diff options
Diffstat (limited to 'fs/btrfs/inode.c')
| -rw-r--r-- | fs/btrfs/inode.c | 1397 |
1 files changed, 682 insertions, 715 deletions
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e064c49c9a9a..89b208201783 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -31,6 +31,7 @@ #include <linux/uio.h> #include <linux/magic.h> #include <linux/iversion.h> +#include <asm/unaligned.h> #include "ctree.h" #include "disk-io.h" #include "transaction.h" @@ -1017,8 +1018,10 @@ static noinline int cow_file_range(struct inode *inode, ram_size, /* ram_bytes */ BTRFS_COMPRESS_NONE, /* compress_type */ BTRFS_ORDERED_REGULAR /* type */); - if (IS_ERR(em)) + if (IS_ERR(em)) { + ret = PTR_ERR(em); goto out_reserve; + } free_extent_map(em); ret = btrfs_add_ordered_extent(inode, start, ins.objectid, @@ -1155,13 +1158,10 @@ static noinline void async_cow_submit(struct btrfs_work *work) nr_pages = (async_cow->end - async_cow->start + PAGE_SIZE) >> PAGE_SHIFT; - /* - * atomic_sub_return implies a barrier for waitqueue_active - */ + /* atomic_sub_return implies a barrier */ if (atomic_sub_return(nr_pages, &fs_info->async_delalloc_pages) < - 5 * SZ_1M && - waitqueue_active(&fs_info->async_submit_wait)) - wake_up(&fs_info->async_submit_wait); + 5 * SZ_1M) + cond_wake_up_nomb(&fs_info->async_submit_wait); if (async_cow->inode) submit_compressed_extents(async_cow->inode, async_cow); @@ -1372,6 +1372,13 @@ next_slot: btrfs_file_extent_encryption(leaf, fi) || btrfs_file_extent_other_encoding(leaf, fi)) goto out_check; + /* + * Do the same check as in btrfs_cross_ref_exist but + * without the unnecessary search. + */ + if (btrfs_file_extent_generation(leaf, fi) <= + btrfs_root_last_snapshot(&root->root_item)) + goto out_check; if (extent_type == BTRFS_FILE_EXTENT_REG && !force) goto out_check; if (btrfs_extent_readonly(fs_info, disk_bytenr)) @@ -1741,24 +1748,32 @@ static void btrfs_add_delalloc_inodes(struct btrfs_root *root, spin_unlock(&root->delalloc_lock); } -static void btrfs_del_delalloc_inode(struct btrfs_root *root, - struct btrfs_inode *inode) + +void __btrfs_del_delalloc_inode(struct btrfs_root *root, + struct btrfs_inode *inode) { struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb); - spin_lock(&root->delalloc_lock); if (!list_empty(&inode->delalloc_inodes)) { list_del_init(&inode->delalloc_inodes); clear_bit(BTRFS_INODE_IN_DELALLOC_LIST, &inode->runtime_flags); root->nr_delalloc_inodes--; if (!root->nr_delalloc_inodes) { + ASSERT(list_empty(&root->delalloc_inodes)); spin_lock(&fs_info->delalloc_root_lock); BUG_ON(list_empty(&root->delalloc_root)); list_del_init(&root->delalloc_root); spin_unlock(&fs_info->delalloc_root_lock); } } +} + +static void btrfs_del_delalloc_inode(struct btrfs_root *root, + struct btrfs_inode *inode) +{ + spin_lock(&root->delalloc_lock); + __btrfs_del_delalloc_inode(root, inode); spin_unlock(&root->delalloc_lock); } @@ -3150,6 +3165,9 @@ out: /* once for the tree */ btrfs_put_ordered_extent(ordered_extent); + /* Try to release some metadata so we don't get an OOM but don't wait */ + btrfs_btree_balance_dirty_nodelay(fs_info); + return ret; } @@ -3292,177 +3310,31 @@ void btrfs_run_delayed_iputs(struct btrfs_fs_info *fs_info) } /* - * This is called in transaction commit time. If there are no orphan - * files in the subvolume, it removes orphan item and frees block_rsv - * structure. - */ -void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, - struct btrfs_root *root) -{ - struct btrfs_fs_info *fs_info = root->fs_info; - struct btrfs_block_rsv *block_rsv; - int ret; - - if (atomic_read(&root->orphan_inodes) || - root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) - return; - - spin_lock(&root->orphan_lock); - if (atomic_read(&root->orphan_inodes)) { - spin_unlock(&root->orphan_lock); - return; - } - - if (root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) { - spin_unlock(&root->orphan_lock); - return; - } - - block_rsv = root->orphan_block_rsv; - root->orphan_block_rsv = NULL; - spin_unlock(&root->orphan_lock); - - if (test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state) && - btrfs_root_refs(&root->root_item) > 0) { - ret = btrfs_del_orphan_item(trans, fs_info->tree_root, - root->root_key.objectid); - if (ret) - btrfs_abort_transaction(trans, ret); - else - clear_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, - &root->state); - } - - if (block_rsv) { - WARN_ON(block_rsv->size > 0); - btrfs_free_block_rsv(fs_info, block_rsv); - } -} - -/* - * This creates an orphan entry for the given inode in case something goes - * wrong in the middle of an unlink/truncate. - * - * NOTE: caller of this function should reserve 5 units of metadata for - * this function. + * This creates an orphan entry for the given inode in case something goes wrong + * in the middle of an unlink. */ int btrfs_orphan_add(struct btrfs_trans_handle *trans, - struct btrfs_inode *inode) + struct btrfs_inode *inode) { - 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 = NULL; - int reserve = 0; - bool insert = false; int ret; - if (!root->orphan_block_rsv) { - block_rsv = btrfs_alloc_block_rsv(fs_info, - BTRFS_BLOCK_RSV_TEMP); - if (!block_rsv) - return -ENOMEM; - } - - if (!test_and_set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, - &inode->runtime_flags)) - insert = true; - - if (!test_and_set_bit(BTRFS_INODE_ORPHAN_META_RESERVED, - &inode->runtime_flags)) - reserve = 1; - - spin_lock(&root->orphan_lock); - /* If someone has created ->orphan_block_rsv, be happy to use it. */ - if (!root->orphan_block_rsv) { - root->orphan_block_rsv = block_rsv; - } else if (block_rsv) { - btrfs_free_block_rsv(fs_info, block_rsv); - block_rsv = NULL; - } - - if (insert) - atomic_inc(&root->orphan_inodes); - spin_unlock(&root->orphan_lock); - - /* grab metadata reservation from transaction handle */ - if (reserve) { - ret = btrfs_orphan_reserve_metadata(trans, inode); - ASSERT(!ret); - if (ret) { - /* - * dec doesn't need spin_lock as ->orphan_block_rsv - * would be released only if ->orphan_inodes is - * zero. - */ - atomic_dec(&root->orphan_inodes); - clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, - &inode->runtime_flags); - if (insert) - clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, - &inode->runtime_flags); - return ret; - } - } - - /* insert an orphan item to track this unlinked/truncated file */ - if (insert) { - ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); - if (ret) { - if (reserve) { - clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, - &inode->runtime_flags); - btrfs_orphan_release_metadata(inode); - } - /* - * btrfs_orphan_commit_root may race with us and set - * ->orphan_block_rsv to zero, in order to avoid that, - * decrease ->orphan_inodes after everything is done. - */ - atomic_dec(&root->orphan_inodes); - if (ret != -EEXIST) { - clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, - &inode->runtime_flags); - btrfs_abort_transaction(trans, ret); - return ret; - } - } - ret = 0; + ret = btrfs_insert_orphan_item(trans, inode->root, btrfs_ino(inode)); + if (ret && ret != -EEXIST) { + btrfs_abort_transaction(trans, ret); + return ret; } return 0; } /* - * We have done the truncate/delete so we can go ahead and remove the orphan - * item for this particular inode. + * We have done the delete so we can go ahead and remove the orphan item for + * this particular inode. */ static int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct btrfs_inode *inode) { - struct btrfs_root *root = inode->root; - int delete_item = 0; - int ret = 0; - - if (test_and_clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, - &inode->runtime_flags)) - delete_item = 1; - - if (delete_item && trans) - ret = btrfs_del_orphan_item(trans, root, btrfs_ino(inode)); - - if (test_and_clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, - &inode->runtime_flags)) - btrfs_orphan_release_metadata(inode); - - /* - * btrfs_orphan_commit_root may race with us and set ->orphan_block_rsv - * to zero, in order to avoid that, decrease ->orphan_inodes after - * everything is done. - */ - if (delete_item) - atomic_dec(&root->orphan_inodes); - - return ret; + return btrfs_del_orphan_item(trans, inode->root, btrfs_ino(inode)); } /* @@ -3478,7 +3350,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) struct btrfs_trans_handle *trans; struct inode *inode; u64 last_objectid = 0; - int ret = 0, nr_unlink = 0, nr_truncate = 0; + int ret = 0, nr_unlink = 0; if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) return 0; @@ -3578,12 +3450,31 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) key.offset = found_key.objectid - 1; continue; } + } + /* - * Inode is already gone but the orphan item is still there, - * kill the orphan item. + * If we have an inode with links, there are a couple of + * possibilities. Old kernels (before v3.12) used to create an + * orphan item for truncate indicating that there were possibly + * extent items past i_size that needed to be deleted. In v3.12, + * truncate was changed to update i_size in sync with the extent + * items, but the (useless) orphan item was still created. Since + * v4.18, we don't create the orphan item for truncate at all. + * + * So, this item could mean that we need to do a truncate, but + * only if this filesystem was last used on a pre-v3.12 kernel + * and was not cleanly unmounted. The odds of that are quite + * slim, and it's a pain to do the truncate now, so just delete + * the orphan item. + * + * It's also possible that this orphan item was supposed to be + * deleted but wasn't. The inode number may have been reused, + * but either way, we can delete the orphan item. */ - if (ret == -ENOENT) { + if (ret == -ENOENT || inode->i_nlink) { + if (!ret) + iput(inode); trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { ret = PTR_ERR(trans); @@ -3599,42 +3490,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) continue; } - /* - * add this inode to the orphan list so btrfs_orphan_del does - * the proper thing when we hit it - */ - set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, - &BTRFS_I(inode)->runtime_flags); - atomic_inc(&root->orphan_inodes); - - /* if we have links, this was a truncate, lets do that */ - if (inode->i_nlink) { - if (WARN_ON(!S_ISREG(inode->i_mode))) { - iput(inode); - continue; - } - nr_truncate++; - - /* 1 for the orphan item deletion. */ - trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) { - iput(inode); - ret = PTR_ERR(trans); - goto out; - } - ret = btrfs_orphan_add(trans, BTRFS_I(inode)); - btrfs_end_transaction(trans); - if (ret) { - iput(inode); - goto out; - } - - ret = btrfs_truncate(inode, false); - if (ret) - btrfs_orphan_del(NULL, BTRFS_I(inode)); - } else { - nr_unlink++; - } + nr_unlink++; /* this will do delete_inode and everything for us */ iput(inode); @@ -3646,12 +3502,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; - if (root->orphan_block_rsv) - btrfs_block_rsv_release(fs_info, root->orphan_block_rsv, - (u64)-1); - - if (root->orphan_block_rsv || - test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state)) { + if (test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state)) { trans = btrfs_join_transaction(root); if (!IS_ERR(trans)) btrfs_end_transaction(trans); @@ -3659,8 +3510,6 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) if (nr_unlink) btrfs_debug(fs_info, "unlinked %d orphans", nr_unlink); - if (nr_truncate) - btrfs_debug(fs_info, "truncated %d orphans", nr_truncate); out: if (ret) @@ -3923,7 +3772,7 @@ cache_acl: break; } - btrfs_update_iflags(inode); + btrfs_sync_inode_flags_to_i_flags(inode); return 0; make_bad: @@ -4237,7 +4086,7 @@ out: return ret; } -int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, +static int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *dir, u64 objectid, const char *name, int name_len) @@ -4318,6 +4167,262 @@ out: return ret; } +/* + * Helper to check if the subvolume references other subvolumes or if it's + * default. + */ +static noinline int may_destroy_subvol(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_path *path; + struct btrfs_dir_item *di; + struct btrfs_key key; + u64 dir_id; + int ret; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + /* Make sure this root isn't set as the default subvol */ + dir_id = btrfs_super_root_dir(fs_info->super_copy); + di = btrfs_lookup_dir_item(NULL, fs_info->tree_root, path, + dir_id, "default", 7, 0); + if (di && !IS_ERR(di)) { + btrfs_dir_item_key_to_cpu(path->nodes[0], di, &key); + if (key.objectid == root->root_key.objectid) { + ret = -EPERM; + btrfs_err(fs_info, + "deleting default subvolume %llu is not allowed", + key.objectid); + goto out; + } + btrfs_release_path(path); + } + + key.objectid = root->root_key.objectid; + key.type = BTRFS_ROOT_REF_KEY; + key.offset = (u64)-1; + + ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, path, 0, 0); + if (ret < 0) + goto out; + BUG_ON(ret == 0); + + ret = 0; + if (path->slots[0] > 0) { + path->slots[0]--; + btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); + if (key.objectid == root->root_key.objectid && + key.type == BTRFS_ROOT_REF_KEY) + ret = -ENOTEMPTY; + } +out: + btrfs_free_path(path); + return ret; +} + +/* Delete all dentries for inodes belonging to the root */ +static void btrfs_prune_dentries(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct rb_node *node; + struct rb_node *prev; + struct btrfs_inode *entry; + struct inode *inode; + u64 objectid = 0; + + if (!test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) + WARN_ON(btrfs_root_refs(&root->root_item) != 0); + + spin_lock(&root->inode_lock); +again: + node = root->inode_tree.rb_node; + prev = NULL; + while (node) { + prev = node; + entry = rb_entry(node, struct btrfs_inode, rb_node); + + if (objectid < btrfs_ino(BTRFS_I(&entry->vfs_inode))) + node = node->rb_left; + else if (objectid > btrfs_ino(BTRFS_I(&entry->vfs_inode))) + node = node->rb_right; + else + break; + } + if (!node) { + while (prev) { + entry = rb_entry(prev, struct btrfs_inode, rb_node); + if (objectid <= btrfs_ino(BTRFS_I(&entry->vfs_inode))) { + node = prev; + break; + } + prev = rb_next(prev); + } + } + while (node) { + entry = rb_entry(node, struct btrfs_inode, rb_node); + objectid = btrfs_ino(BTRFS_I(&entry->vfs_inode)) + 1; + inode = igrab(&entry->vfs_inode); + if (inode) { + spin_unlock(&root->inode_lock); + if (atomic_read(&inode->i_count) > 1) + d_prune_aliases(inode); + /* + * btrfs_drop_inode will have it removed from the inode + * cache when its usage count hits zero. + */ + iput(inode); + cond_resched(); + spin_lock(&root->inode_lock); + goto again; + } + + if (cond_resched_lock(&root->inode_lock)) + goto again; + + node = rb_next(node); + } + spin_unlock(&root->inode_lock); +} + +int btrfs_delete_subvolume(struct inode *dir, struct dentry *dentry) +{ + struct btrfs_fs_info *fs_info = btrfs_sb(dentry->d_sb); + struct btrfs_root *root = BTRFS_I(dir)->root; + struct inode *inode = d_inode(dentry); + struct btrfs_root *dest = BTRFS_I(inode)->root; + struct btrfs_trans_handle *trans; + struct btrfs_block_rsv block_rsv; + u64 root_flags; + int ret; + int err; + + /* + * Don't allow to delete a subvolume with send in progress. This is + * inside the inode lock so the error handling that has to drop the bit + * again is not run concurrently. + */ + spin_lock(&dest->root_item_lock); + root_flags = btrfs_root_flags(&dest->root_item); + if (dest->send_in_progress == 0) { + btrfs_set_root_flags(&dest->root_item, + root_flags | BTRFS_ROOT_SUBVOL_DEAD); + spin_unlock(&dest->root_item_lock); + } else { + spin_unlock(&dest->root_item_lock); + btrfs_warn(fs_info, + "attempt to delete subvolume %llu during send", + dest->root_key.objectid); + return -EPERM; + } + + down_write(&fs_info->subvol_sem); + + err = may_destroy_subvol(dest); + if (err) + goto out_up_write; + + btrfs_init_block_rsv(&block_rsv, BTRFS_BLOCK_RSV_TEMP); + /* + * One for dir inode, + * two for dir entries, + * two for root ref/backref. + */ + err = btrfs_subvolume_reserve_metadata(root, &block_rsv, 5, true); + if (err) + goto out_up_write; + + trans = btrfs_start_transaction(root, 0); + if (IS_ERR(trans)) { + err = PTR_ERR(trans); + goto out_release; + } + trans->block_rsv = &block_rsv; + trans->bytes_reserved = block_rsv.size; + + btrfs_record_snapshot_destroy(trans, BTRFS_I(dir)); + + ret = btrfs_unlink_subvol(trans, root, dir, + dest->root_key.objectid, + dentry->d_name.name, + dentry->d_name.len); + if (ret) { + err = ret; + btrfs_abort_transaction(trans, ret); + goto out_end_trans; + } + + btrfs_record_root_in_trans(trans, dest); + + memset(&dest->root_item.drop_progress, 0, + sizeof(dest->root_item.drop_progress)); + dest->root_item.drop_level = 0; + btrfs_set_root_refs(&dest->root_item, 0); + + if (!test_and_set_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &dest->state)) { + ret = btrfs_insert_orphan_item(trans, + fs_info->tree_root, + dest->root_key.objectid); + if (ret) { + btrfs_abort_transaction(trans, ret); + err = ret; + goto out_end_trans; + } + } + + ret = btrfs_uuid_tree_remove(trans, dest->root_item.uuid, + BTRFS_UUID_KEY_SUBVOL, + dest->root_key.objectid); + if (ret && ret != -ENOENT) { + btrfs_abort_transaction(trans, ret); + err = ret; + goto out_end_trans; + } + if (!btrfs_is_empty_uuid(dest->root_item.received_uuid)) { + ret = btrfs_uuid_tree_remove(trans, + dest->root_item.received_uuid, + BTRFS_UUID_KEY_RECEIVED_SUBVOL, + dest->root_key.objectid); + if (ret && ret != -ENOENT) { + btrfs_abort_transaction(trans, ret); + err = ret; + goto out_end_trans; + } + } + +out_end_trans: + trans->block_rsv = NULL; + trans->bytes_reserved = 0; + ret = btrfs_end_transaction(trans); + if (ret && !err) + err = ret; + inode->i_flags |= S_DEAD; +out_release: + btrfs_subvolume_release_metadata(fs_info, &block_rsv); +out_up_write: + up_write(&fs_info->subvol_sem); + if (err) { + spin_lock(&dest->root_item_lock); + root_flags = btrfs_root_flags(&dest->root_item); + btrfs_set_root_flags(&dest->root_item, + root_flags & ~BTRFS_ROOT_SUBVOL_DEAD); + spin_unlock(&dest->root_item_lock); + } else { + d_invalidate(dentry); + btrfs_prune_dentries(dest); + ASSERT(dest->send_in_progress == 0); + + /* the last ref */ + if (dest->ino_cache_inode) { + iput(dest->ino_cache_inode); + dest->ino_cache_inode = NULL; + } + } + + return err; +} + static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) { struct inode *inode = d_inode(dentry); @@ -4329,7 +4434,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) if (inode->i_size > BTRFS_EMPTY_DIR_SIZE) return -ENOTEMPTY; if (btrfs_ino(BTRFS_I(inode)) == BTRFS_FIRST_FREE_OBJECTID) - return -EPERM; + return btrfs_delete_subvolume(dir, dentry); trans = __unlink_start_trans(dir); if (IS_ERR(trans)) @@ -4441,7 +4546,6 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, int pending_del_slot = 0; int extent_type = -1; int ret; - int err = 0; u64 ino = btrfs_ino(BTRFS_I(inode)); u64 bytes_deleted = 0; bool be_nice = false; @@ -4493,22 +4597,19 @@ search_again: * up a huge file in a single leaf. Most of the time that * bytes_deleted is > 0, it will be huge by the time we get here */ - if (be_nice && bytes_deleted > SZ_32M) { - if (btrfs_should_end_transaction(trans)) { - err = -EAGAIN; - goto error; - } + if (be_nice && bytes_deleted > SZ_32M && + btrfs_should_end_transaction(trans)) { + ret = -EAGAIN; + goto out; } - path->leave_spinning = 1; ret = btrfs_search_slot(trans, root, &key, path, -1, 1); - if (ret < 0) { - err = ret; + if (ret < 0) goto out; - } if (ret > 0) { + ret = 0; /* there are no items in the tree for us to truncate, we're * done */ @@ -4619,7 +4720,7 @@ search_again: * We have to bail so the last_size is set to * just before this extent. */ - err = NEED_TRUNCATE_BLOCK; + ret = NEED_TRUNCATE_BLOCK; break; } @@ -4658,7 +4759,10 @@ delete: extent_num_bytes, 0, btrfs_header_owner(leaf), ino, extent_offset); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, ret); + break; + } if (btrfs_should_throttle_delayed_refs(trans, fs_info)) btrfs_async_run_delayed_refs(fs_info, trans->delayed_ref_updates * 2, @@ -4686,7 +4790,7 @@ delete: pending_del_nr); if (ret) { btrfs_abort_transaction(trans, ret); - goto error; + break; } pending_del_nr = 0; } @@ -4697,8 +4801,8 @@ delete: trans->delayed_ref_updates = 0; ret = btrfs_run_delayed_refs(trans, updates * 2); - if (ret && !err) - err = ret; + if (ret) + break; } } /* @@ -4706,8 +4810,8 @@ delete: * and let the transaction restart */ if (should_end) { - err = -EAGAIN; - goto error; + ret = -EAGAIN; + break; } goto search_again; } else { @@ -4715,32 +4819,37 @@ delete: } } out: - if (pending_del_nr) { - ret = btrfs_del_items(trans, root, path, pending_del_slot, + if (ret >= 0 && pending_del_nr) { + int err; + + err = btrfs_del_items(trans, root, path, pending_del_slot, pending_del_nr); - if (ret) - btrfs_abort_transaction(trans, ret); + if (err) { + btrfs_abort_transaction(trans, err); + ret = err; + } } -error: if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { ASSERT(last_size >= new_size); - if (!err && last_size > new_size) + if (!ret && last_size > new_size) last_size = new_size; btrfs_ordered_update_i_size(inode, last_size, NULL); } btrfs_free_path(path); - if (be_nice && bytes_deleted > SZ_32M) { + if (be_nice && bytes_deleted > SZ_32M && (ret >= 0 || ret == -EAGAIN)) { unsigned long updates = trans->delayed_ref_updates; + int err; + if (updates) { trans->delayed_ref_updates = 0; - ret = btrfs_run_delayed_refs(trans, updates * 2); - if (ret && !err) - err = ret; + err = btrfs_run_delayed_refs(trans, updates * 2); + if (err) + ret = err; } } - return err; + return ret; } /* @@ -5082,30 +5191,6 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr) set_bit(BTRFS_INODE_ORDERED_DATA_CLOSE, &BTRFS_I(inode)->runtime_flags); - /* - * 1 for the orphan item we're going to add - * 1 for the orphan item deletion. - */ - trans = btrfs_start_transaction(root, 2); - if (IS_ERR(trans)) - return PTR_ERR(trans); - - /* - * We need to do this in case we fail at _any_ point during the - * actual truncate. Once we do the truncate_setsize we could - * invalidate pages which forces any outstanding ordered io to - * be instantly completed which will give us extents that need - * to be truncated. If we fail to get an orphan inode down we - * could have left over extents that were never meant to live, - * so we need to guarantee from this point on that everything - * will be consistent. - */ - ret = btrfs_orphan_add(trans, BTRFS_I(inode)); - btrfs_end_transaction(trans); - if (ret) - return ret; - - /* we don't support swapfiles, so vmtruncate shouldn't fail */ truncate_setsize(inode, newsize); /* Disable nonlocked read DIO to avoid the end less truncate */ @@ -5117,29 +5202,16 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr) if (ret && inode->i_nlink) { int err; - /* To get a stable disk_i_size */ - err = btrfs_wait_ordered_range(inode, 0, (u64)-1); - if (err) { - btrfs_orphan_del(NULL, BTRFS_I(inode)); - return err; - } - /* - * failed to truncate, disk_i_size is only adjusted down - * as we remove extents, so it should represent the true - * size of the inode, so reset the in memory size and - * delete our orphan entry. + * Truncate failed, so fix up the in-memory size. We + * adjusted disk_i_size down as we removed extents, so + * wait for disk_i_size to be stable and then update the + * in-memory size to match. */ - trans = btrfs_join_transaction(root); - if (IS_ERR(trans)) { - btrfs_orphan_del(NULL, BTRFS_I(inode)); - return ret; - } - i_size_write(inode, BTRFS_I(inode)->disk_i_size); - err = btrfs_orphan_del(trans, BTRFS_I(inode)); + err = btrfs_wait_ordered_range(inode, 0, (u64)-1); if (err) - btrfs_abort_transaction(trans, err); - btrfs_end_transaction(trans); + return err; + i_size_write(inode, BTRFS_I(inode)->disk_i_size); } } @@ -5269,13 +5341,52 @@ static void evict_inode_truncate_pages(struct inode *inode) spin_unlock(&io_tree->lock); } +static struct btrfs_trans_handle *evict_refill_and_join(struct btrfs_root *root, + struct btrfs_block_rsv *rsv, + u64 min_size) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; + int failures = 0; + + for (;;) { + struct btrfs_trans_handle *trans; + int ret; + + ret = btrfs_block_rsv_refill(root, rsv, min_size, + BTRFS_RESERVE_FLUSH_LIMIT); + + if (ret && ++failures > 2) { + btrfs_warn(fs_info, + "could not allocate space for a delete; will truncate on mount"); + return ERR_PTR(-ENOSPC); + } + + trans = btrfs_join_transaction(root); + if (IS_ERR(trans) || !ret) + return trans; + + /* + * Try to steal from the global reserve if there is space for + * it. + */ + if (!btrfs_check_space_for_delayed_refs(trans, fs_info) && + !btrfs_block_rsv_migrate(global_rsv, rsv, min_size, 0)) + return trans; + + /* If not, commit and try again. */ + ret = btrfs_commit_transaction(trans); + if (ret) + return ERR_PTR(ret); + } +} + void btrfs_evict_inode(struct inode *inode) { struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_block_rsv *rsv, *global_rsv; - int steal_from_global = 0; + struct btrfs_block_rsv *rsv; u64 min_size; int ret; @@ -5296,21 +5407,16 @@ void btrfs_evict_inode(struct inode *inode) btrfs_is_free_space_inode(BTRFS_I(inode)))) goto no_delete; - if (is_bad_inode(inode)) { - btrfs_orphan_del(NULL, BTRFS_I(inode)); + if (is_bad_inode(inode)) goto no_delete; - } /* do we really want it for ->i_nlink > 0 and zero btrfs_root_refs? */ if (!special_file(inode->i_mode)) btrfs_wait_ordered_range(inode, 0, (u64)-1); btrfs_free_io_failure_record(BTRFS_I(inode), 0, (u64)-1); - if (test_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags)) { - BUG_ON(test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, - &BTRFS_I(inode)->runtime_flags)); + if (test_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags)) goto no_delete; - } if (inode->i_nlink > 0) { BUG_ON(btrfs_root_refs(&root->root_item) != 0 && @@ -5319,130 +5425,63 @@ void btrfs_evict_inode(struct inode *inode) } ret = btrfs_commit_inode_delayed_inode(BTRFS_I(inode)); - if (ret) { - btrfs_orphan_del(NULL, BTRFS_I(inode)); + if (ret) goto no_delete; - } rsv = btrfs_alloc_block_rsv(fs_info, BTRFS_BLOCK_RSV_TEMP); - if (!rsv) { - btrfs_orphan_del(NULL, BTRFS_I(inode)); + if (!rsv) goto no_delete; - } rsv->size = min_size; rsv->failfast = 1; - global_rsv = &fs_info->global_block_rsv; btrfs_i_size_write(BTRFS_I(inode), 0); - /* - * This is a bit simpler than btrfs_truncate since we've already - * reserved our space for our orphan item in the unlink, so we just - * need to reserve some slack space in case we add bytes and update - * inode item when doing the truncate. - */ while (1) { - ret = btrfs_block_rsv_refill(root, rsv, min_size, - BTRFS_RESERVE_FLUSH_LIMIT); - - /* - * Try and steal from the global reserve since we will - * likely not use this space anyway, we want to try as - * hard as possible to get this to work. - */ - if (ret) - steal_from_global++; - else - steal_from_global = 0; - ret = 0; - - /* - * steal_from_global == 0: we reserved stuff, hooray! - * steal_from_global == 1: we didn't reserve stuff, boo! - * steal_from_global == 2: we've committed, still not a lot of - * room but maybe we'll have room in the global reserve this - * time. - * steal_from_global == 3: abandon all hope! - */ - if (steal_from_global > 2) { - btrfs_warn(fs_info, - "Could not get space for a delete, will truncate on mount %d", - ret); - btrfs_orphan_del(NULL, BTRFS_I(inode)); - btrfs_free_block_rsv(fs_info, rsv); - goto no_delete; - } - - trans = btrfs_join_transaction(root); - if (IS_ERR(trans)) { - btrfs_orphan_del(NULL, BTRFS_I(inode)); - btrfs_free_block_rsv(fs_info, rsv); - goto no_delete; - } - - /* - * We can't just steal from the global reserve, we need to make - * sure there is room to do it, if not we need to commit and try - * again. - */ - if (steal_from_global) { - if (!btrfs_check_space_for_delayed_refs(trans, fs_info)) - ret = btrfs_block_rsv_migrate(global_rsv, rsv, - min_size, 0); - else - ret = -ENOSPC; - } - - /* - * Couldn't steal from the global reserve, we have too much - * pending stuff built up, commit the transaction and try it - * again. - */ - if (ret) { - ret = btrfs_commit_transaction(trans); - if (ret) { - btrfs_orphan_del(NULL, BTRFS_I(inode)); - btrfs_free_block_rsv(fs_info, rsv); - goto no_delete; - } - continue; - } else { - steal_from_global = 0; - } + trans = evict_refill_and_join(root, rsv, min_size); + if (IS_ERR(trans)) + goto free_rsv; trans->block_rsv = rsv; ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0); - if (ret != -ENOSPC && ret != -EAGAIN) - break; - trans->block_rsv = &fs_info->trans_block_rsv; btrfs_end_transaction(trans); - trans = NULL; btrfs_btree_balance_dirty(fs_info); + if (ret && ret != -ENOSPC && ret != -EAGAIN) + goto free_rsv; + else if (!ret) + break; } - btrfs_free_block_rsv(fs_info, rsv); - /* - * Errors here aren't a big deal, it just means we leave orphan items - * in the tree. They will be cleaned up on the next mount. + * Errors here aren't a big deal, it just means we leave orphan items in + * the tree. They will be cleaned up on the next mount. If the inode + * number gets reused, cleanup deletes the orphan item without doing + * anything, and unlink reuses the existing orphan item. + * + * If it turns out that we are dropping too many of these, we might want + * to add a mechanism for retrying these after a commit. */ - if (ret == 0) { - trans->block_rsv = root->orphan_block_rsv; + trans = evict_refill_and_join(root, rsv, min_size); + if (!IS_ERR(trans)) { + trans->block_rsv = rsv; btrfs_orphan_del(trans, BTRFS_I(inode)); - } else { - btrfs_orphan_del(NULL, BTRFS_I(inode)); + trans->block_rsv = &fs_info->trans_block_rsv; + btrfs_end_transaction(trans); } - trans->block_rsv = &fs_info->trans_block_rsv; if (!(root == fs_info->tree_root || root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) btrfs_return_ino(root, btrfs_ino(BTRFS_I(inode))); - btrfs_end_transaction(trans); - btrfs_btree_balance_dirty(fs_info); +free_rsv: + btrfs_free_block_rsv(fs_info, rsv); no_delete: + /* + * If we didn't successfully delete, the orphan item will still be in + * the tree and we'll retry on the next mount. Again, we might also want + * to retry these periodically in the future. + */ btrfs_remove_delayed_node(BTRFS_I(inode)); clear_inode(inode); } @@ -5618,69 +5657,6 @@ static void inode_tree_del(struct inode *inode) } } -void btrfs_invalidate_inodes(struct btrfs_root *root) -{ - struct btrfs_fs_info *fs_info = root->fs_info; - struct rb_node *node; - struct rb_node *prev; - struct btrfs_inode *entry; - struct inode *inode; - u64 objectid = 0; - - if (!test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) - WARN_ON(btrfs_root_refs(&root->root_item) != 0); - - spin_lock(&root->inode_lock); -again: - node = root->inode_tree.rb_node; - prev = NULL; - while (node) { - prev = node; - entry = rb_entry(node, struct btrfs_inode, rb_node); - - if (objectid < btrfs_ino(BTRFS_I(&entry->vfs_inode))) - node = node->rb_left; - else if (objectid > btrfs_ino(BTRFS_I(&entry->vfs_inode))) - node = node->rb_right; - else - break; - } - if (!node) { - while (prev) { - entry = rb_entry(prev, struct btrfs_inode, rb_node); - if (objectid <= btrfs_ino(BTRFS_I(&entry->vfs_inode))) { - node = prev; - break; - } - prev = rb_next(prev); - } - } - while (node) { - entry = rb_entry(node, struct btrfs_inode, rb_node); - objectid = btrfs_ino(BTRFS_I(&entry->vfs_inode)) + 1; - inode = igrab(&entry->vfs_inode); - if (inode) { - spin_unlock(&root->inode_lock); - if (atomic_read(&inode->i_count) > 1) - d_prune_aliases(inode); - /* - * btrfs_drop_inode will have it removed from - * the inode cache when its usage count - * hits zero. - */ - iput(inode); - cond_resched(); - spin_lock(&root->inode_lock); - goto again; - } - - if (cond_resched_lock(&root->inode_lock)) - goto again; - - node = rb_next(node); - } - spin_unlock(&root->inode_lock); -} static int btrfs_init_locked_inode(struct inode *inode, void *p) { @@ -5842,11 +5818,6 @@ static int btrfs_dentry_delete(const struct dentry *dentry) return 0; } -static void btrfs_dentry_release(struct dentry *dentry) -{ - kfree(dentry->d_fsdata); -} - static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags) { @@ -5905,11 +5876,13 @@ static int btrfs_filldir(void *addr, int entries, struct dir_context *ctx) struct dir_entry *entry = addr; char *name = (char *)(entry + 1); - ctx->pos = entry->offset; - if (!dir_emit(ctx, name, entry->name_len, entry->ino, - entry->type)) + ctx->pos = get_unaligned(&entry->offset); + if (!dir_emit(ctx, name, get_unaligned(&entry->name_len), + get_unaligned(&entry->ino), + get_unaligned(&entry->type))) return 1; - addr += sizeof(struct dir_entry) + entry->name_len; + addr += sizeof(struct dir_entry) + + get_unaligned(&entry->name_len); ctx->pos++; } return 0; @@ -5999,14 +5972,15 @@ again: } entry = addr; - entry->name_len = name_len; + put_unaligned(name_len, &entry->name_len); name_ptr = (char *)(entry + 1); read_extent_buffer(leaf, name_ptr, (unsigned long)(di + 1), name_len); - entry->type = btrfs_filetype_table[btrfs_dir_type(leaf, di)]; + put_unaligned(btrfs_filetype_table[btrfs_dir_type(leaf, di)], + &entry->type); btrfs_dir_item_key_to_cpu(leaf, di, &location); - entry->ino = location.objectid; - entry->offset = found_key.offset; + put_unaligned(location.objectid, &entry->ino); + put_unaligned(found_key.offset, &entry->offset); entries++; addr += sizeof(struct dir_entry) + name_len; total_len += sizeof(struct dir_entry) + name_len; @@ -6259,7 +6233,7 @@ static void btrfs_inherit_iflags(struct inode *inode, struct inode *dir) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; } - btrfs_update_iflags(inode); + btrfs_sync_inode_flags_to_i_flags(inode); } static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, @@ -6575,8 +6549,7 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, goto out_unlock_inode; } else { btrfs_update_inode(trans, root, inode); - unlock_new_inode(inode); - d_instantiate(dentry, inode); + d_instantiate_new(dentry, inode); } out_unlock: @@ -6652,8 +6625,7 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, goto out_unlock_inode; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; - unlock_new_inode(inode); - d_instantiate(dentry, inode); + d_instantiate_new(dentry, inode); out_unlock: btrfs_end_transaction(trans); @@ -6696,8 +6668,9 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, * 2 items for inode and inode ref * 2 items for dir items * 1 item for parent inode + * 1 item for orphan item deletion if O_TMPFILE */ - trans = btrfs_start_transaction(root, 5); + trans = btrfs_start_transaction(root, inode->i_nlink ? 5 : 6); if (IS_ERR(trans)) { err = PTR_ERR(trans); trans = NULL; @@ -6798,12 +6771,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) if (err) goto out_fail_inode; - d_instantiate(dentry, inode); - /* - * mkdir is special. We're unlocking after we call d_instantiate - * to avoid a race with nfsd calling d_instantiate. - */ - unlock_new_inode(inode); + d_instantiate_new(dentry, inode); drop_on_err = 0; out_fail: @@ -7079,7 +7047,7 @@ insert: err = 0; write_lock(&em_tree->lock); - err = btrfs_add_extent_mapping(em_tree, &em, start, len); + err = btrfs_add_extent_mapping(fs_info, em_tree, &em, start, len); write_unlock(&em_tree->lock); out: @@ -7364,6 +7332,14 @@ noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len, btrfs_file_extent_other_encoding(leaf, fi)) goto out; + /* + * Do the same check as in btrfs_cross_ref_exist but without the + * unnecessary search. + */ + if (btrfs_file_extent_generation(leaf, fi) <= + btrfs_root_last_snapshot(&root->root_item)) + goto out; + backref_offset = btrfs_file_extent_offset(leaf, fi); if (orig_start) { @@ -7564,6 +7540,125 @@ static struct extent_map *create_io_em(struct inode *inode, u64 start, u64 len, return em; } + +static int btrfs_get_blocks_direct_read(struct extent_map *em, + struct buffer_head *bh_result, + struct inode *inode, + u64 start, u64 len) +{ + if (em->block_start == EXTENT_MAP_HOLE || + test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) + return -ENOENT; + + len = min(len, em->len - (start - em->start)); + + bh_result->b_blocknr = (em->block_start + (start - em->start)) >> + inode->i_blkbits; + bh_result->b_size = len; + bh_result->b_bdev = em->bdev; + set_buffer_mapped(bh_result); + + return 0; +} + +static int btrfs_get_blocks_direct_write(struct extent_map **map, + struct buffer_head *bh_result, + struct inode *inode, + struct btrfs_dio_data *dio_data, + u64 start, u64 len) +{ + struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); + struct extent_map *em = *map; + int ret = 0; + + /* + * We don't allocate a new extent in the following cases + * + * 1) The inode is marked as NODATACOW. In this case we'll just use the + * existing extent. + * 2) The extent is marked as PREALLOC. We're good to go here and can + * just use the extent. + * + */ + if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) || + ((BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) && + em->block_start != EXTENT_MAP_HOLE)) { + int type; + u64 block_start, orig_start, orig_block_len, ram_bytes; + + if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) + type = BTRFS_ORDERED_PREALLOC; + else + type = BTRFS_ORDERED_NOCOW; + len = min(len, em->len - (start - em->start)); + block_start = em->block_start + (start - em->start); + + if (can_nocow_extent(inode, start, &len, &orig_start, + &orig_block_len, &ram_bytes) == 1 && + btrfs_inc_nocow_writers(fs_info, block_start)) { + struct extent_map *em2; + + em2 = btrfs_create_dio_extent(inode, start, len, + orig_start, block_start, + len, orig_block_len, + ram_bytes, type); + btrfs_dec_nocow_writers(fs_info, block_start); + if (type == BTRFS_ORDERED_PREALLOC) { + free_extent_map(em); + *map = em = em2; + } + + if (em2 && IS_ERR(em2)) { + ret = PTR_ERR(em2); + goto out; + } + /* + * For inode marked NODATACOW or extent marked PREALLOC, + * use the existing or preallocated extent, so does not + * need to adjust btrfs_space_info's bytes_may_use. + */ + btrfs_free_reserved_data_space_noquota(inode, start, + len); + goto skip_cow; + } + } + + /* this will cow the extent */ + len = bh_result->b_size; + free_extent_map(em); + *map = em = btrfs_new_extent_direct(inode, start, len); + if (IS_ERR(em)) { + ret = PTR_ERR(em); + goto out; + } + + len = min(len, em->len - (start - em->start)); + +skip_cow: + bh_result->b_blocknr = (em->block_start + (start - em->start)) >> + inode->i_blkbits; + bh_result->b_size = len; + bh_result->b_bdev = em->bdev; + set_buffer_mapped(bh_result); + + if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) + set_buffer_new(bh_result); + + /* + * Need to update the i_size under the extent lock so buffered + * readers will get the updated i_size when we unlock. + */ + if (!dio_data->overwrite && start + len > i_size_read(inode)) + i_size_write(inode, start + len); + + WARN_ON(dio_data->reserve < len); + dio_data->reserve -= len; + dio_data->unsubmitted_oe_range_end = start + len; + current->journal_info = dio_data; +out: + return ret; +} + static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create) { @@ -7632,116 +7727,36 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, goto unlock_err; } - /* Just a good old fashioned hole, return */ - if (!create && (em->block_start == EXTENT_MAP_HOLE || - test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) { - free_extent_map(em); - goto unlock_err; - } - - /* - * We don't allocate a new extent in the following cases - * - * 1) The inode is marked as NODATACOW. In this case we'll just use the - * existing extent. - * 2) The extent is marked as PREALLOC. We're good to go here and can - * just use the extent. - * - */ - if (!create) { - len = min(len, em->len - (start - em->start)); - lockstart = start + len; - goto unlock; - } - - if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) || - ((BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) && - em->block_start != EXTENT_MAP_HOLE)) { - int type; - u64 block_start, orig_start, orig_block_len, ram_bytes; - - if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) - type = BTRFS_ORDERED_PREALLOC; - else - type = BTRFS_ORDERED_NOCOW; - len = min(len, em->len - (start - em->start)); - block_start = em->block_start + (start - em->start); - - if (can_nocow_extent(inode, start, &len, &orig_start, - &orig_block_len, &ram_bytes) == 1 && - btrfs_inc_nocow_writers(fs_info, block_start)) { - struct extent_map *em2; - - em2 = btrfs_create_dio_extent(inode, start, len, - orig_start, block_start, - len, orig_block_len, - ram_bytes, type); - btrfs_dec_nocow_writers(fs_info, block_start); - if (type == BTRFS_ORDERED_PREALLOC) { - free_extent_map(em); - em = em2; - } - if (em2 && IS_ERR(em2)) { - ret = PTR_ERR(em2); - goto unlock_err; - } - /* - * For inode marked NODATACOW or extent marked PREALLOC, - * use the existing or preallocated extent, so does not - * need to adjust btrfs_space_info's bytes_may_use. - */ - btrfs_free_reserved_data_space_noquota(inode, - start, len); - goto unlock; - } - } - - /* - * this will cow the extent, reset the len in case we changed - * it above - */ - len = bh_result->b_size; - free_extent_map(em); - em = btrfs_new_extent_direct(inode, start, len); - if (IS_ERR(em)) { - ret = PTR_ERR(em); - goto unlock_err; - } - len = min(len, em->len - (start - em->start)); -unlock: - bh_result->b_blocknr = (em->block_start + (start - em->start)) >> - inode->i_blkbits; - bh_result->b_size = len; - bh_result->b_bdev = em->bdev; - set_buffer_mapped(bh_result); if (create) { - if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) - set_buffer_new(bh_result); + ret = btrfs_get_blocks_direct_write(&em, bh_result, inode, + dio_data, start, len); + if (ret < 0) + goto unlock_err; + /* clear and unlock the entire range */ + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, + unlock_bits, 1, 0, &cached_state); + } else { + ret = btrfs_get_blocks_direct_read(em, bh_result, inode, + start, len); + /* Can be negative only if we read from a hole */ + if (ret < 0) { + ret = 0; + free_extent_map(em); + goto unlock_err; + } /* - * Need to update the i_size under the extent lock so buffered - * readers will get the updated i_size when we unlock. + * We need to unlock only the end area that we aren't using. + * The rest is going to be unlocked by the endio routine. */ - if (!dio_data->overwrite && start + len > i_size_read(inode)) - i_size_write(inode, start + len); - - WARN_ON(dio_data->reserve < len); - dio_data->reserve -= len; - dio_data->unsubmitted_oe_range_end = start + len; - current->journal_info = dio_data; - } - - /* - * In the case of write we need to clear and unlock the entire range, - * in the case of read we need to unlock only the end area that we - * aren't using if there is any left over space. - */ - if (lockstart < lockend) { - clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, - lockend, unlock_bits, 1, 0, - &cached_state); - } else { - free_extent_state(cached_state); + lockstart = start + bh_result->b_size; + if (lockstart < lockend) { + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, + lockend, unlock_bits, 1, 0, + &cached_state); + } else { + free_extent_state(cached_state); + } } free_extent_map(em); @@ -8127,7 +8142,6 @@ static void __endio_write_update_ordered(struct inode *inode, u64 ordered_offset = offset; u64 ordered_bytes = bytes; u64 last_offset; - int ret; if (btrfs_is_free_space_inode(BTRFS_I(inode))) { wq = fs_info->endio_freespace_worker; @@ -8137,32 +8151,31 @@ static void __endio_write_update_ordered(struct inode *inode, func = btrfs_endio_write_helper; } -again: - last_offset = ordered_offset; - ret = btrfs_dec_test_first_ordered_pending(inode, &ordered, - &ordered_offset, - ordered_bytes, - uptodate); - if (!ret) - goto out_test; - - btrfs_init_work(&ordered->work, func, finish_ordered_fn, NULL, NULL); - btrfs_queue_work(wq, &ordered->work); -out_test: - /* - * If btrfs_dec_test_ordered_pending does not find any ordered extent - * in the range, we can exit. - */ - if (ordered_offset == last_offset) - return; - /* - * our bio might span multiple ordered extents. If we haven't - * completed the accounting for the whole dio, go back and try again - */ - if (ordered_offset < offset + bytes) { - ordered_bytes = offset + bytes - ordered_offset; - ordered = NULL; - goto again; + while (ordered_offset < offset + bytes) { + last_offset = ordered_offset; + if (btrfs_dec_test_first_ordered_pending(inode, &ordered, + &ordered_offset, + ordered_bytes, + uptodate)) { + btrfs_init_work(&ordered->work, func, + finish_ordered_fn, + NULL, NULL); + btrfs_queue_work(wq, &ordered->work); + } + /* + * If btrfs_dec_test_ordered_pending does not find any ordered + * extent in the range, we can exit. + */ + if (ordered_offset == last_offset) + return; + /* + * Our bio might span multiple ordered extents. In this case + * we keep goin until we have accounted the whole dio. + */ + if (ordered_offset < offset + bytes) { + ordered_bytes = offset + bytes - ordered_offset; + ordered = NULL; + } } } @@ -8701,29 +8714,19 @@ static int btrfs_writepage(struct page *page, struct writeback_control *wbc) static int btrfs_writepages(struct address_space *mapping, struct writeback_control *wbc) { - struct extent_io_tree *tree; - - tree = &BTRFS_I(mapping->host)->io_tree; - return extent_writepages(tree, mapping, wbc); + return extent_writepages(mapping, wbc); } static int btrfs_readpages(struct file *file, struct address_space *mapping, struct list_head *pages, unsigned nr_pages) { - struct extent_io_tree *tree; - tree = &BTRFS_I(mapping->host)->io_tree; - return extent_readpages(tree, mapping, pages, nr_pages); + return extent_readpages(mapping, pages, nr_pages); } + static int __btrfs_releasepage(struct page *page, gfp_t gfp_flags) { - struct extent_io_tree *tree; - struct extent_map_tree *map; - int ret; - - tree = &BTRFS_I(page->mapping->host)->io_tree; - map = &BTRFS_I(page->mapping->host)->extent_tree; - ret = try_release_extent_mapping(map, tree, page, gfp_flags); + int ret = try_release_extent_mapping(page, gfp_flags); if (ret == 1) { ClearPagePrivate(page); set_page_private(page, 0); @@ -8864,8 +8867,8 @@ again: * * We are not allowed to take the i_mutex here so we have to play games to * protect against truncate races as the page could now be beyond EOF. Because - * vmtruncate() writes the inode size before removing pages, once we have the - * page lock we can determine safely if the page is beyond EOF. If it is not + * truncate_setsize() writes the inode size before removing pages, once we have + * the page lock we can determine safely if the page is beyond EOF. If it is not * beyond EOF, then the page is guaranteed safe against truncation until we * unlock the page. */ @@ -9027,8 +9030,7 @@ static int btrfs_truncate(struct inode *inode, bool skip_writeback) struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_block_rsv *rsv; - int ret = 0; - int err = 0; + int ret; struct btrfs_trans_handle *trans; u64 mask = fs_info->sectorsize - 1; u64 min_size = btrfs_calc_trunc_metadata_size(fs_info, 1); @@ -9041,39 +9043,31 @@ static int btrfs_truncate(struct inode *inode, bool skip_writeback) } /* - * Yes ladies and gentlemen, this is indeed ugly. The fact is we have - * 3 things going on here - * - * 1) We need to reserve space for our orphan item and the space to - * delete our orphan item. Lord knows we don't want to have a dangling - * orphan item because we didn't reserve space to remove it. + * Yes ladies and gentlemen, this is indeed ugly. We have a couple of + * things going on here: * - * 2) We need to reserve space to update our inode. + * 1) We need to reserve space to update our inode. * - * 3) We need to have something to cache all the space that is going to + * 2) We need to have something to cache all the space that is going to * be free'd up by the truncate operation, but also have some slack * space reserved in case it uses space during the truncate (thank you * very much snapshotting). * - * And we need these to all be separate. The fact is we can use a lot of + * And we need these to be separate. The fact is we can use a lot of * space doing the truncate, and we have no earthly idea how much space * we will use, so we need the truncate reservation to be separate so it - * doesn't end up using space reserved for updating the inode or - * removing the orphan item. We also need to be able to stop the - * transaction and start a new one, which means we need to be able to - * update the inode several times, and we have no idea of knowing how - * many times that will be, so we can't just reserve 1 item for the - * entirety of the operation, so that has to be done separately as well. - * Then there is the orphan item, which does indeed need to be held on - * to for the whole operation, and we need nobody to touch this reserved - * space except the orphan code. + * doesn't end up using space reserved for updating the inode. We also + * need to be able to stop the transaction and start a new one, which + * means we need to be able to update the inode several times, and we + * have no idea of knowing how many times that will be, so we can't just + * reserve 1 item for the entirety of the operation, so that has to be + * done separately as well. * * So that leaves us with * - * 1) root->orphan_block_rsv - for the orphan deletion. - * 2) rsv - for the truncate reservation, which we will steal from the + * 1) rsv - for the truncate reservation, which we will steal from the * transaction reservation. - * 3) fs_info->trans_block_rsv - this will have 1 items worth left for + * 2) fs_info->trans_block_rsv - this will have 1 items worth left for * updating the inode. */ rsv = btrfs_alloc_block_rsv(fs_info, BTRFS_BLOCK_RSV_TEMP); @@ -9088,7 +9082,7 @@ static int btrfs_truncate(struct inode *inode, bool skip_writeback) */ trans = btrfs_start_transaction(root, 2); if (IS_ERR(trans)) { - err = PTR_ERR(trans); + ret = PTR_ERR(trans); goto out; } @@ -9112,23 +9106,19 @@ static int btrfs_truncate(struct inode *inode, bool skip_writeback) inode->i_size, BTRFS_EXTENT_DATA_KEY); trans->block_rsv = &fs_info->trans_block_rsv; - if (ret != -ENOSPC && ret != -EAGAIN) { - err = ret; + if (ret != -ENOSPC && ret != -EAGAIN) break; - } ret = btrfs_update_inode(trans, root, inode); - if (ret) { - err = ret; + if (ret) break; - } btrfs_end_transaction(trans); btrfs_btree_balance_dirty(fs_info); trans = btrfs_start_transaction(root, 2); if (IS_ERR(trans)) { - ret = err = PTR_ERR(trans); + ret = PTR_ERR(trans); trans = NULL; break; } @@ -9161,29 +9151,23 @@ static int btrfs_truncate(struct inode *inode, bool skip_writeback) btrfs_ordered_update_i_size(inode, inode->i_size, NULL); } - if (ret == 0 && inode->i_nlink > 0) { - trans->block_rsv = root->orphan_block_rsv; - ret = btrfs_orphan_del(trans, BTRFS_I(inode)); - if (ret) - err = ret; - } - if (trans) { + int ret2; + trans->block_rsv = &fs_info->trans_block_rsv; - ret = btrfs_update_inode(trans, root, inode); - if (ret && !err) - err = ret; + ret2 = btrfs_update_inode(trans, root, inode); + if (ret2 && !ret) + ret = ret2; - ret = btrfs_end_transaction(trans); + ret2 = btrfs_end_transaction(trans); + if (ret2 && !ret) + ret = ret2; btrfs_btree_balance_dirty(fs_info); } out: btrfs_free_block_rsv(fs_info, rsv); - if (ret && !err) - err = ret; - - return err; + return ret; } /* @@ -9319,13 +9303,6 @@ void btrfs_destroy_inode(struct inode *inode) if (!root) goto free; - if (test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, - &BTRFS_I(inode)->runtime_flags)) { - btrfs_info(fs_info, "inode %llu still on the orphan list", - btrfs_ino(BTRFS_I(inode))); - atomic_dec(&root->orphan_inodes); - } - while (1) { ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1); if (!ordered) @@ -9959,6 +9936,13 @@ static int btrfs_rename2(struct inode *old_dir, struct dentry *old_dentry, return btrfs_rename(old_dir, old_dentry, new_dir, new_dentry, flags); } +struct btrfs_delalloc_work { + struct inode *inode; + struct completion completion; + struct list_head list; + struct btrfs_work work; +}; + static void btrfs_run_delalloc_work(struct btrfs_work *work) { struct btrfs_delalloc_work *delalloc_work; @@ -9972,15 +9956,11 @@ static void btrfs_run_delalloc_work(struct btrfs_work *work) &BTRFS_I(inode)->runtime_flags)) filemap_flush(inode->i_mapping); - if (delalloc_work->delay_iput) - btrfs_add_delayed_iput(inode); - else - iput(inode); + iput(inode); complete(&delalloc_work->completion); } -struct btrfs_delalloc_work *btrfs_alloc_delalloc_work(struct inode *inode, - int delay_iput) +static struct btrfs_delalloc_work *btrfs_alloc_delalloc_work(struct inode *inode) { struct btrfs_delalloc_work *work; @@ -9991,7 +9971,6 @@ struct btrfs_delalloc_work *btrfs_alloc_delalloc_work(struct inode *inode, init_completion(&work->completion); INIT_LIST_HEAD(&work->list); work->inode = inode; - work->delay_iput = delay_iput; WARN_ON_ONCE(!inode); btrfs_init_work(&work->work, btrfs_flush_delalloc_helper, btrfs_run_delalloc_work, NULL, NULL); @@ -9999,18 +9978,11 @@ struct btrfs_delalloc_work *btrfs_alloc_delalloc_work(struct inode *inode, return work; } -void btrfs_wait_and_free_delalloc_work(struct btrfs_delalloc_work *work) -{ - wait_for_completion(&work->completion); - kfree(work); -} - /* * some fairly slow code that needs optimization. This walks the list * of all the inodes with pending delalloc and forces them to disk. */ -static int __start_delalloc_inodes(struct btrfs_root *root, int delay_iput, - int nr) +static int start_delalloc_inodes(struct btrfs_root *root, int nr) { struct btrfs_inode *binode; struct inode *inode; @@ -10038,12 +10010,9 @@ static int __start_delalloc_inodes(struct btrfs_root *root, int delay_iput, } spin_unlock(&root->delalloc_lock); - work = btrfs_alloc_delalloc_work(inode, delay_iput); + work = btrfs_alloc_delalloc_work(inode); if (!work) { - if (delay_iput) - btrfs_add_delayed_iput(inode); - else - iput(inode); + iput(inode); ret = -ENOMEM; goto out; } @@ -10061,10 +10030,11 @@ static int __start_delalloc_inodes(struct btrfs_root *root, int delay_iput, out: list_for_each_entry_safe(work, next, &works, list) { list_del_init(&work->list); - btrfs_wait_and_free_delalloc_work(work); + wait_for_completion(&work->completion); + kfree(work); } - if (!list_empty_careful(&splice)) { + if (!list_empty(&splice)) { spin_lock(&root->delalloc_lock); list_splice_tail(&splice, &root->delalloc_inodes); spin_unlock(&root->delalloc_lock); @@ -10073,7 +10043,7 @@ out: return ret; } -int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput) +int btrfs_start_delalloc_inodes(struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; int ret; @@ -10081,14 +10051,13 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput) if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) return -EROFS; - ret = __start_delalloc_inodes(root, delay_iput, -1); + ret = start_delalloc_inodes(root, -1); if (ret > 0) ret = 0; return ret; } -int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, - int nr) +int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int nr) { struct btrfs_root *root; struct list_head splice; @@ -10111,7 +10080,7 @@ int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, &fs_info->delalloc_roots); spin_unlock(&fs_info->delalloc_root_lock); - ret = __start_delalloc_inodes(root, delay_iput, nr); + ret = start_delalloc_inodes(root, nr); btrfs_put_fs_root(root); if (ret < 0) goto out; @@ -10126,7 +10095,7 @@ int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, ret = 0; out: - if (!list_empty_careful(&splice)) { + if (!list_empty(&splice)) { spin_lock(&fs_info->delalloc_root_lock); list_splice_tail(&splice, &fs_info->delalloc_roots); spin_unlock(&fs_info->delalloc_root_lock); @@ -10246,8 +10215,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, goto out_unlock_inode; } - unlock_new_inode(inode); - d_instantiate(dentry, inode); + d_instantiate_new(dentry, inode); out_unlock: btrfs_end_transaction(trans); @@ -10665,5 +10633,4 @@ static const struct inode_operations btrfs_symlink_inode_operations = { const struct dentry_operations btrfs_dentry_operations = { .d_delete = btrfs_dentry_delete, - .d_release = btrfs_dentry_release, }; |