level(log->node); orig_level = level; @@ -2098,7 +2109,9 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, * the running transaction open, so a full commit can't hop * in and cause problems either. */ + btrfs_scrub_pause_super(root); write_ctree_super(trans, root->fs_info->tree_root, 1); + btrfs_scrub_continue_super(root); ret = 0; mutex_lock(&root->log_mutex); @@ -2202,6 +2215,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, int ret; int err = 0; int bytes_del = 0; + u64 dir_ino = btrfs_ino(dir); if (BTRFS_I(dir)->logged_trans < trans->transid) return 0; @@ -2214,10 +2228,12 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, log = root->log_root; path = btrfs_alloc_path(); - if (!path) - return -ENOMEM; + if (!path) { + err = -ENOMEM; + goto out_unlock; + } - di = btrfs_lookup_dir_item(trans, log, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, log, path, dir_ino, name, name_len, -1); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2228,8 +2244,8 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, bytes_del += name_len; BUG_ON(ret); } - btrfs_release_path(log, path); - di = btrfs_lookup_dir_index_item(trans, log, path, dir->i_ino, + btrfs_release_path(path); + di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino, index, name, name_len, -1); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2247,10 +2263,10 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, if (bytes_del) { struct btrfs_key key; - key.objectid = dir->i_ino; + key.objectid = dir_ino; key.offset = 0; key.type = BTRFS_INODE_ITEM_KEY; - btrfs_release_path(log, path); + btrfs_release_path(path); ret = btrfs_search_slot(trans, log, &key, path, 0, 1); if (ret < 0) { @@ -2272,10 +2288,11 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(path->nodes[0]); } else ret = 0; - btrfs_release_path(log, path); + btrfs_release_path(path); } fail: btrfs_free_path(path); +out_unlock: mutex_unlock(&BTRFS_I(dir)->log_mutex); if (ret == -ENOSPC) { root->fs_info->last_trans_log_full_commit = trans->transid; @@ -2305,7 +2322,7 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans, log = root->log_root; mutex_lock(&BTRFS_I(inode)->log_mutex); - ret = btrfs_del_inode_ref(trans, log, name, name_len, inode->i_ino, + ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode), dirid, &index); mutex_unlock(&BTRFS_I(inode)->log_mutex); if (ret == -ENOSPC) { @@ -2346,7 +2363,7 @@ static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans, struct btrfs_dir_log_item); btrfs_set_dir_log_end(path->nodes[0], item, last_offset); btrfs_mark_buffer_dirty(path->nodes[0]); - btrfs_release_path(log, path); + btrfs_release_path(path); return 0; } @@ -2371,13 +2388,14 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, int nritems; u64 first_offset = min_offset; u64 last_offset = (u64)-1; + u64 ino = btrfs_ino(inode); log = root->log_root; - max_key.objectid = inode->i_ino; + max_key.objectid = ino; max_key.offset = (u64)-1; max_key.type = key_type; - min_key.objectid = inode->i_ino; + min_key.objectid = ino; min_key.type = key_type; min_key.offset = min_offset; @@ -2390,18 +2408,17 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, * we didn't find anything from this transaction, see if there * is anything at all */ - if (ret != 0 || min_key.objectid != inode->i_ino || - min_key.type != key_type) { - min_key.objectid = inode->i_ino; + if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) { + min_key.objectid = ino; min_key.type = key_type; min_key.offset = (u64)-1; - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0); if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } - ret = btrfs_previous_item(root, path, inode->i_ino, key_type); + ret = btrfs_previous_item(root, path, ino, key_type); /* if ret == 0 there are items for this type, * create a range to tell us the last key of this type. @@ -2419,7 +2436,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, } /* go backward to find any previous key */ - ret = btrfs_previous_item(root, path, inode->i_ino, key_type); + ret = btrfs_previous_item(root, path, ino, key_type); if (ret == 0) { struct btrfs_key tmp; btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]); @@ -2434,7 +2451,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, } } } - btrfs_release_path(root, path); + btrfs_release_path(path); /* find the first key from this transaction again */ ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0); @@ -2454,8 +2471,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, for (i = path->slots[0]; i < nritems; i++) { btrfs_item_key_to_cpu(src, &min_key, i); - if (min_key.objectid != inode->i_ino || - min_key.type != key_type) + if (min_key.objectid != ino || min_key.type != key_type) goto done; ret = overwrite_item(trans, log, dst_path, src, i, &min_key); @@ -2476,7 +2492,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, goto done; } btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]); - if (tmp.objectid != inode->i_ino || tmp.type != key_type) { + if (tmp.objectid != ino || tmp.type != key_type) { last_offset = (u64)-1; goto done; } @@ -2492,8 +2508,8 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, } } done: - btrfs_release_path(root, path); - btrfs_release_path(log, dst_path); + btrfs_release_path(path); + btrfs_release_path(dst_path); if (err == 0) { *last_offset_ret = last_offset; @@ -2502,8 +2518,7 @@ done: * is valid */ ret = insert_dir_log_key(trans, log, path, key_type, - inode->i_ino, first_offset, - last_offset); + ino, first_offset, last_offset); if (ret) err = ret; } @@ -2589,10 +2604,11 @@ static int drop_objectid_items(struct btrfs_trans_handle *trans, break; ret = btrfs_del_item(trans, log, path); - BUG_ON(ret); - btrfs_release_path(log, path); + if (ret) + break; + btrfs_release_path(path); } - btrfs_release_path(log, path); + btrfs_release_path(path); return ret; } @@ -2667,6 +2683,9 @@ static noinline int copy_items(struct btrfs_trans_handle *trans, extent = btrfs_item_ptr(src, start_slot + i, struct btrfs_file_extent_item); + if (btrfs_file_extent_generation(src, extent) < trans->transid) + continue; + found_type = btrfs_file_extent_type(src, extent); if (found_type == BTRFS_FILE_EXTENT_REG || found_type == BTRFS_FILE_EXTENT_PREALLOC) { @@ -2691,14 +2710,14 @@ static noinline int copy_items(struct btrfs_trans_handle *trans, ret = btrfs_lookup_csums_range( log->fs_info->csum_root, ds + cs, ds + cs + cl - 1, - &ordered_sums); + &ordered_sums, 0); BUG_ON(ret); } } } btrfs_mark_buffer_dirty(dst_path->nodes[0]); - btrfs_release_path(log, dst_path); + btrfs_release_path(dst_path); kfree(ins_data); /* @@ -2747,6 +2766,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, int nritems; int ins_start_slot = 0; int ins_nr; + u64 ino = btrfs_ino(inode); log = root->log_root; @@ -2759,11 +2779,11 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, return -ENOMEM; } - min_key.objectid = inode->i_ino; + min_key.objectid = ino; min_key.type = BTRFS_INODE_ITEM_KEY; min_key.offset = 0; - max_key.objectid = inode->i_ino; + max_key.objectid = ino; /* today the code can only do partial logging of directories */ if (!S_ISDIR(inode->i_mode)) @@ -2775,6 +2795,13 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, max_key.type = (u8)-1; max_key.offset = (u64)-1; + ret = btrfs_commit_inode_delayed_items(trans, inode); + if (ret) { + btrfs_free_path(path); + btrfs_free_path(dst_path); + return ret; + } + mutex_lock(&BTRFS_I(inode)->log_mutex); /* @@ -2786,8 +2813,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, if (inode_only == LOG_INODE_EXISTS) max_key_type = BTRFS_XATTR_ITEM_KEY; - ret = drop_objectid_items(trans, log, path, - inode->i_ino, max_key_type); + ret = drop_objectid_items(trans, log, path, ino, max_key_type); } else { ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0); } @@ -2805,7 +2831,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, break; again: /* note, ins_nr might be > 0 here, cleanup outside the loop */ - if (min_key.objectid != inode->i_ino) + if (min_key.objectid != ino) break; if (min_key.type > max_key.type) break; @@ -2847,7 +2873,7 @@ next_slot: } ins_nr = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (min_key.offset < (u64)-1) min_key.offset++; @@ -2870,8 +2896,8 @@ next_slot: } WARN_ON(ins_nr); if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) { - btrfs_release_path(root, path); - btrfs_release_path(log, dst_path); + btrfs_release_path(path); + btrfs_release_path(dst_path); ret = log_directory_changes(trans, root, inode, path, dst_path); if (ret) { err = ret; @@ -3107,9 +3133,11 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree) .stage = 0, }; - fs_info->log_root_recovering = 1; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; + + fs_info->log_root_recovering = 1; trans = btrfs_start_transaction(fs_info->tree_root, 0); BUG_ON(IS_ERR(trans)); @@ -3117,7 +3145,8 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree) wc.trans = trans; wc.pin = 1; - walk_log_tree(trans, log_root_tree, &wc); + ret = walk_log_tree(trans, log_root_tree, &wc); + BUG_ON(ret); again: key.objectid = BTRFS_TREE_LOG_OBJECTID; @@ -3135,14 +3164,13 @@ again: } btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]); - btrfs_release_path(log_root_tree, path); + btrfs_release_path(path); if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID) break; log = btrfs_read_fs_root_no_radix(log_root_tree, &found_key); - BUG_ON(!log); - + BUG_ON(IS_ERR(log)); tmp_key.objectid = found_key.offset; tmp_key.type = BTRFS_ROOT_ITEM_KEY; @@ -3171,7 +3199,7 @@ again: if (found_key.offset == 0) break; } - btrfs_release_path(log_root_tree, path); + btrfs_release_path(path); /* step one is to pin it all, step two is to replay just inodes */ if (wc.pin) { diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h index 3dfae84c8cc8..2270ac58d746 100644 --- a/fs/btrfs/tree-log.h +++ b/fs/btrfs/tree-log.h @@ -38,7 +38,6 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, int name_len, struct inode *inode, u64 dirid); -int btrfs_join_running_log_trans(struct btrfs_root *root); int btrfs_end_log_trans(struct btrfs_root *root); int btrfs_pin_log_trans(struct btrfs_root *root); int btrfs_log_inode_parent(struct btrfs_trans_handle *trans, diff --git a/fs/btrfs/version.sh b/fs/btrfs/version.sh deleted file mode 100644 index 1ca1952fd917..000000000000 --- a/fs/btrfs/version.sh +++ /dev/null @@ -1,43 +0,0 @@ -#!/bin/bash -# -# determine-version -- report a useful version for releases -# -# Copyright 2008, Aron Griffis <agriffis@n01se.net> -# Copyright 2008, Oracle -# Released under the GNU GPLv2 - -v="v0.16" - -which git &> /dev/null -if [ $? == 0 ]; then - git branch >& /dev/null - if [ $? == 0 ]; then - if head=`git rev-parse --verify HEAD 2>/dev/null`; then - if tag=`git describe --tags 2>/dev/null`; then - v="$tag" - fi - - # Are there uncommitted changes? - git update-index --refresh --unmerged > /dev/null - if git diff-index --name-only HEAD | \ - grep -v "^scripts/package" \ - | read dummy; then - v="$v"-dirty - fi - fi - fi -fi - -echo "#ifndef __BUILD_VERSION" > .build-version.h -echo "#define __BUILD_VERSION" >> .build-version.h -echo "#define BTRFS_BUILD_VERSION \"Btrfs $v\"" >> .build-version.h -echo "#endif" >> .build-version.h - -diff -q version.h .build-version.h >& /dev/null - -if [ $? == 0 ]; then - rm .build-version.h - exit 0 -fi - -mv .build-version.h version.h diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 2636a051e4b2..c48214ef5c09 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -33,38 +33,14 @@ #include "volumes.h" #include "async-thread.h" -struct map_lookup { - u64 type; - int io_align; - int io_width; - int stripe_len; - int sector_size; - int num_stripes; - int sub_stripes; - struct btrfs_bio_stripe stripes[]; -}; - static int init_first_rw_device(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_device *device); static int btrfs_relocate_sys_chunks(struct btrfs_root *root); -#define map_lookup_size(n) (sizeof(struct map_lookup) + \ - (sizeof(struct btrfs_bio_stripe) * (n))) - static DEFINE_MUTEX(uuid_mutex); static LIST_HEAD(fs_uuids); -void btrfs_lock_volumes(void) -{ - mutex_lock(&uuid_mutex); -} - -void btrfs_unlock_volumes(void) -{ - mutex_unlock(&uuid_mutex); -} - static void lock_chunks(struct btrfs_root *root) { mutex_lock(&root->fs_info->chunk_mutex); @@ -162,22 +138,25 @@ static noinline int run_scheduled_bios(struct btrfs_device *device) struct bio *cur; int again = 0; unsigned long num_run; - unsigned long num_sync_run; unsigned long batch_run = 0; unsigned long limit; unsigned long last_waited = 0; int force_reg = 0; + struct blk_plug plug; + + /* + * this function runs all the bios we've collected for + * a particular device. We don't want to wander off to + * another device without first sending all of these down. + * So, setup a plug here and finish it off before we return + */ + blk_start_plug(&plug); bdi = blk_get_backing_dev_info(device->bdev); fs_info = device->dev_root->fs_info; limit = btrfs_async_submit_limit(fs_info); limit = limit * 2 / 3; - /* we want to make sure that every time we switch from the sync - * list to the normal list, we unplug - */ - num_sync_run = 0; - loop: spin_lock(&device->io_lock); @@ -223,15 +202,6 @@ loop_lock: spin_unlock(&device->io_lock); - /* - * if we're doing the regular priority list, make sure we unplug - * for any high prio bios we've sent down - */ - if (pending_bios == &device->pending_bios && num_sync_run > 0) { - num_sync_run = 0; - blk_run_backing_dev(bdi, NULL); - } - while (pending) { rmb(); @@ -259,19 +229,11 @@ loop_lock: BUG_ON(atomic_read(&cur->bi_cnt) == 0); - if (cur->bi_rw & REQ_SYNC) - num_sync_run++; - submit_bio(cur->bi_rw, cur); num_run++; batch_run++; - if (need_resched()) { - if (num_sync_run) { - blk_run_backing_dev(bdi, NULL); - num_sync_run = 0; - } + if (need_resched()) cond_resched(); - } /* * we made progress, there is more work to do and the bdi @@ -304,13 +266,8 @@ loop_lock: * against it before looping */ last_waited = ioc->last_waited; - if (need_resched()) { - if (num_sync_run) { - blk_run_backing_dev(bdi, NULL); - num_sync_run = 0; - } + if (need_resched()) cond_resched(); - } continue; } spin_lock(&device->io_lock); @@ -323,22 +280,6 @@ loop_lock: } } - if (num_sync_run) { - num_sync_run = 0; - blk_run_backing_dev(bdi, NULL); - } - /* - * IO has already been through a long path to get here. Checksumming, - * async helper threads, perhaps compression. We've done a pretty - * good job of collecting a batch of IO and should just unplug - * the device right away. - * - * This will help anyone who is waiting on the IO, they might have - * already unplugged, but managed to do so before the bio they - * cared about found its way down here. - */ - blk_run_backing_dev(bdi, NULL); - cond_resched(); if (again) goto loop; @@ -349,6 +290,7 @@ loop_lock: spin_unlock(&device->io_lock); done: + blk_finish_plug(&plug); return 0; } @@ -408,7 +350,7 @@ static noinline int device_list_add(const char *path, INIT_LIST_HEAD(&device->dev_alloc_list); mutex_lock(&fs_devices->device_list_mutex); - list_add(&device->dev_list, &fs_devices->devices); + list_add_rcu(&device->dev_list, &fs_devices->devices); mutex_unlock(&fs_devices->device_list_mutex); device->fs_devices = fs_devices; @@ -451,7 +393,7 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig) fs_devices->latest_trans = orig->latest_trans; memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid)); - mutex_lock(&orig->device_list_mutex); + /* We have held the volume lock, it is safe to get the devices. */ list_for_each_entry(orig_dev, &orig->devices, dev_list) { device = kzalloc(sizeof(*device), GFP_NOFS); if (!device) @@ -474,10 +416,8 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig) device->fs_devices = fs_devices; fs_devices->num_devices++; } - mutex_unlock(&orig->device_list_mutex); return fs_devices; error: - mutex_unlock(&orig->device_list_mutex); free_fs_devices(fs_devices); return ERR_PTR(-ENOMEM); } @@ -488,7 +428,7 @@ int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices) mutex_lock(&uuid_mutex); again: - mutex_lock(&fs_devices->device_list_mutex); + /* This is the initialized path, it is safe to release the devices. */ list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) { if (device->in_fs_metadata) continue; @@ -508,7 +448,6 @@ again: kfree(device->name); kfree(device); } - mutex_unlock(&fs_devices->device_list_mutex); if (fs_devices->seed) { fs_devices = fs_devices->seed; @@ -519,6 +458,29 @@ again: return 0; } +static void __free_device(struct work_struct *work) +{ + struct btrfs_device *device; + + device = container_of(work, struct btrfs_device, rcu_work); + + if (device->bdev) + blkdev_put(device->bdev, device->mode); + + kfree(device->name); + kfree(device); +} + +static void free_device(struct rcu_head *head) +{ + struct btrfs_device *device; + + device = container_of(head, struct btrfs_device, rcu); + + INIT_WORK(&device->rcu_work, __free_device); + schedule_work(&device->rcu_work); +} + static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) { struct btrfs_device *device; @@ -526,20 +488,32 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) if (--fs_devices->opened > 0) return 0; + mutex_lock(&fs_devices->device_list_mutex); list_for_each_entry(device, &fs_devices->devices, dev_list) { - if (device->bdev) { - blkdev_put(device->bdev, device->mode); + struct btrfs_device *new_device; + + if (device->bdev) fs_devices->open_devices--; - } + if (device->writeable) { list_del_init(&device->dev_alloc_list); fs_devices->rw_devices--; } - device->bdev = NULL; - device->writeable = 0; - device->in_fs_metadata = 0; + new_device = kmalloc(sizeof(*new_device), GFP_NOFS); + BUG_ON(!new_device); + memcpy(new_device, device, sizeof(*new_device)); + new_device->name = kstrdup(device->name, GFP_NOFS); + BUG_ON(!new_device->name); + new_device->bdev = NULL; + new_device->writeable = 0; + new_device->in_fs_metadata = 0; + list_replace_rcu(&device->dev_list, &new_device->dev_list); + + call_rcu(&device->rcu, free_device); } + mutex_unlock(&fs_devices->device_list_mutex); + WARN_ON(fs_devices->open_devices); WARN_ON(fs_devices->rw_devices); fs_devices->opened = 0; @@ -642,6 +616,7 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, list_add(&device->dev_alloc_list, &fs_devices->alloc_list); } + brelse(bh); continue; error_brelse: @@ -860,10 +835,7 @@ int find_free_dev_extent(struct btrfs_trans_handle *trans, /* we don't want to overwrite the superblock on the drive, * so we make sure to start at an offset of at least 1MB */ - search_start = 1024 * 1024; - - if (root->fs_info->alloc_start + num_bytes <= search_end) - search_start = max(root->fs_info->alloc_start, search_start); + search_start = max(root->fs_info->alloc_start, 1024ull * 1024); max_hole_start = search_start; max_hole_size = 0; @@ -994,14 +966,14 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, if (ret > 0) { ret = btrfs_previous_item(root, path, key.objectid, BTRFS_DEV_EXTENT_KEY); - BUG_ON(ret); + if (ret) + goto out; leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); extent = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_extent); BUG_ON(found_key.offset > start || found_key.offset + btrfs_dev_extent_length(leaf, extent) < start); - ret = 0; } else if (ret == 0) { leaf = path->nodes[0]; extent = btrfs_item_ptr(leaf, path->slots[0], @@ -1012,8 +984,8 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, if (device->bytes_used > 0) device->bytes_used -= btrfs_dev_extent_length(leaf, extent); ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); +out: btrfs_free_path(path); return ret; } @@ -1248,11 +1220,13 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) struct block_device *bdev; struct buffer_head *bh = NULL; struct btrfs_super_block *disk_super; + struct btrfs_fs_devices *cur_devices; u64 all_avail; u64 devid; u64 num_devices; u8 *dev_uuid; int ret = 0; + bool clear_super = false; mutex_lock(&uuid_mutex); mutex_lock(&root->fs_info->volume_mutex); @@ -1283,14 +1257,16 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) device = NULL; devices = &root->fs_info->fs_devices->devices; - mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + /* + * It is safe to read the devices since the volume_mutex + * is held. + */ list_for_each_entry(tmp, devices, dev_list) { if (tmp->in_fs_metadata && !tmp->bdev) { device = tmp; break; } } - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); bdev = NULL; bh = NULL; disk_super = NULL; @@ -1332,28 +1308,33 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) } if (device->writeable) { + lock_chunks(root); list_del_init(&device->dev_alloc_list); + unlock_chunks(root); root->fs_info->fs_devices->rw_devices--; + clear_super = true; } ret = btrfs_shrink_device(device, 0); if (ret) - goto error_brelse; + goto error_undo; ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device); if (ret) - goto error_brelse; + goto error_undo; device->in_fs_metadata = 0; + btrfs_scrub_cancel_dev(root, device); /* * the device list mutex makes sure that we don't change * the device list while someone else is writing out all * the device supers. */ + + cur_devices = device->fs_devices; mutex_lock(&root->fs_info->fs_devices->device_list_mutex); - list_del_init(&device->dev_list); - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + list_del_rcu(&device->dev_list); device->fs_devices->num_devices--; @@ -1367,34 +1348,36 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (device->bdev == root->fs_info->fs_devices->latest_bdev) root->fs_info->fs_devices->latest_bdev = next_device->bdev; - if (device->bdev) { - blkdev_put(device->bdev, device->mode); - device->bdev = NULL; + if (device->bdev) device->fs_devices->open_devices--; - } + + call_rcu(&device->rcu, free_device); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1; btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices); - if (device->fs_devices->open_devices == 0) { + if (cur_devices->open_devices == 0) { struct btrfs_fs_devices *fs_devices; fs_devices = root->fs_info->fs_devices; while (fs_devices) { - if (fs_devices->seed == device->fs_devices) + if (fs_devices->seed == cur_devices) break; fs_devices = fs_devices->seed; } - fs_devices->seed = device->fs_devices->seed; - device->fs_devices->seed = NULL; - __btrfs_close_devices(device->fs_devices); - free_fs_devices(device->fs_devices); + fs_devices->seed = cur_devices->seed; + cur_devices->seed = NULL; + lock_chunks(root); + __btrfs_close_devices(cur_devices); + unlock_chunks(root); + free_fs_devices(cur_devices); } /* * at this point, the device is zero sized. We want to * remove it from the devices list and zero out the old super */ - if (device->writeable) { + if (clear_super) { /* make sure this device isn't detected as part of * the FS anymore */ @@ -1403,8 +1386,6 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) sync_dirty_buffer(bh); } - kfree(device->name); - kfree(device); ret = 0; error_brelse: @@ -1416,6 +1397,15 @@ out: mutex_unlock(&root->fs_info->volume_mutex); mutex_unlock(&uuid_mutex); return ret; +error_undo: + if (device->writeable) { + lock_chunks(root); + list_add(&device->dev_alloc_list, + &root->fs_info->fs_devices->alloc_list); + unlock_chunks(root); + root->fs_info->fs_devices->rw_devices++; + } + goto error_brelse; } /* @@ -1452,7 +1442,12 @@ static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans, INIT_LIST_HEAD(&seed_devices->devices); INIT_LIST_HEAD(&seed_devices->alloc_list); mutex_init(&seed_devices->device_list_mutex); - list_splice_init(&fs_devices->devices, &seed_devices->devices); + + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + list_splice_init_rcu(&fs_devices->devices, &seed_devices->devices, + synchronize_rcu); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list); list_for_each_entry(device, &seed_devices->devices, dev_list) { device->fs_devices = seed_devices; @@ -1513,7 +1508,7 @@ next_slot: goto error; leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } @@ -1605,12 +1600,14 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) ret = find_next_devid(root, &device->devid); if (ret) { + kfree(device->name); kfree(device); goto error; } trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) { + kfree(device->name); kfree(device); ret = PTR_ERR(trans); goto error; @@ -1631,7 +1628,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) device->dev_root = root->fs_info->dev_root; device->bdev = bdev; device->in_fs_metadata = 1; - device->mode = 0; + device->mode = FMODE_EXCL; set_blocksize(device->bdev, 4096); if (seeding_dev) { @@ -1647,7 +1644,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) * half setup */ mutex_lock(&root->fs_info->fs_devices->device_list_mutex); - list_add(&device->dev_list, &root->fs_info->fs_devices->devices); + list_add_rcu(&device->dev_list, &root->fs_info->fs_devices->devices); list_add(&device->dev_alloc_list, &root->fs_info->fs_devices->alloc_list); root->fs_info->fs_devices->num_devices++; @@ -1805,10 +1802,9 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans, BUG_ON(ret); ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); btrfs_free_path(path); - return 0; + return ret; } static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64 @@ -1914,6 +1910,8 @@ static int btrfs_relocate_chunk(struct btrfs_root *root, BUG_ON(ret); + trace_btrfs_chunk_free(root, map, chunk_offset, em->len); + if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) { ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset); BUG_ON(ret); @@ -1981,7 +1979,7 @@ again: chunk = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_chunk); chunk_type = btrfs_chunk_type(leaf, chunk); - btrfs_release_path(chunk_root, path); + btrfs_release_path(path); if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) { ret = btrfs_relocate_chunk(chunk_root, chunk_tree, @@ -2099,7 +2097,7 @@ int btrfs_balance(struct btrfs_root *dev_root) if (found_key.offset == 0) break; - btrfs_release_path(chunk_root, path); + btrfs_release_path(path); ret = btrfs_relocate_chunk(chunk_root, chunk_root->root_key.objectid, found_key.objectid, @@ -2171,7 +2169,7 @@ again: goto done; if (ret) { ret = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); break; } @@ -2180,7 +2178,7 @@ again: btrfs_item_key_to_cpu(l, &key, path->slots[0]); if (key.objectid != device->devid) { - btrfs_release_path(root, path); + btrfs_release_path(path); break; } @@ -2188,14 +2186,14 @@ again: length = btrfs_dev_extent_length(l, dev_extent); if (key.offset + length <= new_size) { - btrfs_release_path(root, path); + btrfs_release_path(path); break; } chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent); chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent); chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_relocate_chunk(root, chunk_tree, chunk_objectid, chunk_offset); @@ -2271,275 +2269,204 @@ static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans, return 0; } -static noinline u64 chunk_bytes_by_type(u64 type, u64 calc_size, - int num_stripes, int sub_stripes) +/* + * sort the devices in descending order by max_avail, total_avail + */ +static int btrfs_cmp_device_info(const void *a, const void *b) { - if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP)) - return calc_size; - else if (type & BTRFS_BLOCK_GROUP_RAID10) - return calc_size * (num_stripes / sub_stripes); - else - return calc_size * num_stripes; -} + const struct btrfs_device_info *di_a = a; + const struct btrfs_device_info *di_b = b; -/* Used to sort the devices by max_avail(descending sort) */ -int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2) -{ - if (((struct btrfs_device_info *)dev_info1)->max_avail > - ((struct btrfs_device_info *)dev_info2)->max_avail) + if (di_a->max_avail > di_b->max_avail) return -1; - else if (((struct btrfs_device_info *)dev_info1)->max_avail < - ((struct btrfs_device_info *)dev_info2)->max_avail) + if (di_a->max_avail < di_b->max_avail) return 1; - else - return 0; + if (di_a->total_avail > di_b->total_avail) + return -1; + if (di_a->total_avail < di_b->total_avail) + return 1; + return 0; } -static int __btrfs_calc_nstripes(struct btrfs_fs_devices *fs_devices, u64 type, - int *num_stripes, int *min_stripes, - int *sub_stripes) +static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, + struct btrfs_root *extent_root, + struct map_lookup **map_ret, + u64 *num_bytes_out, u64 *stripe_size_out, + u64 start, u64 type) { - *num_stripes = 1; - *min_stripes = 1; - *sub_stripes = 0; + struct btrfs_fs_info *info = extent_root->fs_info; + struct btrfs_fs_devices *fs_devices = info->fs_devices; + struct list_head *cur; + struct map_lookup *map = NULL; + struct extent_map_tree *em_tree; + struct extent_map *em; + struct btrfs_device_info *devices_info = NULL; + u64 total_avail; + int num_stripes; /* total number of stripes to allocate */ + int sub_stripes; /* sub_stripes info for map */ + int dev_stripes; /* stripes per dev */ + int devs_max; /* max devs to use */ + int devs_min; /* min devs needed */ + int devs_increment; /* ndevs has to be a multiple of this */ + int ncopies; /* how many copies to data has */ + int ret; + u64 max_stripe_size; + u64 max_chunk_size; + u64 stripe_size; + u64 num_bytes; + int ndevs; + int i; + int j; - if (type & (BTRFS_BLOCK_GROUP_RAID0)) { - *num_stripes = fs_devices->rw_devices; - *min_stripes = 2; - } - if (type & (BTRFS_BLOCK_GROUP_DUP)) { - *num_stripes = 2; - *min_stripes = 2; - } - if (type & (BTRFS_BLOCK_GROUP_RAID1)) { - if (fs_devices->rw_devices < 2) - return -ENOSPC; - *num_stripes = 2; - *min_stripes = 2; - } - if (type & (BTRFS_BLOCK_GROUP_RAID10)) { - *num_stripes = fs_devices->rw_devices; - if (*num_stripes < 4) - return -ENOSPC; - *num_stripes &= ~(u32)1; - *sub_stripes = 2; - *min_stripes = 4; + if ((type & BTRFS_BLOCK_GROUP_RAID1) && + (type & BTRFS_BLOCK_GROUP_DUP)) { + WARN_ON(1); + type &= ~BTRFS_BLOCK_GROUP_DUP; } - return 0; -} + if (list_empty(&fs_devices->alloc_list)) + return -ENOSPC; -static u64 __btrfs_calc_stripe_size(struct btrfs_fs_devices *fs_devices, - u64 proposed_size, u64 type, - int num_stripes, int small_stripe) -{ - int min_stripe_size = 1 * 1024 * 1024; - u64 calc_size = proposed_size; - u64 max_chunk_size = calc_size; - int ncopies = 1; + sub_stripes = 1; + dev_stripes = 1; + devs_increment = 1; + ncopies = 1; + devs_max = 0; /* 0 == as many as possible */ + devs_min = 1; - if (type & (BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_DUP | - BTRFS_BLOCK_GROUP_RAID10)) + /* + * define the properties of each RAID type. + * FIXME: move this to a global table and use it in all RAID + * calculation code + */ + if (type & (BTRFS_BLOCK_GROUP_DUP)) { + dev_stripes = 2; + ncopies = 2; + devs_max = 1; + } else if (type & (BTRFS_BLOCK_GROUP_RAID0)) { + devs_min = 2; + } else if (type & (BTRFS_BLOCK_GROUP_RAID1)) { + devs_increment = 2; + ncopies = 2; + devs_max = 2; + devs_min = 2; + } else if (type & (BTRFS_BLOCK_GROUP_RAID10)) { + sub_stripes = 2; + devs_increment = 2; ncopies = 2; + devs_min = 4; + } else { + devs_max = 1; + } if (type & BTRFS_BLOCK_GROUP_DATA) { - max_chunk_size = 10 * calc_size; - min_stripe_size = 64 * 1024 * 1024; + max_stripe_size = 1024 * 1024 * 1024; + max_chunk_size = 10 * max_stripe_size; } else if (type & BTRFS_BLOCK_GROUP_METADATA) { - max_chunk_size = 256 * 1024 * 1024; - min_stripe_size = 32 * 1024 * 1024; + max_stripe_size = 256 * 1024 * 1024; + max_chunk_size = max_stripe_size; } else if (type & BTRFS_BLOCK_GROUP_SYSTEM) { - calc_size = 8 * 1024 * 1024; - max_chunk_size = calc_size * 2; - min_stripe_size = 1 * 1024 * 1024; + max_stripe_size = 8 * 1024 * 1024; + max_chunk_size = 2 * max_stripe_size; + } else { + printk(KERN_ERR "btrfs: invalid chunk type 0x%llx requested\n", + type); + BUG_ON(1); } /* we don't want a chunk larger than 10% of writeable space */ max_chunk_size = min(div_factor(fs_devices->total_rw_bytes, 1), max_chunk_size); - if (calc_size * num_stripes > max_chunk_size * ncopies) { - calc_size = max_chunk_size * ncopies; - do_div(calc_size, num_stripes); - do_div(calc_size, BTRFS_STRIPE_LEN); - calc_size *= BTRFS_STRIPE_LEN; - } + devices_info = kzalloc(sizeof(*devices_info) * fs_devices->rw_devices, + GFP_NOFS); + if (!devices_info) + return -ENOMEM; - /* we don't want tiny stripes */ - if (!small_stripe) - calc_size = max_t(u64, min_stripe_size, calc_size); + cur = fs_devices->alloc_list.next; /* - * we're about to do_div by the BTRFS_STRIPE_LEN so lets make sure - * we end up with something bigger than a stripe + * in the first pass through the devices list, we gather information + * about the available holes on each device. */ - calc_size = max_t(u64, calc_size, BTRFS_STRIPE_LEN); + ndevs = 0; + while (cur != &fs_devices->alloc_list) { + struct btrfs_device *device; + u64 max_avail; + u64 dev_offset; - do_div(calc_size, BTRFS_STRIPE_LEN); - calc_size *= BTRFS_STRIPE_LEN; - - return calc_size; -} - -static struct map_lookup *__shrink_map_lookup_stripes(struct map_lookup *map, - int num_stripes) -{ - struct map_lookup *new; - size_t len = map_lookup_size(num_stripes); - - BUG_ON(map->num_stripes < num_stripes); - - if (map->num_stripes == num_stripes) - return map; - - new = kmalloc(len, GFP_NOFS); - if (!new) { - /* just change map->num_stripes */ - map->num_stripes = num_stripes; - return map; - } - - memcpy(new, map, len); - new->num_stripes = num_stripes; - kfree(map); - return new; -} - -/* - * helper to allocate device space from btrfs_device_info, in which we stored - * max free space information of every device. It is used when we can not - * allocate chunks by default size. - * - * By this helper, we can allocate a new chunk as larger as possible. - */ -static int __btrfs_alloc_tiny_space(struct btrfs_trans_handle *trans, - struct btrfs_fs_devices *fs_devices, - struct btrfs_device_info *devices, - int nr_device, u64 type, - struct map_lookup **map_lookup, - int min_stripes, u64 *stripe_size) -{ - int i, index, sort_again = 0; - int min_devices = min_stripes; - u64 max_avail, min_free; - struct map_lookup *map = *map_lookup; - int ret; + device = list_entry(cur, struct btrfs_device, dev_alloc_list); - if (nr_device < min_stripes) - return -ENOSPC; + cur = cur->next; - btrfs_descending_sort_devices(devices, nr_device); + if (!device->writeable) { + printk(KERN_ERR + "btrfs: read-only device in alloc_list\n"); + WARN_ON(1); + continue; + } - max_avail = devices[0].max_avail; - if (!max_avail) - return -ENOSPC; + if (!device->in_fs_metadata) + continue; - for (i = 0; i < nr_device; i++) { - /* - * if dev_offset = 0, it means the free space of this device - * is less than what we need, and we didn't search max avail - * extent on this device, so do it now. + if (device->total_bytes > device->bytes_used) + total_avail = device->total_bytes - device->bytes_used; + else + total_avail = 0; + /* avail is off by max(alloc_start, 1MB), but that is the same + * for all devices, so it doesn't hurt the sorting later on */ - if (!devices[i].dev_offset) { - ret = find_free_dev_extent(trans, devices[i].dev, - max_avail, - &devices[i].dev_offset, - &devices[i].max_avail); - if (ret != 0 && ret != -ENOSPC) - return ret; - sort_again = 1; - } - } - /* we update the max avail free extent of each devices, sort again */ - if (sort_again) - btrfs_descending_sort_devices(devices, nr_device); - - if (type & BTRFS_BLOCK_GROUP_DUP) - min_devices = 1; + ret = find_free_dev_extent(trans, device, + max_stripe_size * dev_stripes, + &dev_offset, &max_avail); + if (ret && ret != -ENOSPC) + goto error; - if (!devices[min_devices - 1].max_avail) - return -ENOSPC; + if (ret == 0) + max_avail = max_stripe_size * dev_stripes; - max_avail = devices[min_devices - 1].max_avail; - if (type & BTRFS_BLOCK_GROUP_DUP) - do_div(max_avail, 2); + if (max_avail < BTRFS_STRIPE_LEN * dev_stripes) + continue; - max_avail = __btrfs_calc_stripe_size(fs_devices, max_avail, type, - min_stripes, 1); - if (type & BTRFS_BLOCK_GROUP_DUP) - min_free = max_avail * 2; - else - min_free = max_avail; + devices_info[ndevs].dev_offset = dev_offset; + devices_info[ndevs].max_avail = max_avail; + devices_info[ndevs].total_avail = total_avail; + devices_info[ndevs].dev = device; + ++ndevs; + } - if (min_free > devices[min_devices - 1].max_avail) - return -ENOSPC; + /* + * now sort the devices by hole size / available space + */ + sort(devices_info, ndevs, sizeof(struct btrfs_device_info), + btrfs_cmp_device_info, NULL); - map = __shrink_map_lookup_stripes(map, min_stripes); - *stripe_size = max_avail; + /* round down to number of usable stripes */ + ndevs -= ndevs % devs_increment; - index = 0; - for (i = 0; i < min_stripes; i++) { - map->stripes[i].dev = devices[index].dev; - map->stripes[i].physical = devices[index].dev_offset; - if (type & BTRFS_BLOCK_GROUP_DUP) { - i++; - map->stripes[i].dev = devices[index].dev; - map->stripes[i].physical = devices[index].dev_offset + - max_avail; - } - index++; + if (ndevs < devs_increment * sub_stripes || ndevs < devs_min) { + ret = -ENOSPC; + goto error; } - *map_lookup = map; - - return 0; -} -static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct map_lookup **map_ret, - u64 *num_bytes, u64 *stripe_size, - u64 start, u64 type) -{ - struct btrfs_fs_info *info = extent_root->fs_info; - struct btrfs_device *device = NULL; - struct btrfs_fs_devices *fs_devices = info->fs_devices; - struct list_head *cur; - struct map_lookup *map; - struct extent_map_tree *em_tree; - struct extent_map *em; - struct btrfs_device_info *devices_info; - struct list_head private_devs; - u64 calc_size = 1024 * 1024 * 1024; - u64 min_free; - u64 avail; - u64 dev_offset; - int num_stripes; - int min_stripes; - int sub_stripes; - int min_devices; /* the min number of devices we need */ - int i; - int ret; - int index; + if (devs_max && ndevs > devs_max) + ndevs = devs_max; + /* + * the primary goal is to maximize the number of stripes, so use as many + * devices as possible, even if the stripes are not maximum sized. + */ + stripe_size = devices_info[ndevs-1].max_avail; + num_stripes = ndevs * dev_stripes; - if ((type & BTRFS_BLOCK_GROUP_RAID1) && - (type & BTRFS_BLOCK_GROUP_DUP)) { - WARN_ON(1); - type &= ~BTRFS_BLOCK_GROUP_DUP; + if (stripe_size * num_stripes > max_chunk_size * ncopies) { + stripe_size = max_chunk_size * ncopies; + do_div(stripe_size, num_stripes); } - if (list_empty(&fs_devices->alloc_list)) - return -ENOSPC; - ret = __btrfs_calc_nstripes(fs_devices, type, &num_stripes, - &min_stripes, &sub_stripes); - if (ret) - return ret; - - devices_info = kzalloc(sizeof(*devices_info) * fs_devices->rw_devices, - GFP_NOFS); - if (!devices_info) - return -ENOMEM; + do_div(stripe_size, dev_stripes); + do_div(stripe_size, BTRFS_STRIPE_LEN); + stripe_size *= BTRFS_STRIPE_LEN; map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS); if (!map) { @@ -2548,85 +2475,12 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, } map->num_stripes = num_stripes; - cur = fs_devices->alloc_list.next; - index = 0; - i = 0; - - calc_size = __btrfs_calc_stripe_size(fs_devices, calc_size, type, - num_stripes, 0); - - if (type & BTRFS_BLOCK_GROUP_DUP) { - min_free = calc_size * 2; - min_devices = 1; - } else { - min_free = calc_size; - min_devices = min_stripes; - } - - INIT_LIST_HEAD(&private_devs); - while (index < num_stripes) { - device = list_entry(cur, struct btrfs_device, dev_alloc_list); - BUG_ON(!device->writeable); - if (device->total_bytes > device->bytes_used) - avail = device->total_bytes - device->bytes_used; - else - avail = 0; - cur = cur->next; - - if (device->in_fs_metadata && avail >= min_free) { - ret = find_free_dev_extent(trans, device, min_free, - &devices_info[i].dev_offset, - &devices_info[i].max_avail); - if (ret == 0) { - list_move_tail(&device->dev_alloc_list, - &private_devs); - map->stripes[index].dev = device; - map->stripes[index].physical = - devices_info[i].dev_offset; - index++; - if (type & BTRFS_BLOCK_GROUP_DUP) { - map->stripes[index].dev = device; - map->stripes[index].physical = - devices_info[i].dev_offset + - calc_size; - index++; - } - } else if (ret != -ENOSPC) - goto error; - - devices_info[i].dev = device; - i++; - } else if (device->in_fs_metadata && - avail >= BTRFS_STRIPE_LEN) { - devices_info[i].dev = device; - devices_info[i].max_avail = avail; - i++; - } - - if (cur == &fs_devices->alloc_list) - break; - } - - list_splice(&private_devs, &fs_devices->alloc_list); - if (index < num_stripes) { - if (index >= min_stripes) { - num_stripes = index; - if (type & (BTRFS_BLOCK_GROUP_RAID10)) { - num_stripes /= sub_stripes; - num_stripes *= sub_stripes; - } - - map = __shrink_map_lookup_stripes(map, num_stripes); - } else if (i >= min_devices) { - ret = __btrfs_alloc_tiny_space(trans, fs_devices, - devices_info, i, type, - &map, min_stripes, - &calc_size); - if (ret) - goto error; - } else { - ret = -ENOSPC; - goto error; + for (i = 0; i < ndevs; ++i) { + for (j = 0; j < dev_stripes; ++j) { + int s = i * dev_stripes + j; + map->stripes[s].dev = devices_info[i].dev; + map->stripes[s].physical = devices_info[i].dev_offset + + j * stripe_size; } } map->sector_size = extent_root->sectorsize; @@ -2637,18 +2491,21 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, map->sub_stripes = sub_stripes; *map_ret = map; - *stripe_size = calc_size; - *num_bytes = chunk_bytes_by_type(type, calc_size, - map->num_stripes, sub_stripes); + num_bytes = stripe_size * (num_stripes / ncopies); + + *stripe_size_out = stripe_size; + *num_bytes_out = num_bytes; + + trace_btrfs_chunk_alloc(info->chunk_root, map, start, num_bytes); - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) { ret = -ENOMEM; goto error; } em->bdev = (struct block_device *)map; em->start = start; - em->len = *num_bytes; + em->len = num_bytes; em->block_start = 0; em->block_len = em->len; @@ -2661,20 +2518,21 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, ret = btrfs_make_block_group(trans, extent_root, 0, type, BTRFS_FIRST_CHUNK_TREE_OBJECTID, - start, *num_bytes); + start, num_bytes); BUG_ON(ret); - index = 0; - while (index < map->num_stripes) { - device = map->stripes[index].dev; - dev_offset = map->stripes[index].physical; + for (i = 0; i < map->num_stripes; ++i) { + struct btrfs_device *device; + u64 dev_offset; + + device = map->stripes[i].dev; + dev_offset = map->stripes[i].physical; ret = btrfs_alloc_dev_extent(trans, device, info->chunk_root->root_key.objectid, BTRFS_FIRST_CHUNK_TREE_OBJECTID, - start, dev_offset, calc_size); + start, dev_offset, stripe_size); BUG_ON(ret); - index++; } kfree(devices_info); @@ -2749,6 +2607,7 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans, item_size); BUG_ON(ret); } + kfree(chunk); return 0; } @@ -2880,7 +2739,7 @@ int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset) void btrfs_mapping_init(struct btrfs_mapping_tree *tree) { - extent_map_tree_init(&tree->map_tree, GFP_NOFS); + extent_map_tree_init(&tree->map_tree); } void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree) @@ -2946,14 +2805,17 @@ static int find_live_mirror(struct map_lookup *map, int first, int num, static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 logical, u64 *length, struct btrfs_multi_bio **multi_ret, - int mirror_num, struct page *unplug_page) + int mirror_num) { struct extent_map *em; struct map_lookup *map; struct extent_map_tree *em_tree = &map_tree->map_tree; u64 offset; u64 stripe_offset; + u64 stripe_end_offset; u64 stripe_nr; + u64 stripe_nr_orig; + u64 stripe_nr_end; int stripes_allocated = 8; int stripes_required = 1; int stripe_index; @@ -2962,7 +2824,7 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, int max_errors = 0; struct btrfs_multi_bio *multi = NULL; - if (multi_ret && !(rw & REQ_WRITE)) + if (multi_ret && !(rw & (REQ_WRITE | REQ_DISCARD))) stripes_allocated = 1; again: if (multi_ret) { @@ -2978,11 +2840,6 @@ again: em = lookup_extent_mapping(em_tree, logical, *length); read_unlock(&em_tree->lock); - if (!em && unplug_page) { - kfree(multi); - return 0; - } - if (!em) { printk(KERN_CRIT "unable to find logical %llu len %llu\n", (unsigned long long)logical, @@ -3008,7 +2865,15 @@ again: max_errors = 1; } } - if (multi_ret && (rw & REQ_WRITE) && + if (rw & REQ_DISCARD) { + if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | + BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_DUP | + BTRFS_BLOCK_GROUP_RAID10)) { + stripes_required = map->num_stripes; + } + } + if (multi_ret && (rw & (REQ_WRITE | REQ_DISCARD)) && stripes_allocated < stripes_required) { stripes_allocated = map->num_stripes; free_extent_map(em); @@ -3028,23 +2893,37 @@ again: /* stripe_offset is the offset of this block in its stripe*/ stripe_offset = offset - stripe_offset; - if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_RAID10 | - BTRFS_BLOCK_GROUP_DUP)) { + if (rw & REQ_DISCARD) + *length = min_t(u64, em->len - offset, *length); + else if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | + BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10 | + BTRFS_BLOCK_GROUP_DUP)) { /* we limit the length of each bio to what fits in a stripe */ *length = min_t(u64, em->len - offset, - map->stripe_len - stripe_offset); + map->stripe_len - stripe_offset); } else { *length = em->len - offset; } - if (!multi_ret && !unplug_page) + if (!multi_ret) goto out; num_stripes = 1; stripe_index = 0; - if (map->type & BTRFS_BLOCK_GROUP_RAID1) { - if (unplug_page || (rw & REQ_WRITE)) + stripe_nr_orig = stripe_nr; + stripe_nr_end = (offset + *length + map->stripe_len - 1) & + (~(map->stripe_len - 1)); + do_div(stripe_nr_end, map->stripe_len); + stripe_end_offset = stripe_nr_end * map->stripe_len - + (offset + *length); + if (map->type & BTRFS_BLOCK_GROUP_RAID0) { + if (rw & REQ_DISCARD) + num_stripes = min_t(u64, map->num_stripes, + stripe_nr_end - stripe_nr_orig); + stripe_index = do_div(stripe_nr, map->num_stripes); + } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { + if (rw & (REQ_WRITE | REQ_DISCARD)) num_stripes = map->num_stripes; else if (mirror_num) stripe_index = mirror_num - 1; @@ -3055,7 +2934,7 @@ again: } } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { - if (rw & REQ_WRITE) + if (rw & (REQ_WRITE | REQ_DISCARD)) num_stripes = map->num_stripes; else if (mirror_num) stripe_index = mirror_num - 1; @@ -3066,8 +2945,12 @@ again: stripe_index = do_div(stripe_nr, factor); stripe_index *= map->sub_stripes; - if (unplug_page || (rw & REQ_WRITE)) + if (rw & REQ_WRITE) num_stripes = map->sub_stripes; + else if (rw & REQ_DISCARD) + num_stripes = min_t(u64, map->sub_stripes * + (stripe_nr_end - stripe_nr_orig), + map->num_stripes); else if (mirror_num) stripe_index += mirror_num - 1; else { @@ -3085,24 +2968,101 @@ again: } BUG_ON(stripe_index >= map->num_stripes); - for (i = 0; i < num_stripes; i++) { - if (unplug_page) { - struct btrfs_device *device; - struct backing_dev_info *bdi; - - device = map->stripes[stripe_index].dev; - if (device->bdev) { - bdi = blk_get_backing_dev_info(device->bdev); - if (bdi->unplug_io_fn) - bdi->unplug_io_fn(bdi, unplug_page); - } - } else { + if (rw & REQ_DISCARD) { + for (i = 0; i < num_stripes; i++) { multi->stripes[i].physical = map->stripes[stripe_index].physical + stripe_offset + stripe_nr * map->stripe_len; multi->stripes[i].dev = map->stripes[stripe_index].dev; + + if (map->type & BTRFS_BLOCK_GROUP_RAID0) { + u64 stripes; + u32 last_stripe = 0; + int j; + + div_u64_rem(stripe_nr_end - 1, + map->num_stripes, + &last_stripe); + + for (j = 0; j < map->num_stripes; j++) { + u32 test; + + div_u64_rem(stripe_nr_end - 1 - j, + map->num_stripes, &test); + if (test == stripe_index) + break; + } + stripes = stripe_nr_end - 1 - j; + do_div(stripes, map->num_stripes); + multi->stripes[i].length = map->stripe_len * + (stripes - stripe_nr + 1); + + if (i == 0) { + multi->stripes[i].length -= + stripe_offset; + stripe_offset = 0; + } + if (stripe_index == last_stripe) + multi->stripes[i].length -= + stripe_end_offset; + } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { + u64 stripes; + int j; + int factor = map->num_stripes / + map->sub_stripes; + u32 last_stripe = 0; + + div_u64_rem(stripe_nr_end - 1, + factor, &last_stripe); + last_stripe *= map->sub_stripes; + + for (j = 0; j < factor; j++) { + u32 test; + + div_u64_rem(stripe_nr_end - 1 - j, + factor, &test); + + if (test == + stripe_index / map->sub_stripes) + break; + } + stripes = stripe_nr_end - 1 - j; + do_div(stripes, factor); + multi->stripes[i].length = map->stripe_len * + (stripes - stripe_nr + 1); + + if (i < map->sub_stripes) { + multi->stripes[i].length -= + stripe_offset; + if (i == map->sub_stripes - 1) + stripe_offset = 0; + } + if (stripe_index >= last_stripe && + stripe_index <= (last_stripe + + map->sub_stripes - 1)) { + multi->stripes[i].length -= + stripe_end_offset; + } + } else + multi->stripes[i].length = *length; + + stripe_index++; + if (stripe_index == map->num_stripes) { + /* This could only happen for RAID0/10 */ + stripe_index = 0; + stripe_nr++; + } + } + } else { + for (i = 0; i < num_stripes; i++) { + multi->stripes[i].physical = + map->stripes[stripe_index].physical + + stripe_offset + + stripe_nr * map->stripe_len; + multi->stripes[i].dev = + map->stripes[stripe_index].dev; + stripe_index++; } - stripe_index++; } if (multi_ret) { *multi_ret = multi; @@ -3119,7 +3079,7 @@ int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, struct btrfs_multi_bio **multi_ret, int mirror_num) { return __btrfs_map_block(map_tree, rw, logical, length, multi_ret, - mirror_num, NULL); + mirror_num); } int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, @@ -3187,14 +3147,6 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, return 0; } -int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree, - u64 logical, struct page *page) -{ - u64 length = PAGE_CACHE_SIZE; - return __btrfs_map_block(map_tree, READ, logical, &length, - NULL, 0, page); -} - static void end_bio_multi_stripe(struct bio *bio, int err) { struct btrfs_multi_bio *multi = bio->bi_private; @@ -3437,7 +3389,7 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, free_extent_map(em); } - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) return -ENOMEM; num_stripes = btrfs_chunk_num_stripes(leaf, chunk); @@ -3626,15 +3578,6 @@ static int read_one_dev(struct btrfs_root *root, return ret; } -int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf) -{ - struct btrfs_dev_item *dev_item; - - dev_item = (struct btrfs_dev_item *)offsetof(struct btrfs_super_block, - dev_item); - return read_one_dev(root, buf, dev_item); -} - int btrfs_read_sys_array(struct btrfs_root *root) { struct btrfs_super_block *super_copy = &root->fs_info->super_copy; @@ -3751,7 +3694,7 @@ again: } if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) { key.objectid = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } ret = 0; diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 7fb59d45fe8c..7c12d61ae7ae 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -85,7 +85,12 @@ struct btrfs_device { /* physical drive uuid (or lvm uuid) */ u8 uuid[BTRFS_UUID_SIZE]; + /* per-device scrub information */ + struct scrub_dev *scrub_device; + struct btrfs_work work; + struct rcu_head rcu; + struct work_struct rcu_work; }; struct btrfs_fs_devices { @@ -126,6 +131,7 @@ struct btrfs_fs_devices { struct btrfs_bio_stripe { struct btrfs_device *dev; u64 physical; + u64 length; /* only used for discard mappings */ }; struct btrfs_multi_bio { @@ -143,22 +149,22 @@ struct btrfs_device_info { struct btrfs_device *dev; u64 dev_offset; u64 max_avail; + u64 total_avail; }; -/* Used to sort the devices by max_avail(descending sort) */ -int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2); +struct map_lookup { + u64 type; + int io_align; + int io_width; + int stripe_len; + int sector_size; + int num_stripes; + int sub_stripes; + struct btrfs_bio_stripe stripes[]; +}; -/* - * sort the devices by max_avail, in which max free extent size of each device - * is stored.(Descending Sort) - */ -static inline void btrfs_descending_sort_devices( - struct btrfs_device_info *devices, - size_t nr_devices) -{ - sort(devices, nr_devices, sizeof(struct btrfs_device_info), - btrfs_cmp_device_free_bytes, NULL); -} +#define map_lookup_size(n) (sizeof(struct map_lookup) + \ + (sizeof(struct btrfs_bio_stripe) * (n))) int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start, u64 end, u64 *length); @@ -184,7 +190,6 @@ void btrfs_mapping_init(struct btrfs_mapping_tree *tree); void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree); int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, int mirror_num, int async_submit); -int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf); int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, fmode_t flags, void *holder); int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, @@ -197,8 +202,6 @@ int btrfs_add_device(struct btrfs_trans_handle *trans, int btrfs_rm_device(struct btrfs_root *root, char *device_path); int btrfs_cleanup_fs_uuids(void); int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len); -int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree, - u64 logical, struct page *page); int btrfs_grow_device(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 new_size); struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, @@ -206,8 +209,6 @@ struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, int btrfs_shrink_device(struct btrfs_device *device, u64 new_size); int btrfs_init_new_device(struct btrfs_root *root, char *path); int btrfs_balance(struct btrfs_root *dev_root); -void btrfs_unlock_volumes(void); -void btrfs_lock_volumes(void); int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset); int find_free_dev_extent(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 num_bytes, diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index a5776531dc2b..f3107e4b4d56 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -44,7 +44,7 @@ ssize_t __btrfs_getxattr(struct inode *inode, const char *name, return -ENOMEM; /* lookup the xattr by name */ - di = btrfs_lookup_xattr(NULL, root, path, inode->i_ino, name, + di = btrfs_lookup_xattr(NULL, root, path, btrfs_ino(inode), name, strlen(name), 0); if (!di) { ret = -ENODATA; @@ -103,7 +103,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans, return -ENOMEM; /* first lets see if we already have this xattr */ - di = btrfs_lookup_xattr(trans, root, path, inode->i_ino, name, + di = btrfs_lookup_xattr(trans, root, path, btrfs_ino(inode), name, strlen(name), -1); if (IS_ERR(di)) { ret = PTR_ERR(di); @@ -120,13 +120,13 @@ static int do_setxattr(struct btrfs_trans_handle *trans, ret = btrfs_delete_one_dir_name(trans, root, path, di); BUG_ON(ret); - btrfs_release_path(root, path); + btrfs_release_path(path); /* if we don't have a value then we are removing the xattr */ if (!value) goto out; } else { - btrfs_release_path(root, path); + btrfs_release_path(path); if (flags & XATTR_REPLACE) { /* we couldn't find the attr to replace */ @@ -136,7 +136,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans, } /* ok we have to create a completely new xattr */ - ret = btrfs_insert_xattr_item(trans, root, path, inode->i_ino, + ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode), name, name_len, value, size); BUG_ON(ret); out: @@ -180,18 +180,17 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size) struct btrfs_path *path; struct extent_buffer *leaf; struct btrfs_dir_item *di; - int ret = 0, slot, advance; + int ret = 0, slot; size_t total_size = 0, size_left = size; unsigned long name_ptr; size_t name_len; - u32 nritems; /* * ok we want all objects associated with this id. * NOTE: we set key.offset = 0; because we want to start with the * first xattr that we find and walk forward */ - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY); key.offset = 0; @@ -204,34 +203,24 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size) ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) goto err; - advance = 0; + while (1) { leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); slot = path->slots[0]; /* this is where we start walking through the path */ - if (advance || slot >= nritems) { + if (slot >= btrfs_header_nritems(leaf)) { /* * if we've reached the last slot in this leaf we need * to go to the next leaf and reset everything */ - if (slot >= nritems-1) { - ret = btrfs_next_leaf(root, path); - if (ret) - break; - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - slot = path->slots[0]; - } else { - /* - * just walking through the slots on this leaf - */ - slot++; - path->slots[0]++; - } + ret = btrfs_next_leaf(root, path); + if (ret < 0) + goto err; + else if (ret > 0) + break; + continue; } - advance = 1; btrfs_item_key_to_cpu(leaf, &found_key, slot); @@ -242,13 +231,15 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size) break; di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item); + if (verify_dir_item(root, leaf, di)) + continue; name_len = btrfs_dir_name_len(leaf, di); total_size += name_len + 1; /* we are just looking for how big our buffer needs to be */ if (!size) - continue; + goto next; if (!buffer || (name_len + 1) > size_left) { ret = -ERANGE; @@ -261,6 +252,8 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size) size_left -= name_len + 1; buffer += name_len + 1; +next: + path->slots[0]++; } ret = total_size; @@ -370,7 +363,8 @@ int btrfs_removexattr(struct dentry *dentry, const char *name) } int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, - struct inode *inode, struct inode *dir) + struct inode *inode, struct inode *dir, + const struct qstr *qstr) { int err; size_t len; @@ -378,7 +372,8 @@ int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, char *suffix; char *name; - err = security_inode_init_security(inode, dir, &suffix, &value, &len); + err = security_inode_init_security(inode, dir, qstr, &suffix, &value, + &len); if (err) { if (err == -EOPNOTSUPP) return 0; diff --git a/fs/btrfs/xattr.h b/fs/btrfs/xattr.h index 7a43fd640bbb..b3cc8039134b 100644 --- a/fs/btrfs/xattr.h +++ b/fs/btrfs/xattr.h @@ -37,6 +37,7 @@ extern int btrfs_setxattr(struct dentry *dentry, const char *name, extern int btrfs_removexattr(struct dentry *dentry, const char *name); extern int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, - struct inode *inode, struct inode *dir); + struct inode *inode, struct inode *dir, + const struct qstr *qstr); #endif /* __XATTR__ */ diff --git a/fs/btrfs/zlib.c b/fs/btrfs/zlib.c index f5ec2d44150d..faccd47c6c46 100644 --- a/fs/btrfs/zlib.c +++ b/fs/btrfs/zlib.c @@ -57,7 +57,8 @@ static struct list_head *zlib_alloc_workspace(void) if (!workspace) return ERR_PTR(-ENOMEM); - workspace->def_strm.workspace = vmalloc(zlib_deflate_workspacesize()); + workspace->def_strm.workspace = vmalloc(zlib_deflate_workspacesize( + MAX_WBITS, MAX_MEM_LEVEL)); workspace->inf_strm.workspace = vmalloc(zlib_inflate_workspacesize()); workspace->buf = kmalloc(PAGE_CACHE_SIZE, GFP_NOFS); if (!workspace->def_strm.workspace || |