diff options
Diffstat (limited to 'fs/btrfs/extent-tree.c')
| -rw-r--r-- | fs/btrfs/extent-tree.c | 934 |
1 files changed, 485 insertions, 449 deletions
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index dcd2e798767e..375f8c728d91 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -16,6 +16,7 @@ * Boston, MA 021110-1307, USA. */ #include <linux/sched.h> +#include <linux/sched/signal.h> #include <linux/pagemap.h> #include <linux/writeback.h> #include <linux/blkdev.h> @@ -96,10 +97,11 @@ static int btrfs_free_reserved_bytes(struct btrfs_block_group_cache *cache, u64 num_bytes, int delalloc); static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, u64 num_bytes); -static int __reserve_metadata_bytes(struct btrfs_root *root, +static int __reserve_metadata_bytes(struct btrfs_fs_info *fs_info, struct btrfs_space_info *space_info, u64 orig_bytes, - enum btrfs_reserve_flush_enum flush); + enum btrfs_reserve_flush_enum flush, + bool system_chunk); static void space_info_add_new_bytes(struct btrfs_fs_info *fs_info, struct btrfs_space_info *space_info, u64 num_bytes); @@ -130,6 +132,16 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache) if (atomic_dec_and_test(&cache->count)) { WARN_ON(cache->pinned > 0); WARN_ON(cache->reserved > 0); + + /* + * If not empty, someone is still holding mutex of + * full_stripe_lock, which can only be released by caller. + * And it will definitely cause use-after-free when caller + * tries to release full stripe lock. + * + * No better way to resolve, but only to warn. + */ + WARN_ON(!RB_EMPTY_ROOT(&cache->full_stripe_locks_root.root)); kfree(cache->free_space_ctl); kfree(cache); } @@ -315,14 +327,14 @@ get_caching_control(struct btrfs_block_group_cache *cache) } ctl = cache->caching_ctl; - atomic_inc(&ctl->count); + refcount_inc(&ctl->count); spin_unlock(&cache->lock); return ctl; } static void put_caching_control(struct btrfs_caching_control *ctl) { - if (atomic_dec_and_test(&ctl->count)) + if (refcount_dec_and_test(&ctl->count)) kfree(ctl); } @@ -598,7 +610,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, init_waitqueue_head(&caching_ctl->wait); caching_ctl->block_group = cache; caching_ctl->progress = cache->key.objectid; - atomic_set(&caching_ctl->count, 1); + refcount_set(&caching_ctl->count, 1); btrfs_init_work(&caching_ctl->work, btrfs_cache_helper, caching_thread, NULL, NULL); @@ -619,7 +631,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, struct btrfs_caching_control *ctl; ctl = cache->caching_ctl; - atomic_inc(&ctl->count); + refcount_inc(&ctl->count); prepare_to_wait(&ctl->wait, &wait, TASK_UNINTERRUPTIBLE); spin_unlock(&cache->lock); @@ -706,7 +718,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, } down_write(&fs_info->commit_root_sem); - atomic_inc(&caching_ctl->count); + refcount_inc(&caching_ctl->count); list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups); up_write(&fs_info->commit_root_sem); @@ -755,6 +767,26 @@ static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info, return NULL; } +static void add_pinned_bytes(struct btrfs_fs_info *fs_info, s64 num_bytes, + u64 owner, u64 root_objectid) +{ + struct btrfs_space_info *space_info; + u64 flags; + + if (owner < BTRFS_FIRST_FREE_OBJECTID) { + if (root_objectid == BTRFS_CHUNK_TREE_OBJECTID) + flags = BTRFS_BLOCK_GROUP_SYSTEM; + else + flags = BTRFS_BLOCK_GROUP_METADATA; + } else { + flags = BTRFS_BLOCK_GROUP_DATA; + } + + space_info = __find_space_info(fs_info, flags); + ASSERT(space_info); + percpu_counter_add(&space_info->total_bytes_pinned, num_bytes); +} + /* * after adding space to the filesystem, we need to clear the full flags * on all the space infos. @@ -888,10 +920,10 @@ search_again: delayed_refs = &trans->transaction->delayed_refs; spin_lock(&delayed_refs->lock); - head = btrfs_find_delayed_ref_head(trans, bytenr); + head = btrfs_find_delayed_ref_head(delayed_refs, bytenr); if (head) { if (!mutex_trylock(&head->mutex)) { - atomic_inc(&head->node.refs); + refcount_inc(&head->node.refs); spin_unlock(&delayed_refs->lock); btrfs_release_path(path); @@ -1035,10 +1067,11 @@ out_free: #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 static int convert_extent_item_v0(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, u64 owner, u32 extra_size) { + struct btrfs_root *root = fs_info->extent_root; struct btrfs_extent_item *item; struct btrfs_extent_item_v0 *ei0; struct btrfs_extent_ref_v0 *ref0; @@ -1092,7 +1125,7 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans, return ret; BUG_ON(ret); /* Corruption */ - btrfs_extend_item(root->fs_info, path, new_size); + btrfs_extend_item(fs_info, path, new_size); leaf = path->nodes[0]; item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); @@ -1151,12 +1184,13 @@ static int match_extent_data_ref(struct extent_buffer *leaf, } static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, u64 bytenr, u64 parent, u64 root_objectid, u64 owner, u64 offset) { + struct btrfs_root *root = fs_info->extent_root; struct btrfs_key key; struct btrfs_extent_data_ref *ref; struct extent_buffer *leaf; @@ -1238,12 +1272,13 @@ fail: } static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, u64 bytenr, u64 parent, u64 root_objectid, u64 owner, u64 offset, int refs_to_add) { + struct btrfs_root *root = fs_info->extent_root; struct btrfs_key key; struct extent_buffer *leaf; u32 size; @@ -1317,7 +1352,7 @@ fail: } static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, int refs_to_drop, int *last_ref) { @@ -1354,7 +1389,7 @@ static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans, num_refs -= refs_to_drop; if (num_refs == 0) { - ret = btrfs_del_item(trans, root, path); + ret = btrfs_del_item(trans, fs_info->extent_root, path); *last_ref = 1; } else { if (key.type == BTRFS_EXTENT_DATA_REF_KEY) @@ -1416,11 +1451,12 @@ static noinline u32 extent_data_ref_count(struct btrfs_path *path, } static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, u64 bytenr, u64 parent, u64 root_objectid) { + struct btrfs_root *root = fs_info->extent_root; struct btrfs_key key; int ret; @@ -1449,7 +1485,7 @@ static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans, } static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, u64 bytenr, u64 parent, u64 root_objectid) @@ -1466,7 +1502,8 @@ static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans, key.offset = root_objectid; } - ret = btrfs_insert_empty_item(trans, root, path, &key, 0); + ret = btrfs_insert_empty_item(trans, fs_info->extent_root, + path, &key, 0); btrfs_release_path(path); return ret; } @@ -1524,14 +1561,14 @@ static int find_next_key(struct btrfs_path *path, int level, */ static noinline_for_stack int lookup_inline_extent_backref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, struct btrfs_extent_inline_ref **ref_ret, u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, u64 owner, u64 offset, int insert) { - struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_root *root = fs_info->extent_root; struct btrfs_key key; struct extent_buffer *leaf; struct btrfs_extent_item *ei; @@ -1614,7 +1651,7 @@ again: err = -ENOENT; goto out; } - ret = convert_extent_item_v0(trans, root, path, owner, + ret = convert_extent_item_v0(trans, fs_info, path, owner, extra_size); if (ret < 0) { err = ret; @@ -1716,7 +1753,7 @@ out: * helper to add new inline back ref */ static noinline_for_stack -void setup_inline_extent_backref(struct btrfs_root *root, +void setup_inline_extent_backref(struct btrfs_fs_info *fs_info, struct btrfs_path *path, struct btrfs_extent_inline_ref *iref, u64 parent, u64 root_objectid, @@ -1739,7 +1776,7 @@ void setup_inline_extent_backref(struct btrfs_root *root, type = extent_ref_type(parent, owner); size = btrfs_extent_inline_ref_size(type); - btrfs_extend_item(root->fs_info, path, size); + btrfs_extend_item(fs_info, path, size); ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); refs = btrfs_extent_refs(leaf, ei); @@ -1777,7 +1814,7 @@ void setup_inline_extent_backref(struct btrfs_root *root, } static int lookup_extent_backref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, struct btrfs_extent_inline_ref **ref_ret, u64 bytenr, u64 num_bytes, u64 parent, @@ -1785,7 +1822,7 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans, { int ret; - ret = lookup_inline_extent_backref(trans, root, path, ref_ret, + ret = lookup_inline_extent_backref(trans, fs_info, path, ref_ret, bytenr, num_bytes, parent, root_objectid, owner, offset, 0); if (ret != -ENOENT) @@ -1795,11 +1832,12 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans, *ref_ret = NULL; if (owner < BTRFS_FIRST_FREE_OBJECTID) { - ret = lookup_tree_block_ref(trans, root, path, bytenr, parent, - root_objectid); + ret = lookup_tree_block_ref(trans, fs_info, path, bytenr, + parent, root_objectid); } else { - ret = lookup_extent_data_ref(trans, root, path, bytenr, parent, - root_objectid, owner, offset); + ret = lookup_extent_data_ref(trans, fs_info, path, bytenr, + parent, root_objectid, owner, + offset); } return ret; } @@ -1808,7 +1846,7 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans, * helper to update/remove inline back ref */ static noinline_for_stack -void update_inline_extent_backref(struct btrfs_root *root, +void update_inline_extent_backref(struct btrfs_fs_info *fs_info, struct btrfs_path *path, struct btrfs_extent_inline_ref *iref, int refs_to_mod, @@ -1866,14 +1904,14 @@ void update_inline_extent_backref(struct btrfs_root *root, memmove_extent_buffer(leaf, ptr, ptr + size, end - ptr - size); item_size -= size; - btrfs_truncate_item(root->fs_info, path, item_size, 1); + btrfs_truncate_item(fs_info, path, item_size, 1); } btrfs_mark_buffer_dirty(leaf); } static noinline_for_stack int insert_inline_extent_backref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, u64 owner, @@ -1883,15 +1921,15 @@ int insert_inline_extent_backref(struct btrfs_trans_handle *trans, struct btrfs_extent_inline_ref *iref; int ret; - ret = lookup_inline_extent_backref(trans, root, path, &iref, + ret = lookup_inline_extent_backref(trans, fs_info, path, &iref, bytenr, num_bytes, parent, root_objectid, owner, offset, 1); if (ret == 0) { BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID); - update_inline_extent_backref(root, path, iref, + update_inline_extent_backref(fs_info, path, iref, refs_to_add, extent_op, NULL); } else if (ret == -ENOENT) { - setup_inline_extent_backref(root, path, iref, parent, + setup_inline_extent_backref(fs_info, path, iref, parent, root_objectid, owner, offset, refs_to_add, extent_op); ret = 0; @@ -1900,7 +1938,7 @@ int insert_inline_extent_backref(struct btrfs_trans_handle *trans, } static int insert_extent_backref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, u64 bytenr, u64 parent, u64 root_objectid, u64 owner, u64 offset, int refs_to_add) @@ -1908,10 +1946,10 @@ static int insert_extent_backref(struct btrfs_trans_handle *trans, int ret; if (owner < BTRFS_FIRST_FREE_OBJECTID) { BUG_ON(refs_to_add != 1); - ret = insert_tree_block_ref(trans, root, path, bytenr, + ret = insert_tree_block_ref(trans, fs_info, path, bytenr, parent, root_objectid); } else { - ret = insert_extent_data_ref(trans, root, path, bytenr, + ret = insert_extent_data_ref(trans, fs_info, path, bytenr, parent, root_objectid, owner, offset, refs_to_add); } @@ -1919,7 +1957,7 @@ static int insert_extent_backref(struct btrfs_trans_handle *trans, } static int remove_extent_backref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, + struct btrfs_fs_info *fs_info, struct btrfs_path *path, struct btrfs_extent_inline_ref *iref, int refs_to_drop, int is_data, int *last_ref) @@ -1928,14 +1966,14 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, BUG_ON(!is_data && refs_to_drop != 1); if (iref) { - update_inline_extent_backref(root, path, iref, + update_inline_extent_backref(fs_info, path, iref, -refs_to_drop, NULL, last_ref); } else if (is_data) { - ret = remove_extent_data_ref(trans, root, path, refs_to_drop, + ret = remove_extent_data_ref(trans, fs_info, path, refs_to_drop, last_ref); } else { *last_ref = 1; - ret = btrfs_del_item(trans, root, path); + ret = btrfs_del_item(trans, fs_info->extent_root, path); } return ret; } @@ -2075,6 +2113,7 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, u64 owner, u64 offset) { + int old_ref_mod, new_ref_mod; int ret; BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID && @@ -2082,15 +2121,21 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, if (owner < BTRFS_FIRST_FREE_OBJECTID) { ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, - num_bytes, - parent, root_objectid, (int)owner, - BTRFS_ADD_DELAYED_REF, NULL); + num_bytes, parent, + root_objectid, (int)owner, + BTRFS_ADD_DELAYED_REF, NULL, + &old_ref_mod, &new_ref_mod); } else { ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, - num_bytes, parent, root_objectid, - owner, offset, 0, - BTRFS_ADD_DELAYED_REF, NULL); + num_bytes, parent, + root_objectid, owner, offset, + 0, BTRFS_ADD_DELAYED_REF, + &old_ref_mod, &new_ref_mod); } + + if (ret == 0 && old_ref_mod < 0 && new_ref_mod >= 0) + add_pinned_bytes(fs_info, -num_bytes, owner, root_objectid); + return ret; } @@ -2117,9 +2162,9 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, path->reada = READA_FORWARD; path->leave_spinning = 1; /* this will setup the path even if it fails to insert the back ref */ - ret = insert_inline_extent_backref(trans, fs_info->extent_root, path, - bytenr, num_bytes, parent, - root_objectid, owner, offset, + ret = insert_inline_extent_backref(trans, fs_info, path, bytenr, + num_bytes, parent, root_objectid, + owner, offset, refs_to_add, extent_op); if ((ret < 0 && ret != -EAGAIN) || !ret) goto out; @@ -2143,9 +2188,8 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, path->reada = READA_FORWARD; path->leave_spinning = 1; /* now insert the actual backref */ - ret = insert_extent_backref(trans, fs_info->extent_root, - path, bytenr, parent, root_objectid, - owner, offset, refs_to_add); + ret = insert_extent_backref(trans, fs_info, path, bytenr, parent, + root_objectid, owner, offset, refs_to_add); if (ret) btrfs_abort_transaction(trans, ret); out: @@ -2290,8 +2334,7 @@ again: item_size = btrfs_item_size_nr(leaf, path->slots[0]); #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 if (item_size < sizeof(*ei)) { - ret = convert_extent_item_v0(trans, fs_info->extent_root, - path, (u64)-1, 0); + ret = convert_extent_item_v0(trans, fs_info, path, (u64)-1, 0); if (ret < 0) { err = ret; goto out; @@ -2396,6 +2439,16 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans, head = btrfs_delayed_node_to_head(node); trace_run_delayed_ref_head(fs_info, node, head, node->action); + if (head->total_ref_mod < 0) { + struct btrfs_block_group_cache *cache; + + cache = btrfs_lookup_block_group(fs_info, node->bytenr); + ASSERT(cache); + percpu_counter_add(&cache->space_info->total_bytes_pinned, + -node->num_bytes); + btrfs_put_block_group(cache); + } + if (insert_reserved) { btrfs_pin_extent(fs_info, node->bytenr, node->num_bytes, 1); @@ -2975,7 +3028,7 @@ again: struct btrfs_delayed_ref_node *ref; ref = &head->node; - atomic_inc(&ref->refs); + refcount_inc(&ref->refs); spin_unlock(&delayed_refs->lock); /* @@ -2998,7 +3051,6 @@ again: goto again; } out: - assert_qgroups_uptodate(trans); trans->can_flush_pending_bgs = can_flush_pending_bgs; return 0; } @@ -3028,8 +3080,7 @@ int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans, return ret; } -static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +static noinline int check_delayed_ref(struct btrfs_root *root, struct btrfs_path *path, u64 objectid, u64 offset, u64 bytenr) { @@ -3037,18 +3088,23 @@ static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_node *ref; struct btrfs_delayed_data_ref *data_ref; struct btrfs_delayed_ref_root *delayed_refs; + struct btrfs_transaction *cur_trans; int ret = 0; - delayed_refs = &trans->transaction->delayed_refs; + cur_trans = root->fs_info->running_transaction; + if (!cur_trans) + return 0; + + delayed_refs = &cur_trans->delayed_refs; spin_lock(&delayed_refs->lock); - head = btrfs_find_delayed_ref_head(trans, bytenr); + head = btrfs_find_delayed_ref_head(delayed_refs, bytenr); if (!head) { spin_unlock(&delayed_refs->lock); return 0; } if (!mutex_trylock(&head->mutex)) { - atomic_inc(&head->node.refs); + refcount_inc(&head->node.refs); spin_unlock(&delayed_refs->lock); btrfs_release_path(path); @@ -3090,8 +3146,7 @@ static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, return ret; } -static noinline int check_committed_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +static noinline int check_committed_ref(struct btrfs_root *root, struct btrfs_path *path, u64 objectid, u64 offset, u64 bytenr) { @@ -3162,9 +3217,8 @@ out: return ret; } -int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - u64 objectid, u64 offset, u64 bytenr) +int btrfs_cross_ref_exist(struct btrfs_root *root, u64 objectid, u64 offset, + u64 bytenr) { struct btrfs_path *path; int ret; @@ -3175,12 +3229,12 @@ int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, return -ENOENT; do { - ret = check_committed_ref(trans, root, path, objectid, + ret = check_committed_ref(root, path, objectid, offset, bytenr); if (ret && ret != -ENOENT) goto out; - ret2 = check_delayed_ref(trans, root, path, objectid, + ret2 = check_delayed_ref(root, path, objectid, offset, bytenr); } while (ret2 == -EAGAIN); @@ -3348,6 +3402,7 @@ static int cache_save_setup(struct btrfs_block_group_cache *block_group, struct btrfs_fs_info *fs_info = block_group->fs_info; struct btrfs_root *root = fs_info->tree_root; struct inode *inode = NULL; + struct extent_changeset *data_reserved = NULL; u64 alloc_hint = 0; int dcs = BTRFS_DC_ERROR; u64 num_pages = 0; @@ -3368,7 +3423,7 @@ static int cache_save_setup(struct btrfs_block_group_cache *block_group, if (trans->aborted) return 0; again: - inode = lookup_free_space_inode(root, block_group, path); + inode = lookup_free_space_inode(fs_info, block_group, path); if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { ret = PTR_ERR(inode); btrfs_release_path(path); @@ -3382,7 +3437,8 @@ again: if (block_group->ro) goto out_free; - ret = create_free_space_inode(root, trans, block_group, path); + ret = create_free_space_inode(fs_info, trans, block_group, + path); if (ret) goto out_free; goto again; @@ -3424,7 +3480,7 @@ again: if (ret) goto out_put; - ret = btrfs_truncate_free_space_cache(root, trans, NULL, inode); + ret = btrfs_truncate_free_space_cache(trans, NULL, inode); if (ret) goto out_put; } @@ -3435,7 +3491,8 @@ again: /* * don't bother trying to write stuff out _if_ * a) we're not cached, - * b) we're with nospace_cache mount option. + * b) we're with nospace_cache mount option, + * c) we're with v2 space_cache (FREE_SPACE_TREE). */ dcs = BTRFS_DC_WRITTEN; spin_unlock(&block_group->lock); @@ -3465,7 +3522,7 @@ again: num_pages *= 16; num_pages *= PAGE_SIZE; - ret = btrfs_check_data_free_space(inode, 0, num_pages); + ret = btrfs_check_data_free_space(inode, &data_reserved, 0, num_pages); if (ret) goto out_put; @@ -3496,6 +3553,7 @@ out: block_group->disk_cache_state = dcs; spin_unlock(&block_group->lock); + extent_changeset_free(data_reserved); return ret; } @@ -3906,87 +3964,83 @@ static const char *alloc_name(u64 flags) }; } -static int update_space_info(struct btrfs_fs_info *info, u64 flags, - u64 total_bytes, u64 bytes_used, - u64 bytes_readonly, - struct btrfs_space_info **space_info) +static int create_space_info(struct btrfs_fs_info *info, u64 flags, + struct btrfs_space_info **new) { - struct btrfs_space_info *found; + + struct btrfs_space_info *space_info; int i; - int factor; int ret; - if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_RAID10)) - factor = 2; - else - factor = 1; - - found = __find_space_info(info, flags); - if (found) { - spin_lock(&found->lock); - found->total_bytes += total_bytes; - found->disk_total += total_bytes * factor; - found->bytes_used += bytes_used; - found->disk_used += bytes_used * factor; - found->bytes_readonly += bytes_readonly; - if (total_bytes > 0) - found->full = 0; - space_info_add_new_bytes(info, found, total_bytes - - bytes_used - bytes_readonly); - spin_unlock(&found->lock); - *space_info = found; - return 0; - } - found = kzalloc(sizeof(*found), GFP_NOFS); - if (!found) + space_info = kzalloc(sizeof(*space_info), GFP_NOFS); + if (!space_info) return -ENOMEM; - ret = percpu_counter_init(&found->total_bytes_pinned, 0, GFP_KERNEL); + ret = percpu_counter_init(&space_info->total_bytes_pinned, 0, + GFP_KERNEL); if (ret) { - kfree(found); + kfree(space_info); return ret; } for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) - INIT_LIST_HEAD(&found->block_groups[i]); - init_rwsem(&found->groups_sem); - spin_lock_init(&found->lock); - found->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK; - found->total_bytes = total_bytes; - found->disk_total = total_bytes * factor; - found->bytes_used = bytes_used; - found->disk_used = bytes_used * factor; - found->bytes_pinned = 0; - found->bytes_reserved = 0; - found->bytes_readonly = bytes_readonly; - found->bytes_may_use = 0; - found->full = 0; - found->max_extent_size = 0; - found->force_alloc = CHUNK_ALLOC_NO_FORCE; - found->chunk_alloc = 0; - found->flush = 0; - init_waitqueue_head(&found->wait); - INIT_LIST_HEAD(&found->ro_bgs); - INIT_LIST_HEAD(&found->tickets); - INIT_LIST_HEAD(&found->priority_tickets); - - ret = kobject_init_and_add(&found->kobj, &space_info_ktype, + INIT_LIST_HEAD(&space_info->block_groups[i]); + init_rwsem(&space_info->groups_sem); + spin_lock_init(&space_info->lock); + space_info->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK; + space_info->force_alloc = CHUNK_ALLOC_NO_FORCE; + init_waitqueue_head(&space_info->wait); + INIT_LIST_HEAD(&space_info->ro_bgs); + INIT_LIST_HEAD(&space_info->tickets); + INIT_LIST_HEAD(&space_info->priority_tickets); + + ret = kobject_init_and_add(&space_info->kobj, &space_info_ktype, info->space_info_kobj, "%s", - alloc_name(found->flags)); + alloc_name(space_info->flags)); if (ret) { - kfree(found); + percpu_counter_destroy(&space_info->total_bytes_pinned); + kfree(space_info); return ret; } - *space_info = found; - list_add_rcu(&found->list, &info->space_info); + *new = space_info; + list_add_rcu(&space_info->list, &info->space_info); if (flags & BTRFS_BLOCK_GROUP_DATA) - info->data_sinfo = found; + info->data_sinfo = space_info; return ret; } +static void update_space_info(struct btrfs_fs_info *info, u64 flags, + u64 total_bytes, u64 bytes_used, + u64 bytes_readonly, + struct btrfs_space_info **space_info) +{ + struct btrfs_space_info *found; + int factor; + + if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10)) + factor = 2; + else + factor = 1; + + found = __find_space_info(info, flags); + ASSERT(found); + spin_lock(&found->lock); + found->total_bytes += total_bytes; + found->disk_total += total_bytes * factor; + found->bytes_used += bytes_used; + found->disk_used += bytes_used * factor; + found->bytes_readonly += bytes_readonly; + if (total_bytes > 0) + found->full = 0; + space_info_add_new_bytes(info, found, total_bytes - + bytes_used - bytes_readonly); + spin_unlock(&found->lock); + *space_info = found; +} + static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags) { u64 extra_flags = chunk_to_extended(flags) & @@ -4102,7 +4156,7 @@ static u64 get_alloc_profile(struct btrfs_fs_info *fs_info, u64 orig_flags) return btrfs_reduce_alloc_profile(fs_info, flags); } -u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data) +static u64 get_alloc_profile_by_root(struct btrfs_root *root, int data) { struct btrfs_fs_info *fs_info = root->fs_info; u64 flags; @@ -4119,10 +4173,34 @@ u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data) return ret; } -int btrfs_alloc_data_chunk_ondemand(struct inode *inode, u64 bytes) +u64 btrfs_data_alloc_profile(struct btrfs_fs_info *fs_info) +{ + return get_alloc_profile(fs_info, BTRFS_BLOCK_GROUP_DATA); +} + +u64 btrfs_metadata_alloc_profile(struct btrfs_fs_info *fs_info) +{ + return get_alloc_profile(fs_info, BTRFS_BLOCK_GROUP_METADATA); +} + +u64 btrfs_system_alloc_profile(struct btrfs_fs_info *fs_info) +{ + return get_alloc_profile(fs_info, BTRFS_BLOCK_GROUP_SYSTEM); +} + +static u64 btrfs_space_info_used(struct btrfs_space_info *s_info, + bool may_use_included) +{ + ASSERT(s_info); + return s_info->bytes_used + s_info->bytes_reserved + + s_info->bytes_pinned + s_info->bytes_readonly + + (may_use_included ? s_info->bytes_may_use : 0); +} + +int btrfs_alloc_data_chunk_ondemand(struct btrfs_inode *inode, u64 bytes) { struct btrfs_space_info *data_sinfo; - struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_root *root = inode->root; struct btrfs_fs_info *fs_info = root->fs_info; u64 used; int ret = 0; @@ -4144,9 +4222,7 @@ int btrfs_alloc_data_chunk_ondemand(struct inode *inode, u64 bytes) again: /* make sure we have enough space to handle the data first */ spin_lock(&data_sinfo->lock); - used = data_sinfo->bytes_used + data_sinfo->bytes_reserved + - data_sinfo->bytes_pinned + data_sinfo->bytes_readonly + - data_sinfo->bytes_may_use; + used = btrfs_space_info_used(data_sinfo, true); if (used + bytes > data_sinfo->total_bytes) { struct btrfs_trans_handle *trans; @@ -4161,7 +4237,7 @@ again: data_sinfo->force_alloc = CHUNK_ALLOC_FORCE; spin_unlock(&data_sinfo->lock); alloc: - alloc_target = btrfs_get_alloc_profile(root, 1); + alloc_target = btrfs_data_alloc_profile(fs_info); /* * It is ugly that we don't call nolock join * transaction for the free space inode case here. @@ -4212,7 +4288,7 @@ commit_trans: if (need_commit > 0) { btrfs_start_delalloc_roots(fs_info, 0, -1); - btrfs_wait_ordered_roots(fs_info, -1, 0, + btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1); } @@ -4252,12 +4328,8 @@ commit_trans: return ret; } -/* - * New check_data_free_space() with ability for precious data reservation - * Will replace old btrfs_check_data_free_space(), but for patch split, - * add a new function first and then replace it. - */ -int btrfs_check_data_free_space(struct inode *inode, u64 start, u64 len) +int btrfs_check_data_free_space(struct inode *inode, + struct extent_changeset **reserved, u64 start, u64 len) { struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); int ret; @@ -4267,14 +4339,16 @@ int btrfs_check_data_free_space(struct inode *inode, u64 start, u64 len) round_down(start, fs_info->sectorsize); start = round_down(start, fs_info->sectorsize); - ret = btrfs_alloc_data_chunk_ondemand(inode, len); + ret = btrfs_alloc_data_chunk_ondemand(BTRFS_I(inode), len); if (ret < 0) return ret; /* Use new btrfs_qgroup_reserve_data to reserve precious data space. */ - ret = btrfs_qgroup_reserve_data(inode, start, len); - if (ret) + ret = btrfs_qgroup_reserve_data(inode, reserved, start, len); + if (ret < 0) btrfs_free_reserved_data_space_noquota(inode, start, len); + else + ret = 0; return ret; } @@ -4315,7 +4389,8 @@ void btrfs_free_reserved_data_space_noquota(struct inode *inode, u64 start, * This one will handle the per-inode data rsv map for accurate reserved * space framework. */ -void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len) +void btrfs_free_reserved_data_space(struct inode *inode, + struct extent_changeset *reserved, u64 start, u64 len) { struct btrfs_root *root = BTRFS_I(inode)->root; @@ -4325,7 +4400,7 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len) start = round_down(start, root->fs_info->sectorsize); btrfs_free_reserved_data_space_noquota(inode, start, len); - btrfs_qgroup_free_data(inode, start, len); + btrfs_qgroup_free_data(inode, reserved, start, len); } static void force_metadata_allocation(struct btrfs_fs_info *info) @@ -4421,9 +4496,7 @@ void check_system_chunk(struct btrfs_trans_handle *trans, info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM); spin_lock(&info->lock); - left = info->total_bytes - info->bytes_used - info->bytes_pinned - - info->bytes_reserved - info->bytes_readonly - - info->bytes_may_use; + left = info->total_bytes - btrfs_space_info_used(info, true); spin_unlock(&info->lock); num_devs = get_profile_num_devs(fs_info, type); @@ -4439,9 +4512,8 @@ void check_system_chunk(struct btrfs_trans_handle *trans, } if (left < thresh) { - u64 flags; + u64 flags = btrfs_system_alloc_profile(fs_info); - flags = btrfs_get_alloc_profile(fs_info->chunk_root, 0); /* * Ignore failure to create system chunk. We might end up not * needing it, as we might not need to COW all nodes/leafs from @@ -4482,10 +4554,10 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, space_info = __find_space_info(fs_info, flags); if (!space_info) { - ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); - BUG_ON(ret); /* -ENOMEM */ + ret = create_space_info(fs_info, flags, &space_info); + if (ret) + return ret; } - BUG_ON(!space_info); /* Logic error */ again: spin_lock(&space_info->lock); @@ -4590,11 +4662,11 @@ out: return ret; } -static int can_overcommit(struct btrfs_root *root, +static int can_overcommit(struct btrfs_fs_info *fs_info, struct btrfs_space_info *space_info, u64 bytes, - enum btrfs_reserve_flush_enum flush) + enum btrfs_reserve_flush_enum flush, + bool system_chunk) { - struct btrfs_fs_info *fs_info = root->fs_info; struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; u64 profile; u64 space_size; @@ -4605,9 +4677,12 @@ static int can_overcommit(struct btrfs_root *root, if (space_info->flags & BTRFS_BLOCK_GROUP_DATA) return 0; - profile = btrfs_get_alloc_profile(root, 0); - used = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly; + if (system_chunk) + profile = btrfs_system_alloc_profile(fs_info); + else + profile = btrfs_metadata_alloc_profile(fs_info); + + used = btrfs_space_info_used(space_info, false); /* * We only want to allow over committing if we have lots of actual space @@ -4623,9 +4698,7 @@ static int can_overcommit(struct btrfs_root *root, used += space_info->bytes_may_use; - spin_lock(&fs_info->free_chunk_lock); - avail = fs_info->free_chunk_space; - spin_unlock(&fs_info->free_chunk_lock); + avail = atomic64_read(&fs_info->free_chunk_space); /* * If we have dup, raid1 or raid10 then only half of the free @@ -4675,14 +4748,14 @@ static void btrfs_writeback_inodes_sb_nr(struct btrfs_fs_info *fs_info, } } -static inline int calc_reclaim_items_nr(struct btrfs_fs_info *fs_info, +static inline u64 calc_reclaim_items_nr(struct btrfs_fs_info *fs_info, u64 to_reclaim) { u64 bytes; - int nr; + u64 nr; bytes = btrfs_calc_trans_metadata_size(fs_info, 1); - nr = (int)div64_u64(to_reclaim, bytes); + nr = div64_u64(to_reclaim, bytes); if (!nr) nr = 1; return nr; @@ -4693,24 +4766,23 @@ static inline int calc_reclaim_items_nr(struct btrfs_fs_info *fs_info, /* * shrink metadata reservation for delalloc */ -static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, - bool wait_ordered) +static void shrink_delalloc(struct btrfs_fs_info *fs_info, u64 to_reclaim, + u64 orig, bool wait_ordered) { - struct btrfs_fs_info *fs_info = root->fs_info; struct btrfs_block_rsv *block_rsv; struct btrfs_space_info *space_info; struct btrfs_trans_handle *trans; u64 delalloc_bytes; u64 max_reclaim; + u64 items; long time_left; unsigned long nr_pages; int loops; - int items; enum btrfs_reserve_flush_enum flush; /* Calc the number of the pages we need flush for space reservation */ items = calc_reclaim_items_nr(fs_info, to_reclaim); - to_reclaim = (u64)items * EXTENT_SIZE_PER_ITEM; + to_reclaim = items * EXTENT_SIZE_PER_ITEM; trans = (struct btrfs_trans_handle *)current->journal_info; block_rsv = &fs_info->delalloc_block_rsv; @@ -4753,7 +4825,7 @@ skip_async: else flush = BTRFS_RESERVE_NO_FLUSH; spin_lock(&space_info->lock); - if (can_overcommit(root, space_info, orig, flush)) { + if (can_overcommit(fs_info, space_info, orig, flush, false)) { spin_unlock(&space_info->lock); break; } @@ -4787,11 +4859,10 @@ skip_async: * get us somewhere and then commit the transaction if it does. Otherwise it * will return -ENOSPC. */ -static int may_commit_transaction(struct btrfs_root *root, +static int may_commit_transaction(struct btrfs_fs_info *fs_info, struct btrfs_space_info *space_info, u64 bytes, int force) { - struct btrfs_fs_info *fs_info = root->fs_info; struct btrfs_block_rsv *delayed_rsv = &fs_info->delayed_block_rsv; struct btrfs_trans_handle *trans; @@ -4816,14 +4887,14 @@ static int may_commit_transaction(struct btrfs_root *root, spin_lock(&delayed_rsv->lock); if (percpu_counter_compare(&space_info->total_bytes_pinned, - bytes - delayed_rsv->size) >= 0) { + bytes - delayed_rsv->size) < 0) { spin_unlock(&delayed_rsv->lock); return -ENOSPC; } spin_unlock(&delayed_rsv->lock); commit: - trans = btrfs_join_transaction(root); + trans = btrfs_join_transaction(fs_info->extent_root); if (IS_ERR(trans)) return -ENOSPC; @@ -4837,11 +4908,11 @@ struct reserve_ticket { wait_queue_head_t wait; }; -static int flush_space(struct btrfs_root *root, +static int flush_space(struct btrfs_fs_info *fs_info, struct btrfs_space_info *space_info, u64 num_bytes, u64 orig_bytes, int state) { - struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_root *root = fs_info->extent_root; struct btrfs_trans_handle *trans; int nr; int ret = 0; @@ -4864,7 +4935,7 @@ static int flush_space(struct btrfs_root *root, break; case FLUSH_DELALLOC: case FLUSH_DELALLOC_WAIT: - shrink_delalloc(root, num_bytes * 2, orig_bytes, + shrink_delalloc(fs_info, num_bytes * 2, orig_bytes, state == FLUSH_DELALLOC_WAIT); break; case ALLOC_CHUNK: @@ -4874,14 +4945,15 @@ static int flush_space(struct btrfs_root *root, break; } ret = do_chunk_alloc(trans, fs_info, - btrfs_get_alloc_profile(root, 0), + btrfs_metadata_alloc_profile(fs_info), CHUNK_ALLOC_NO_FORCE); btrfs_end_transaction(trans); if (ret > 0 || ret == -ENOSPC) ret = 0; break; case COMMIT_TRANS: - ret = may_commit_transaction(root, space_info, orig_bytes, 0); + ret = may_commit_transaction(fs_info, space_info, + orig_bytes, 0); break; default: ret = -ENOSPC; @@ -4894,8 +4966,9 @@ static int flush_space(struct btrfs_root *root, } static inline u64 -btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, - struct btrfs_space_info *space_info) +btrfs_calc_reclaim_metadata_size(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + bool system_chunk) { struct reserve_ticket *ticket; u64 used; @@ -4910,14 +4983,14 @@ btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, return to_reclaim; to_reclaim = min_t(u64, num_online_cpus() * SZ_1M, SZ_16M); - if (can_overcommit(root, space_info, to_reclaim, - BTRFS_RESERVE_FLUSH_ALL)) + if (can_overcommit(fs_info, space_info, to_reclaim, + BTRFS_RESERVE_FLUSH_ALL, system_chunk)) return 0; - used = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; - if (can_overcommit(root, space_info, SZ_1M, BTRFS_RESERVE_FLUSH_ALL)) + used = btrfs_space_info_used(space_info, true); + + if (can_overcommit(fs_info, space_info, SZ_1M, + BTRFS_RESERVE_FLUSH_ALL, system_chunk)) expected = div_factor_fine(space_info->total_bytes, 95); else expected = div_factor_fine(space_info->total_bytes, 90); @@ -4931,17 +5004,18 @@ btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, return to_reclaim; } -static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, - struct btrfs_root *root, u64 used) +static inline int need_do_async_reclaim(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 used, bool system_chunk) { - struct btrfs_fs_info *fs_info = root->fs_info; u64 thresh = div_factor_fine(space_info->total_bytes, 98); /* If we're just plain full then async reclaim just slows us down. */ if ((space_info->bytes_used + space_info->bytes_reserved) >= thresh) return 0; - if (!btrfs_calc_reclaim_metadata_size(root, space_info)) + if (!btrfs_calc_reclaim_metadata_size(fs_info, space_info, + system_chunk)) return 0; return (used >= thresh && !btrfs_fs_closing(fs_info) && @@ -4978,8 +5052,8 @@ static void btrfs_async_reclaim_metadata_space(struct work_struct *work) space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); spin_lock(&space_info->lock); - to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, - space_info); + to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info, space_info, + false); if (!to_reclaim) { space_info->flush = 0; spin_unlock(&space_info->lock); @@ -4993,16 +5067,17 @@ static void btrfs_async_reclaim_metadata_space(struct work_struct *work) struct reserve_ticket *ticket; int ret; - ret = flush_space(fs_info->fs_root, space_info, to_reclaim, - to_reclaim, flush_state); + ret = flush_space(fs_info, space_info, to_reclaim, to_reclaim, + flush_state); spin_lock(&space_info->lock); if (list_empty(&space_info->tickets)) { space_info->flush = 0; spin_unlock(&space_info->lock); return; } - to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, - space_info); + to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info, + space_info, + false); ticket = list_first_entry(&space_info->tickets, struct reserve_ticket, list); if (last_tickets_id == space_info->tickets_id) { @@ -5040,8 +5115,8 @@ static void priority_reclaim_metadata_space(struct btrfs_fs_info *fs_info, int flush_state = FLUSH_DELAYED_ITEMS_NR; spin_lock(&space_info->lock); - to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, - space_info); + to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info, space_info, + false); if (!to_reclaim) { spin_unlock(&space_info->lock); return; @@ -5049,8 +5124,8 @@ static void priority_reclaim_metadata_space(struct btrfs_fs_info *fs_info, spin_unlock(&space_info->lock); do { - flush_space(fs_info->fs_root, space_info, to_reclaim, - to_reclaim, flush_state); + flush_space(fs_info, space_info, to_reclaim, to_reclaim, + flush_state); flush_state++; spin_lock(&space_info->lock); if (ticket->bytes == 0) { @@ -5120,12 +5195,12 @@ static int wait_reserve_ticket(struct btrfs_fs_info *fs_info, * regain reservations will be made and this will fail if there is not enough * space already. */ -static int __reserve_metadata_bytes(struct btrfs_root *root, +static int __reserve_metadata_bytes(struct btrfs_fs_info *fs_info, struct btrfs_space_info *space_info, u64 orig_bytes, - enum btrfs_reserve_flush_enum flush) + enum btrfs_reserve_flush_enum flush, + bool system_chunk) { - struct btrfs_fs_info *fs_info = root->fs_info; struct reserve_ticket ticket; u64 used; int ret = 0; @@ -5135,9 +5210,7 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, spin_lock(&space_info->lock); ret = -ENOSPC; - used = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; + used = btrfs_space_info_used(space_info, true); /* * If we have enough space then hooray, make our reservation and carry @@ -5149,7 +5222,8 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, trace_btrfs_space_reservation(fs_info, "space_info", space_info->flags, orig_bytes, 1); ret = 0; - } else if (can_overcommit(root, space_info, orig_bytes, flush)) { + } else if (can_overcommit(fs_info, space_info, orig_bytes, flush, + system_chunk)) { space_info->bytes_may_use += orig_bytes; trace_btrfs_space_reservation(fs_info, "space_info", space_info->flags, orig_bytes, 1); @@ -5176,7 +5250,7 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, orig_bytes, flush, "enospc"); queue_work(system_unbound_wq, - &root->fs_info->async_reclaim_work); + &fs_info->async_reclaim_work); } } else { list_add_tail(&ticket.list, @@ -5190,7 +5264,8 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, * the async reclaim as we will panic. */ if (!test_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags) && - need_do_async_reclaim(space_info, root, used) && + need_do_async_reclaim(fs_info, space_info, + used, system_chunk) && !work_busy(&fs_info->async_reclaim_work)) { trace_btrfs_trigger_flush(fs_info, space_info->flags, orig_bytes, flush, "preempt"); @@ -5248,9 +5323,10 @@ static int reserve_metadata_bytes(struct btrfs_root *root, struct btrfs_fs_info *fs_info = root->fs_info; struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; int ret; + bool system_chunk = (root == fs_info->chunk_root); - ret = __reserve_metadata_bytes(root, block_rsv->space_info, orig_bytes, - flush); + ret = __reserve_metadata_bytes(fs_info, block_rsv->space_info, + orig_bytes, flush, system_chunk); if (ret == -ENOSPC && unlikely(root->orphan_cleanup_state == ORPHAN_CLEANUP_STARTED)) { if (block_rsv != global_rsv && @@ -5359,9 +5435,7 @@ static void space_info_add_old_bytes(struct btrfs_fs_info *fs_info, * overcommit, and if we can't then we just need to free up our space * and not satisfy any requests. */ - used = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; + used = btrfs_space_info_used(space_info, true); if (used - num_bytes >= space_info->total_bytes) check_overcommit = true; again: @@ -5373,8 +5447,7 @@ again: * adding the ticket space would be a double count. */ if (check_overcommit && - !can_overcommit(fs_info->extent_root, space_info, 0, - flush)) + !can_overcommit(fs_info, space_info, 0, flush, false)) break; if (num_bytes >= ticket->bytes) { list_del_init(&ticket->list); @@ -5630,9 +5703,7 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info) block_rsv->size = min_t(u64, num_bytes, SZ_512M); if (block_rsv->reserved < block_rsv->size) { - num_bytes = sinfo->bytes_used + sinfo->bytes_pinned + - sinfo->bytes_reserved + sinfo->bytes_readonly + - sinfo->bytes_may_use; + num_bytes = btrfs_space_info_used(sinfo, true); if (sinfo->total_bytes > num_bytes) { num_bytes = sinfo->total_bytes - num_bytes; num_bytes = min(num_bytes, @@ -5735,10 +5806,10 @@ void btrfs_trans_release_chunk_metadata(struct btrfs_trans_handle *trans) /* Can only return 0 or -ENOSPC */ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, - struct inode *inode) + struct btrfs_inode *inode) { - struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); - struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb); + struct btrfs_root *root = inode->root; /* * We always use trans->block_rsv here as we will have reserved space * for our orphan when starting the transaction, using get_block_rsv() @@ -5755,19 +5826,19 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, */ u64 num_bytes = btrfs_calc_trans_metadata_size(fs_info, 1); - trace_btrfs_space_reservation(fs_info, "orphan", - btrfs_ino(inode), num_bytes, 1); + trace_btrfs_space_reservation(fs_info, "orphan", btrfs_ino(inode), + num_bytes, 1); return btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); } -void btrfs_orphan_release_metadata(struct inode *inode) +void btrfs_orphan_release_metadata(struct btrfs_inode *inode) { - struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); - struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb); + struct btrfs_root *root = inode->root; u64 num_bytes = btrfs_calc_trans_metadata_size(fs_info, 1); - trace_btrfs_space_reservation(fs_info, "orphan", - btrfs_ino(inode), num_bytes, 0); + trace_btrfs_space_reservation(fs_info, "orphan", btrfs_ino(inode), + num_bytes, 0); btrfs_block_rsv_release(fs_info, root->orphan_block_rsv, num_bytes); } @@ -5799,7 +5870,7 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, if (test_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags)) { /* One for parent inode, two for dir entries */ num_bytes = 3 * fs_info->nodesize; - ret = btrfs_qgroup_reserve_meta(root, num_bytes); + ret = btrfs_qgroup_reserve_meta(root, num_bytes, true); if (ret) return ret; } else { @@ -5824,8 +5895,7 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, } void btrfs_subvolume_release_metadata(struct btrfs_fs_info *fs_info, - struct btrfs_block_rsv *rsv, - u64 qgroup_reserved) + struct btrfs_block_rsv *rsv) { btrfs_block_rsv_release(fs_info, rsv, (u64)-1); } @@ -5840,35 +5910,32 @@ void btrfs_subvolume_release_metadata(struct btrfs_fs_info *fs_info, * reserved extents that need to be freed. This must be called with * BTRFS_I(inode)->lock held. */ -static unsigned drop_outstanding_extent(struct inode *inode, u64 num_bytes) +static unsigned drop_outstanding_extent(struct btrfs_inode *inode, + u64 num_bytes) { unsigned drop_inode_space = 0; unsigned dropped_extents = 0; - unsigned num_extents = 0; + unsigned num_extents; - num_extents = (unsigned)div64_u64(num_bytes + - BTRFS_MAX_EXTENT_SIZE - 1, - BTRFS_MAX_EXTENT_SIZE); + num_extents = count_max_extents(num_bytes); ASSERT(num_extents); - ASSERT(BTRFS_I(inode)->outstanding_extents >= num_extents); - BTRFS_I(inode)->outstanding_extents -= num_extents; + ASSERT(inode->outstanding_extents >= num_extents); + inode->outstanding_extents -= num_extents; - if (BTRFS_I(inode)->outstanding_extents == 0 && + if (inode->outstanding_extents == 0 && test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags)) + &inode->runtime_flags)) drop_inode_space = 1; /* * If we have more or the same amount of outstanding extents than we have * reserved then we need to leave the reserved extents count alone. */ - if (BTRFS_I(inode)->outstanding_extents >= - BTRFS_I(inode)->reserved_extents) + if (inode->outstanding_extents >= inode->reserved_extents) return drop_inode_space; - dropped_extents = BTRFS_I(inode)->reserved_extents - - BTRFS_I(inode)->outstanding_extents; - BTRFS_I(inode)->reserved_extents -= dropped_extents; + dropped_extents = inode->reserved_extents - inode->outstanding_extents; + inode->reserved_extents -= dropped_extents; return dropped_extents + drop_inode_space; } @@ -5890,24 +5957,21 @@ static unsigned drop_outstanding_extent(struct inode *inode, u64 num_bytes) * * This must be called with BTRFS_I(inode)->lock held. */ -static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes, +static u64 calc_csum_metadata_size(struct btrfs_inode *inode, u64 num_bytes, int reserve) { - struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); + struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb); u64 old_csums, num_csums; - if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM && - BTRFS_I(inode)->csum_bytes == 0) + if (inode->flags & BTRFS_INODE_NODATASUM && inode->csum_bytes == 0) return 0; - old_csums = btrfs_csum_bytes_to_leaves(fs_info, - BTRFS_I(inode)->csum_bytes); + old_csums = btrfs_csum_bytes_to_leaves(fs_info, inode->csum_bytes); if (reserve) - BTRFS_I(inode)->csum_bytes += num_bytes; + inode->csum_bytes += num_bytes; else - BTRFS_I(inode)->csum_bytes -= num_bytes; - num_csums = btrfs_csum_bytes_to_leaves(fs_info, - BTRFS_I(inode)->csum_bytes); + inode->csum_bytes -= num_bytes; + num_csums = btrfs_csum_bytes_to_leaves(fs_info, inode->csum_bytes); /* No change, no need to reserve more */ if (old_csums == num_csums) @@ -5920,14 +5984,14 @@ static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes, return btrfs_calc_trans_metadata_size(fs_info, old_csums - num_csums); } -int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) +int btrfs_delalloc_reserve_metadata(struct btrfs_inode *inode, u64 num_bytes) { - struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); - struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb); + struct btrfs_root *root = inode->root; struct btrfs_block_rsv *block_rsv = &fs_info->delalloc_block_rsv; u64 to_reserve = 0; u64 csum_bytes; - unsigned nr_extents = 0; + unsigned nr_extents; enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_FLUSH_ALL; int ret = 0; bool delalloc_lock = true; @@ -5955,31 +6019,28 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) schedule_timeout(1); if (delalloc_lock) - mutex_lock(&BTRFS_I(inode)->delalloc_mutex); + mutex_lock(&inode->delalloc_mutex); num_bytes = ALIGN(num_bytes, fs_info->sectorsize); - spin_lock(&BTRFS_I(inode)->lock); - nr_extents = (unsigned)div64_u64(num_bytes + - BTRFS_MAX_EXTENT_SIZE - 1, - BTRFS_MAX_EXTENT_SIZE); - BTRFS_I(inode)->outstanding_extents += nr_extents; + spin_lock(&inode->lock); + nr_extents = count_max_extents(num_bytes); + inode->outstanding_extents += nr_extents; nr_extents = 0; - if (BTRFS_I(inode)->outstanding_extents > - BTRFS_I(inode)->reserved_extents) - nr_extents += BTRFS_I(inode)->outstanding_extents - - BTRFS_I(inode)->reserved_extents; + if (inode->outstanding_extents > inode->reserved_extents) + nr_extents += inode->outstanding_extents - + inode->reserved_extents; /* We always want to reserve a slot for updating the inode. */ to_reserve = btrfs_calc_trans_metadata_size(fs_info, nr_extents + 1); to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); - csum_bytes = BTRFS_I(inode)->csum_bytes; - spin_unlock(&BTRFS_I(inode)->lock); + csum_bytes = inode->csum_bytes; + spin_unlock(&inode->lock); if (test_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags)) { ret = btrfs_qgroup_reserve_meta(root, - nr_extents * fs_info->nodesize); + nr_extents * fs_info->nodesize, true); if (ret) goto out_fail; } @@ -5991,17 +6052,17 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) goto out_fail; } - spin_lock(&BTRFS_I(inode)->lock); + spin_lock(&inode->lock); if (test_and_set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags)) { + &inode->runtime_flags)) { to_reserve -= btrfs_calc_trans_metadata_size(fs_info, 1); release_extra = true; } - BTRFS_I(inode)->reserved_extents += nr_extents; - spin_unlock(&BTRFS_I(inode)->lock); + inode->reserved_extents += nr_extents; + spin_unlock(&inode->lock); if (delalloc_lock) - mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); + mutex_unlock(&inode->delalloc_mutex); if (to_reserve) trace_btrfs_space_reservation(fs_info, "delalloc", @@ -6012,17 +6073,17 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) return 0; out_fail: - spin_lock(&BTRFS_I(inode)->lock); + spin_lock(&inode->lock); dropped = drop_outstanding_extent(inode, num_bytes); /* * If the inodes csum_bytes is the same as the original * csum_bytes then we know we haven't raced with any free()ers * so we can just reduce our inodes csum bytes and carry on. */ - if (BTRFS_I(inode)->csum_bytes == csum_bytes) { + if (inode->csum_bytes == csum_bytes) { calc_csum_metadata_size(inode, num_bytes, 0); } else { - u64 orig_csum_bytes = BTRFS_I(inode)->csum_bytes; + u64 orig_csum_bytes = inode->csum_bytes; u64 bytes; /* @@ -6033,8 +6094,8 @@ out_fail: * number of bytes that were freed while we were trying our * reservation. */ - bytes = csum_bytes - BTRFS_I(inode)->csum_bytes; - BTRFS_I(inode)->csum_bytes = csum_bytes; + bytes = csum_bytes - inode->csum_bytes; + inode->csum_bytes = csum_bytes; to_free = calc_csum_metadata_size(inode, bytes, 0); @@ -6043,7 +6104,7 @@ out_fail: * been making this reservation and our ->csum_bytes were not * artificially inflated. */ - BTRFS_I(inode)->csum_bytes = csum_bytes - num_bytes; + inode->csum_bytes = csum_bytes - num_bytes; bytes = csum_bytes - orig_csum_bytes; bytes = calc_csum_metadata_size(inode, bytes, 0); @@ -6055,13 +6116,13 @@ out_fail: * need to do anything, the other free-ers did the correct * thing. */ - BTRFS_I(inode)->csum_bytes = orig_csum_bytes - num_bytes; + inode->csum_bytes = orig_csum_bytes - num_bytes; if (bytes > to_free) to_free = bytes - to_free; else to_free = 0; } - spin_unlock(&BTRFS_I(inode)->lock); + spin_unlock(&inode->lock); if (dropped) to_free += btrfs_calc_trans_metadata_size(fs_info, dropped); @@ -6071,7 +6132,7 @@ out_fail: btrfs_ino(inode), to_free, 0); } if (delalloc_lock) - mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); + mutex_unlock(&inode->delalloc_mutex); return ret; } @@ -6084,27 +6145,27 @@ out_fail: * once we complete IO for a given set of bytes to release their metadata * reservations. */ -void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) +void btrfs_delalloc_release_metadata(struct btrfs_inode *inode, u64 num_bytes) { - struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); + struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb); u64 to_free = 0; unsigned dropped; num_bytes = ALIGN(num_bytes, fs_info->sectorsize); - spin_lock(&BTRFS_I(inode)->lock); + spin_lock(&inode->lock); dropped = drop_outstanding_extent(inode, num_bytes); if (num_bytes) to_free = calc_csum_metadata_size(inode, num_bytes, 0); - spin_unlock(&BTRFS_I(inode)->lock); + spin_unlock(&inode->lock); if (dropped > 0) to_free += btrfs_calc_trans_metadata_size(fs_info, dropped); if (btrfs_is_testing(fs_info)) return; - trace_btrfs_space_reservation(fs_info, "delalloc", - btrfs_ino(inode), to_free, 0); + trace_btrfs_space_reservation(fs_info, "delalloc", btrfs_ino(inode), + to_free, 0); btrfs_block_rsv_release(fs_info, &fs_info->delalloc_block_rsv, to_free); } @@ -6115,6 +6176,8 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) * @inode: inode we're writing to * @start: start range we are writing to * @len: how long the range we are writing to + * @reserved: mandatory parameter, record actually reserved qgroup ranges of + * current reservation. * * This will do the following things * @@ -6132,16 +6195,17 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) * Return 0 for success * Return <0 for error(-ENOSPC or -EQUOT) */ -int btrfs_delalloc_reserve_space(struct inode *inode, u64 start, u64 len) +int btrfs_delalloc_reserve_space(struct inode *inode, + struct extent_changeset **reserved, u64 start, u64 len) { int ret; - ret = btrfs_check_data_free_space(inode, start, len); + ret = btrfs_check_data_free_space(inode, reserved, start, len); if (ret < 0) return ret; - ret = btrfs_delalloc_reserve_metadata(inode, len); + ret = btrfs_delalloc_reserve_metadata(BTRFS_I(inode), len); if (ret < 0) - btrfs_free_reserved_data_space(inode, start, len); + btrfs_free_reserved_data_space(inode, *reserved, start, len); return ret; } @@ -6160,10 +6224,11 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 start, u64 len) * list if there are no delalloc bytes left. * Also it will handle the qgroup reserved space. */ -void btrfs_delalloc_release_space(struct inode *inode, u64 start, u64 len) +void btrfs_delalloc_release_space(struct inode *inode, + struct extent_changeset *reserved, u64 start, u64 len) { - btrfs_delalloc_release_metadata(inode, len); - btrfs_free_reserved_data_space(inode, start, len); + btrfs_delalloc_release_metadata(BTRFS_I(inode), len); + btrfs_free_reserved_data_space(inode, reserved, start, len); } static int update_block_group(struct btrfs_trans_handle *trans, @@ -6239,6 +6304,8 @@ static int update_block_group(struct btrfs_trans_handle *trans, trace_btrfs_space_reservation(info, "pinned", cache->space_info->flags, num_bytes, 1); + percpu_counter_add(&cache->space_info->total_bytes_pinned, + num_bytes); set_extent_dirty(info->pinned_extents, bytenr, bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); @@ -6315,6 +6382,7 @@ static int pin_down_extent(struct btrfs_fs_info *fs_info, trace_btrfs_space_reservation(fs_info, "pinned", cache->space_info->flags, num_bytes, 1); + percpu_counter_add(&cache->space_info->total_bytes_pinned, num_bytes); set_extent_dirty(fs_info->pinned_extents, bytenr, bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); return 0; @@ -6561,8 +6629,7 @@ static int btrfs_free_reserved_bytes(struct btrfs_block_group_cache *cache, spin_unlock(&space_info->lock); return ret; } -void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, - struct btrfs_fs_info *fs_info) +void btrfs_prepare_extent_commit(struct btrfs_fs_info *fs_info) { struct btrfs_caching_control *next; struct btrfs_caching_control *caching_ctl; @@ -6786,27 +6853,6 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, return 0; } -static void add_pinned_bytes(struct btrfs_fs_info *fs_info, u64 num_bytes, - u64 owner, u64 root_objectid) -{ - struct btrfs_space_info *space_info; - u64 flags; - - if (owner < BTRFS_FIRST_FREE_OBJECTID) { - if (root_objectid == BTRFS_CHUNK_TREE_OBJECTID) - flags = BTRFS_BLOCK_GROUP_SYSTEM; - else - flags = BTRFS_BLOCK_GROUP_METADATA; - } else { - flags = BTRFS_BLOCK_GROUP_DATA; - } - - space_info = __find_space_info(fs_info, flags); - BUG_ON(!space_info); /* Logic bug */ - percpu_counter_add(&space_info->total_bytes_pinned, num_bytes); -} - - static int __btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_fs_info *info, struct btrfs_delayed_ref_node *node, u64 parent, @@ -6845,7 +6891,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, if (is_data) skinny_metadata = 0; - ret = lookup_extent_backref(trans, extent_root, path, &iref, + ret = lookup_extent_backref(trans, info, path, &iref, bytenr, num_bytes, parent, root_objectid, owner_objectid, owner_offset); @@ -6877,8 +6923,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, #endif if (!found_extent) { BUG_ON(iref); - ret = remove_extent_backref(trans, extent_root, path, - NULL, refs_to_drop, + ret = remove_extent_backref(trans, info, path, NULL, + refs_to_drop, is_data, &last_ref); if (ret) { btrfs_abort_transaction(trans, ret); @@ -6953,8 +6999,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 if (item_size < sizeof(*ei)) { BUG_ON(found_extent || extent_slot != path->slots[0]); - ret = convert_extent_item_v0(trans, extent_root, path, - owner_objectid, 0); + ret = convert_extent_item_v0(trans, info, path, owner_objectid, + 0); if (ret < 0) { btrfs_abort_transaction(trans, ret); goto out; @@ -7021,7 +7067,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(leaf); } if (found_extent) { - ret = remove_extent_backref(trans, extent_root, path, + ret = remove_extent_backref(trans, info, path, iref, refs_to_drop, is_data, &last_ref); if (ret) { @@ -7029,8 +7075,6 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, goto out; } } - add_pinned_bytes(info, -num_bytes, owner_objectid, - root_objectid); } else { if (found_extent) { BUG_ON(is_data && refs_to_drop != @@ -7095,7 +7139,7 @@ static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans, delayed_refs = &trans->transaction->delayed_refs; spin_lock(&delayed_refs->lock); - head = btrfs_find_delayed_ref_head(trans, bytenr); + head = btrfs_find_delayed_ref_head(delayed_refs, bytenr); if (!head) goto out_delayed_unlock; @@ -7162,19 +7206,19 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, int ret; if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { - ret = btrfs_add_delayed_tree_ref(fs_info, trans, - buf->start, buf->len, - parent, + int old_ref_mod, new_ref_mod; + + ret = btrfs_add_delayed_tree_ref(fs_info, trans, buf->start, + buf->len, parent, root->root_key.objectid, btrfs_header_level(buf), - BTRFS_DROP_DELAYED_REF, NULL); + BTRFS_DROP_DELAYED_REF, NULL, + &old_ref_mod, &new_ref_mod); BUG_ON(ret); /* -ENOMEM */ + pin = old_ref_mod >= 0 && new_ref_mod < 0; } - if (!last_ref) - return; - - if (btrfs_header_generation(buf) == trans->transid) { + if (last_ref && btrfs_header_generation(buf) == trans->transid) { struct btrfs_block_group_cache *cache; if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { @@ -7183,6 +7227,7 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, goto out; } + pin = 0; cache = btrfs_lookup_block_group(fs_info, buf->start); if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) { @@ -7198,18 +7243,19 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, btrfs_free_reserved_bytes(cache, buf->len, 0); btrfs_put_block_group(cache); trace_btrfs_reserved_extent_free(fs_info, buf->start, buf->len); - pin = 0; } out: if (pin) add_pinned_bytes(fs_info, buf->len, btrfs_header_level(buf), root->root_key.objectid); - /* - * Deleting the buffer, clear the corrupt flag since it doesn't matter - * anymore. - */ - clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags); + if (last_ref) { + /* + * Deleting the buffer, clear the corrupt flag since it doesn't + * matter anymore. + */ + clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags); + } } /* Can return -ENOMEM */ @@ -7218,12 +7264,12 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, u64 owner, u64 offset) { + int old_ref_mod, new_ref_mod; int ret; if (btrfs_is_testing(fs_info)) return 0; - add_pinned_bytes(fs_info, num_bytes, owner, root_objectid); /* * tree log blocks never actually go into the extent allocation @@ -7233,19 +7279,25 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, WARN_ON(owner >= BTRFS_FIRST_FREE_OBJECTID); /* unlocks the pinned mutex */ btrfs_pin_extent(fs_info, bytenr, num_bytes, 1); + old_ref_mod = new_ref_mod = 0; ret = 0; } else if (owner < BTRFS_FIRST_FREE_OBJECTID) { ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, - num_bytes, - parent, root_objectid, (int)owner, - BTRFS_DROP_DELAYED_REF, NULL); + num_bytes, parent, + root_objectid, (int)owner, + BTRFS_DROP_DELAYED_REF, NULL, + &old_ref_mod, &new_ref_mod); } else { ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, - num_bytes, - parent, root_objectid, owner, - offset, 0, - BTRFS_DROP_DELAYED_REF, NULL); + num_bytes, parent, + root_objectid, owner, offset, + 0, BTRFS_DROP_DELAYED_REF, + &old_ref_mod, &new_ref_mod); } + + if (ret == 0 && old_ref_mod >= 0 && new_ref_mod < 0) + add_pinned_bytes(fs_info, num_bytes, owner, root_objectid); + return ret; } @@ -7419,12 +7471,11 @@ btrfs_release_block_group(struct btrfs_block_group_cache *cache, * If there is no suitable free space, we will record the max size of * the free space extent currently. */ -static noinline int find_free_extent(struct btrfs_root *orig_root, +static noinline int find_free_extent(struct btrfs_fs_info *fs_info, u64 ram_bytes, u64 num_bytes, u64 empty_size, u64 hint_byte, struct btrfs_key *ins, u64 flags, int delalloc) { - struct btrfs_fs_info *fs_info = orig_root->fs_info; int ret = 0; struct btrfs_root *root = fs_info->extent_root; struct btrfs_free_cluster *last_ptr = NULL; @@ -7716,18 +7767,20 @@ unclustered_alloc: last_ptr->fragmented = 1; spin_unlock(&last_ptr->lock); } - spin_lock(&block_group->free_space_ctl->tree_lock); - if (cached && - block_group->free_space_ctl->free_space < - num_bytes + empty_cluster + empty_size) { - if (block_group->free_space_ctl->free_space > - max_extent_size) - max_extent_size = - block_group->free_space_ctl->free_space; - spin_unlock(&block_group->free_space_ctl->tree_lock); - goto loop; + if (cached) { + struct btrfs_free_space_ctl *ctl = + block_group->free_space_ctl; + + spin_lock(&ctl->tree_lock); + if (ctl->free_space < + num_bytes + empty_cluster + empty_size) { + if (ctl->free_space > max_extent_size) + max_extent_size = ctl->free_space; + spin_unlock(&ctl->tree_lock); + goto loop; + } + spin_unlock(&ctl->tree_lock); } - spin_unlock(&block_group->free_space_ctl->tree_lock); offset = btrfs_find_space_for_alloc(block_group, search_start, num_bytes, empty_size, @@ -7908,9 +7961,8 @@ static void dump_space_info(struct btrfs_fs_info *fs_info, spin_lock(&info->lock); btrfs_info(fs_info, "space_info %llu has %llu free, is %sfull", info->flags, - info->total_bytes - info->bytes_used - info->bytes_pinned - - info->bytes_reserved - info->bytes_readonly - - info->bytes_may_use, (info->full) ? "" : "not "); + info->total_bytes - btrfs_space_info_used(info, true), + info->full ? "" : "not "); btrfs_info(fs_info, "space_info total=%llu, used=%llu, pinned=%llu, reserved=%llu, may_use=%llu, readonly=%llu", info->total_bytes, info->bytes_used, info->bytes_pinned, @@ -7948,10 +8000,10 @@ int btrfs_reserve_extent(struct btrfs_root *root, u64 ram_bytes, u64 flags; int ret; - flags = btrfs_get_alloc_profile(root, is_data); + flags = get_alloc_profile_by_root(root, is_data); again: WARN_ON(num_bytes < fs_info->sectorsize); - ret = find_free_extent(root, ram_bytes, num_bytes, empty_size, + ret = find_free_extent(fs_info, ram_bytes, num_bytes, empty_size, hint_byte, ins, flags, delalloc); if (!ret && !is_data) { btrfs_dec_block_group_reservations(fs_info, ins->objectid); @@ -8192,10 +8244,9 @@ int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans, BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID); ret = btrfs_add_delayed_data_ref(fs_info, trans, ins->objectid, - ins->offset, 0, - root_objectid, owner, offset, - ram_bytes, BTRFS_ADD_DELAYED_EXTENT, - NULL); + ins->offset, 0, root_objectid, owner, + offset, ram_bytes, + BTRFS_ADD_DELAYED_EXTENT, NULL, NULL); return ret; } @@ -8256,7 +8307,7 @@ btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root, btrfs_set_header_generation(buf, trans->transid); btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level); btrfs_tree_lock(buf); - clean_tree_block(trans, fs_info, buf); + clean_tree_block(fs_info, buf); clear_bit(EXTENT_BUFFER_STALE, &buf->bflags); btrfs_set_lock_blocking(buf); @@ -8351,10 +8402,11 @@ static void unuse_block_rsv(struct btrfs_fs_info *fs_info, * returns the tree buffer or an ERR_PTR on error. */ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - u64 parent, u64 root_objectid, - struct btrfs_disk_key *key, int level, - u64 hint, u64 empty_size) + struct btrfs_root *root, + u64 parent, u64 root_objectid, + const struct btrfs_disk_key *key, + int level, u64 hint, + u64 empty_size) { struct btrfs_fs_info *fs_info = root->fs_info; struct btrfs_key ins; @@ -8414,11 +8466,11 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, extent_op->is_data = false; extent_op->level = level; - ret = btrfs_add_delayed_tree_ref(fs_info, trans, - ins.objectid, ins.offset, - parent, root_objectid, level, + ret = btrfs_add_delayed_tree_ref(fs_info, trans, ins.objectid, + ins.offset, parent, + root_objectid, level, BTRFS_ADD_DELAYED_EXTENT, - extent_op); + extent_op, NULL, NULL); if (ret) goto out_free_delayed; } @@ -8876,7 +8928,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans, btrfs_set_lock_blocking(eb); path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; } - clean_tree_block(trans, fs_info, eb); + clean_tree_block(fs_info, eb); } if (eb == root->node) { @@ -9346,8 +9398,7 @@ static int inc_block_group_ro(struct btrfs_block_group_cache *cache, int force) num_bytes = cache->key.offset - cache->reserved - cache->pinned - cache->bytes_super - btrfs_block_group_used(&cache->item); - if (sinfo->bytes_used + sinfo->bytes_reserved + sinfo->bytes_pinned + - sinfo->bytes_may_use + sinfo->bytes_readonly + num_bytes + + if (btrfs_space_info_used(sinfo, true) + num_bytes + min_allocable_bytes <= sinfo->total_bytes) { sinfo->bytes_readonly += num_bytes; cache->ro++; @@ -9360,17 +9411,16 @@ out: return ret; } -int btrfs_inc_block_group_ro(struct btrfs_root *root, +int btrfs_inc_block_group_ro(struct btrfs_fs_info *fs_info, struct btrfs_block_group_cache *cache) { - struct btrfs_fs_info *fs_info = root->fs_info; struct btrfs_trans_handle *trans; u64 alloc_flags; int ret; again: - trans = btrfs_join_transaction(root); + trans = btrfs_join_transaction(fs_info->extent_root); if (IS_ERR(trans)) return PTR_ERR(trans); @@ -9557,9 +9607,8 @@ int btrfs_can_relocate(struct btrfs_fs_info *fs_info, u64 bytenr) * all of the extents from this block group. If we can, we're good */ if ((space_info->total_bytes != block_group->key.offset) && - (space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - min_free < space_info->total_bytes)) { + (btrfs_space_info_used(space_info, false) + min_free < + space_info->total_bytes)) { spin_unlock(&space_info->lock); goto out; } @@ -9742,6 +9791,11 @@ void btrfs_put_block_group_cache(struct btrfs_fs_info *info) } } +/* + * Must be called only after stopping all workers, since we could have block + * group caching kthreads running, and therefore they could race with us if we + * freed the block groups before stopping them. + */ int btrfs_free_block_groups(struct btrfs_fs_info *info) { struct btrfs_block_group_cache *block_group; @@ -9781,9 +9835,6 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info) list_del(&block_group->list); up_write(&block_group->space_info->groups_sem); - if (block_group->cached == BTRFS_CACHE_STARTED) - wait_block_group_cache_done(block_group); - /* * We haven't cached this block group, which means we could * possibly have excluded extents on this block group. @@ -9793,6 +9844,7 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info) free_excluded_extents(info, block_group); btrfs_remove_free_space_cache(block_group); + ASSERT(block_group->cached != BTRFS_CACHE_STARTED); ASSERT(list_empty(&block_group->dirty_list)); ASSERT(list_empty(&block_group->io_list)); ASSERT(list_empty(&block_group->bg_list)); @@ -9920,6 +9972,7 @@ btrfs_create_block_group_cache(struct btrfs_fs_info *fs_info, btrfs_init_free_space_ctl(cache); atomic_set(&cache->trimming, 0); mutex_init(&cache->free_space_lock); + btrfs_init_full_stripe_locks_tree(&cache->full_stripe_locks_root); return cache; } @@ -10050,19 +10103,9 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info) } trace_btrfs_add_block_group(info, cache, 0); - ret = update_space_info(info, cache->flags, found_key.offset, - btrfs_block_group_used(&cache->item), - cache->bytes_super, &space_info); - if (ret) { - btrfs_remove_free_space_cache(cache); - spin_lock(&info->block_group_cache_lock); - rb_erase(&cache->cache_node, - &info->block_group_cache_tree); - RB_CLEAR_NODE(&cache->cache_node); - spin_unlock(&info->block_group_cache_lock); - btrfs_put_block_group(cache); - goto error; - } + update_space_info(info, cache->flags, found_key.offset, + btrfs_block_group_used(&cache->item), + cache->bytes_super, &space_info); cache->space_info = space_info; @@ -10194,16 +10237,19 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, } #endif /* - * Call to ensure the corresponding space_info object is created and - * assigned to our block group, but don't update its counters just yet. - * We want our bg to be added to the rbtree with its ->space_info set. + * Ensure the corresponding space_info object is created and + * assigned to our block group. We want our bg to be added to the rbtree + * with its ->space_info set. */ - ret = update_space_info(fs_info, cache->flags, 0, 0, 0, - &cache->space_info); - if (ret) { - btrfs_remove_free_space_cache(cache); - btrfs_put_block_group(cache); - return ret; + cache->space_info = __find_space_info(fs_info, cache->flags); + if (!cache->space_info) { + ret = create_space_info(fs_info, cache->flags, + &cache->space_info); + if (ret) { + btrfs_remove_free_space_cache(cache); + btrfs_put_block_group(cache); + return ret; + } } ret = btrfs_add_block_group_cache(fs_info, cache); @@ -10218,18 +10264,8 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, * the rbtree, update the space info's counters. */ trace_btrfs_add_block_group(fs_info, cache, 1); - ret = update_space_info(fs_info, cache->flags, size, bytes_used, + update_space_info(fs_info, cache->flags, size, bytes_used, cache->bytes_super, &cache->space_info); - if (ret) { - btrfs_remove_free_space_cache(cache); - spin_lock(&fs_info->block_group_cache_lock); - rb_erase(&cache->cache_node, - &fs_info->block_group_cache_tree); - RB_CLEAR_NODE(&cache->cache_node); - spin_unlock(&fs_info->block_group_cache_lock); - btrfs_put_block_group(cache); - return ret; - } update_global_block_rsv(fs_info); __link_block_group(cache->space_info, cache); @@ -10317,7 +10353,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, * get the inode first so any iput calls done for the io_list * aren't the final iput (no unlinks allowed now) */ - inode = lookup_free_space_inode(tree_root, block_group, path); + inode = lookup_free_space_inode(fs_info, block_group, path); mutex_lock(&trans->transaction->cache_write_mutex); /* @@ -10344,7 +10380,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, mutex_unlock(&trans->transaction->cache_write_mutex); if (!IS_ERR(inode)) { - ret = btrfs_orphan_add(trans, inode); + ret = btrfs_orphan_add(trans, BTRFS_I(inode)); if (ret) { btrfs_add_delayed_iput(inode); goto out; @@ -10419,7 +10455,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, &fs_info->caching_block_groups, list) if (ctl->block_group == block_group) { caching_ctl = ctl; - atomic_inc(&caching_ctl->count); + refcount_inc(&caching_ctl->count); break; } } @@ -10777,21 +10813,21 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info) mixed = 1; flags = BTRFS_BLOCK_GROUP_SYSTEM; - ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); + ret = create_space_info(fs_info, flags, &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, 0, &space_info); + ret = create_space_info(fs_info, flags, &space_info); } else { flags = BTRFS_BLOCK_GROUP_METADATA; - ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); + ret = create_space_info(fs_info, flags, &space_info); if (ret) goto out; flags = BTRFS_BLOCK_GROUP_DATA; - ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); + ret = create_space_info(fs_info, flags, &space_info); } out: return ret; @@ -10853,7 +10889,7 @@ static int btrfs_trim_free_extents(struct btrfs_device *device, spin_lock(&fs_info->trans_lock); trans = fs_info->running_transaction; if (trans) - atomic_inc(&trans->use_count); + refcount_inc(&trans->use_count); spin_unlock(&fs_info->trans_lock); ret = find_free_dev_extent_start(trans, device, minlen, start, |