diff options
Diffstat (limited to 'fs')
215 files changed, 12242 insertions, 8379 deletions
diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile index 9b72dcf..c0ddfd2 100644 --- a/fs/btrfs/Makefile +++ b/fs/btrfs/Makefile @@ -6,5 +6,8 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \ transaction.o inode.o file.o tree-defrag.o \ extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \ extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \ - export.o tree-log.o acl.o free-space-cache.o zlib.o lzo.o \ - compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o + export.o tree-log.o free-space-cache.o zlib.o lzo.o \ + compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \ + reada.o backref.o + +btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c index f66fc99..89b156d 100644 --- a/fs/btrfs/acl.c +++ b/fs/btrfs/acl.c @@ -28,9 +28,7 @@ #include "btrfs_inode.h" #include "xattr.h" -#ifdef CONFIG_BTRFS_FS_POSIX_ACL - -static struct posix_acl *btrfs_get_acl(struct inode *inode, int type) +struct posix_acl *btrfs_get_acl(struct inode *inode, int type) { int size; const char *name; @@ -61,22 +59,19 @@ static struct posix_acl *btrfs_get_acl(struct inode *inode, int type) if (!value) return ERR_PTR(-ENOMEM); size = __btrfs_getxattr(inode, name, value, size); - if (size > 0) { - acl = posix_acl_from_xattr(value, size); - if (IS_ERR(acl)) { - kfree(value); - return acl; - } - set_cached_acl(inode, type, acl); - } - kfree(value); + } + if (size > 0) { + acl = posix_acl_from_xattr(value, size); } else if (size == -ENOENT || size == -ENODATA || size == 0) { /* FIXME, who returns -ENOENT? I think nobody */ acl = NULL; - set_cached_acl(inode, type, acl); } else { acl = ERR_PTR(-EIO); } + kfree(value); + + if (!IS_ERR(acl)) + set_cached_acl(inode, type, acl); return acl; } @@ -111,7 +106,6 @@ static int btrfs_set_acl(struct btrfs_trans_handle *trans, int ret, size = 0; const char *name; char *value = NULL; - mode_t mode; if (acl) { ret = posix_acl_valid(acl); @@ -122,13 +116,11 @@ static int btrfs_set_acl(struct btrfs_trans_handle *trans, switch (type) { case ACL_TYPE_ACCESS: - mode = inode->i_mode; name = POSIX_ACL_XATTR_ACCESS; if (acl) { - ret = posix_acl_equiv_mode(acl, &mode); + ret = posix_acl_equiv_mode(acl, &inode->i_mode); if (ret < 0) return ret; - inode->i_mode = mode; } ret = 0; break; @@ -195,28 +187,6 @@ out: return ret; } -int btrfs_check_acl(struct inode *inode, int mask, unsigned int flags) -{ - int error = -EAGAIN; - - if (flags & IPERM_FLAG_RCU) { - if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) - error = -ECHILD; - - } else { - struct posix_acl *acl; - acl = btrfs_get_acl(inode, ACL_TYPE_ACCESS); - if (IS_ERR(acl)) - return PTR_ERR(acl); - if (acl) { - error = posix_acl_permission(inode, acl, mask); - posix_acl_release(acl); - } - } - - return error; -} - /* * btrfs_init_acl is already generally called under fs_mutex, so the locking * stuff has been fixed to work with that. If the locking stuff changes, we @@ -244,31 +214,20 @@ int btrfs_init_acl(struct btrfs_trans_handle *trans, } if (IS_POSIXACL(dir) && acl) { - struct posix_acl *clone; - mode_t mode; - if (S_ISDIR(inode->i_mode)) { ret = btrfs_set_acl(trans, inode, acl, ACL_TYPE_DEFAULT); if (ret) goto failed; } - clone = posix_acl_clone(acl, GFP_NOFS); - ret = -ENOMEM; - if (!clone) - goto failed; - - mode = inode->i_mode; - ret = posix_acl_create_masq(clone, &mode); - if (ret >= 0) { - inode->i_mode = mode; - if (ret > 0) { - /* we need an acl */ - ret = btrfs_set_acl(trans, inode, clone, - ACL_TYPE_ACCESS); - } + ret = posix_acl_create(&acl, GFP_NOFS, &inode->i_mode); + if (ret < 0) + return ret; + + if (ret > 0) { + /* we need an acl */ + ret = btrfs_set_acl(trans, inode, acl, ACL_TYPE_ACCESS); } - posix_acl_release(clone); } failed: posix_acl_release(acl); @@ -278,7 +237,7 @@ failed: int btrfs_acl_chmod(struct inode *inode) { - struct posix_acl *acl, *clone; + struct posix_acl *acl; int ret = 0; if (S_ISLNK(inode->i_mode)) @@ -291,17 +250,11 @@ int btrfs_acl_chmod(struct inode *inode) if (IS_ERR_OR_NULL(acl)) return PTR_ERR(acl); - clone = posix_acl_clone(acl, GFP_KERNEL); + ret = posix_acl_chmod(&acl, GFP_KERNEL, inode->i_mode); + if (ret) + return ret; + ret = btrfs_set_acl(NULL, inode, acl, ACL_TYPE_ACCESS); posix_acl_release(acl); - if (!clone) - return -ENOMEM; - - ret = posix_acl_chmod_masq(clone, inode->i_mode); - if (!ret) - ret = btrfs_set_acl(NULL, inode, clone, ACL_TYPE_ACCESS); - - posix_acl_release(clone); - return ret; } @@ -318,18 +271,3 @@ const struct xattr_handler btrfs_xattr_acl_access_handler = { .get = btrfs_xattr_acl_get, .set = btrfs_xattr_acl_set, }; - -#else /* CONFIG_BTRFS_FS_POSIX_ACL */ - -int btrfs_acl_chmod(struct inode *inode) -{ - return 0; -} - -int btrfs_init_acl(struct btrfs_trans_handle *trans, - struct inode *inode, struct inode *dir) -{ - return 0; -} - -#endif /* CONFIG_BTRFS_FS_POSIX_ACL */ diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c index 8006a28..03321e5 100644 --- a/fs/btrfs/async-thread.c +++ b/fs/btrfs/async-thread.c @@ -64,6 +64,8 @@ struct btrfs_worker_thread { int idle; }; +static int __btrfs_start_workers(struct btrfs_workers *workers); + /* * btrfs_start_workers uses kthread_run, which can block waiting for memory * for a very long time. It will actually throttle on page writeback, @@ -88,27 +90,10 @@ static void start_new_worker_func(struct btrfs_work *work) { struct worker_start *start; start = container_of(work, struct worker_start, work); - btrfs_start_workers(start->queue, 1); + __btrfs_start_workers(start->queue); kfree(start); } -static int start_new_worker(struct btrfs_workers *queue) -{ - struct worker_start *start; - int ret; - - start = kzalloc(sizeof(*start), GFP_NOFS); - if (!start) - return -ENOMEM; - - start->work.func = start_new_worker_func; - start->queue = queue; - ret = btrfs_queue_worker(queue->atomic_worker_start, &start->work); - if (ret) - kfree(start); - return ret; -} - /* * helper function to move a thread onto the idle list after it * has finished some requests. @@ -153,12 +138,20 @@ static void check_busy_worker(struct btrfs_worker_thread *worker) static void check_pending_worker_creates(struct btrfs_worker_thread *worker) { struct btrfs_workers *workers = worker->workers; + struct worker_start *start; unsigned long flags; rmb(); if (!workers->atomic_start_pending) return; + start = kzalloc(sizeof(*start), GFP_NOFS); + if (!start) + return; + + start->work.func = start_new_worker_func; + start->queue = workers; + spin_lock_irqsave(&workers->lock, flags); if (!workers->atomic_start_pending) goto out; @@ -170,10 +163,11 @@ static void check_pending_worker_creates(struct btrfs_worker_thread *worker) workers->num_workers_starting += 1; spin_unlock_irqrestore(&workers->lock, flags); - start_new_worker(workers); + btrfs_queue_worker(workers->atomic_worker_start, &start->work); return; out: + kfree(start); spin_unlock_irqrestore(&workers->lock, flags); } @@ -338,7 +332,7 @@ again: run_ordered_completions(worker->workers, work); check_pending_worker_creates(worker); - + cond_resched(); } spin_lock_irq(&worker->lock); @@ -469,56 +463,55 @@ void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max, * starts new worker threads. This does not enforce the max worker * count in case you need to temporarily go past it. */ -static int __btrfs_start_workers(struct btrfs_workers *workers, - int num_workers) +static int __btrfs_start_workers(struct btrfs_workers *workers) { struct btrfs_worker_thread *worker; int ret = 0; - int i; - for (i = 0; i < num_workers; i++) { - worker = kzalloc(sizeof(*worker), GFP_NOFS); - if (!worker) { - ret = -ENOMEM; - goto fail; - } + worker = kzalloc(sizeof(*worker), GFP_NOFS); + if (!worker) { + ret = -ENOMEM; + goto fail; + } - INIT_LIST_HEAD(&worker->pending); - INIT_LIST_HEAD(&worker->prio_pending); - INIT_LIST_HEAD(&worker->worker_list); - spin_lock_init(&worker->lock); - - atomic_set(&worker->num_pending, 0); - atomic_set(&worker->refs, 1); - worker->workers = workers; - worker->task = kthread_run(worker_loop, worker, - "btrfs-%s-%d", workers->name, - workers->num_workers + i); - if (IS_ERR(worker->task)) { - ret = PTR_ERR(worker->task); - kfree(worker); - goto fail; - } - spin_lock_irq(&workers->lock); - list_add_tail(&worker->worker_list, &workers->idle_list); - worker->idle = 1; - workers->num_workers++; - workers->num_workers_starting--; - WARN_ON(workers->num_workers_starting < 0); - spin_unlock_irq(&workers->lock); + INIT_LIST_HEAD(&worker->pending); + INIT_LIST_HEAD(&worker->prio_pending); + INIT_LIST_HEAD(&worker->worker_list); + spin_lock_init(&worker->lock); + + atomic_set(&worker->num_pending, 0); + atomic_set(&worker->refs, 1); + worker->workers = workers; + worker->task = kthread_run(worker_loop, worker, + "btrfs-%s-%d", workers->name, + workers->num_workers + 1); + if (IS_ERR(worker->task)) { + ret = PTR_ERR(worker->task); + kfree(worker); + goto fail; } + spin_lock_irq(&workers->lock); + list_add_tail(&worker->worker_list, &workers->idle_list); + worker->idle = 1; + workers->num_workers++; + workers->num_workers_starting--; + WARN_ON(workers->num_workers_starting < 0); + spin_unlock_irq(&workers->lock); + return 0; fail: - btrfs_stop_workers(workers); + spin_lock_irq(&workers->lock); + workers->num_workers_starting--; + spin_unlock_irq(&workers->lock); return ret; } -int btrfs_start_workers(struct btrfs_workers *workers, int num_workers) +int btrfs_start_workers(struct btrfs_workers *workers) { spin_lock_irq(&workers->lock); - workers->num_workers_starting += num_workers; + workers->num_workers_starting++; spin_unlock_irq(&workers->lock); - return __btrfs_start_workers(workers, num_workers); + return __btrfs_start_workers(workers); } /* @@ -575,9 +568,10 @@ static struct btrfs_worker_thread *find_worker(struct btrfs_workers *workers) struct btrfs_worker_thread *worker; unsigned long flags; struct list_head *fallback; + int ret; -again: spin_lock_irqsave(&workers->lock, flags); +again: worker = next_worker(workers); if (!worker) { @@ -591,7 +585,10 @@ again: workers->num_workers_starting++; spin_unlock_irqrestore(&workers->lock, flags); /* we're below the limit, start another worker */ - __btrfs_start_workers(workers, 1); + ret = __btrfs_start_workers(workers); + spin_lock_irqsave(&workers->lock, flags); + if (ret) + goto fallback; goto again; } } @@ -672,7 +669,7 @@ void btrfs_set_work_high_prio(struct btrfs_work *work) /* * places a struct btrfs_work into the pending queue of one of the kthreads */ -int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) +void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) { struct btrfs_worker_thread *worker; unsigned long flags; @@ -680,7 +677,7 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) /* don't requeue something already on a list */ if (test_and_set_bit(WORK_QUEUED_BIT, &work->flags)) - goto out; + return; worker = find_worker(workers); if (workers->ordered) { @@ -719,7 +716,4 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) if (wake) wake_up_process(worker->task); spin_unlock_irqrestore(&worker->lock, flags); - -out: - return 0; } diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h index 5077746..f34cc31 100644 --- a/fs/btrfs/async-thread.h +++ b/fs/btrfs/async-thread.h @@ -109,8 +109,8 @@ struct btrfs_workers { char *name; }; -int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work); -int btrfs_start_workers(struct btrfs_workers *workers, int num_workers); +void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work); +int btrfs_start_workers(struct btrfs_workers *workers); int btrfs_stop_workers(struct btrfs_workers *workers); void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max, struct btrfs_workers *async_starter); diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index 52d7eca..634608d 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -34,6 +34,9 @@ struct btrfs_inode { */ struct btrfs_key location; + /* Lock for counters */ + spinlock_t lock; + /* the extent_tree has caches of all the extent mappings to disk */ struct extent_map_tree extent_tree; @@ -100,11 +103,6 @@ struct btrfs_inode { */ u64 delalloc_bytes; - /* total number of bytes that may be used for this inode for - * delalloc - */ - u64 reserved_bytes; - /* * the size of the file stored in the metadata on disk. data=ordered * means the in-memory i_size might be larger than the size on disk @@ -112,9 +110,6 @@ struct btrfs_inode { */ u64 disk_i_size; - /* flags field from the on disk inode */ - u32 flags; - /* * if this is a directory then index_cnt is the counter for the index * number for new files that are created @@ -129,13 +124,22 @@ struct btrfs_inode { u64 last_unlink_trans; /* + * Number of bytes outstanding that are going to need csums. This is + * used in ENOSPC accounting. + */ + u64 csum_bytes; + + /* flags field from the on disk inode */ + u32 flags; + + /* * Counters to keep track of the number of extent item's we may use due * to delalloc and such. outstanding_extents is the number of extent * items we think we'll end up using, and reserved_extents is the number * of extent items we've reserved metadata for. */ - atomic_t outstanding_extents; - atomic_t reserved_extents; + unsigned outstanding_extents; + unsigned reserved_extents; /* * ordered_data_close is set by truncate when a file that used @@ -143,14 +147,12 @@ struct btrfs_inode { * the btrfs file release call will add this inode to the * ordered operations list so that we make sure to flush out any * new data the application may have written before commit. - * - * yes, its silly to have a single bitflag, but we might grow more - * of these. */ unsigned ordered_data_close:1; unsigned orphan_meta_reserved:1; unsigned dummy_inode:1; unsigned in_defrag:1; + unsigned delalloc_meta_reserved:1; /* * always compress this one file @@ -173,7 +175,11 @@ static inline u64 btrfs_ino(struct inode *inode) { u64 ino = BTRFS_I(inode)->location.objectid; - if (ino <= BTRFS_FIRST_FREE_OBJECTID) + /* + * !ino: btree_inode + * type == BTRFS_ROOT_ITEM_KEY: subvol dir + */ + if (!ino || BTRFS_I(inode)->location.type == BTRFS_ROOT_ITEM_KEY) ino = inode->i_ino; return ino; } @@ -184,4 +190,13 @@ static inline void btrfs_i_size_write(struct inode *inode, u64 size) BTRFS_I(inode)->disk_i_size = size; } +static inline bool btrfs_is_free_space_inode(struct btrfs_root *root, + struct inode *inode) +{ + if (root == root->fs_info->tree_root || + BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) + return true; + return false; +} + #endif diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index bfe42b0..14f1c5a 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -85,7 +85,8 @@ struct compressed_bio { static inline int compressed_bio_size(struct btrfs_root *root, unsigned long disk_size) { - u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); + return sizeof(struct compressed_bio) + ((disk_size + root->sectorsize - 1) / root->sectorsize) * csum_size; @@ -338,6 +339,7 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, u64 first_byte = disk_start; struct block_device *bdev; int ret; + int skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; WARN_ON(start & ((u64)PAGE_CACHE_SIZE - 1)); cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS); @@ -392,8 +394,11 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); BUG_ON(ret); - ret = btrfs_csum_one_bio(root, inode, bio, start, 1); - BUG_ON(ret); + if (!skip_sum) { + ret = btrfs_csum_one_bio(root, inode, bio, + start, 1); + BUG_ON(ret); + } ret = btrfs_map_bio(root, WRITE, bio, 0, 1); BUG_ON(ret); @@ -418,8 +423,10 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); BUG_ON(ret); - ret = btrfs_csum_one_bio(root, inode, bio, start, 1); - BUG_ON(ret); + if (!skip_sum) { + ret = btrfs_csum_one_bio(root, inode, bio, start, 1); + BUG_ON(ret); + } ret = btrfs_map_bio(root, WRITE, bio, 0, 1); BUG_ON(ret); diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 2e66786..dede441 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -54,8 +54,13 @@ noinline void btrfs_set_path_blocking(struct btrfs_path *p) { int i; for (i = 0; i < BTRFS_MAX_LEVEL; i++) { - if (p->nodes[i] && p->locks[i]) - btrfs_set_lock_blocking(p->nodes[i]); + if (!p->nodes[i] || !p->locks[i]) + continue; + btrfs_set_lock_blocking_rw(p->nodes[i], p->locks[i]); + if (p->locks[i] == BTRFS_READ_LOCK) + p->locks[i] = BTRFS_READ_LOCK_BLOCKING; + else if (p->locks[i] == BTRFS_WRITE_LOCK) + p->locks[i] = BTRFS_WRITE_LOCK_BLOCKING; } } @@ -68,7 +73,7 @@ noinline void btrfs_set_path_blocking(struct btrfs_path *p) * for held */ noinline void btrfs_clear_path_blocking(struct btrfs_path *p, - struct extent_buffer *held) + struct extent_buffer *held, int held_rw) { int i; @@ -79,19 +84,29 @@ noinline void btrfs_clear_path_blocking(struct btrfs_path *p, * really sure by forcing the path to blocking before we clear * the path blocking. */ - if (held) - btrfs_set_lock_blocking(held); + if (held) { + btrfs_set_lock_blocking_rw(held, held_rw); + if (held_rw == BTRFS_WRITE_LOCK) + held_rw = BTRFS_WRITE_LOCK_BLOCKING; + else if (held_rw == BTRFS_READ_LOCK) + held_rw = BTRFS_READ_LOCK_BLOCKING; + } btrfs_set_path_blocking(p); #endif for (i = BTRFS_MAX_LEVEL - 1; i >= 0; i--) { - if (p->nodes[i] && p->locks[i]) - btrfs_clear_lock_blocking(p->nodes[i]); + if (p->nodes[i] && p->locks[i]) { + btrfs_clear_lock_blocking_rw(p->nodes[i], p->locks[i]); + if (p->locks[i] == BTRFS_WRITE_LOCK_BLOCKING) + p->locks[i] = BTRFS_WRITE_LOCK; + else if (p->locks[i] == BTRFS_READ_LOCK_BLOCKING) + p->locks[i] = BTRFS_READ_LOCK; + } } #ifdef CONFIG_DEBUG_LOCK_ALLOC if (held) - btrfs_clear_lock_blocking(held); + btrfs_clear_lock_blocking_rw(held, held_rw); #endif } @@ -119,7 +134,7 @@ noinline void btrfs_release_path(struct btrfs_path *p) if (!p->nodes[i]) continue; if (p->locks[i]) { - btrfs_tree_unlock(p->nodes[i]); + btrfs_tree_unlock_rw(p->nodes[i], p->locks[i]); p->locks[i] = 0; } free_extent_buffer(p->nodes[i]); @@ -167,6 +182,25 @@ struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root) return eb; } +/* loop around taking references on and locking the root node of the + * tree until you end up with a lock on the root. A locked buffer + * is returned, with a reference held. + */ +struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root) +{ + struct extent_buffer *eb; + + while (1) { + eb = btrfs_root_node(root); + btrfs_tree_read_lock(eb); + if (eb == root->node) + break; + btrfs_tree_read_unlock(eb); + free_extent_buffer(eb); + } + return eb; +} + /* cowonly root (everything not a reference counted cow subvolume), just get * put onto a simple dirty list. transaction.c walks this to make sure they * get properly updated on disk. @@ -480,10 +514,25 @@ static inline int should_cow_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf) { + /* ensure we can see the force_cow */ + smp_rmb(); + + /* + * We do not need to cow a block if + * 1) this block is not created or changed in this transaction; + * 2) this block does not belong to TREE_RELOC tree; + * 3) the root is not forced COW. + * + * What is forced COW: + * when we create snapshot during commiting the transaction, + * after we've finished coping src root, we must COW the shared + * block to ensure the metadata consistency. + */ if (btrfs_header_generation(buf) == trans->transid && !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN) && !(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID && - btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC))) + btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)) && + !root->force_cow) return 0; return 1; } @@ -626,14 +675,6 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, for (i = start_slot; i < end_slot; i++) { int close = 1; - if (!parent->map_token) { - map_extent_buffer(parent, - btrfs_node_key_ptr_offset(i), - sizeof(struct btrfs_key_ptr), - &parent->map_token, &parent->kaddr, - &parent->map_start, &parent->map_len, - KM_USER1); - } btrfs_node_key(parent, &disk_key, i); if (!progress_passed && comp_keys(&disk_key, progress) < 0) continue; @@ -656,11 +697,6 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, last_block = blocknr; continue; } - if (parent->map_token) { - unmap_extent_buffer(parent, parent->map_token, - KM_USER1); - parent->map_token = NULL; - } cur = btrfs_find_tree_block(root, blocknr, blocksize); if (cur) @@ -701,11 +737,6 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, btrfs_tree_unlock(cur); free_extent_buffer(cur); } - if (parent->map_token) { - unmap_extent_buffer(parent, parent->map_token, - KM_USER1); - parent->map_token = NULL; - } return err; } @@ -746,7 +777,6 @@ static noinline int generic_bin_search(struct extent_buffer *eb, struct btrfs_disk_key *tmp = NULL; struct btrfs_disk_key unaligned; unsigned long offset; - char *map_token = NULL; char *kaddr = NULL; unsigned long map_start = 0; unsigned long map_len = 0; @@ -756,18 +786,13 @@ static noinline int generic_bin_search(struct extent_buffer *eb, mid = (low + high) / 2; offset = p + mid * item_size; - if (!map_token || offset < map_start || + if (!kaddr || offset < map_start || (offset + sizeof(struct btrfs_disk_key)) > map_start + map_len) { - if (map_token) { - unmap_extent_buffer(eb, map_token, KM_USER0); - map_token = NULL; - } err = map_private_extent_buffer(eb, offset, sizeof(struct btrfs_disk_key), - &map_token, &kaddr, - &map_start, &map_len, KM_USER0); + &kaddr, &map_start, &map_len); if (!err) { tmp = (struct btrfs_disk_key *)(kaddr + offset - @@ -790,14 +815,10 @@ static noinline int generic_bin_search(struct extent_buffer *eb, high = mid; else { *slot = mid; - if (map_token) - unmap_extent_buffer(eb, map_token, KM_USER0); return 0; } } *slot = low; - if (map_token) - unmap_extent_buffer(eb, map_token, KM_USER0); return 1; } @@ -890,14 +911,16 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, mid = path->nodes[level]; - WARN_ON(!path->locks[level]); + WARN_ON(path->locks[level] != BTRFS_WRITE_LOCK && + path->locks[level] != BTRFS_WRITE_LOCK_BLOCKING); WARN_ON(btrfs_header_generation(mid) != trans->transid); orig_ptr = btrfs_node_blockptr(mid, orig_slot); - if (level < BTRFS_MAX_LEVEL - 1) + if (level < BTRFS_MAX_LEVEL - 1) { parent = path->nodes[level + 1]; - pslot = path->slots[level + 1]; + pslot = path->slots[level + 1]; + } /* * deal with the case where there is only one pointer in the root @@ -1100,9 +1123,10 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans, mid = path->nodes[level]; WARN_ON(btrfs_header_generation(mid) != trans->transid); - if (level < BTRFS_MAX_LEVEL - 1) + if (level < BTRFS_MAX_LEVEL - 1) { parent = path->nodes[level + 1]; - pslot = path->slots[level + 1]; + pslot = path->slots[level + 1]; + } if (!parent) return 1; @@ -1228,7 +1252,6 @@ static void reada_for_search(struct btrfs_root *root, u32 nr; u32 blocksize; u32 nscan = 0; - bool map = true; if (level != 1) return; @@ -1250,19 +1273,8 @@ static void reada_for_search(struct btrfs_root *root, nritems = btrfs_header_nritems(node); nr = slot; - if (node->map_token || path->skip_locking) - map = false; while (1) { - if (map && !node->map_token) { - unsigned long offset = btrfs_node_key_ptr_offset(nr); - map_private_extent_buffer(node, offset, - sizeof(struct btrfs_key_ptr), - &node->map_token, - &node->kaddr, - &node->map_start, - &node->map_len, KM_USER1); - } if (direction < 0) { if (nr == 0) break; @@ -1281,11 +1293,6 @@ static void reada_for_search(struct btrfs_root *root, if ((search <= target && target - search <= 65536) || (search > target && search - target <= 65536)) { gen = btrfs_node_ptr_generation(node, nr); - if (map && node->map_token) { - unmap_extent_buffer(node, node->map_token, - KM_USER1); - node->map_token = NULL; - } readahead_tree_block(root, search, blocksize, gen); nread += blocksize; } @@ -1293,10 +1300,6 @@ static void reada_for_search(struct btrfs_root *root, if ((nread > 65536 || nscan > 32)) break; } - if (map && node->map_token) { - unmap_extent_buffer(node, node->map_token, KM_USER1); - node->map_token = NULL; - } } /* @@ -1409,7 +1412,7 @@ static noinline void unlock_up(struct btrfs_path *path, int level, t = path->nodes[i]; if (i >= lowest_unlock && i > skip_level && path->locks[i]) { - btrfs_tree_unlock(t); + btrfs_tree_unlock_rw(t, path->locks[i]); path->locks[i] = 0; } } @@ -1436,7 +1439,7 @@ noinline void btrfs_unlock_up_safe(struct btrfs_path *path, int level) continue; if (!path->locks[i]) continue; - btrfs_tree_unlock(path->nodes[i]); + btrfs_tree_unlock_rw(path->nodes[i], path->locks[i]); path->locks[i] = 0; } } @@ -1485,6 +1488,8 @@ read_block_for_search(struct btrfs_trans_handle *trans, * we can trust our generation number */ free_extent_buffer(tmp); + btrfs_set_path_blocking(p); + tmp = read_tree_block(root, blocknr, blocksize, gen); if (tmp && btrfs_buffer_uptodate(tmp, gen)) { *eb_ret = tmp; @@ -1540,20 +1545,27 @@ read_block_for_search(struct btrfs_trans_handle *trans, static int setup_nodes_for_search(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *p, - struct extent_buffer *b, int level, int ins_len) + struct extent_buffer *b, int level, int ins_len, + int *write_lock_level) { int ret; if ((p->search_for_split || ins_len > 0) && btrfs_header_nritems(b) >= BTRFS_NODEPTRS_PER_BLOCK(root) - 3) { int sret; + if (*write_lock_level < level + 1) { + *write_lock_level = level + 1; + btrfs_release_path(p); + goto again; + } + sret = reada_for_balance(root, p, level); if (sret) goto again; btrfs_set_path_blocking(p); sret = split_node(trans, root, p, level); - btrfs_clear_path_blocking(p, NULL); + btrfs_clear_path_blocking(p, NULL, 0); BUG_ON(sret > 0); if (sret) { @@ -1565,13 +1577,19 @@ setup_nodes_for_search(struct btrfs_trans_handle *trans, BTRFS_NODEPTRS_PER_BLOCK(root) / 2) { int sret; + if (*write_lock_level < level + 1) { + *write_lock_level = level + 1; + btrfs_release_path(p); + goto again; + } + sret = reada_for_balance(root, p, level); if (sret) goto again; btrfs_set_path_blocking(p); sret = balance_level(trans, root, p, level); - btrfs_clear_path_blocking(p, NULL); + btrfs_clear_path_blocking(p, NULL, 0); if (sret) { ret = sret; @@ -1615,27 +1633,78 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root int err; int level; int lowest_unlock = 1; + int root_lock; + /* everything at write_lock_level or lower must be write locked */ + int write_lock_level = 0; u8 lowest_level = 0; lowest_level = p->lowest_level; WARN_ON(lowest_level && ins_len > 0); WARN_ON(p->nodes[0] != NULL); - if (ins_len < 0) + if (ins_len < 0) { lowest_unlock = 2; + /* when we are removing items, we might have to go up to level + * two as we update tree pointers Make sure we keep write + * for those levels as well + */ + write_lock_level = 2; + } else if (ins_len > 0) { + /* + * for inserting items, make sure we have a write lock on + * level 1 so we can update keys + */ + write_lock_level = 1; + } + + if (!cow) + write_lock_level = -1; + + if (cow && (p->keep_locks || p->lowest_level)) + write_lock_level = BTRFS_MAX_LEVEL; + again: + /* + * we try very hard to do read locks on the root + */ + root_lock = BTRFS_READ_LOCK; + level = 0; if (p->search_commit_root) { + /* + * the commit roots are read only + * so we always do read locks + */ b = root->commit_root; extent_buffer_get(b); + level = btrfs_header_level(b); if (!p->skip_locking) - btrfs_tree_lock(b); + btrfs_tree_read_lock(b); } else { - if (p->skip_locking) + if (p->skip_locking) { b = btrfs_root_node(root); - else - b = btrfs_lock_root_node(root); + level = btrfs_header_level(b); + } else { + /* we don't know the level of the root node + * until we actually have it read locked + */ + b = btrfs_read_lock_root_node(root); + level = btrfs_header_level(b); + if (level <= write_lock_level) { + /* whoops, must trade for write lock */ + btrfs_tree_read_unlock(b); + free_extent_buffer(b); + b = btrfs_lock_root_node(root); + root_lock = BTRFS_WRITE_LOCK; + + /* the level might have changed, check again */ + level = btrfs_header_level(b); + } + } } + p->nodes[level] = b; + if (!p->skip_locking) + p->locks[level] = root_lock; while (b) { level = btrfs_header_level(b); @@ -1644,10 +1713,6 @@ again: * setup the path here so we can release it under lock * contention with the cow code */ - p->nodes[level] = b; - if (!p->skip_locking) - p->locks[level] = 1; - if (cow) { /* * if we don't really need to cow this block @@ -1659,6 +1724,16 @@ again: btrfs_set_path_blocking(p); + /* + * must have write locks on this node and the + * parent + */ + if (level + 1 > write_lock_level) { + write_lock_level = level + 1; + btrfs_release_path(p); + goto again; + } + err = btrfs_cow_block(trans, root, b, p->nodes[level + 1], p->slots[level + 1], &b); @@ -1671,10 +1746,7 @@ cow_done: BUG_ON(!cow && ins_len); p->nodes[level] = b; - if (!p->skip_locking) - p->locks[level] = 1; - - btrfs_clear_path_blocking(p, NULL); + btrfs_clear_path_blocking(p, NULL, 0); /* * we have a lock on b and as long as we aren't changing @@ -1700,7 +1772,7 @@ cow_done: } p->slots[level] = slot; err = setup_nodes_for_search(trans, root, p, b, level, - ins_len); + ins_len, &write_lock_level); if (err == -EAGAIN) goto again; if (err) { @@ -1710,6 +1782,19 @@ cow_done: b = p->nodes[level]; slot = p->slots[level]; + /* + * slot 0 is special, if we change the key + * we have to update the parent pointer + * which means we must have a write lock + * on the parent + */ + if (slot == 0 && cow && + write_lock_level < level + 1) { + write_lock_level = level + 1; + btrfs_release_path(p); + goto again; + } + unlock_up(p, level, lowest_unlock); if (level == lowest_level) { @@ -1728,23 +1813,42 @@ cow_done: } if (!p->skip_locking) { - btrfs_clear_path_blocking(p, NULL); - err = btrfs_try_spin_lock(b); - - if (!err) { - btrfs_set_path_blocking(p); - btrfs_tree_lock(b); - btrfs_clear_path_blocking(p, b); + level = btrfs_header_level(b); + if (level <= write_lock_level) { + err = btrfs_try_tree_write_lock(b); + if (!err) { + btrfs_set_path_blocking(p); + btrfs_tree_lock(b); + btrfs_clear_path_blocking(p, b, + BTRFS_WRITE_LOCK); + } + p->locks[level] = BTRFS_WRITE_LOCK; + } else { + err = btrfs_try_tree_read_lock(b); + if (!err) { + btrfs_set_path_blocking(p); + btrfs_tree_read_lock(b); + btrfs_clear_path_blocking(p, b, + BTRFS_READ_LOCK); + } + p->locks[level] = BTRFS_READ_LOCK; } + p->nodes[level] = b; } } else { p->slots[level] = slot; if (ins_len > 0 && btrfs_leaf_free_space(root, b) < ins_len) { + if (write_lock_level < 1) { + write_lock_level = 1; + btrfs_release_path(p); + goto again; + } + btrfs_set_path_blocking(p); err = split_leaf(trans, root, key, p, ins_len, ret == 0); - btrfs_clear_path_blocking(p, NULL); + btrfs_clear_path_blocking(p, NULL, 0); BUG_ON(err > 0); if (err) { @@ -2025,7 +2129,7 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans, add_root_to_dirty_list(root); extent_buffer_get(c); path->nodes[level] = c; - path->locks[level] = 1; + path->locks[level] = BTRFS_WRITE_LOCK; path->slots[level] = 0; return 0; } @@ -2253,14 +2357,6 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans, if (path->slots[0] == i) push_space += data_size; - if (!left->map_token) { - map_extent_buffer(left, (unsigned long)item, - sizeof(struct btrfs_item), - &left->map_token, &left->kaddr, - &left->map_start, &left->map_len, - KM_USER1); - } - this_item_size = btrfs_item_size(left, item); if (this_item_size + sizeof(*item) + push_space > free_space) break; @@ -2271,10 +2367,6 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans, break; i--; } - if (left->map_token) { - unmap_extent_buffer(left, left->map_token, KM_USER1); - left->map_token = NULL; - } if (push_items == 0) goto out_unlock; @@ -2316,21 +2408,10 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans, push_space = BTRFS_LEAF_DATA_SIZE(root); for (i = 0; i < right_nritems; i++) { item = btrfs_item_nr(right, i); - if (!right->map_token) { - map_extent_buffer(right, (unsigned long)item, - sizeof(struct btrfs_item), - &right->map_token, &right->kaddr, - &right->map_start, &right->map_len, - KM_USER1); - } push_space -= btrfs_item_size(right, item); btrfs_set_item_offset(right, item, push_space); } - if (right->map_token) { - unmap_extent_buffer(right, right->map_token, KM_USER1); - right->map_token = NULL; - } left_nritems -= push_items; btrfs_set_header_nritems(left, left_nritems); @@ -2467,13 +2548,6 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, for (i = 0; i < nr; i++) { item = btrfs_item_nr(right, i); - if (!right->map_token) { - map_extent_buffer(right, (unsigned long)item, - sizeof(struct btrfs_item), - &right->map_token, &right->kaddr, - &right->map_start, &right->map_len, - KM_USER1); - } if (!empty && push_items > 0) { if (path->slots[0] < i) @@ -2496,11 +2570,6 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, push_space += this_item_size + sizeof(*item); } - if (right->map_token) { - unmap_extent_buffer(right, right->map_token, KM_USER1); - right->map_token = NULL; - } - if (push_items == 0) { ret = 1; goto out; @@ -2530,23 +2599,12 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, u32 ioff; item = btrfs_item_nr(left, i); - if (!left->map_token) { - map_extent_buffer(left, (unsigned long)item, - sizeof(struct btrfs_item), - &left->map_token, &left->kaddr, - &left->map_start, &left->map_len, - KM_USER1); - } ioff = btrfs_item_offset(left, item); btrfs_set_item_offset(left, item, ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size)); } btrfs_set_header_nritems(left, old_left_nritems + push_items); - if (left->map_token) { - unmap_extent_buffer(left, left->map_token, KM_USER1); - left->map_token = NULL; - } /* fixup right node */ if (push_items > right_nritems) { @@ -2574,21 +2632,9 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, for (i = 0; i < right_nritems; i++) { item = btrfs_item_nr(right, i); - if (!right->map_token) { - map_extent_buffer(right, (unsigned long)item, - sizeof(struct btrfs_item), - &right->map_token, &right->kaddr, - &right->map_start, &right->map_len, - KM_USER1); - } - push_space = push_space - btrfs_item_size(right, item); btrfs_set_item_offset(right, item, push_space); } - if (right->map_token) { - unmap_extent_buffer(right, right->map_token, KM_USER1); - right->map_token = NULL; - } btrfs_mark_buffer_dirty(left); if (right_nritems) @@ -2729,23 +2775,10 @@ static noinline int copy_for_split(struct btrfs_trans_handle *trans, struct btrfs_item *item = btrfs_item_nr(right, i); u32 ioff; - if (!right->map_token) { - map_extent_buffer(right, (unsigned long)item, - sizeof(struct btrfs_item), - &right->map_token, &right->kaddr, - &right->map_start, &right->map_len, - KM_USER1); - } - ioff = btrfs_item_offset(right, item); btrfs_set_item_offset(right, item, ioff + rt_data_off); } - if (right->map_token) { - unmap_extent_buffer(right, right->map_token, KM_USER1); - right->map_token = NULL; - } - btrfs_set_header_nritems(l, mid); ret = 0; btrfs_item_key(right, &disk_key, 0); @@ -3264,23 +3297,10 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, u32 ioff; item = btrfs_item_nr(leaf, i); - if (!leaf->map_token) { - map_extent_buffer(leaf, (unsigned long)item, - sizeof(struct btrfs_item), - &leaf->map_token, &leaf->kaddr, - &leaf->map_start, &leaf->map_len, - KM_USER1); - } - ioff = btrfs_item_offset(leaf, item); btrfs_set_item_offset(leaf, item, ioff + size_diff); } - if (leaf->map_token) { - unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); - leaf->map_token = NULL; - } - /* shift the data */ if (from_end) { memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) + @@ -3377,22 +3397,10 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, u32 ioff; item = btrfs_item_nr(leaf, i); - if (!leaf->map_token) { - map_extent_buffer(leaf, (unsigned long)item, - sizeof(struct btrfs_item), - &leaf->map_token, &leaf->kaddr, - &leaf->map_start, &leaf->map_len, - KM_USER1); - } ioff = btrfs_item_offset(leaf, item); btrfs_set_item_offset(leaf, item, ioff - data_size); } - if (leaf->map_token) { - unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); - leaf->map_token = NULL; - } - /* shift the data */ memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) + data_end - data_size, btrfs_leaf_data(leaf) + @@ -3494,27 +3502,13 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans, * item0..itemN ... dataN.offset..dataN.size .. data0.size */ /* first correct the data pointers */ - WARN_ON(leaf->map_token); for (i = slot; i < nritems; i++) { u32 ioff; item = btrfs_item_nr(leaf, i); - if (!leaf->map_token) { - map_extent_buffer(leaf, (unsigned long)item, - sizeof(struct btrfs_item), - &leaf->map_token, &leaf->kaddr, - &leaf->map_start, &leaf->map_len, - KM_USER1); - } - ioff = btrfs_item_offset(leaf, item); btrfs_set_item_offset(leaf, item, ioff - total_data); } - if (leaf->map_token) { - unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); - leaf->map_token = NULL; - } - /* shift the items */ memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr), btrfs_item_nr_offset(slot), @@ -3608,27 +3602,13 @@ int setup_items_for_insert(struct btrfs_trans_handle *trans, * item0..itemN ... dataN.offset..dataN.size .. data0.size */ /* first correct the data pointers */ - WARN_ON(leaf->map_token); for (i = slot; i < nritems; i++) { u32 ioff; item = btrfs_item_nr(leaf, i); - if (!leaf->map_token) { - map_extent_buffer(leaf, (unsigned long)item, - sizeof(struct btrfs_item), - &leaf->map_token, &leaf->kaddr, - &leaf->map_start, &leaf->map_len, - KM_USER1); - } - ioff = btrfs_item_offset(leaf, item); btrfs_set_item_offset(leaf, item, ioff - total_data); } - if (leaf->map_token) { - unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); - leaf->map_token = NULL; - } - /* shift the items */ memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr), btrfs_item_nr_offset(slot), @@ -3840,22 +3820,10 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, u32 ioff; item = btrfs_item_nr(leaf, i); - if (!leaf->map_token) { - map_extent_buffer(leaf, (unsigned long)item, - sizeof(struct btrfs_item), - &leaf->map_token, &leaf->kaddr, - &leaf->map_start, &leaf->map_len, - KM_USER1); - } ioff = btrfs_item_offset(leaf, item); btrfs_set_item_offset(leaf, item, ioff + dsize); } - if (leaf->map_token) { - unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); - leaf->map_token = NULL; - } - memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot), btrfs_item_nr_offset(slot + nr), sizeof(struct btrfs_item) * @@ -4004,11 +3972,11 @@ int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key, WARN_ON(!path->keep_locks); again: - cur = btrfs_lock_root_node(root); + cur = btrfs_read_lock_root_node(root); level = btrfs_header_level(cur); WARN_ON(path->nodes[level]); path->nodes[level] = cur; - path->locks[level] = 1; + path->locks[level] = BTRFS_READ_LOCK; if (btrfs_header_generation(cur) < min_trans) { ret = 1; @@ -4098,12 +4066,12 @@ find_next_key: cur = read_node_slot(root, cur, slot); BUG_ON(!cur); - btrfs_tree_lock(cur); + btrfs_tree_read_lock(cur); - path->locks[level - 1] = 1; + path->locks[level - 1] = BTRFS_READ_LOCK; path->nodes[level - 1] = cur; unlock_up(path, level, 1); - btrfs_clear_path_blocking(path, NULL); + btrfs_clear_path_blocking(path, NULL, 0); } out: if (ret == 0) @@ -4218,30 +4186,21 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path) u32 nritems; int ret; int old_spinning = path->leave_spinning; - int force_blocking = 0; + int next_rw_lock = 0; nritems = btrfs_header_nritems(path->nodes[0]); if (nritems == 0) return 1; - /* - * we take the blocks in an order that upsets lockdep. Using - * blocking mode is the only way around it. - */ -#ifdef CONFIG_DEBUG_LOCK_ALLOC - force_blocking = 1; -#endif - btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1); again: level = 1; next = NULL; + next_rw_lock = 0; btrfs_release_path(path); path->keep_locks = 1; - - if (!force_blocking) - path->leave_spinning = 1; + path->leave_spinning = 1; ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); path->keep_locks = 0; @@ -4281,11 +4240,12 @@ again: } if (next) { - btrfs_tree_unlock(next); + btrfs_tree_unlock_rw(next, next_rw_lock); free_extent_buffer(next); } next = c; + next_rw_lock = path->locks[level]; ret = read_block_for_search(NULL, root, path, &next, level, slot, &key); if (ret == -EAGAIN) @@ -4297,15 +4257,14 @@ again: } if (!path->skip_locking) { - ret = btrfs_try_spin_lock(next); + ret = btrfs_try_tree_read_lock(next); if (!ret) { btrfs_set_path_blocking(path); - btrfs_tree_lock(next); - if (!force_blocking) - btrfs_clear_path_blocking(path, next); + btrfs_tree_read_lock(next); + btrfs_clear_path_blocking(path, next, + BTRFS_READ_LOCK); } - if (force_blocking) - btrfs_set_lock_blocking(next); + next_rw_lock = BTRFS_READ_LOCK; } break; } @@ -4314,14 +4273,13 @@ again: level--; c = path->nodes[level]; if (path->locks[level]) - btrfs_tree_unlock(c); + btrfs_tree_unlock_rw(c, path->locks[level]); free_extent_buffer(c); path->nodes[level] = next; path->slots[level] = 0; if (!path->skip_locking) - path->locks[level] = 1; - + path->locks[level] = next_rw_lock; if (!level) break; @@ -4336,16 +4294,14 @@ again: } if (!path->skip_locking) { - btrfs_assert_tree_locked(path->nodes[level]); - ret = btrfs_try_spin_lock(next); + ret = btrfs_try_tree_read_lock(next); if (!ret) { btrfs_set_path_blocking(path); - btrfs_tree_lock(next); - if (!force_blocking) - btrfs_clear_path_blocking(path, next); + btrfs_tree_read_lock(next); + btrfs_clear_path_blocking(path, next, + BTRFS_READ_LOCK); } - if (force_blocking) - btrfs_set_lock_blocking(next); + next_rw_lock = BTRFS_READ_LOCK; } } ret = 0; diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 66179bc..83a871f 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -30,6 +30,7 @@ #include <linux/kobject.h> #include <trace/events/btrfs.h> #include <asm/kmap_types.h> +#include <linux/pagemap.h> #include "extent_io.h" #include "extent_map.h" #include "async-thread.h" @@ -360,6 +361,47 @@ struct btrfs_header { #define BTRFS_LABEL_SIZE 256 /* + * just in case we somehow lose the roots and are not able to mount, + * we store an array of the roots from previous transactions + * in the super. + */ +#define BTRFS_NUM_BACKUP_ROOTS 4 +struct btrfs_root_backup { + __le64 tree_root; + __le64 tree_root_gen; + + __le64 chunk_root; + __le64 chunk_root_gen; + + __le64 extent_root; + __le64 extent_root_gen; + + __le64 fs_root; + __le64 fs_root_gen; + + __le64 dev_root; + __le64 dev_root_gen; + + __le64 csum_root; + __le64 csum_root_gen; + + __le64 total_bytes; + __le64 bytes_used; + __le64 num_devices; + /* future */ + __le64 unsed_64[4]; + + u8 tree_root_level; + u8 chunk_root_level; + u8 extent_root_level; + u8 fs_root_level; + u8 dev_root_level; + u8 csum_root_level; + /* future and to align */ + u8 unused_8[10]; +} __attribute__ ((__packed__)); + +/* * the super block basically lists the main trees of the FS * it currently lacks any block count etc etc */ @@ -405,6 +447,7 @@ struct btrfs_super_block { /* future expansion */ __le64 reserved[31]; u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE]; + struct btrfs_root_backup super_roots[BTRFS_NUM_BACKUP_ROOTS]; } __attribute__ ((__packed__)); /* @@ -755,6 +798,8 @@ struct btrfs_space_info { chunks for this space */ unsigned int chunk_alloc:1; /* set if we are allocating a chunk */ + unsigned int flush:1; /* set if we are trying to make space */ + unsigned int force_alloc; /* set if we need to force a chunk alloc for this space */ @@ -764,20 +809,14 @@ struct btrfs_space_info { struct list_head block_groups[BTRFS_NR_RAID_TYPES]; spinlock_t lock; struct rw_semaphore groups_sem; - atomic_t caching_threads; + wait_queue_head_t wait; }; struct btrfs_block_rsv { u64 size; u64 reserved; - u64 freed[2]; struct btrfs_space_info *space_info; - struct list_head list; spinlock_t lock; - atomic_t usage; - unsigned int priority:8; - unsigned int durable:1; - unsigned int refill_used:1; unsigned int full:1; }; @@ -809,7 +848,8 @@ struct btrfs_free_cluster { enum btrfs_caching_type { BTRFS_CACHE_NO = 0, BTRFS_CACHE_STARTED = 1, - BTRFS_CACHE_FINISHED = 2, + BTRFS_CACHE_FAST = 2, + BTRFS_CACHE_FINISHED = 3, }; enum btrfs_disk_cache_state { @@ -824,6 +864,7 @@ struct btrfs_caching_control { struct list_head list; struct mutex mutex; wait_queue_head_t wait; + struct btrfs_work work; struct btrfs_block_group_cache *block_group; u64 progress; atomic_t count; @@ -837,10 +878,10 @@ struct btrfs_block_group_cache { spinlock_t lock; u64 pinned; u64 reserved; - u64 reserved_pinned; u64 bytes_super; u64 flags; u64 sectorsize; + u64 cache_generation; unsigned int ro:1; unsigned int dirty:1; unsigned int iref:1; @@ -896,6 +937,10 @@ struct btrfs_fs_info { spinlock_t block_group_cache_lock; struct rb_root block_group_cache_tree; + /* keep track of unallocated space */ + spinlock_t free_chunk_lock; + u64 free_chunk_space; + struct extent_io_tree freed_extents[2]; struct extent_io_tree *pinned_extents; @@ -913,14 +958,11 @@ struct btrfs_fs_info { struct btrfs_block_rsv trans_block_rsv; /* block reservation for chunk tree */ struct btrfs_block_rsv chunk_block_rsv; + /* block reservation for delayed operations */ + struct btrfs_block_rsv delayed_block_rsv; struct btrfs_block_rsv empty_block_rsv; - /* list of block reservations that cross multiple transactions */ - struct list_head durable_block_rsv_list; - - struct mutex durable_block_rsv_mutex; - u64 generation; u64 last_trans_committed; @@ -939,8 +981,8 @@ struct btrfs_fs_info { wait_queue_head_t transaction_blocked_wait; wait_queue_head_t async_submit_wait; - struct btrfs_super_block super_copy; - struct btrfs_super_block super_for_commit; + struct btrfs_super_block *super_copy; + struct btrfs_super_block *super_for_commit; struct block_device *__bdev; struct super_block *sb; struct inode *btree_inode; @@ -1032,6 +1074,9 @@ struct btrfs_fs_info { struct btrfs_workers endio_write_workers; struct btrfs_workers endio_freespace_worker; struct btrfs_workers submit_workers; + struct btrfs_workers caching_workers; + struct btrfs_workers readahead_workers; + /* * fixup workers take dirty pages that didn't properly go through * the cow mechanism and make them safe to write. It happens @@ -1114,6 +1159,13 @@ struct btrfs_fs_info { u64 fs_state; struct btrfs_delayed_root *delayed_root; + + /* readahead tree */ + spinlock_t reada_lock; + struct radix_tree_root reada_tree; + + /* next backup root to be overwritten */ + int backup_root_index; }; /* @@ -1219,7 +1271,9 @@ struct btrfs_root { * right now this just gets used so that a root has its own devid * for stat. It may be used for more later */ - struct super_block anon_super; + dev_t anon_dev; + + int force_cow; }; struct btrfs_ioctl_defrag_range_args { @@ -1358,6 +1412,7 @@ struct btrfs_ioctl_defrag_range_args { #define BTRFS_MOUNT_ENOSPC_DEBUG (1 << 15) #define BTRFS_MOUNT_AUTO_DEFRAG (1 << 16) #define BTRFS_MOUNT_INODE_MAP_CACHE (1 << 17) +#define BTRFS_MOUNT_RECOVERY (1 << 18) #define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt) #define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt) @@ -1410,17 +1465,15 @@ void btrfs_set_##name(struct extent_buffer *eb, type *s, u##bits val); #define BTRFS_SETGET_HEADER_FUNCS(name, type, member, bits) \ static inline u##bits btrfs_##name(struct extent_buffer *eb) \ { \ - type *p = kmap_atomic(eb->first_page, KM_USER0); \ + type *p = page_address(eb->first_page); \ u##bits res = le##bits##_to_cpu(p->member); \ - kunmap_atomic(p, KM_USER0); \ return res; \ } \ static inline void btrfs_set_##name(struct extent_buffer *eb, \ u##bits val) \ { \ - type *p = kmap_atomic(eb->first_page, KM_USER0); \ + type *p = page_address(eb->first_page); \ p->member = cpu_to_le##bits(val); \ - kunmap_atomic(p, KM_USER0); \ } #define BTRFS_SETGET_STACK_FUNCS(name, type, member, bits) \ @@ -1975,6 +2028,55 @@ static inline bool btrfs_root_readonly(struct btrfs_root *root) return (root->root_item.flags & cpu_to_le64(BTRFS_ROOT_SUBVOL_RDONLY)) != 0; } +/* struct btrfs_root_backup */ +BTRFS_SETGET_STACK_FUNCS(backup_tree_root, struct btrfs_root_backup, + tree_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_tree_root_gen, struct btrfs_root_backup, + tree_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_tree_root_level, struct btrfs_root_backup, + tree_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_chunk_root, struct btrfs_root_backup, + chunk_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_chunk_root_gen, struct btrfs_root_backup, + chunk_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_chunk_root_level, struct btrfs_root_backup, + chunk_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_extent_root, struct btrfs_root_backup, + extent_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_extent_root_gen, struct btrfs_root_backup, + extent_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_extent_root_level, struct btrfs_root_backup, + extent_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_fs_root, struct btrfs_root_backup, + fs_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_fs_root_gen, struct btrfs_root_backup, + fs_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_fs_root_level, struct btrfs_root_backup, + fs_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_dev_root, struct btrfs_root_backup, + dev_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_dev_root_gen, struct btrfs_root_backup, + dev_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_dev_root_level, struct btrfs_root_backup, + dev_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_csum_root, struct btrfs_root_backup, + csum_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_csum_root_gen, struct btrfs_root_backup, + csum_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_csum_root_level, struct btrfs_root_backup, + csum_root_level, 8); +BTRFS_SETGET_STACK_FUNCS(backup_total_bytes, struct btrfs_root_backup, + total_bytes, 64); +BTRFS_SETGET_STACK_FUNCS(backup_bytes_used, struct btrfs_root_backup, + bytes_used, 64); +BTRFS_SETGET_STACK_FUNCS(backup_num_devices, struct btrfs_root_backup, + num_devices, 64); + /* struct btrfs_super_block */ BTRFS_SETGET_STACK_FUNCS(super_bytenr, struct btrfs_super_block, bytenr, 64); @@ -2126,14 +2228,30 @@ static inline bool btrfs_mixed_space_info(struct btrfs_space_info *space_info) (space_info->flags & BTRFS_BLOCK_GROUP_DATA)); } +static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping) +{ + return mapping_gfp_mask(mapping) & ~__GFP_FS; +} + /* extent-tree.c */ static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root, - int num_items) + unsigned num_items) { return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) * 3 * num_items; } +/* + * Doing a truncate won't result in new nodes or leaves, just what we need for + * COW. + */ +static inline u64 btrfs_calc_trunc_metadata_size(struct btrfs_root *root, + unsigned num_items) +{ + return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) * + num_items; +} + void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); @@ -2143,6 +2261,9 @@ int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, u64 num_bytes, u64 *refs, u64 *flags); int btrfs_pin_extent(struct btrfs_root *root, u64 bytenr, u64 num, int reserved); +int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + u64 bytenr, u64 num_bytes); int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 objectid, u64 offset, u64 bytenr); @@ -2193,8 +2314,8 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, u64 root_objectid, u64 owner, u64 offset); int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len); -int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo); +int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root, + u64 start, u64 len); int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, @@ -2222,9 +2343,6 @@ void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *ionde); void btrfs_clear_space_info_full(struct btrfs_fs_info *info); int btrfs_check_data_free_space(struct inode *inode, u64 bytes); void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes); -int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - int num_items); void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, @@ -2240,25 +2358,26 @@ void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv); struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root); void btrfs_free_block_rsv(struct btrfs_root *root, struct btrfs_block_rsv *rsv); -void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info, - struct btrfs_block_rsv *rsv); -int btrfs_block_rsv_add(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +int btrfs_block_rsv_add(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 num_bytes); -int btrfs_block_rsv_check(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +int btrfs_block_rsv_add_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes); +int btrfs_block_rsv_check(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, int min_factor); +int btrfs_block_rsv_refill(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, - u64 min_reserved, int min_factor); + u64 min_reserved); +int btrfs_block_rsv_refill_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved); int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, struct btrfs_block_rsv *dst_rsv, u64 num_bytes); void btrfs_block_rsv_release(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 num_bytes); -int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *rsv); int btrfs_set_block_group_ro(struct btrfs_root *root, struct btrfs_block_group_cache *cache); int btrfs_set_block_group_rw(struct btrfs_root *root, @@ -2330,7 +2449,7 @@ struct btrfs_path *btrfs_alloc_path(void); void btrfs_free_path(struct btrfs_path *p); void btrfs_set_path_blocking(struct btrfs_path *p); void btrfs_clear_path_blocking(struct btrfs_path *p, - struct extent_buffer *held); + struct extent_buffer *held, int held_rw); void btrfs_unlock_up_safe(struct btrfs_path *p, int level); int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2365,8 +2484,8 @@ static inline int btrfs_insert_empty_item(struct btrfs_trans_handle *trans, int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path); int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path); int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf); -int btrfs_drop_snapshot(struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, int update_ref); +void btrfs_drop_snapshot(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, int update_ref); int btrfs_drop_subtree(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *node, @@ -2379,6 +2498,18 @@ static inline int btrfs_fs_closing(struct btrfs_fs_info *fs_info) smp_mb(); return fs_info->closing; } +static inline void free_fs_info(struct btrfs_fs_info *fs_info) +{ + kfree(fs_info->delayed_root); + kfree(fs_info->extent_root); + kfree(fs_info->tree_root); + kfree(fs_info->chunk_root); + kfree(fs_info->dev_root); + kfree(fs_info->csum_root); + kfree(fs_info->super_copy); + kfree(fs_info->super_for_commit); + kfree(fs_info); +} /* root-item.c */ int btrfs_find_root_ref(struct btrfs_root *tree_root, @@ -2404,8 +2535,8 @@ int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct btrfs_root_item *item, struct btrfs_key *key); int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid); int btrfs_find_orphan_roots(struct btrfs_root *tree_root); -int btrfs_set_root_node(struct btrfs_root_item *item, - struct extent_buffer *node); +void btrfs_set_root_node(struct btrfs_root_item *item, + struct extent_buffer *node); void btrfs_check_and_init_root_item(struct btrfs_root_item *item); /* dir-item.c */ @@ -2510,6 +2641,9 @@ int btrfs_csum_truncate(struct btrfs_trans_handle *trans, int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, struct list_head *list, int search_commit); /* inode.c */ +struct extent_map *btrfs_get_extent_fiemap(struct inode *inode, struct page *page, + size_t pg_offset, u64 start, u64 len, + int create); /* RHEL and EL kernels have a patch that renames PG_checked to FsMisc */ #if defined(ClearPageFsMisc) && !defined(ClearPageChecked) @@ -2518,6 +2652,14 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, #define PageChecked PageFsMisc #endif +/* This forces readahead on a given range of bytes in an inode */ +static inline void btrfs_force_ra(struct address_space *mapping, + struct file_ra_state *ra, struct file *file, + pgoff_t offset, unsigned long req_size) +{ + page_cache_sync_readahead(mapping, ra, file, offset, req_size); +} + struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry); int btrfs_set_inode_index(struct inode *dir, u64 *index); int btrfs_unlink_inode(struct btrfs_trans_handle *trans, @@ -2546,14 +2688,12 @@ int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, int btrfs_merge_bio_hook(struct page *page, unsigned long offset, size_t size, struct bio *bio, unsigned long bio_flags); -unsigned long btrfs_force_ra(struct address_space *mapping, - struct file_ra_state *ra, struct file *file, - pgoff_t offset, pgoff_t last_index); int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf); int btrfs_readpage(struct file *file, struct page *page); void btrfs_evict_inode(struct inode *inode); int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc); -void btrfs_dirty_inode(struct inode *inode, int flags); +int btrfs_dirty_inode(struct inode *inode); +int btrfs_update_time(struct file *file); struct inode *btrfs_alloc_inode(struct super_block *sb); void btrfs_destroy_inode(struct inode *inode); int btrfs_drop_inode(struct inode *inode); @@ -2571,11 +2711,6 @@ int btrfs_update_inode(struct btrfs_trans_handle *trans, int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_orphan_cleanup(struct btrfs_root *root); -void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending, - u64 *bytes_to_reserve); -void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending); void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size); @@ -2602,7 +2737,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info); -int btrfs_sync_file(struct file *file, int datasync); +int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync); int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, int skip_pinned); extern const struct file_operations btrfs_file_operations; @@ -2642,13 +2777,22 @@ do { \ /* acl.c */ #ifdef CONFIG_BTRFS_FS_POSIX_ACL -int btrfs_check_acl(struct inode *inode, int mask, unsigned int flags); -#else -#define btrfs_check_acl NULL -#endif +struct posix_acl *btrfs_get_acl(struct inode *inode, int type); int btrfs_init_acl(struct btrfs_trans_handle *trans, struct inode *inode, struct inode *dir); int btrfs_acl_chmod(struct inode *inode); +#else +#define btrfs_get_acl NULL +static inline int btrfs_init_acl(struct btrfs_trans_handle *trans, + struct inode *inode, struct inode *dir) +{ + return 0; +} +static inline int btrfs_acl_chmod(struct inode *inode) +{ + return 0; +} +#endif /* relocation.c */ int btrfs_relocate_block_group(struct btrfs_root *root, u64 group_start); @@ -2680,4 +2824,20 @@ int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid); int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, struct btrfs_scrub_progress *progress); +/* reada.c */ +struct reada_control { + struct btrfs_root *root; /* tree to prefetch */ + struct btrfs_key key_start; + struct btrfs_key key_end; /* exclusive */ + atomic_t elems; + struct kref refcnt; + wait_queue_head_t wait; +}; +struct reada_control *btrfs_reada_add(struct btrfs_root *root, + struct btrfs_key *start, struct btrfs_key *end); +int btrfs_reada_wait(void *handle); +void btrfs_reada_detach(void *handle); +int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb, + u64 start, int err); + #endif diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 98c68e6..9c1eccc 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -591,7 +591,7 @@ static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, return 0; src_rsv = trans->block_rsv; - dst_rsv = &root->fs_info->global_block_rsv; + dst_rsv = &root->fs_info->delayed_block_rsv; num_bytes = btrfs_calc_trans_metadata_size(root, 1); ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); @@ -609,7 +609,7 @@ static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, if (!item->bytes_reserved) return; - rsv = &root->fs_info->global_block_rsv; + rsv = &root->fs_info->delayed_block_rsv; btrfs_block_rsv_release(root, rsv, item->bytes_reserved); } @@ -617,24 +617,102 @@ static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, static int btrfs_delayed_inode_reserve_metadata( struct btrfs_trans_handle *trans, struct btrfs_root *root, + struct inode *inode, struct btrfs_delayed_node *node) { struct btrfs_block_rsv *src_rsv; struct btrfs_block_rsv *dst_rsv; u64 num_bytes; int ret; - - if (!trans->bytes_reserved) - return 0; + int release = false; src_rsv = trans->block_rsv; - dst_rsv = &root->fs_info->global_block_rsv; + dst_rsv = &root->fs_info->delayed_block_rsv; num_bytes = btrfs_calc_trans_metadata_size(root, 1); + + /* + * btrfs_dirty_inode will update the inode under btrfs_join_transaction + * which doesn't reserve space for speed. This is a problem since we + * still need to reserve space for this update, so try to reserve the + * space. + * + * Now if src_rsv == delalloc_block_rsv we'll let it just steal since + * we're accounted for. + */ + if (!src_rsv || (!trans->bytes_reserved && + src_rsv != &root->fs_info->delalloc_block_rsv)) { + ret = btrfs_block_rsv_add_noflush(root, dst_rsv, num_bytes); + /* + * Since we're under a transaction reserve_metadata_bytes could + * try to commit the transaction which will make it return + * EAGAIN to make us stop the transaction we have, so return + * ENOSPC instead so that btrfs_dirty_inode knows what to do. + */ + if (ret == -EAGAIN) + ret = -ENOSPC; + if (!ret) + node->bytes_reserved = num_bytes; + return ret; + } else if (src_rsv == &root->fs_info->delalloc_block_rsv) { + spin_lock(&BTRFS_I(inode)->lock); + if (BTRFS_I(inode)->delalloc_meta_reserved) { + BTRFS_I(inode)->delalloc_meta_reserved = 0; + spin_unlock(&BTRFS_I(inode)->lock); + release = true; + goto migrate; + } + spin_unlock(&BTRFS_I(inode)->lock); + + /* Ok we didn't have space pre-reserved. This shouldn't happen + * too often but it can happen if we do delalloc to an existing + * inode which gets dirtied because of the time update, and then + * isn't touched again until after the transaction commits and + * then we try to write out the data. First try to be nice and + * reserve something strictly for us. If not be a pain and try + * to steal from the delalloc block rsv. + */ + ret = btrfs_block_rsv_add_noflush(root, dst_rsv, num_bytes); + if (!ret) + goto out; + + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + if (!ret) + goto out; + + /* + * Ok this is a problem, let's just steal from the global rsv + * since this really shouldn't happen that often. + */ + WARN_ON(1); + ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv, + dst_rsv, num_bytes); + goto out; + } + +migrate: ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + +out: + /* + * Migrate only takes a reservation, it doesn't touch the size of the + * block_rsv. This is to simplify people who don't normally have things + * migrated from their block rsv. If they go to release their + * reservation, that will decrease the size as well, so if migrate + * reduced size we'd end up with a negative size. But for the + * delalloc_meta_reserved stuff we will only know to drop 1 reservation, + * but we could in fact do this reserve/migrate dance several times + * between the time we did the original reservation and we'd clean it + * up. So to take care of this, release the space for the meta + * reservation here. I think it may be time for a documentation page on + * how block rsvs. work. + */ if (!ret) node->bytes_reserved = num_bytes; + if (release) + btrfs_block_rsv_release(root, src_rsv, num_bytes); + return ret; } @@ -646,7 +724,7 @@ static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root, if (!node->bytes_reserved) return; - rsv = &root->fs_info->global_block_rsv; + rsv = &root->fs_info->delayed_block_rsv; btrfs_block_rsv_release(root, rsv, node->bytes_reserved); node->bytes_reserved = 0; @@ -735,7 +813,7 @@ static int btrfs_batch_insert_items(struct btrfs_trans_handle *trans, } /* reset all the locked nodes in the patch to spinning locks. */ - btrfs_clear_path_blocking(path, NULL); + btrfs_clear_path_blocking(path, NULL, 0); /* insert the keys of the items */ ret = setup_items_for_insert(trans, root, path, keys, data_size, @@ -1026,7 +1104,7 @@ int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, path->leave_spinning = 1; block_rsv = trans->block_rsv; - trans->block_rsv = &root->fs_info->global_block_rsv; + trans->block_rsv = &root->fs_info->delayed_block_rsv; delayed_root = btrfs_get_delayed_root(root); @@ -1069,7 +1147,7 @@ static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, path->leave_spinning = 1; block_rsv = trans->block_rsv; - trans->block_rsv = &node->root->fs_info->global_block_rsv; + trans->block_rsv = &node->root->fs_info->delayed_block_rsv; ret = btrfs_insert_delayed_items(trans, path, node->root, node); if (!ret) @@ -1149,7 +1227,7 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work) goto free_path; block_rsv = trans->block_rsv; - trans->block_rsv = &root->fs_info->global_block_rsv; + trans->block_rsv = &root->fs_info->delayed_block_rsv; ret = btrfs_insert_delayed_items(trans, path, root, delayed_node); if (!ret) @@ -1641,7 +1719,7 @@ int btrfs_fill_inode(struct inode *inode, u32 *rdev) inode->i_gid = btrfs_stack_inode_gid(inode_item); btrfs_i_size_write(inode, btrfs_stack_inode_size(inode_item)); inode->i_mode = btrfs_stack_inode_mode(inode_item); - inode->i_nlink = btrfs_stack_inode_nlink(inode_item); + set_nlink(inode, btrfs_stack_inode_nlink(inode_item)); inode_set_bytes(inode, btrfs_stack_inode_nbytes(inode_item)); BTRFS_I(inode)->generation = btrfs_stack_inode_generation(inode_item); BTRFS_I(inode)->sequence = btrfs_stack_inode_sequence(inode_item); @@ -1685,12 +1763,10 @@ int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, goto release_node; } - ret = btrfs_delayed_inode_reserve_metadata(trans, root, delayed_node); - /* - * we must reserve enough space when we start a new transaction, - * so reserving metadata failure is impossible - */ - BUG_ON(ret); + ret = btrfs_delayed_inode_reserve_metadata(trans, root, inode, + delayed_node); + if (ret) + goto release_node; fill_stack_inode_item(trans, &delayed_node->inode_item, inode); delayed_node->inode_dirty = 1; diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h index 8d27af4..7083d08 100644 --- a/fs/btrfs/delayed-inode.h +++ b/fs/btrfs/delayed-inode.h @@ -25,7 +25,7 @@ #include <linux/mutex.h> #include <linux/list.h> #include <linux/wait.h> -#include <asm/atomic.h> +#include <linux/atomic.h> #include "ctree.h" diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c index 685f259..31d84e7 100644 --- a/fs/btrfs/dir-item.c +++ b/fs/btrfs/dir-item.c @@ -89,13 +89,8 @@ int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans, data_size = sizeof(*dir_item) + name_len + data_len; dir_item = insert_with_overflow(trans, root, path, &key, data_size, name, name_len); - /* - * FIXME: at some point we should handle xattr's that are larger than - * what we can fit in our leaf. We set location to NULL b/c we arent - * pointing at anything else, that will change if we store the xattr - * data in a separate inode. - */ - BUG_ON(IS_ERR(dir_item)); + if (IS_ERR(dir_item)) + return PTR_ERR(dir_item); memset(&location, 0, sizeof(location)); leaf = path->nodes[0]; @@ -203,8 +198,6 @@ struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans, struct btrfs_key key; int ins_len = mod < 0 ? -1 : 0; int cow = mod != 0; - struct btrfs_key found_key; - struct extent_buffer *leaf; key.objectid = dir; btrfs_set_key_type(&key, BTRFS_DIR_ITEM_KEY); @@ -214,18 +207,7 @@ struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans, ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow); if (ret < 0) return ERR_PTR(ret); - if (ret > 0) { - if (path->slots[0] == 0) - return NULL; - path->slots[0]--; - } - - leaf = path->nodes[0]; - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - - if (found_key.objectid != dir || - btrfs_key_type(&found_key) != BTRFS_DIR_ITEM_KEY || - found_key.offset != key.offset) + if (ret > 0) return NULL; return btrfs_match_dir_item_name(root, path, name, name_len); @@ -320,8 +302,6 @@ struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans, struct btrfs_key key; int ins_len = mod < 0 ? -1 : 0; int cow = mod != 0; - struct btrfs_key found_key; - struct extent_buffer *leaf; key.objectid = dir; btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY); @@ -329,18 +309,7 @@ struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans, ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow); if (ret < 0) return ERR_PTR(ret); - if (ret > 0) { - if (path->slots[0] == 0) - return NULL; - path->slots[0]--; - } - - leaf = path->nodes[0]; - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - - if (found_key.objectid != dir || - btrfs_key_type(&found_key) != BTRFS_XATTR_ITEM_KEY || - found_key.offset != key.offset) + if (ret > 0) return NULL; return btrfs_match_dir_item_name(root, path, name, name_len); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 57106a9..b8fc473 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -100,38 +100,83 @@ struct async_submit_bio { struct btrfs_work work; }; -/* These are used to set the lockdep class on the extent buffer locks. - * The class is set by the readpage_end_io_hook after the buffer has - * passed csum validation but before the pages are unlocked. +/* + * Lockdep class keys for extent_buffer->lock's in this root. For a given + * eb, the lockdep key is determined by the btrfs_root it belongs to and + * the level the eb occupies in the tree. + * + * Different roots are used for different purposes and may nest inside each + * other and they require separate keysets. As lockdep keys should be + * static, assign keysets according to the purpose of the root as indicated + * by btrfs_root->objectid. This ensures that all special purpose roots + * have separate keysets. * - * The lockdep class is also set by btrfs_init_new_buffer on freshly - * allocated blocks. + * Lock-nesting across peer nodes is always done with the immediate parent + * node locked thus preventing deadlock. As lockdep doesn't know this, use + * subclass to avoid triggering lockdep warning in such cases. * - * The class is based on the level in the tree block, which allows lockdep - * to know that lower nodes nest inside the locks of higher nodes. + * The key is set by the readpage_end_io_hook after the buffer has passed + * csum validation but before the pages are unlocked. It is also set by + * btrfs_init_new_buffer on freshly allocated blocks. * - * We also add a check to make sure the highest level of the tree is - * the same as our lockdep setup here. If BTRFS_MAX_LEVEL changes, this - * code needs update as well. + * We also add a check to make sure the highest level of the tree is the + * same as our lockdep setup here. If BTRFS_MAX_LEVEL changes, this code + * needs update as well. */ #ifdef CONFIG_DEBUG_LOCK_ALLOC # if BTRFS_MAX_LEVEL != 8 # error # endif -static struct lock_class_key btrfs_eb_class[BTRFS_MAX_LEVEL + 1]; -static const char *btrfs_eb_name[BTRFS_MAX_LEVEL + 1] = { - /* leaf */ - "btrfs-extent-00", - "btrfs-extent-01", - "btrfs-extent-02", - "btrfs-extent-03", - "btrfs-extent-04", - "btrfs-extent-05", - "btrfs-extent-06", - "btrfs-extent-07", - /* highest possible level */ - "btrfs-extent-08", + +static struct btrfs_lockdep_keyset { + u64 id; /* root objectid */ + const char *name_stem; /* lock name stem */ + char names[BTRFS_MAX_LEVEL + 1][20]; + struct lock_class_key keys[BTRFS_MAX_LEVEL + 1]; +} btrfs_lockdep_keysets[] = { + { .id = BTRFS_ROOT_TREE_OBJECTID, .name_stem = "root" }, + { .id = BTRFS_EXTENT_TREE_OBJECTID, .name_stem = "extent" }, + { .id = BTRFS_CHUNK_TREE_OBJECTID, .name_stem = "chunk" }, + { .id = BTRFS_DEV_TREE_OBJECTID, .name_stem = "dev" }, + { .id = BTRFS_FS_TREE_OBJECTID, .name_stem = "fs" }, + { .id = BTRFS_CSUM_TREE_OBJECTID, .name_stem = "csum" }, + { .id = BTRFS_ORPHAN_OBJECTID, .name_stem = "orphan" }, + { .id = BTRFS_TREE_LOG_OBJECTID, .name_stem = "log" }, + { .id = BTRFS_TREE_RELOC_OBJECTID, .name_stem = "treloc" }, + { .id = BTRFS_DATA_RELOC_TREE_OBJECTID, .name_stem = "dreloc" }, + { .id = 0, .name_stem = "tree" }, }; + +void __init btrfs_init_lockdep(void) +{ + int i, j; + + /* initialize lockdep class names */ + for (i = 0; i < ARRAY_SIZE(btrfs_lockdep_keysets); i++) { + struct btrfs_lockdep_keyset *ks = &btrfs_lockdep_keysets[i]; + + for (j = 0; j < ARRAY_SIZE(ks->names); j++) + snprintf(ks->names[j], sizeof(ks->names[j]), + "btrfs-%s-%02d", ks->name_stem, j); + } +} + +void btrfs_set_buffer_lockdep_class(u64 objectid, struct extent_buffer *eb, + int level) +{ + struct btrfs_lockdep_keyset *ks; + + BUG_ON(level >= ARRAY_SIZE(ks->keys)); + + /* find the matching keyset, id 0 is the default entry */ + for (ks = btrfs_lockdep_keysets; ks->id; ks++) + if (ks->id == objectid) + break; + + lockdep_set_class_and_name(&eb->lock, + &ks->keys[level], ks->names[level]); +} + #endif /* @@ -211,13 +256,11 @@ void btrfs_csum_final(u32 crc, char *result) static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf, int verify) { - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); char *result = NULL; unsigned long len; unsigned long cur_len; unsigned long offset = BTRFS_CSUM_SIZE; - char *map_token = NULL; char *kaddr; unsigned long map_start; unsigned long map_len; @@ -228,8 +271,7 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf, len = buf->len - offset; while (len > 0) { err = map_private_extent_buffer(buf, offset, 32, - &map_token, &kaddr, - &map_start, &map_len, KM_USER0); + &kaddr, &map_start, &map_len); if (err) return 1; cur_len = min(len, map_len - (offset - map_start)); @@ -237,7 +279,6 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf, crc, cur_len); len -= cur_len; offset += cur_len; - unmap_extent_buffer(buf, map_token, KM_USER0); } if (csum_size > sizeof(inline_result)) { result = kzalloc(csum_size * sizeof(char), GFP_NOFS); @@ -325,7 +366,8 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags); io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree; while (1) { - ret = read_extent_buffer_pages(io_tree, eb, start, 1, + ret = read_extent_buffer_pages(io_tree, eb, start, + WAIT_COMPLETE, btree_get_extent, mirror_num); if (!ret && !verify_parent_transid(io_tree, eb, parent_transid)) @@ -494,15 +536,6 @@ static noinline int check_leaf(struct btrfs_root *root, return 0; } -#ifdef CONFIG_DEBUG_LOCK_ALLOC -void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, int level) -{ - lockdep_set_class_and_name(&eb->lock, - &btrfs_eb_class[level], - btrfs_eb_name[level]); -} -#endif - static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, struct extent_state *state) { @@ -553,7 +586,8 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, } found_level = btrfs_header_level(eb); - btrfs_set_buffer_lockdep_class(eb, found_level); + btrfs_set_buffer_lockdep_class(btrfs_header_owner(eb), + eb, found_level); ret = csum_tree_block(root, eb, 1); if (ret) { @@ -574,11 +608,48 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, end = min_t(u64, eb->len, PAGE_CACHE_SIZE); end = eb->start + end - 1; err: + if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) { + clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags); + btree_readahead_hook(root, eb, eb->start, ret); + } + free_extent_buffer(eb); out: return ret; } +static int btree_io_failed_hook(struct bio *failed_bio, + struct page *page, u64 start, u64 end, + int mirror_num, struct extent_state *state) +{ + struct extent_io_tree *tree; + unsigned long len; + struct extent_buffer *eb; + struct btrfs_root *root = BTRFS_I(page->mapping->host)->root; + + tree = &BTRFS_I(page->mapping->host)->io_tree; + if (page->private == EXTENT_PAGE_PRIVATE) + goto out; + if (!page->private) + goto out; + + len = page->private >> 2; + WARN_ON(len == 0); + + eb = alloc_extent_buffer(tree, start, len, page); + if (eb == NULL) + goto out; + + if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) { + clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags); + btree_readahead_hook(root, eb, eb->start, -EIO); + } + free_extent_buffer(eb); + +out: + return -EIO; /* we fixed nothing */ +} + static void end_workqueue_bio(struct bio *bio, int err) { struct end_io_wq *end_io_wq = bio->bi_private; @@ -875,7 +946,7 @@ static int btree_readpage(struct file *file, struct page *page) { struct extent_io_tree *tree; tree = &BTRFS_I(page->mapping->host)->io_tree; - return extent_read_full_page(tree, page, btree_get_extent); + return extent_read_full_page(tree, page, btree_get_extent, 0); } static int btree_releasepage(struct page *page, gfp_t gfp_flags) @@ -941,11 +1012,43 @@ int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, if (!buf) return 0; read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree, - buf, 0, 0, btree_get_extent, 0); + buf, 0, WAIT_NONE, btree_get_extent, 0); free_extent_buffer(buf); return ret; } +int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize, + int mirror_num, struct extent_buffer **eb) +{ + struct extent_buffer *buf = NULL; + struct inode *btree_inode = root->fs_info->btree_inode; + struct extent_io_tree *io_tree = &BTRFS_I(btree_inode)->io_tree; + int ret; + + buf = btrfs_find_create_tree_block(root, bytenr, blocksize); + if (!buf) + return 0; + + set_bit(EXTENT_BUFFER_READAHEAD, &buf->bflags); + + ret = read_extent_buffer_pages(io_tree, buf, 0, WAIT_PAGE_LOCK, + btree_get_extent, mirror_num); + if (ret) { + free_extent_buffer(buf); + return ret; + } + + if (test_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags)) { + free_extent_buffer(buf); + return -EIO; + } else if (extent_buffer_uptodate(io_tree, buf, NULL)) { + *eb = buf; + } else { + free_extent_buffer(buf); + } + return 0; +} + struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize) { @@ -1078,12 +1181,7 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, init_completion(&root->kobj_unregister); root->defrag_running = 0; root->root_key.objectid = objectid; - root->anon_super.s_root = NULL; - root->anon_super.s_dev = 0; - INIT_LIST_HEAD(&root->anon_super.s_list); - INIT_LIST_HEAD(&root->anon_super.s_instances); - init_rwsem(&root->anon_super.s_umount); - + root->anon_dev = 0; return 0; } @@ -1107,10 +1205,12 @@ static int find_and_setup_root(struct btrfs_root *tree_root, generation = btrfs_root_generation(&root->root_item); blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item)); + root->commit_root = NULL; root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item), blocksize, generation); if (!root->node || !btrfs_buffer_uptodate(root->node, generation)) { free_extent_buffer(root->node); + root->node = NULL; return -EIO; } root->commit_root = btrfs_root_node(root); @@ -1312,7 +1412,7 @@ again: spin_lock_init(&root->cache_lock); init_waitqueue_head(&root->cache_wait); - ret = set_anon_super(&root->anon_super, NULL); + ret = get_anon_bdev(&root->anon_dev); if (ret) goto fail; @@ -1549,6 +1649,235 @@ sleep: return 0; } +/* + * this will find the highest generation in the array of + * root backups. The index of the highest array is returned, + * or -1 if we can't find anything. + * + * We check to make sure the array is valid by comparing the + * generation of the latest root in the array with the generation + * in the super block. If they don't match we pitch it. + */ +static int find_newest_super_backup(struct btrfs_fs_info *info, u64 newest_gen) +{ + u64 cur; + int newest_index = -1; + struct btrfs_root_backup *root_backup; + int i; + + for (i = 0; i < BTRFS_NUM_BACKUP_ROOTS; i++) { + root_backup = info->super_copy->super_roots + i; + cur = btrfs_backup_tree_root_gen(root_backup); + if (cur == newest_gen) + newest_index = i; + } + + /* check to see if we actually wrapped around */ + if (newest_index == BTRFS_NUM_BACKUP_ROOTS - 1) { + root_backup = info->super_copy->super_roots; + cur = btrfs_backup_tree_root_gen(root_backup); + if (cur == newest_gen) + newest_index = 0; + } + return newest_index; +} + + +/* + * find the oldest backup so we know where to store new entries + * in the backup array. This will set the backup_root_index + * field in the fs_info struct + */ +static void find_oldest_super_backup(struct btrfs_fs_info *info, + u64 newest_gen) +{ + int newest_index = -1; + + newest_index = find_newest_super_backup(info, newest_gen); + /* if there was garbage in there, just move along */ + if (newest_index == -1) { + info->backup_root_index = 0; + } else { + info->backup_root_index = (newest_index + 1) % BTRFS_NUM_BACKUP_ROOTS; + } +} + +/* + * copy all the root pointers into the super backup array. + * this will bump the backup pointer by one when it is + * done + */ +static void backup_super_roots(struct btrfs_fs_info *info) +{ + int next_backup; + struct btrfs_root_backup *root_backup; + int last_backup; + + next_backup = info->backup_root_index; + last_backup = (next_backup + BTRFS_NUM_BACKUP_ROOTS - 1) % + BTRFS_NUM_BACKUP_ROOTS; + + /* + * just overwrite the last backup if we're at the same generation + * this happens only at umount + */ + root_backup = info->super_for_commit->super_roots + last_backup; + if (btrfs_backup_tree_root_gen(root_backup) == + btrfs_header_generation(info->tree_root->node)) + next_backup = last_backup; + + root_backup = info->super_for_commit->super_roots + next_backup; + + /* + * make sure all of our padding and empty slots get zero filled + * regardless of which ones we use today + */ + memset(root_backup, 0, sizeof(*root_backup)); + + info->backup_root_index = (next_backup + 1) % BTRFS_NUM_BACKUP_ROOTS; + + btrfs_set_backup_tree_root(root_backup, info->tree_root->node->start); + btrfs_set_backup_tree_root_gen(root_backup, + btrfs_header_generation(info->tree_root->node)); + + btrfs_set_backup_tree_root_level(root_backup, + btrfs_header_level(info->tree_root->node)); + + btrfs_set_backup_chunk_root(root_backup, info->chunk_root->node->start); + btrfs_set_backup_chunk_root_gen(root_backup, + btrfs_header_generation(info->chunk_root->node)); + btrfs_set_backup_chunk_root_level(root_backup, + btrfs_header_level(info->chunk_root->node)); + + btrfs_set_backup_extent_root(root_backup, info->extent_root->node->start); + btrfs_set_backup_extent_root_gen(root_backup, + btrfs_header_generation(info->extent_root->node)); + btrfs_set_backup_extent_root_level(root_backup, + btrfs_header_level(info->extent_root->node)); + + /* + * we might commit during log recovery, which happens before we set + * the fs_root. Make sure it is valid before we fill it in. + */ + if (info->fs_root && info->fs_root->node) { + btrfs_set_backup_fs_root(root_backup, + info->fs_root->node->start); + btrfs_set_backup_fs_root_gen(root_backup, + btrfs_header_generation(info->fs_root->node)); + btrfs_set_backup_fs_root_level(root_backup, + btrfs_header_level(info->fs_root->node)); + } + + btrfs_set_backup_dev_root(root_backup, info->dev_root->node->start); + btrfs_set_backup_dev_root_gen(root_backup, + btrfs_header_generation(info->dev_root->node)); + btrfs_set_backup_dev_root_level(root_backup, + btrfs_header_level(info->dev_root->node)); + + btrfs_set_backup_csum_root(root_backup, info->csum_root->node->start); + btrfs_set_backup_csum_root_gen(root_backup, + btrfs_header_generation(info->csum_root->node)); + btrfs_set_backup_csum_root_level(root_backup, + btrfs_header_level(info->csum_root->node)); + + btrfs_set_backup_total_bytes(root_backup, + btrfs_super_total_bytes(info->super_copy)); + btrfs_set_backup_bytes_used(root_backup, + btrfs_super_bytes_used(info->super_copy)); + btrfs_set_backup_num_devices(root_backup, + btrfs_super_num_devices(info->super_copy)); + + /* + * if we don't copy this out to the super_copy, it won't get remembered + * for the next commit + */ + memcpy(&info->super_copy->super_roots, + &info->super_for_commit->super_roots, + sizeof(*root_backup) * BTRFS_NUM_BACKUP_ROOTS); +} + +/* + * this copies info out of the root backup array and back into + * the in-memory super block. It is meant to help iterate through + * the array, so you send it the number of backups you've already + * tried and the last backup index you used. + * + * this returns -1 when it has tried all the backups + */ +static noinline int next_root_backup(struct btrfs_fs_info *info, + struct btrfs_super_block *super, + int *num_backups_tried, int *backup_index) +{ + struct btrfs_root_backup *root_backup; + int newest = *backup_index; + + if (*num_backups_tried == 0) { + u64 gen = btrfs_super_generation(super); + + newest = find_newest_super_backup(info, gen); + if (newest == -1) + return -1; + + *backup_index = newest; + *num_backups_tried = 1; + } else if (*num_backups_tried == BTRFS_NUM_BACKUP_ROOTS) { + /* we've tried all the backups, all done */ + return -1; + } else { + /* jump to the next oldest backup */ + newest = (*backup_index + BTRFS_NUM_BACKUP_ROOTS - 1) % + BTRFS_NUM_BACKUP_ROOTS; + *backup_index = newest; + *num_backups_tried += 1; + } + root_backup = super->super_roots + newest; + + btrfs_set_super_generation(super, + btrfs_backup_tree_root_gen(root_backup)); + btrfs_set_super_root(super, btrfs_backup_tree_root(root_backup)); + btrfs_set_super_root_level(super, + btrfs_backup_tree_root_level(root_backup)); + btrfs_set_super_bytes_used(super, btrfs_backup_bytes_used(root_backup)); + + /* + * fixme: the total bytes and num_devices need to match or we should + * need a fsck + */ + btrfs_set_super_total_bytes(super, btrfs_backup_total_bytes(root_backup)); + btrfs_set_super_num_devices(super, btrfs_backup_num_devices(root_backup)); + return 0; +} + +/* helper to cleanup tree roots */ +static void free_root_pointers(struct btrfs_fs_info *info, int chunk_root) +{ + free_extent_buffer(info->tree_root->node); + free_extent_buffer(info->tree_root->commit_root); + free_extent_buffer(info->dev_root->node); + free_extent_buffer(info->dev_root->commit_root); + free_extent_buffer(info->extent_root->node); + free_extent_buffer(info->extent_root->commit_root); + free_extent_buffer(info->csum_root->node); + free_extent_buffer(info->csum_root->commit_root); + + info->tree_root->node = NULL; + info->tree_root->commit_root = NULL; + info->dev_root->node = NULL; + info->dev_root->commit_root = NULL; + info->extent_root->node = NULL; + info->extent_root->commit_root = NULL; + info->csum_root->node = NULL; + info->csum_root->commit_root = NULL; + + if (chunk_root) { + free_extent_buffer(info->chunk_root->node); + free_extent_buffer(info->chunk_root->commit_root); + info->chunk_root->node = NULL; + info->chunk_root->commit_root = NULL; + } +} + + struct btrfs_root *open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_devices, char *options) @@ -1562,29 +1891,32 @@ struct btrfs_root *open_ctree(struct super_block *sb, u64 features; struct btrfs_key location; struct buffer_head *bh; - struct btrfs_root *extent_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); - struct btrfs_root *csum_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); + struct btrfs_super_block *disk_super; struct btrfs_root *tree_root = btrfs_sb(sb); - struct btrfs_fs_info *fs_info = NULL; - struct btrfs_root *chunk_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); - struct btrfs_root *dev_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); + struct btrfs_fs_info *fs_info = tree_root->fs_info; + struct btrfs_root *extent_root; + struct btrfs_root *csum_root; + struct btrfs_root *chunk_root; + struct btrfs_root *dev_root; struct btrfs_root *log_tree_root; - int ret; int err = -EINVAL; - - struct btrfs_super_block *disk_super; - - if (!extent_root || !tree_root || !tree_root->fs_info || - !chunk_root || !dev_root || !csum_root) { + int num_backups_tried = 0; + int backup_index = 0; + + extent_root = fs_info->extent_root = + kzalloc(sizeof(struct btrfs_root), GFP_NOFS); + csum_root = fs_info->csum_root = + kzalloc(sizeof(struct btrfs_root), GFP_NOFS); + chunk_root = fs_info->chunk_root = + kzalloc(sizeof(struct btrfs_root), GFP_NOFS); + dev_root = fs_info->dev_root = + kzalloc(sizeof(struct btrfs_root), GFP_NOFS); + + if (!extent_root || !csum_root || !chunk_root || !dev_root) { err = -ENOMEM; goto fail; } - fs_info = tree_root->fs_info; ret = init_srcu_struct(&fs_info->subvol_srcu); if (ret) { @@ -1604,7 +1936,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, goto fail_bdi; } - fs_info->btree_inode->i_mapping->flags &= ~__GFP_FS; + mapping_set_gfp_mask(fs_info->btree_inode->i_mapping, GFP_NOFS); INIT_RADIX_TREE(&fs_info->fs_roots_radix, GFP_ATOMIC); INIT_LIST_HEAD(&fs_info->trans_list); @@ -1620,15 +1952,10 @@ struct btrfs_root *open_ctree(struct super_block *sb, spin_lock_init(&fs_info->fs_roots_radix_lock); spin_lock_init(&fs_info->delayed_iput_lock); spin_lock_init(&fs_info->defrag_inodes_lock); + spin_lock_init(&fs_info->free_chunk_lock); mutex_init(&fs_info->reloc_mutex); init_completion(&fs_info->kobj_unregister); - fs_info->tree_root = tree_root; - fs_info->extent_root = extent_root; - fs_info->csum_root = csum_root; - fs_info->chunk_root = chunk_root; - fs_info->dev_root = dev_root; - fs_info->fs_devices = fs_devices; INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots); INIT_LIST_HEAD(&fs_info->space_info); btrfs_mapping_init(&fs_info->mapping_tree); @@ -1637,8 +1964,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_init_block_rsv(&fs_info->trans_block_rsv); btrfs_init_block_rsv(&fs_info->chunk_block_rsv); btrfs_init_block_rsv(&fs_info->empty_block_rsv); - INIT_LIST_HEAD(&fs_info->durable_block_rsv_list); - mutex_init(&fs_info->durable_block_rsv_mutex); + btrfs_init_block_rsv(&fs_info->delayed_block_rsv); atomic_set(&fs_info->nr_async_submits, 0); atomic_set(&fs_info->async_delalloc_pages, 0); atomic_set(&fs_info->async_submit_draining, 0); @@ -1649,6 +1975,11 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->metadata_ratio = 0; fs_info->defrag_inodes = RB_ROOT; fs_info->trans_no_join = 0; + fs_info->free_chunk_space = 0; + + /* readahead state */ + INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_WAIT); + spin_lock_init(&fs_info->reada_lock); fs_info->thread_pool_size = min_t(unsigned long, num_online_cpus() + 2, 8); @@ -1677,7 +2008,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, sb->s_bdi = &fs_info->bdi; fs_info->btree_inode->i_ino = BTRFS_BTREE_INODE_OBJECTID; - fs_info->btree_inode->i_nlink = 1; + set_nlink(fs_info->btree_inode, 1); /* * we set the i_size on the btree inode to the max possible int. * the real end of the address space is determined by all of @@ -1738,14 +2069,14 @@ struct btrfs_root *open_ctree(struct super_block *sb, goto fail_alloc; } - memcpy(&fs_info->super_copy, bh->b_data, sizeof(fs_info->super_copy)); - memcpy(&fs_info->super_for_commit, &fs_info->super_copy, - sizeof(fs_info->super_for_commit)); + memcpy(fs_info->super_copy, bh->b_data, sizeof(*fs_info->super_copy)); + memcpy(fs_info->super_for_commit, fs_info->super_copy, + sizeof(*fs_info->super_for_commit)); brelse(bh); - memcpy(fs_info->fsid, fs_info->super_copy.fsid, BTRFS_FSID_SIZE); + memcpy(fs_info->fsid, fs_info->super_copy->fsid, BTRFS_FSID_SIZE); - disk_super = &fs_info->super_copy; + disk_super = fs_info->super_copy; if (!btrfs_super_root(disk_super)) goto fail_alloc; @@ -1755,6 +2086,13 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY); /* + * run through our array of backup supers and setup + * our ring pointer to the oldest one + */ + generation = btrfs_super_generation(disk_super); + find_oldest_super_backup(fs_info, generation); + + /* * In the long term, we'll store the compression type in the super * block, and it'll be used for per file compression control. */ @@ -1808,6 +2146,9 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->thread_pool_size), &fs_info->generic_worker); + btrfs_init_workers(&fs_info->caching_workers, "cache", + 2, &fs_info->generic_worker); + /* a higher idle thresh on the submit workers makes it much more * likely that bios will be send down in a sane order to the * devices @@ -1839,6 +2180,9 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_init_workers(&fs_info->delayed_workers, "delayed-meta", fs_info->thread_pool_size, &fs_info->generic_worker); + btrfs_init_workers(&fs_info->readahead_workers, "readahead", + fs_info->thread_pool_size, + &fs_info->generic_worker); /* * endios are largely parallel and should have a very @@ -1849,18 +2193,29 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->endio_write_workers.idle_thresh = 2; fs_info->endio_meta_write_workers.idle_thresh = 2; + fs_info->readahead_workers.idle_thresh = 2; - btrfs_start_workers(&fs_info->workers, 1); - btrfs_start_workers(&fs_info->generic_worker, 1); - btrfs_start_workers(&fs_info->submit_workers, 1); - btrfs_start_workers(&fs_info->delalloc_workers, 1); - btrfs_start_workers(&fs_info->fixup_workers, 1); - btrfs_start_workers(&fs_info->endio_workers, 1); - btrfs_start_workers(&fs_info->endio_meta_workers, 1); - btrfs_start_workers(&fs_info->endio_meta_write_workers, 1); - btrfs_start_workers(&fs_info->endio_write_workers, 1); - btrfs_start_workers(&fs_info->endio_freespace_worker, 1); - btrfs_start_workers(&fs_info->delayed_workers, 1); + /* + * btrfs_start_workers can really only fail because of ENOMEM so just + * return -ENOMEM if any of these fail. + */ + ret = btrfs_start_workers(&fs_info->workers); + ret |= btrfs_start_workers(&fs_info->generic_worker); + ret |= btrfs_start_workers(&fs_info->submit_workers); + ret |= btrfs_start_workers(&fs_info->delalloc_workers); + ret |= btrfs_start_workers(&fs_info->fixup_workers); + ret |= btrfs_start_workers(&fs_info->endio_workers); + ret |= btrfs_start_workers(&fs_info->endio_meta_workers); + ret |= btrfs_start_workers(&fs_info->endio_meta_write_workers); + ret |= btrfs_start_workers(&fs_info->endio_write_workers); + ret |= btrfs_start_workers(&fs_info->endio_freespace_worker); + ret |= btrfs_start_workers(&fs_info->delayed_workers); + ret |= btrfs_start_workers(&fs_info->caching_workers); + ret |= btrfs_start_workers(&fs_info->readahead_workers); + if (ret) { + ret = -ENOMEM; + goto fail_sb_buffer; + } fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super); fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages, @@ -1907,7 +2262,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, if (!test_bit(EXTENT_BUFFER_UPTODATE, &chunk_root->node->bflags)) { printk(KERN_WARNING "btrfs: failed to read chunk root on %s\n", sb->s_id); - goto fail_chunk_root; + goto fail_tree_roots; } btrfs_set_root_node(&chunk_root->root_item, chunk_root->node); chunk_root->commit_root = btrfs_root_node(chunk_root); @@ -1922,11 +2277,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, if (ret) { printk(KERN_WARNING "btrfs: failed to read chunk tree on %s\n", sb->s_id); - goto fail_chunk_root; + goto fail_tree_roots; } btrfs_close_extra_devices(fs_devices); +retry_root_backup: blocksize = btrfs_level_size(tree_root, btrfs_super_root_level(disk_super)); generation = btrfs_super_generation(disk_super); @@ -1934,32 +2290,33 @@ struct btrfs_root *open_ctree(struct super_block *sb, tree_root->node = read_tree_block(tree_root, btrfs_super_root(disk_super), blocksize, generation); - if (!tree_root->node) - goto fail_chunk_root; - if (!test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) { + if (!tree_root->node || + !test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) { printk(KERN_WARNING "btrfs: failed to read tree root on %s\n", sb->s_id); - goto fail_tree_root; + + goto recovery_tree_root; } + btrfs_set_root_node(&tree_root->root_item, tree_root->node); tree_root->commit_root = btrfs_root_node(tree_root); ret = find_and_setup_root(tree_root, fs_info, BTRFS_EXTENT_TREE_OBJECTID, extent_root); if (ret) - goto fail_tree_root; + goto recovery_tree_root; extent_root->track_dirty = 1; ret = find_and_setup_root(tree_root, fs_info, BTRFS_DEV_TREE_OBJECTID, dev_root); if (ret) - goto fail_extent_root; + goto recovery_tree_root; dev_root->track_dirty = 1; ret = find_and_setup_root(tree_root, fs_info, BTRFS_CSUM_TREE_OBJECTID, csum_root); if (ret) - goto fail_dev_root; + goto recovery_tree_root; csum_root->track_dirty = 1; @@ -2092,22 +2449,13 @@ fail_cleaner: fail_block_groups: btrfs_free_block_groups(fs_info); - free_extent_buffer(csum_root->node); - free_extent_buffer(csum_root->commit_root); -fail_dev_root: - free_extent_buffer(dev_root->node); - free_extent_buffer(dev_root->commit_root); -fail_extent_root: - free_extent_buffer(extent_root->node); - free_extent_buffer(extent_root->commit_root); -fail_tree_root: - free_extent_buffer(tree_root->node); - free_extent_buffer(tree_root->commit_root); -fail_chunk_root: - free_extent_buffer(chunk_root->node); - free_extent_buffer(chunk_root->commit_root); + +fail_tree_roots: + free_root_pointers(fs_info, 1); + fail_sb_buffer: btrfs_stop_workers(&fs_info->generic_worker); + btrfs_stop_workers(&fs_info->readahead_workers); btrfs_stop_workers(&fs_info->fixup_workers); btrfs_stop_workers(&fs_info->delalloc_workers); btrfs_stop_workers(&fs_info->workers); @@ -2118,26 +2466,39 @@ fail_sb_buffer: btrfs_stop_workers(&fs_info->endio_freespace_worker); btrfs_stop_workers(&fs_info->submit_workers); btrfs_stop_workers(&fs_info->delayed_workers); + btrfs_stop_workers(&fs_info->caching_workers); fail_alloc: - kfree(fs_info->delayed_root); fail_iput: + btrfs_mapping_tree_free(&fs_info->mapping_tree); + invalidate_inode_pages2(fs_info->btree_inode->i_mapping); iput(fs_info->btree_inode); - - btrfs_close_devices(fs_info->fs_devices); - btrfs_mapping_tree_free(&fs_info->mapping_tree); fail_bdi: bdi_destroy(&fs_info->bdi); fail_srcu: cleanup_srcu_struct(&fs_info->subvol_srcu); fail: - kfree(extent_root); - kfree(tree_root); - kfree(fs_info); - kfree(chunk_root); - kfree(dev_root); - kfree(csum_root); + btrfs_close_devices(fs_info->fs_devices); + free_fs_info(fs_info); return ERR_PTR(err); + +recovery_tree_root: + if (!btrfs_test_opt(tree_root, RECOVERY)) + goto fail_tree_roots; + + free_root_pointers(fs_info, 0); + + /* don't use the log in recovery mode, it won't be valid */ + btrfs_set_super_log_root(disk_super, 0); + + /* we can't trust the free space cache either */ + btrfs_set_opt(fs_info->mount_opt, CLEAR_CACHE); + + ret = next_root_backup(fs_info, fs_info->super_copy, + &num_backups_tried, &backup_index); + if (ret == -1) + goto fail_block_groups; + goto retry_root_backup; } static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate) @@ -2221,22 +2582,10 @@ static int write_dev_supers(struct btrfs_device *device, int errors = 0; u32 crc; u64 bytenr; - int last_barrier = 0; if (max_mirrors == 0) max_mirrors = BTRFS_SUPER_MIRROR_MAX; - /* make sure only the last submit_bh does a barrier */ - if (do_barriers) { - for (i = 0; i < max_mirrors; i++) { - bytenr = btrfs_sb_offset(i); - if (bytenr + BTRFS_SUPER_INFO_SIZE >= - device->total_bytes) - break; - last_barrier = i; - } - } - for (i = 0; i < max_mirrors; i++) { bytenr = btrfs_sb_offset(i); if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes) @@ -2282,17 +2631,140 @@ static int write_dev_supers(struct btrfs_device *device, bh->b_end_io = btrfs_end_buffer_write_sync; } - if (i == last_barrier && do_barriers) - ret = submit_bh(WRITE_FLUSH_FUA, bh); - else - ret = submit_bh(WRITE_SYNC, bh); - + /* + * we fua the first super. The others we allow + * to go down lazy. + */ + ret = submit_bh(WRITE_FUA, bh); if (ret) errors++; } return errors < i ? 0 : -1; } +/* + * endio for the write_dev_flush, this will wake anyone waiting + * for the barrier when it is done + */ +static void btrfs_end_empty_barrier(struct bio *bio, int err) +{ + if (err) { + if (err == -EOPNOTSUPP) + set_bit(BIO_EOPNOTSUPP, &bio->bi_flags); + clear_bit(BIO_UPTODATE, &bio->bi_flags); + } + if (bio->bi_private) + complete(bio->bi_private); + bio_put(bio); +} + +/* + * trigger flushes for one the devices. If you pass wait == 0, the flushes are + * sent down. With wait == 1, it waits for the previous flush. + * + * any device where the flush fails with eopnotsupp are flagged as not-barrier + * capable + */ +static int write_dev_flush(struct btrfs_device *device, int wait) +{ + struct bio *bio; + int ret = 0; + + if (device->nobarriers) + return 0; + + if (wait) { + bio = device->flush_bio; + if (!bio) + return 0; + + wait_for_completion(&device->flush_wait); + + if (bio_flagged(bio, BIO_EOPNOTSUPP)) { + printk("btrfs: disabling barriers on dev %s\n", + device->name); + device->nobarriers = 1; + } + if (!bio_flagged(bio, BIO_UPTODATE)) { + ret = -EIO; + } + + /* drop the reference from the wait == 0 run */ + bio_put(bio); + device->flush_bio = NULL; + + return ret; + } + + /* + * one reference for us, and we leave it for the + * caller + */ + device->flush_bio = NULL;; + bio = bio_alloc(GFP_NOFS, 0); + if (!bio) + return -ENOMEM; + + bio->bi_end_io = btrfs_end_empty_barrier; + bio->bi_bdev = device->bdev; + init_completion(&device->flush_wait); + bio->bi_private = &device->flush_wait; + device->flush_bio = bio; + + bio_get(bio); + submit_bio(WRITE_FLUSH, bio); + + return 0; +} + +/* + * send an empty flush down to each device in parallel, + * then wait for them + */ +static int barrier_all_devices(struct btrfs_fs_info *info) +{ + struct list_head *head; + struct btrfs_device *dev; + int errors = 0; + int ret; + + /* send down all the barriers */ + head = &info->fs_devices->devices; + list_for_each_entry_rcu(dev, head, dev_list) { + if (dev->missing) + continue; + if (!dev->bdev) { + errors++; + continue; + } + if (!dev->in_fs_metadata || !dev->writeable) + continue; + + ret = write_dev_flush(dev, 0); + if (ret) + errors++; + } + + /* wait for all the barriers */ + list_for_each_entry_rcu(dev, head, dev_list) { + if (dev->missing) + continue; + if (!dev->bdev) { + errors++; + continue; + } + if (!dev->in_fs_metadata || !dev->writeable) + continue; + + ret = write_dev_flush(dev, 1); + if (ret) + errors++; + } + if (errors) + return -EIO; + return 0; +} + int write_all_supers(struct btrfs_root *root, int max_mirrors) { struct list_head *head; @@ -2305,14 +2777,19 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) int total_errors = 0; u64 flags; - max_errors = btrfs_super_num_devices(&root->fs_info->super_copy) - 1; + max_errors = btrfs_super_num_devices(root->fs_info->super_copy) - 1; do_barriers = !btrfs_test_opt(root, NOBARRIER); + backup_super_roots(root->fs_info); - sb = &root->fs_info->super_for_commit; + sb = root->fs_info->super_for_commit; dev_item = &sb->dev_item; mutex_lock(&root->fs_info->fs_devices->device_list_mutex); head = &root->fs_info->fs_devices->devices; + + if (do_barriers) + barrier_all_devices(root->fs_info); + list_for_each_entry_rcu(dev, head, dev_list) { if (!dev->bdev) { total_errors++; @@ -2394,10 +2871,8 @@ static void free_fs_root(struct btrfs_root *root) { iput(root->cache_inode); WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree)); - if (root->anon_super.s_dev) { - down_write(&root->anon_super.s_umount); - kill_anon_super(&root->anon_super); - } + if (root->anon_dev) + free_anon_bdev(root->anon_dev); free_extent_buffer(root->node); free_extent_buffer(root->commit_root); kfree(root->free_ino_ctl); @@ -2514,8 +2989,6 @@ int close_ctree(struct btrfs_root *root) /* clear out the rbtree of defraggable inodes */ btrfs_run_defrag_inodes(root->fs_info); - btrfs_put_block_group_cache(fs_info); - /* * Here come 2 situations when btrfs is broken to flip readonly: * @@ -2541,6 +3014,8 @@ int close_ctree(struct btrfs_root *root) printk(KERN_ERR "btrfs: commit super ret %d\n", ret); } + btrfs_put_block_group_cache(fs_info); + kthread_stop(root->fs_info->transaction_kthread); kthread_stop(root->fs_info->cleaner_kthread); @@ -2572,7 +3047,6 @@ int close_ctree(struct btrfs_root *root) del_fs_roots(fs_info); iput(fs_info->btree_inode); - kfree(fs_info->delayed_root); btrfs_stop_workers(&fs_info->generic_worker); btrfs_stop_workers(&fs_info->fixup_workers); @@ -2585,6 +3059,8 @@ int close_ctree(struct btrfs_root *root) btrfs_stop_workers(&fs_info->endio_freespace_worker); btrfs_stop_workers(&fs_info->submit_workers); btrfs_stop_workers(&fs_info->delayed_workers); + btrfs_stop_workers(&fs_info->caching_workers); + btrfs_stop_workers(&fs_info->readahead_workers); btrfs_close_devices(fs_info->fs_devices); btrfs_mapping_tree_free(&fs_info->mapping_tree); @@ -2592,12 +3068,7 @@ int close_ctree(struct btrfs_root *root) bdi_destroy(&fs_info->bdi); cleanup_srcu_struct(&fs_info->subvol_srcu); - kfree(fs_info->extent_root); - kfree(fs_info->tree_root); - kfree(fs_info->chunk_root); - kfree(fs_info->dev_root); - kfree(fs_info->csum_root); - kfree(fs_info); + free_fs_info(fs_info); return 0; } @@ -2703,7 +3174,8 @@ int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid) return ret; } -int btree_lock_page_hook(struct page *page) +static int btree_lock_page_hook(struct page *page, void *data, + void (*flush_fn)(void *)) { struct inode *inode = page->mapping->host; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -2720,7 +3192,10 @@ int btree_lock_page_hook(struct page *page) if (!eb) goto out; - btrfs_tree_lock(eb); + if (!btrfs_try_tree_write_lock(eb)) { + flush_fn(data); + btrfs_tree_lock(eb); + } btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN); if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) { @@ -2735,7 +3210,10 @@ int btree_lock_page_hook(struct page *page) btrfs_tree_unlock(eb); free_extent_buffer(eb); out: - lock_page(page); + if (!trylock_page(page)) { + flush_fn(data); + lock_page(page); + } return 0; } @@ -3003,12 +3481,6 @@ static int btrfs_destroy_pinned_extent(struct btrfs_root *root, if (ret) break; - /* opt_discard */ - if (btrfs_test_opt(root, DISCARD)) - ret = btrfs_error_discard_extent(root, start, - end + 1 - start, - NULL); - clear_extent_dirty(unpin, start, end, GFP_NOFS); btrfs_error_unpin_extent_range(root, start, end); cond_resched(); @@ -3091,6 +3563,7 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) static struct extent_io_ops btree_extent_io_ops = { .write_cache_pages_lock_hook = btree_lock_page_hook, .readpage_end_io_hook = btree_readpage_end_io_hook, + .readpage_io_failed_hook = btree_io_failed_hook, .submit_bio_hook = btree_submit_bio_hook, /* note we're sharing with inode.c for the merge bio hook */ .merge_bio_hook = btrfs_merge_bio_hook, diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index a0b610a..c99d0a8f 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -40,6 +40,8 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, u64 parent_transid); int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, u64 parent_transid); +int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize, + int mirror_num, struct extent_buffer **eb); struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize); int clean_tree_block(struct btrfs_trans_handle *trans, @@ -83,14 +85,16 @@ int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info); int btrfs_add_log_tree(struct btrfs_trans_handle *trans, struct btrfs_root *root); -int btree_lock_page_hook(struct page *page); - #ifdef CONFIG_DEBUG_LOCK_ALLOC -void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, int level); +void btrfs_init_lockdep(void); +void btrfs_set_buffer_lockdep_class(u64 objectid, + struct extent_buffer *eb, int level); #else -static inline void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, - int level) +static inline void btrfs_init_lockdep(void) +{ } +static inline void btrfs_set_buffer_lockdep_class(u64 objectid, + struct extent_buffer *eb, int level) { } #endif diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 01220b7..da528f8 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -23,6 +23,7 @@ #include <linux/rcupdate.h> #include <linux/kthread.h> #include <linux/slab.h> +#include <linux/ratelimit.h> #include "compat.h" #include "hash.h" #include "ctree.h" @@ -52,6 +53,21 @@ enum { CHUNK_ALLOC_LIMITED = 2, }; +/* + * Control how reservations are dealt with. + * + * RESERVE_FREE - freeing a reservation. + * RESERVE_ALLOC - allocating space and we need to update bytes_may_use for + * ENOSPC accounting + * RESERVE_ALLOC_NO_ACCOUNT - allocating space and we should not update + * bytes_may_use as the ENOSPC accounting is done elsewhere + */ +enum { + RESERVE_FREE = 0, + RESERVE_ALLOC = 1, + RESERVE_ALLOC_NO_ACCOUNT = 2, +}; + static int update_block_group(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, int alloc); @@ -81,6 +97,8 @@ static int find_next_key(struct btrfs_path *path, int level, struct btrfs_key *key); static void dump_space_info(struct btrfs_space_info *info, u64 bytes, int dump_block_groups); +static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve); static noinline int block_group_cache_done(struct btrfs_block_group_cache *cache) @@ -104,7 +122,6 @@ 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); - WARN_ON(cache->reserved_pinned > 0); kfree(cache->free_space_ctl); kfree(cache); } @@ -320,12 +337,12 @@ static u64 add_new_free_space(struct btrfs_block_group_cache *block_group, return total_added; } -static int caching_kthread(void *data) +static noinline void caching_thread(struct btrfs_work *work) { - struct btrfs_block_group_cache *block_group = data; - struct btrfs_fs_info *fs_info = block_group->fs_info; - struct btrfs_caching_control *caching_ctl = block_group->caching_ctl; - struct btrfs_root *extent_root = fs_info->extent_root; + struct btrfs_block_group_cache *block_group; + struct btrfs_fs_info *fs_info; + struct btrfs_caching_control *caching_ctl; + struct btrfs_root *extent_root; struct btrfs_path *path; struct extent_buffer *leaf; struct btrfs_key key; @@ -334,9 +351,14 @@ static int caching_kthread(void *data) u32 nritems; int ret = 0; + caching_ctl = container_of(work, struct btrfs_caching_control, work); + block_group = caching_ctl->block_group; + fs_info = block_group->fs_info; + extent_root = fs_info->extent_root; + path = btrfs_alloc_path(); if (!path) - return -ENOMEM; + goto out; last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET); @@ -433,13 +455,11 @@ err: free_excluded_extents(extent_root, block_group); mutex_unlock(&caching_ctl->mutex); +out: wake_up(&caching_ctl->wait); put_caching_control(caching_ctl); - atomic_dec(&block_group->space_info->caching_threads); btrfs_put_block_group(block_group); - - return 0; } static int cache_block_group(struct btrfs_block_group_cache *cache, @@ -447,14 +467,59 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, struct btrfs_root *root, int load_cache_only) { + DEFINE_WAIT(wait); struct btrfs_fs_info *fs_info = cache->fs_info; struct btrfs_caching_control *caching_ctl; - struct task_struct *tsk; int ret = 0; - smp_mb(); - if (cache->cached != BTRFS_CACHE_NO) + caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); + BUG_ON(!caching_ctl); + + INIT_LIST_HEAD(&caching_ctl->list); + mutex_init(&caching_ctl->mutex); + init_waitqueue_head(&caching_ctl->wait); + caching_ctl->block_group = cache; + caching_ctl->progress = cache->key.objectid; + atomic_set(&caching_ctl->count, 1); + caching_ctl->work.func = caching_thread; + + spin_lock(&cache->lock); + /* + * This should be a rare occasion, but this could happen I think in the + * case where one thread starts to load the space cache info, and then + * some other thread starts a transaction commit which tries to do an + * allocation while the other thread is still loading the space cache + * info. The previous loop should have kept us from choosing this block + * group, but if we've moved to the state where we will wait on caching + * block groups we need to first check if we're doing a fast load here, + * so we can wait for it to finish, otherwise we could end up allocating + * from a block group who's cache gets evicted for one reason or + * another. + */ + while (cache->cached == BTRFS_CACHE_FAST) { + struct btrfs_caching_control *ctl; + + ctl = cache->caching_ctl; + atomic_inc(&ctl->count); + prepare_to_wait(&ctl->wait, &wait, TASK_UNINTERRUPTIBLE); + spin_unlock(&cache->lock); + + schedule(); + + finish_wait(&ctl->wait, &wait); + put_caching_control(ctl); + spin_lock(&cache->lock); + } + + if (cache->cached != BTRFS_CACHE_NO) { + spin_unlock(&cache->lock); + kfree(caching_ctl); return 0; + } + WARN_ON(cache->caching_ctl); + cache->caching_ctl = caching_ctl; + cache->cached = BTRFS_CACHE_FAST; + spin_unlock(&cache->lock); /* * We can't do the read from on-disk cache during a commit since we need @@ -463,69 +528,59 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, * we likely hold important locks. */ if (trans && (!trans->transaction->in_commit) && - (root && root != root->fs_info->tree_root)) { - spin_lock(&cache->lock); - if (cache->cached != BTRFS_CACHE_NO) { - spin_unlock(&cache->lock); - return 0; - } - cache->cached = BTRFS_CACHE_STARTED; - spin_unlock(&cache->lock); - + (root && root != root->fs_info->tree_root) && + btrfs_test_opt(root, SPACE_CACHE)) { ret = load_free_space_cache(fs_info, cache); spin_lock(&cache->lock); if (ret == 1) { + cache->caching_ctl = NULL; cache->cached = BTRFS_CACHE_FINISHED; cache->last_byte_to_unpin = (u64)-1; } else { - cache->cached = BTRFS_CACHE_NO; + if (load_cache_only) { + cache->caching_ctl = NULL; + cache->cached = BTRFS_CACHE_NO; + } else { + cache->cached = BTRFS_CACHE_STARTED; + } } spin_unlock(&cache->lock); + wake_up(&caching_ctl->wait); if (ret == 1) { + put_caching_control(caching_ctl); free_excluded_extents(fs_info->extent_root, cache); return 0; } + } else { + /* + * We are not going to do the fast caching, set cached to the + * appropriate value and wakeup any waiters. + */ + spin_lock(&cache->lock); + if (load_cache_only) { + cache->caching_ctl = NULL; + cache->cached = BTRFS_CACHE_NO; + } else { + cache->cached = BTRFS_CACHE_STARTED; + } + spin_unlock(&cache->lock); + wake_up(&caching_ctl->wait); } - if (load_cache_only) - return 0; - - caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); - BUG_ON(!caching_ctl); - - INIT_LIST_HEAD(&caching_ctl->list); - mutex_init(&caching_ctl->mutex); - init_waitqueue_head(&caching_ctl->wait); - caching_ctl->block_group = cache; - caching_ctl->progress = cache->key.objectid; - /* one for caching kthread, one for caching block group list */ - atomic_set(&caching_ctl->count, 2); - - spin_lock(&cache->lock); - if (cache->cached != BTRFS_CACHE_NO) { - spin_unlock(&cache->lock); - kfree(caching_ctl); + if (load_cache_only) { + put_caching_control(caching_ctl); return 0; } - cache->caching_ctl = caching_ctl; - cache->cached = BTRFS_CACHE_STARTED; - spin_unlock(&cache->lock); down_write(&fs_info->extent_commit_sem); + atomic_inc(&caching_ctl->count); list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups); up_write(&fs_info->extent_commit_sem); - atomic_inc(&cache->space_info->caching_threads); btrfs_get_block_group(cache); - tsk = kthread_run(caching_kthread, cache, "btrfs-cache-%llu\n", - cache->key.objectid); - if (IS_ERR(tsk)) { - ret = PTR_ERR(tsk); - printk(KERN_ERR "error running thread %d\n", ret); - BUG(); - } + btrfs_queue_worker(&fs_info->caching_workers, &caching_ctl->work); return ret; } @@ -667,7 +722,9 @@ int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len) struct btrfs_path *path; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; + key.objectid = start; key.offset = len; btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY); @@ -1772,18 +1829,18 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, { int ret; u64 discarded_bytes = 0; - struct btrfs_multi_bio *multi = NULL; + struct btrfs_bio *bbio = NULL; /* Tell the block device(s) that the sectors can be discarded */ ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD, - bytenr, &num_bytes, &multi, 0); + bytenr, &num_bytes, &bbio, 0); if (!ret) { - struct btrfs_bio_stripe *stripe = multi->stripes; + struct btrfs_bio_stripe *stripe = bbio->stripes; int i; - for (i = 0; i < multi->num_stripes; i++, stripe++) { + for (i = 0; i < bbio->num_stripes; i++, stripe++) { if (!stripe->dev->can_discard) continue; @@ -1802,7 +1859,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, */ ret = 0; } - kfree(multi); + kfree(bbio); } if (actual_bytes) @@ -2702,6 +2759,13 @@ again: goto again; } + /* We've already setup this transaction, go ahead and exit */ + if (block_group->cache_generation == trans->transid && + i_size_read(inode)) { + dcs = BTRFS_DC_SETUP; + goto out_put; + } + /* * We want to set the generation to 0, that way if anything goes wrong * from here on out we know not to trust this cache when we load up next @@ -2751,12 +2815,15 @@ again: if (!ret) dcs = BTRFS_DC_SETUP; btrfs_free_reserved_data_space(inode, num_pages); + out_put: iput(inode); out_free: btrfs_release_path(path); out: spin_lock(&block_group->lock); + if (!ret && dcs == BTRFS_DC_SETUP) + block_group->cache_generation = trans->transid; block_group->disk_cache_state = dcs; spin_unlock(&block_group->lock); @@ -2940,9 +3007,10 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->full = 0; found->force_alloc = CHUNK_ALLOC_NO_FORCE; found->chunk_alloc = 0; + found->flush = 0; + init_waitqueue_head(&found->wait); *space_info = found; list_add_rcu(&found->list, &info->space_info); - atomic_set(&found->caching_threads, 0); return 0; } @@ -3123,16 +3191,13 @@ commit_trans: return -ENOSPC; } data_sinfo->bytes_may_use += bytes; - BTRFS_I(inode)->reserved_bytes += bytes; spin_unlock(&data_sinfo->lock); return 0; } /* - * called when we are clearing an delalloc extent from the - * inode's io_tree or there was an error for whatever reason - * after calling btrfs_check_data_free_space + * Called if we need to clear a data reservation for this inode. */ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes) { @@ -3145,7 +3210,6 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes) data_sinfo = BTRFS_I(inode)->space_info; spin_lock(&data_sinfo->lock); data_sinfo->bytes_may_use -= bytes; - BTRFS_I(inode)->reserved_bytes -= bytes; spin_unlock(&data_sinfo->lock); } @@ -3166,6 +3230,7 @@ static int should_alloc_chunk(struct btrfs_root *root, struct btrfs_space_info *sinfo, u64 alloc_bytes, int force) { + struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly; u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved; u64 thresh; @@ -3174,11 +3239,18 @@ static int should_alloc_chunk(struct btrfs_root *root, return 1; /* + * We need to take into account the global rsv because for all intents + * and purposes it's used space. Don't worry about locking the + * global_rsv, it doesn't change except when the transaction commits. + */ + num_allocated += global_rsv->size; + + /* * in limited mode, we want to have some free space up to * about 1% of the FS size. */ if (force == CHUNK_ALLOC_LIMITED) { - thresh = btrfs_super_total_bytes(&root->fs_info->super_copy); + thresh = btrfs_super_total_bytes(root->fs_info->super_copy); thresh = max_t(u64, 64 * 1024 * 1024, div_factor_fine(thresh, 1)); @@ -3200,7 +3272,7 @@ static int should_alloc_chunk(struct btrfs_root *root, if (num_allocated + alloc_bytes < div_factor(num_bytes, 8)) return 0; - thresh = btrfs_super_total_bytes(&root->fs_info->super_copy); + thresh = btrfs_super_total_bytes(root->fs_info->super_copy); /* 256MB or 5% of the FS */ thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 5)); @@ -3283,6 +3355,9 @@ again: } ret = btrfs_alloc_chunk(trans, extent_root, flags); + if (ret < 0 && ret != -ENOSPC) + goto out; + spin_lock(&space_info->lock); if (ret) space_info->full = 1; @@ -3292,6 +3367,7 @@ again: space_info->force_alloc = CHUNK_ALLOC_NO_FORCE; space_info->chunk_alloc = 0; spin_unlock(&space_info->lock); +out: mutex_unlock(&extent_root->fs_info->chunk_mutex); return ret; } @@ -3299,42 +3375,54 @@ again: /* * shrink metadata reservation for delalloc */ -static int shrink_delalloc(struct btrfs_trans_handle *trans, - struct btrfs_root *root, u64 to_reclaim, int sync) +static int shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, + bool wait_ordered) { struct btrfs_block_rsv *block_rsv; struct btrfs_space_info *space_info; + struct btrfs_trans_handle *trans; u64 reserved; u64 max_reclaim; u64 reclaimed = 0; long time_left; - int nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT; + unsigned long nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT; int loops = 0; unsigned long progress; + trans = (struct btrfs_trans_handle *)current->journal_info; block_rsv = &root->fs_info->delalloc_block_rsv; space_info = block_rsv->space_info; smp_mb(); - reserved = space_info->bytes_reserved; + reserved = space_info->bytes_may_use; progress = space_info->reservation_progress; if (reserved == 0) return 0; - max_reclaim = min(reserved, to_reclaim); + smp_mb(); + if (root->fs_info->delalloc_bytes == 0) { + if (trans) + return 0; + btrfs_wait_ordered_extents(root, 0, 0); + return 0; + } + max_reclaim = min(reserved, to_reclaim); + nr_pages = max_t(unsigned long, nr_pages, + max_reclaim >> PAGE_CACHE_SHIFT); while (loops < 1024) { /* have the flusher threads jump in and do some IO */ smp_mb(); nr_pages = min_t(unsigned long, nr_pages, root->fs_info->delalloc_bytes >> PAGE_CACHE_SHIFT); - writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages); + writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages, + WB_REASON_FS_FREE_SPACE); spin_lock(&space_info->lock); - if (reserved > space_info->bytes_reserved) - reclaimed += reserved - space_info->bytes_reserved; - reserved = space_info->bytes_reserved; + if (reserved > space_info->bytes_may_use) + reclaimed += reserved - space_info->bytes_may_use; + reserved = space_info->bytes_may_use; spin_unlock(&space_info->lock); loops++; @@ -3345,11 +3433,15 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans, if (trans && trans->transaction->blocked) return -EAGAIN; - time_left = schedule_timeout_interruptible(1); + if (wait_ordered && !trans) { + btrfs_wait_ordered_extents(root, 0, 0); + } else { + time_left = schedule_timeout_interruptible(1); - /* We were interrupted, exit */ - if (time_left) - break; + /* We were interrupted, exit */ + if (time_left) + break; + } /* we've kicked the IO a few times, if anything has been freed, * exit. There is no sense in looping here for a long time @@ -3364,42 +3456,121 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans, } } + return reclaimed >= to_reclaim; } -/* - * Retries tells us how many times we've called reserve_metadata_bytes. The - * idea is if this is the first call (retries == 0) then we will add to our - * reserved count if we can't make the allocation in order to hold our place - * while we go and try and free up space. That way for retries > 1 we don't try - * and add space, we just check to see if the amount of unused space is >= the - * total space, meaning that our reservation is valid. +/** + * maybe_commit_transaction - possibly commit the transaction if its ok to + * @root - the root we're allocating for + * @bytes - the number of bytes we want to reserve + * @force - force the commit * - * However if we don't intend to retry this reservation, pass -1 as retries so - * that it short circuits this logic. + * This will check to make sure that committing the transaction will actually + * get us somewhere and then commit the transaction if it does. Otherwise it + * will return -ENOSPC. */ -static int reserve_metadata_bytes(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +static int may_commit_transaction(struct btrfs_root *root, + struct btrfs_space_info *space_info, + u64 bytes, int force) +{ + struct btrfs_block_rsv *delayed_rsv = &root->fs_info->delayed_block_rsv; + struct btrfs_trans_handle *trans; + + trans = (struct btrfs_trans_handle *)current->journal_info; + if (trans) + return -EAGAIN; + + if (force) + goto commit; + + /* See if there is enough pinned space to make this reservation */ + spin_lock(&space_info->lock); + if (space_info->bytes_pinned >= bytes) { + spin_unlock(&space_info->lock); + goto commit; + } + spin_unlock(&space_info->lock); + + /* + * See if there is some space in the delayed insertion reservation for + * this reservation. + */ + if (space_info != delayed_rsv->space_info) + return -ENOSPC; + + spin_lock(&delayed_rsv->lock); + if (delayed_rsv->size < bytes) { + spin_unlock(&delayed_rsv->lock); + return -ENOSPC; + } + spin_unlock(&delayed_rsv->lock); + +commit: + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) + return -ENOSPC; + + return btrfs_commit_transaction(trans, root); +} + +/** + * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space + * @root - the root we're allocating for + * @block_rsv - the block_rsv we're allocating for + * @orig_bytes - the number of bytes we want + * @flush - wether or not we can flush to make our reservation + * + * This will reserve orgi_bytes number of bytes from the space info associated + * with the block_rsv. If there is not enough space it will make an attempt to + * flush out space to make room. It will do this by flushing delalloc if + * possible or committing the transaction. If flush is 0 then no attempts to + * 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, struct btrfs_block_rsv *block_rsv, u64 orig_bytes, int flush) { struct btrfs_space_info *space_info = block_rsv->space_info; - u64 unused; + u64 used; u64 num_bytes = orig_bytes; int retries = 0; int ret = 0; - bool reserved = false; bool committed = false; + bool flushing = false; + bool wait_ordered = false; again: - ret = -ENOSPC; - if (reserved) - num_bytes = 0; - + ret = 0; spin_lock(&space_info->lock); - unused = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; + /* + * We only want to wait if somebody other than us is flushing and we are + * actually alloed to flush. + */ + while (flush && !flushing && space_info->flush) { + spin_unlock(&space_info->lock); + /* + * If we have a trans handle we can't wait because the flusher + * may have to commit the transaction, which would mean we would + * deadlock since we are waiting for the flusher to finish, but + * hold the current transaction open. + */ + if (current->journal_info) + return -EAGAIN; + ret = wait_event_interruptible(space_info->wait, + !space_info->flush); + /* Must have been interrupted, return */ + if (ret) + return -EINTR; + + 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; /* * The idea here is that we've not already over-reserved the block group @@ -3408,11 +3579,9 @@ again: * lets start flushing stuff first and then come back and try to make * our reservation. */ - if (unused <= space_info->total_bytes) { - unused = space_info->total_bytes - unused; - if (unused >= num_bytes) { - if (!reserved) - space_info->bytes_reserved += orig_bytes; + if (used <= space_info->total_bytes) { + if (used + orig_bytes <= space_info->total_bytes) { + space_info->bytes_may_use += orig_bytes; ret = 0; } else { /* @@ -3428,91 +3597,129 @@ again: * amount plus the amount of bytes that we need for this * reservation. */ - num_bytes = unused - space_info->total_bytes + + wait_ordered = true; + num_bytes = used - space_info->total_bytes + (orig_bytes * (retries + 1)); } + if (ret) { + u64 profile = btrfs_get_alloc_profile(root, 0); + u64 avail; + + /* + * If we have a lot of space that's pinned, don't bother doing + * the overcommit dance yet and just commit the transaction. + */ + avail = (space_info->total_bytes - space_info->bytes_used) * 8; + do_div(avail, 10); + if (space_info->bytes_pinned >= avail && flush && !committed) { + space_info->flush = 1; + flushing = true; + spin_unlock(&space_info->lock); + ret = may_commit_transaction(root, space_info, + orig_bytes, 1); + if (ret) + goto out; + committed = true; + goto again; + } + + spin_lock(&root->fs_info->free_chunk_lock); + avail = root->fs_info->free_chunk_space; + + /* + * If we have dup, raid1 or raid10 then only half of the free + * space is actually useable. + */ + if (profile & (BTRFS_BLOCK_GROUP_DUP | + BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10)) + avail >>= 1; + + /* + * If we aren't flushing don't let us overcommit too much, say + * 1/8th of the space. If we can flush, let it overcommit up to + * 1/2 of the space. + */ + if (flush) + avail >>= 3; + else + avail >>= 1; + spin_unlock(&root->fs_info->free_chunk_lock); + + if (used + num_bytes < space_info->total_bytes + avail) { + space_info->bytes_may_use += orig_bytes; + ret = 0; + } else { + wait_ordered = true; + } + } + /* * Couldn't make our reservation, save our place so while we're trying * to reclaim space we can actually use it instead of somebody else * stealing it from us. */ - if (ret && !reserved) { - space_info->bytes_reserved += orig_bytes; - reserved = true; + if (ret && flush) { + flushing = true; + space_info->flush = 1; } spin_unlock(&space_info->lock); - if (!ret) - return 0; - - if (!flush) + if (!ret || !flush) goto out; /* * We do synchronous shrinking since we don't actually unreserve * metadata until after the IO is completed. */ - ret = shrink_delalloc(trans, root, num_bytes, 1); - if (ret > 0) - return 0; - else if (ret < 0) + ret = shrink_delalloc(root, num_bytes, wait_ordered); + if (ret < 0) goto out; + ret = 0; + /* * So if we were overcommitted it's possible that somebody else flushed * out enough space and we simply didn't have enough space to reclaim, * so go back around and try again. */ if (retries < 2) { + wait_ordered = true; retries++; goto again; } - spin_lock(&space_info->lock); - /* - * Not enough space to be reclaimed, don't bother committing the - * transaction. - */ - if (space_info->bytes_pinned < orig_bytes) - ret = -ENOSPC; - spin_unlock(&space_info->lock); - if (ret) - goto out; - - ret = -EAGAIN; - if (trans || committed) - goto out; - ret = -ENOSPC; - trans = btrfs_join_transaction(root); - if (IS_ERR(trans)) + if (committed) goto out; - ret = btrfs_commit_transaction(trans, root); + + ret = may_commit_transaction(root, space_info, orig_bytes, 0); if (!ret) { - trans = NULL; committed = true; goto again; } out: - if (reserved) { + if (flushing) { spin_lock(&space_info->lock); - space_info->bytes_reserved -= orig_bytes; + space_info->flush = 0; + wake_up_all(&space_info->wait); spin_unlock(&space_info->lock); } - return ret; } static struct btrfs_block_rsv *get_block_rsv(struct btrfs_trans_handle *trans, struct btrfs_root *root) { - struct btrfs_block_rsv *block_rsv; - if (root->ref_cows) + struct btrfs_block_rsv *block_rsv = NULL; + + if (root->ref_cows || root == root->fs_info->csum_root) block_rsv = trans->block_rsv; - else + + if (!block_rsv) block_rsv = root->block_rsv; if (!block_rsv) @@ -3583,7 +3790,7 @@ static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, } if (num_bytes) { spin_lock(&space_info->lock); - space_info->bytes_reserved -= num_bytes; + space_info->bytes_may_use -= num_bytes; space_info->reservation_progress++; spin_unlock(&space_info->lock); } @@ -3607,9 +3814,6 @@ void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv) { memset(rsv, 0, sizeof(*rsv)); spin_lock_init(&rsv->lock); - atomic_set(&rsv->usage, 1); - rsv->priority = 6; - INIT_LIST_HEAD(&rsv->list); } struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root) @@ -3630,38 +3834,20 @@ struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root) void btrfs_free_block_rsv(struct btrfs_root *root, struct btrfs_block_rsv *rsv) { - if (rsv && atomic_dec_and_test(&rsv->usage)) { - btrfs_block_rsv_release(root, rsv, (u64)-1); - if (!rsv->durable) - kfree(rsv); - } -} - -/* - * make the block_rsv struct be able to capture freed space. - * the captured space will re-add to the the block_rsv struct - * after transaction commit - */ -void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info, - struct btrfs_block_rsv *block_rsv) -{ - block_rsv->durable = 1; - mutex_lock(&fs_info->durable_block_rsv_mutex); - list_add_tail(&block_rsv->list, &fs_info->durable_block_rsv_list); - mutex_unlock(&fs_info->durable_block_rsv_mutex); + btrfs_block_rsv_release(root, rsv, (u64)-1); + kfree(rsv); } -int btrfs_block_rsv_add(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, - u64 num_bytes) +static inline int __block_rsv_add(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes, int flush) { int ret; if (num_bytes == 0) return 0; - ret = reserve_metadata_bytes(trans, root, block_rsv, num_bytes, 1); + ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); if (!ret) { block_rsv_add_bytes(block_rsv, num_bytes, 1); return 0; @@ -3670,56 +3856,80 @@ int btrfs_block_rsv_add(struct btrfs_trans_handle *trans, return ret; } -int btrfs_block_rsv_check(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, - u64 min_reserved, int min_factor) +int btrfs_block_rsv_add(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes) +{ + return __block_rsv_add(root, block_rsv, num_bytes, 1); +} + +int btrfs_block_rsv_add_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes) +{ + return __block_rsv_add(root, block_rsv, num_bytes, 0); +} + +int btrfs_block_rsv_check(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, int min_factor) { u64 num_bytes = 0; - int commit_trans = 0; int ret = -ENOSPC; if (!block_rsv) return 0; spin_lock(&block_rsv->lock); - if (min_factor > 0) - num_bytes = div_factor(block_rsv->size, min_factor); - if (min_reserved > num_bytes) - num_bytes = min_reserved; + num_bytes = div_factor(block_rsv->size, min_factor); + if (block_rsv->reserved >= num_bytes) + ret = 0; + spin_unlock(&block_rsv->lock); - if (block_rsv->reserved >= num_bytes) { + return ret; +} + +static inline int __btrfs_block_rsv_refill(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved, int flush) +{ + u64 num_bytes = 0; + int ret = -ENOSPC; + + if (!block_rsv) + return 0; + + spin_lock(&block_rsv->lock); + num_bytes = min_reserved; + if (block_rsv->reserved >= num_bytes) ret = 0; - } else { + else num_bytes -= block_rsv->reserved; - if (block_rsv->durable && - block_rsv->freed[0] + block_rsv->freed[1] >= num_bytes) - commit_trans = 1; - } spin_unlock(&block_rsv->lock); + if (!ret) return 0; - if (block_rsv->refill_used) { - ret = reserve_metadata_bytes(trans, root, block_rsv, - num_bytes, 0); - if (!ret) { - block_rsv_add_bytes(block_rsv, num_bytes, 0); - return 0; - } + ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); + if (!ret) { + block_rsv_add_bytes(block_rsv, num_bytes, 0); + return 0; } - if (commit_trans) { - if (trans) - return -EAGAIN; + return ret; +} - trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); - ret = btrfs_commit_transaction(trans, root); - return 0; - } +int btrfs_block_rsv_refill(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved) +{ + return __btrfs_block_rsv_refill(root, block_rsv, min_reserved, 1); +} - return -ENOSPC; +int btrfs_block_rsv_refill_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved) +{ + return __btrfs_block_rsv_refill(root, block_rsv, min_reserved, 0); } int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, @@ -3751,7 +3961,7 @@ static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) u64 num_bytes; u64 meta_used; u64 data_used; - int csum_size = btrfs_super_csum_size(&fs_info->super_copy); + int csum_size = btrfs_super_csum_size(fs_info->super_copy); sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); spin_lock(&sinfo->lock); @@ -3795,12 +4005,12 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info) if (sinfo->total_bytes > num_bytes) { num_bytes = sinfo->total_bytes - num_bytes; block_rsv->reserved += num_bytes; - sinfo->bytes_reserved += num_bytes; + sinfo->bytes_may_use += num_bytes; } if (block_rsv->reserved >= block_rsv->size) { num_bytes = block_rsv->reserved - block_rsv->size; - sinfo->bytes_reserved -= num_bytes; + sinfo->bytes_may_use -= num_bytes; sinfo->reservation_progress++; block_rsv->reserved = block_rsv->size; block_rsv->full = 1; @@ -3816,16 +4026,13 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info) space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM); fs_info->chunk_block_rsv.space_info = space_info; - fs_info->chunk_block_rsv.priority = 10; space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); fs_info->global_block_rsv.space_info = space_info; - fs_info->global_block_rsv.priority = 10; - fs_info->global_block_rsv.refill_used = 1; fs_info->delalloc_block_rsv.space_info = space_info; fs_info->trans_block_rsv.space_info = space_info; fs_info->empty_block_rsv.space_info = space_info; - fs_info->empty_block_rsv.priority = 10; + fs_info->delayed_block_rsv.space_info = space_info; fs_info->extent_root->block_rsv = &fs_info->global_block_rsv; fs_info->csum_root->block_rsv = &fs_info->global_block_rsv; @@ -3833,10 +4040,6 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info) fs_info->tree_root->block_rsv = &fs_info->global_block_rsv; fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv; - btrfs_add_durable_block_rsv(fs_info, &fs_info->global_block_rsv); - - btrfs_add_durable_block_rsv(fs_info, &fs_info->delalloc_block_rsv); - update_global_block_rsv(fs_info); } @@ -3849,57 +4052,8 @@ static void release_global_block_rsv(struct btrfs_fs_info *fs_info) WARN_ON(fs_info->trans_block_rsv.reserved > 0); WARN_ON(fs_info->chunk_block_rsv.size > 0); WARN_ON(fs_info->chunk_block_rsv.reserved > 0); -} - -int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *rsv) -{ - struct btrfs_block_rsv *trans_rsv = &root->fs_info->trans_block_rsv; - u64 num_bytes; - int ret; - - /* - * Truncate should be freeing data, but give us 2 items just in case it - * needs to use some space. We may want to be smarter about this in the - * future. - */ - num_bytes = btrfs_calc_trans_metadata_size(root, 2); - - /* We already have enough bytes, just return */ - if (rsv->reserved >= num_bytes) - return 0; - - num_bytes -= rsv->reserved; - - /* - * You should have reserved enough space before hand to do this, so this - * should not fail. - */ - ret = block_rsv_migrate_bytes(trans_rsv, rsv, num_bytes); - BUG_ON(ret); - - return 0; -} - -int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - int num_items) -{ - u64 num_bytes; - int ret; - - if (num_items == 0 || root->fs_info->chunk_root == root) - return 0; - - num_bytes = btrfs_calc_trans_metadata_size(root, num_items); - ret = btrfs_block_rsv_add(trans, root, &root->fs_info->trans_block_rsv, - num_bytes); - if (!ret) { - trans->bytes_reserved += num_bytes; - trans->block_rsv = &root->fs_info->trans_block_rsv; - } - return ret; + WARN_ON(fs_info->delayed_block_rsv.size > 0); + WARN_ON(fs_info->delayed_block_rsv.reserved > 0); } void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans, @@ -3908,9 +4062,7 @@ void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans, if (!trans->bytes_reserved) return; - BUG_ON(trans->block_rsv != &root->fs_info->trans_block_rsv); - btrfs_block_rsv_release(root, trans->block_rsv, - trans->bytes_reserved); + btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved); trans->bytes_reserved = 0; } @@ -3952,90 +4104,229 @@ int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans, return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); } -static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes) +/** + * drop_outstanding_extent - drop an outstanding extent + * @inode: the inode we're dropping the extent for + * + * This is called when we are freeing up an outstanding extent, either called + * after an error or after an extent is written. This will return the number of + * 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) { - return num_bytes >>= 3; + unsigned drop_inode_space = 0; + unsigned dropped_extents = 0; + + BUG_ON(!BTRFS_I(inode)->outstanding_extents); + BTRFS_I(inode)->outstanding_extents--; + + if (BTRFS_I(inode)->outstanding_extents == 0 && + BTRFS_I(inode)->delalloc_meta_reserved) { + drop_inode_space = 1; + BTRFS_I(inode)->delalloc_meta_reserved = 0; + } + + /* + * If we have more or the same amount of outsanding 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) + return drop_inode_space; + + dropped_extents = BTRFS_I(inode)->reserved_extents - + BTRFS_I(inode)->outstanding_extents; + BTRFS_I(inode)->reserved_extents -= dropped_extents; + return dropped_extents + drop_inode_space; +} + +/** + * calc_csum_metadata_size - return the amount of metada space that must be + * reserved/free'd for the given bytes. + * @inode: the inode we're manipulating + * @num_bytes: the number of bytes in question + * @reserve: 1 if we are reserving space, 0 if we are freeing space + * + * This adjusts the number of csum_bytes in the inode and then returns the + * correct amount of metadata that must either be reserved or freed. We + * calculate how many checksums we can fit into one leaf and then divide the + * number of bytes that will need to be checksumed by this value to figure out + * how many checksums will be required. If we are adding bytes then the number + * may go up and we will return the number of additional bytes that must be + * reserved. If it is going down we will return the number of bytes that must + * be freed. + * + * This must be called with BTRFS_I(inode)->lock held. + */ +static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes, + int reserve) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + u64 csum_size; + int num_csums_per_leaf; + int num_csums; + int old_csums; + + if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM && + BTRFS_I(inode)->csum_bytes == 0) + return 0; + + old_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); + if (reserve) + BTRFS_I(inode)->csum_bytes += num_bytes; + else + BTRFS_I(inode)->csum_bytes -= num_bytes; + csum_size = BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item); + num_csums_per_leaf = (int)div64_u64(csum_size, + sizeof(struct btrfs_csum_item) + + sizeof(struct btrfs_disk_key)); + num_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); + num_csums = num_csums + num_csums_per_leaf - 1; + num_csums = num_csums / num_csums_per_leaf; + + old_csums = old_csums + num_csums_per_leaf - 1; + old_csums = old_csums / num_csums_per_leaf; + + /* No change, no need to reserve more */ + if (old_csums == num_csums) + return 0; + + if (reserve) + return btrfs_calc_trans_metadata_size(root, + num_csums - old_csums); + + return btrfs_calc_trans_metadata_size(root, old_csums - num_csums); } int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv; - u64 to_reserve; - int nr_extents; - int reserved_extents; + u64 to_reserve = 0; + u64 csum_bytes; + unsigned nr_extents = 0; + int extra_reserve = 0; + int flush = 1; int ret; - if (btrfs_transaction_in_commit(root->fs_info)) + /* Need to be holding the i_mutex here if we aren't free space cache */ + if (btrfs_is_free_space_inode(root, inode)) + flush = 0; + else + WARN_ON(!mutex_is_locked(&inode->i_mutex)); + + if (flush && btrfs_transaction_in_commit(root->fs_info)) schedule_timeout(1); num_bytes = ALIGN(num_bytes, root->sectorsize); - nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents) + 1; - reserved_extents = atomic_read(&BTRFS_I(inode)->reserved_extents); + spin_lock(&BTRFS_I(inode)->lock); + BTRFS_I(inode)->outstanding_extents++; - if (nr_extents > reserved_extents) { - nr_extents -= reserved_extents; - to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); - } else { - nr_extents = 0; - to_reserve = 0; + if (BTRFS_I(inode)->outstanding_extents > + BTRFS_I(inode)->reserved_extents) + nr_extents = BTRFS_I(inode)->outstanding_extents - + BTRFS_I(inode)->reserved_extents; + + /* + * Add an item to reserve for updating the inode when we complete the + * delalloc io. + */ + if (!BTRFS_I(inode)->delalloc_meta_reserved) { + nr_extents++; + extra_reserve = 1; } - to_reserve += calc_csum_metadata_size(inode, num_bytes); - ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1); - if (ret) + to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); + to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); + csum_bytes = BTRFS_I(inode)->csum_bytes; + spin_unlock(&BTRFS_I(inode)->lock); + + ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush); + if (ret) { + u64 to_free = 0; + unsigned dropped; + + spin_lock(&BTRFS_I(inode)->lock); + dropped = drop_outstanding_extent(inode); + /* + * 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. + * Otherwise we have to do the normal free thing to account for + * the case that the free side didn't free up its reserve + * because of this outstanding reservation. + */ + if (BTRFS_I(inode)->csum_bytes == csum_bytes) + calc_csum_metadata_size(inode, num_bytes, 0); + else + to_free = calc_csum_metadata_size(inode, num_bytes, 0); + spin_unlock(&BTRFS_I(inode)->lock); + if (dropped) + to_free += btrfs_calc_trans_metadata_size(root, dropped); + + if (to_free) + btrfs_block_rsv_release(root, block_rsv, to_free); return ret; + } - atomic_add(nr_extents, &BTRFS_I(inode)->reserved_extents); - atomic_inc(&BTRFS_I(inode)->outstanding_extents); + spin_lock(&BTRFS_I(inode)->lock); + if (extra_reserve) { + BTRFS_I(inode)->delalloc_meta_reserved = 1; + nr_extents--; + } + BTRFS_I(inode)->reserved_extents += nr_extents; + spin_unlock(&BTRFS_I(inode)->lock); block_rsv_add_bytes(block_rsv, to_reserve, 1); - if (block_rsv->size > 512 * 1024 * 1024) - shrink_delalloc(NULL, root, to_reserve, 0); - return 0; } +/** + * btrfs_delalloc_release_metadata - release a metadata reservation for an inode + * @inode: the inode to release the reservation for + * @num_bytes: the number of bytes we're releasing + * + * This will release the metadata reservation for an inode. This can be called + * 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) { struct btrfs_root *root = BTRFS_I(inode)->root; - u64 to_free; - int nr_extents; - int reserved_extents; + u64 to_free = 0; + unsigned dropped; num_bytes = ALIGN(num_bytes, root->sectorsize); - atomic_dec(&BTRFS_I(inode)->outstanding_extents); - WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents) < 0); - - reserved_extents = atomic_read(&BTRFS_I(inode)->reserved_extents); - do { - int old, new; - - nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents); - if (nr_extents >= reserved_extents) { - nr_extents = 0; - break; - } - old = reserved_extents; - nr_extents = reserved_extents - nr_extents; - new = reserved_extents - nr_extents; - old = atomic_cmpxchg(&BTRFS_I(inode)->reserved_extents, - reserved_extents, new); - if (likely(old == reserved_extents)) - break; - reserved_extents = old; - } while (1); + spin_lock(&BTRFS_I(inode)->lock); + dropped = drop_outstanding_extent(inode); - to_free = calc_csum_metadata_size(inode, num_bytes); - if (nr_extents > 0) - to_free += btrfs_calc_trans_metadata_size(root, nr_extents); + to_free = calc_csum_metadata_size(inode, num_bytes, 0); + spin_unlock(&BTRFS_I(inode)->lock); + if (dropped > 0) + to_free += btrfs_calc_trans_metadata_size(root, dropped); btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv, to_free); } +/** + * btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc + * @inode: inode we're writing to + * @num_bytes: the number of bytes we want to allocate + * + * This will do the following things + * + * o reserve space in the data space info for num_bytes + * o reserve space in the metadata space info based on number of outstanding + * extents and how much csums will be needed + * o add to the inodes ->delalloc_bytes + * o add it to the fs_info's delalloc inodes list. + * + * This will return 0 for success and -ENOSPC if there is no space left. + */ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes) { int ret; @@ -4053,6 +4344,19 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes) return 0; } +/** + * btrfs_delalloc_release_space - release data and metadata space for delalloc + * @inode: inode we're releasing space for + * @num_bytes: the number of bytes we want to free up + * + * This must be matched with a call to btrfs_delalloc_reserve_space. This is + * called in the case that we don't need the metadata AND data reservations + * anymore. So if there is an error or we insert an inline extent. + * + * This function will release the metadata space that was not used and will + * decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes + * list if there are no delalloc bytes left. + */ void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes) { btrfs_delalloc_release_metadata(inode, num_bytes); @@ -4072,12 +4376,12 @@ static int update_block_group(struct btrfs_trans_handle *trans, /* block accounting for super block */ spin_lock(&info->delalloc_lock); - old_val = btrfs_super_bytes_used(&info->super_copy); + old_val = btrfs_super_bytes_used(info->super_copy); if (alloc) old_val += num_bytes; else old_val -= num_bytes; - btrfs_set_super_bytes_used(&info->super_copy, old_val); + btrfs_set_super_bytes_used(info->super_copy, old_val); spin_unlock(&info->delalloc_lock); while (total) { @@ -4105,7 +4409,7 @@ static int update_block_group(struct btrfs_trans_handle *trans, spin_lock(&cache->space_info->lock); spin_lock(&cache->lock); - if (btrfs_super_cache_generation(&info->super_copy) != 0 && + if (btrfs_test_opt(root, SPACE_CACHE) && cache->disk_cache_state < BTRFS_DC_CLEAR) cache->disk_cache_state = BTRFS_DC_CLEAR; @@ -4117,7 +4421,6 @@ static int update_block_group(struct btrfs_trans_handle *trans, btrfs_set_block_group_used(&cache->item, old_val); cache->reserved -= num_bytes; cache->space_info->bytes_reserved -= num_bytes; - cache->space_info->reservation_progress++; cache->space_info->bytes_used += num_bytes; cache->space_info->disk_used += num_bytes * factor; spin_unlock(&cache->lock); @@ -4169,7 +4472,6 @@ static int pin_down_extent(struct btrfs_root *root, if (reserved) { cache->reserved -= num_bytes; cache->space_info->bytes_reserved -= num_bytes; - cache->space_info->reservation_progress++; } spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); @@ -4197,45 +4499,82 @@ int btrfs_pin_extent(struct btrfs_root *root, } /* - * update size of reserved extents. this function may return -EAGAIN - * if 'reserve' is true or 'sinfo' is false. + * this function must be called within transaction */ -int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo) +int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + u64 bytenr, u64 num_bytes) { + struct btrfs_block_group_cache *cache; + + cache = btrfs_lookup_block_group(root->fs_info, bytenr); + BUG_ON(!cache); + + /* + * pull in the free space cache (if any) so that our pin + * removes the free space from the cache. We have load_only set + * to one because the slow code to read in the free extents does check + * the pinned extents. + */ + cache_block_group(cache, trans, root, 1); + + pin_down_extent(root, cache, bytenr, num_bytes, 0); + + /* remove us from the free space cache (if we're there at all) */ + btrfs_remove_free_space(cache, bytenr, num_bytes); + btrfs_put_block_group(cache); + return 0; +} + +/** + * btrfs_update_reserved_bytes - update the block_group and space info counters + * @cache: The cache we are manipulating + * @num_bytes: The number of bytes in question + * @reserve: One of the reservation enums + * + * This is called by the allocator when it reserves space, or by somebody who is + * freeing space that was never actually used on disk. For example if you + * reserve some space for a new leaf in transaction A and before transaction A + * commits you free that leaf, you call this with reserve set to 0 in order to + * clear the reservation. + * + * Metadata reservations should be called with RESERVE_ALLOC so we do the proper + * ENOSPC accounting. For data we handle the reservation through clearing the + * delalloc bits in the io_tree. We have to do this since we could end up + * allocating less disk space for the amount of data we have reserved in the + * case of compression. + * + * If this is a reservation and the block group has become read only we cannot + * make the reservation and return -EAGAIN, otherwise this function always + * succeeds. + */ +static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve) +{ + struct btrfs_space_info *space_info = cache->space_info; int ret = 0; - if (sinfo) { - struct btrfs_space_info *space_info = cache->space_info; - spin_lock(&space_info->lock); - spin_lock(&cache->lock); - if (reserve) { - if (cache->ro) { - ret = -EAGAIN; - } else { - cache->reserved += num_bytes; - space_info->bytes_reserved += num_bytes; - } - } else { - if (cache->ro) - space_info->bytes_readonly += num_bytes; - cache->reserved -= num_bytes; - space_info->bytes_reserved -= num_bytes; - space_info->reservation_progress++; - } - spin_unlock(&cache->lock); - spin_unlock(&space_info->lock); - } else { - spin_lock(&cache->lock); + spin_lock(&space_info->lock); + spin_lock(&cache->lock); + if (reserve != RESERVE_FREE) { if (cache->ro) { ret = -EAGAIN; } else { - if (reserve) - cache->reserved += num_bytes; - else - cache->reserved -= num_bytes; + cache->reserved += num_bytes; + space_info->bytes_reserved += num_bytes; + if (reserve == RESERVE_ALLOC) { + BUG_ON(space_info->bytes_may_use < num_bytes); + space_info->bytes_may_use -= num_bytes; + } } - spin_unlock(&cache->lock); + } else { + if (cache->ro) + space_info->bytes_readonly += num_bytes; + cache->reserved -= num_bytes; + space_info->bytes_reserved -= num_bytes; + space_info->reservation_progress++; } + spin_unlock(&cache->lock); + spin_unlock(&space_info->lock); return ret; } @@ -4272,7 +4611,8 @@ int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, return 0; } -static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) +static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end, + const bool return_free_space) { struct btrfs_fs_info *fs_info = root->fs_info; struct btrfs_block_group_cache *cache = NULL; @@ -4292,7 +4632,8 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) if (start < cache->last_byte_to_unpin) { len = min(len, cache->last_byte_to_unpin - start); - btrfs_add_free_space(cache, start, len); + if (return_free_space) + btrfs_add_free_space(cache, start, len); } start += len; @@ -4301,13 +4642,8 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) spin_lock(&cache->lock); cache->pinned -= len; cache->space_info->bytes_pinned -= len; - if (cache->ro) { + if (cache->ro) cache->space_info->bytes_readonly += len; - } else if (cache->reserved_pinned > 0) { - len = min(len, cache->reserved_pinned); - cache->reserved_pinned -= len; - cache->space_info->bytes_reserved += len; - } spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); } @@ -4322,11 +4658,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, { struct btrfs_fs_info *fs_info = root->fs_info; struct extent_io_tree *unpin; - struct btrfs_block_rsv *block_rsv; - struct btrfs_block_rsv *next_rsv; u64 start; u64 end; - int idx; int ret; if (fs_info->pinned_extents == &fs_info->freed_extents[0]) @@ -4345,34 +4678,10 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, end + 1 - start, NULL); clear_extent_dirty(unpin, start, end, GFP_NOFS); - unpin_extent_range(root, start, end); + unpin_extent_range(root, start, end, true); cond_resched(); } - mutex_lock(&fs_info->durable_block_rsv_mutex); - list_for_each_entry_safe(block_rsv, next_rsv, - &fs_info->durable_block_rsv_list, list) { - - idx = trans->transid & 0x1; - if (block_rsv->freed[idx] > 0) { - block_rsv_add_bytes(block_rsv, - block_rsv->freed[idx], 0); - block_rsv->freed[idx] = 0; - } - if (atomic_read(&block_rsv->usage) == 0) { - btrfs_block_rsv_release(root, block_rsv, (u64)-1); - - if (block_rsv->freed[0] == 0 && - block_rsv->freed[1] == 0) { - list_del_init(&block_rsv->list); - kfree(block_rsv); - } - } else { - btrfs_block_rsv_release(root, block_rsv, 0); - } - } - mutex_unlock(&fs_info->durable_block_rsv_mutex); - return 0; } @@ -4452,7 +4761,9 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, printk(KERN_ERR "umm, got %d back from search" ", was looking for %llu\n", ret, (unsigned long long)bytenr); - btrfs_print_leaf(extent_root, path->nodes[0]); + if (ret > 0) + btrfs_print_leaf(extent_root, + path->nodes[0]); } BUG_ON(ret); extent_slot = path->slots[0]; @@ -4648,7 +4959,6 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, struct extent_buffer *buf, u64 parent, int last_ref) { - struct btrfs_block_rsv *block_rsv; struct btrfs_block_group_cache *cache = NULL; int ret; @@ -4663,64 +4973,24 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, if (!last_ref) return; - block_rsv = get_block_rsv(trans, root); cache = btrfs_lookup_block_group(root->fs_info, buf->start); - if (block_rsv->space_info != cache->space_info) - goto out; if (btrfs_header_generation(buf) == trans->transid) { if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { ret = check_ref_cleanup(trans, root, buf->start); if (!ret) - goto pin; + goto out; } if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) { pin_down_extent(root, cache, buf->start, buf->len, 1); - goto pin; + goto out; } WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)); btrfs_add_free_space(cache, buf->start, buf->len); - ret = btrfs_update_reserved_bytes(cache, buf->len, 0, 0); - if (ret == -EAGAIN) { - /* block group became read-only */ - btrfs_update_reserved_bytes(cache, buf->len, 0, 1); - goto out; - } - - ret = 1; - spin_lock(&block_rsv->lock); - if (block_rsv->reserved < block_rsv->size) { - block_rsv->reserved += buf->len; - ret = 0; - } - spin_unlock(&block_rsv->lock); - - if (ret) { - spin_lock(&cache->space_info->lock); - cache->space_info->bytes_reserved -= buf->len; - cache->space_info->reservation_progress++; - spin_unlock(&cache->space_info->lock); - } - goto out; - } -pin: - if (block_rsv->durable && !cache->ro) { - ret = 0; - spin_lock(&cache->lock); - if (!cache->ro) { - cache->reserved_pinned += buf->len; - ret = 1; - } - spin_unlock(&cache->lock); - - if (ret) { - spin_lock(&block_rsv->lock); - block_rsv->freed[trans->transid & 0x1] += buf->len; - spin_unlock(&block_rsv->lock); - } + btrfs_update_reserved_bytes(cache, buf->len, RESERVE_FREE); } out: /* @@ -4856,17 +5126,20 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root = orig_root->fs_info->extent_root; struct btrfs_free_cluster *last_ptr = NULL; struct btrfs_block_group_cache *block_group = NULL; + struct btrfs_block_group_cache *used_block_group; int empty_cluster = 2 * 1024 * 1024; int allowed_chunk_alloc = 0; int done_chunk_alloc = 0; struct btrfs_space_info *space_info; - int last_ptr_loop = 0; int loop = 0; int index = 0; + int alloc_type = (data & BTRFS_BLOCK_GROUP_DATA) ? + RESERVE_ALLOC_NO_ACCOUNT : RESERVE_ALLOC; bool found_uncached_bg = false; bool failed_cluster_refill = false; bool failed_alloc = false; bool use_cluster = true; + bool have_caching_bg = false; u64 ideal_cache_percent = 0; u64 ideal_cache_offset = 0; @@ -4919,6 +5192,7 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, ideal_cache: block_group = btrfs_lookup_block_group(root->fs_info, search_start); + used_block_group = block_group; /* * we don't want to use the block group if it doesn't match our * allocation bits, or if its not cached. @@ -4949,12 +5223,14 @@ ideal_cache: } } search: + have_caching_bg = false; down_read(&space_info->groups_sem); list_for_each_entry(block_group, &space_info->block_groups[index], list) { u64 offset; int cached; + used_block_group = block_group; btrfs_get_block_group(block_group); search_start = block_group->key.objectid; @@ -4978,13 +5254,15 @@ search: } have_block_group: - if (unlikely(block_group->cached == BTRFS_CACHE_NO)) { + cached = block_group_cache_done(block_group); + if (unlikely(!cached)) { u64 free_percent; + found_uncached_bg = true; ret = cache_block_group(block_group, trans, orig_root, 1); if (block_group->cached == BTRFS_CACHE_FINISHED) - goto have_block_group; + goto alloc; free_percent = btrfs_block_group_used(&block_group->item); free_percent *= 100; @@ -4998,19 +5276,14 @@ have_block_group: } /* - * We only want to start kthread caching if we are at - * the point where we will wait for caching to make - * progress, or if our ideal search is over and we've - * found somebody to start caching. + * The caching workers are limited to 2 threads, so we + * can queue as much work as we care to. */ - if (loop > LOOP_CACHING_NOWAIT || - (loop > LOOP_FIND_IDEAL && - atomic_read(&space_info->caching_threads) < 2)) { + if (loop > LOOP_FIND_IDEAL) { ret = cache_block_group(block_group, trans, orig_root, 0); BUG_ON(ret); } - found_uncached_bg = true; /* * If loop is set for cached only, try the next block @@ -5020,92 +5293,80 @@ have_block_group: goto loop; } - cached = block_group_cache_done(block_group); - if (unlikely(!cached)) - found_uncached_bg = true; - +alloc: if (unlikely(block_group->ro)) goto loop; spin_lock(&block_group->free_space_ctl->tree_lock); if (cached && block_group->free_space_ctl->free_space < - num_bytes + empty_size) { + num_bytes + empty_cluster + empty_size) { spin_unlock(&block_group->free_space_ctl->tree_lock); goto loop; } spin_unlock(&block_group->free_space_ctl->tree_lock); /* - * Ok we want to try and use the cluster allocator, so lets look - * there, unless we are on LOOP_NO_EMPTY_SIZE, since we will - * have tried the cluster allocator plenty of times at this - * point and not have found anything, so we are likely way too - * fragmented for the clustering stuff to find anything, so lets - * just skip it and let the allocator find whatever block it can - * find + * Ok we want to try and use the cluster allocator, so + * lets look there */ - if (last_ptr && loop < LOOP_NO_EMPTY_SIZE) { + if (last_ptr) { /* * the refill lock keeps out other * people trying to start a new cluster */ spin_lock(&last_ptr->refill_lock); - if (last_ptr->block_group && - (last_ptr->block_group->ro || - !block_group_bits(last_ptr->block_group, data))) { - offset = 0; + used_block_group = last_ptr->block_group; + if (used_block_group != block_group && + (!used_block_group || + used_block_group->ro || + !block_group_bits(used_block_group, data))) { + used_block_group = block_group; goto refill_cluster; } - offset = btrfs_alloc_from_cluster(block_group, last_ptr, - num_bytes, search_start); + if (used_block_group != block_group) + btrfs_get_block_group(used_block_group); + + offset = btrfs_alloc_from_cluster(used_block_group, + last_ptr, num_bytes, used_block_group->key.objectid); if (offset) { /* we have a block, we're done */ spin_unlock(&last_ptr->refill_lock); goto checks; } - spin_lock(&last_ptr->lock); - /* - * whoops, this cluster doesn't actually point to - * this block group. Get a ref on the block - * group is does point to and try again - */ - if (!last_ptr_loop && last_ptr->block_group && - last_ptr->block_group != block_group) { - - btrfs_put_block_group(block_group); - block_group = last_ptr->block_group; - btrfs_get_block_group(block_group); - spin_unlock(&last_ptr->lock); - spin_unlock(&last_ptr->refill_lock); - - last_ptr_loop = 1; - search_start = block_group->key.objectid; - /* - * we know this block group is properly - * in the list because - * btrfs_remove_block_group, drops the - * cluster before it removes the block - * group from the list - */ - goto have_block_group; + WARN_ON(last_ptr->block_group != used_block_group); + if (used_block_group != block_group) { + btrfs_put_block_group(used_block_group); + used_block_group = block_group; } - spin_unlock(&last_ptr->lock); refill_cluster: + BUG_ON(used_block_group != block_group); + /* If we are on LOOP_NO_EMPTY_SIZE, we can't + * set up a new clusters, so lets just skip it + * and let the allocator find whatever block + * it can find. If we reach this point, we + * will have tried the cluster allocator + * plenty of times and not have found + * anything, so we are likely way too + * fragmented for the clustering stuff to find + * anything. */ + if (loop >= LOOP_NO_EMPTY_SIZE) { + spin_unlock(&last_ptr->refill_lock); + goto unclustered_alloc; + } + /* * this cluster didn't work out, free it and * start over */ btrfs_return_cluster_to_free_space(NULL, last_ptr); - last_ptr_loop = 0; - /* allocate a cluster in this block group */ ret = btrfs_find_space_cluster(trans, root, block_group, last_ptr, - offset, num_bytes, + search_start, num_bytes, empty_cluster + empty_size); if (ret == 0) { /* @@ -5141,6 +5402,7 @@ refill_cluster: goto loop; } +unclustered_alloc: offset = btrfs_find_space_for_alloc(block_group, search_start, num_bytes, empty_size); /* @@ -5159,20 +5421,22 @@ refill_cluster: failed_alloc = true; goto have_block_group; } else if (!offset) { + if (!cached) + have_caching_bg = true; goto loop; } checks: search_start = stripe_align(root, offset); /* move on to the next group */ if (search_start + num_bytes >= search_end) { - btrfs_add_free_space(block_group, offset, num_bytes); + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } /* move on to the next group */ if (search_start + num_bytes > - block_group->key.objectid + block_group->key.offset) { - btrfs_add_free_space(block_group, offset, num_bytes); + used_block_group->key.objectid + used_block_group->key.offset) { + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } @@ -5180,14 +5444,14 @@ checks: ins->offset = num_bytes; if (offset < search_start) - btrfs_add_free_space(block_group, offset, + btrfs_add_free_space(used_block_group, offset, search_start - offset); BUG_ON(offset > search_start); - ret = btrfs_update_reserved_bytes(block_group, num_bytes, 1, - (data & BTRFS_BLOCK_GROUP_DATA)); + ret = btrfs_update_reserved_bytes(used_block_group, num_bytes, + alloc_type); if (ret == -EAGAIN) { - btrfs_add_free_space(block_group, offset, num_bytes); + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } @@ -5196,19 +5460,26 @@ checks: ins->offset = num_bytes; if (offset < search_start) - btrfs_add_free_space(block_group, offset, + btrfs_add_free_space(used_block_group, offset, search_start - offset); BUG_ON(offset > search_start); + if (used_block_group != block_group) + btrfs_put_block_group(used_block_group); btrfs_put_block_group(block_group); break; loop: failed_cluster_refill = false; failed_alloc = false; BUG_ON(index != get_block_group_index(block_group)); + if (used_block_group != block_group) + btrfs_put_block_group(used_block_group); btrfs_put_block_group(block_group); } up_read(&space_info->groups_sem); + if (!ins->objectid && loop >= LOOP_CACHING_WAIT && have_caching_bg) + goto search; + if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES) goto search; @@ -5227,8 +5498,7 @@ loop: if (loop == LOOP_FIND_IDEAL && found_uncached_bg) { found_uncached_bg = false; loop++; - if (!ideal_cache_percent && - atomic_read(&space_info->caching_threads)) + if (!ideal_cache_percent) goto search; /* @@ -5308,7 +5578,8 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes, int index = 0; spin_lock(&info->lock); - printk(KERN_INFO "space_info has %llu free, is %sfull\n", + printk(KERN_INFO "space_info %llu has %llu free, is %sfull\n", + (unsigned long long)info->flags, (unsigned long long)(info->total_bytes - info->bytes_used - info->bytes_pinned - info->bytes_reserved - info->bytes_readonly), @@ -5394,7 +5665,8 @@ again: return ret; } -int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) +static int __btrfs_free_reserved_extent(struct btrfs_root *root, + u64 start, u64 len, int pin) { struct btrfs_block_group_cache *cache; int ret = 0; @@ -5406,11 +5678,14 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) return -ENOSPC; } - if (btrfs_test_opt(root, DISCARD)) - ret = btrfs_discard_extent(root, start, len, NULL); - - btrfs_add_free_space(cache, start, len); - btrfs_update_reserved_bytes(cache, len, 0, 1); + if (pin) + pin_down_extent(root, cache, start, len, 1); + else { + if (btrfs_test_opt(root, DISCARD)) + ret = btrfs_discard_extent(root, start, len, NULL); + btrfs_add_free_space(cache, start, len); + btrfs_update_reserved_bytes(cache, len, RESERVE_FREE); + } btrfs_put_block_group(cache); trace_btrfs_reserved_extent_free(root, start, len); @@ -5418,6 +5693,18 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) return ret; } +int btrfs_free_reserved_extent(struct btrfs_root *root, + u64 start, u64 len) +{ + return __btrfs_free_reserved_extent(root, start, len, 0); +} + +int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root, + u64 start, u64 len) +{ + return __btrfs_free_reserved_extent(root, start, len, 1); +} + static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 parent, u64 root_objectid, @@ -5502,7 +5789,8 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans, u32 size = sizeof(*extent_item) + sizeof(*block_info) + sizeof(*iref); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; path->leave_spinning = 1; ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, @@ -5612,7 +5900,8 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, put_caching_control(caching_ctl); } - ret = btrfs_update_reserved_bytes(block_group, ins->offset, 1, 1); + ret = btrfs_update_reserved_bytes(block_group, ins->offset, + RESERVE_ALLOC_NO_ACCOUNT); BUG_ON(ret); btrfs_put_block_group(block_group); ret = alloc_reserved_file_extent(trans, root, 0, root_objectid, @@ -5631,7 +5920,7 @@ struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans, if (!buf) return ERR_PTR(-ENOMEM); btrfs_set_header_generation(buf, trans->transid); - btrfs_set_buffer_lockdep_class(buf, level); + btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level); btrfs_tree_lock(buf); clean_tree_block(trans, root, buf); @@ -5669,8 +5958,7 @@ use_block_rsv(struct btrfs_trans_handle *trans, block_rsv = get_block_rsv(trans, root); if (block_rsv->size == 0) { - ret = reserve_metadata_bytes(trans, root, block_rsv, - blocksize, 0); + ret = reserve_metadata_bytes(root, block_rsv, blocksize, 0); /* * If we couldn't reserve metadata bytes try and use some from * the global reserve. @@ -5690,13 +5978,15 @@ use_block_rsv(struct btrfs_trans_handle *trans, if (!ret) return block_rsv; if (ret) { - WARN_ON(1); - ret = reserve_metadata_bytes(trans, root, block_rsv, blocksize, - 0); + static DEFINE_RATELIMIT_STATE(_rs, + DEFAULT_RATELIMIT_INTERVAL, + /*DEFAULT_RATELIMIT_BURST*/ 2); + if (__ratelimit(&_rs)) { + printk(KERN_DEBUG "btrfs: block rsv returned %d\n", ret); + WARN_ON(1); + } + ret = reserve_metadata_bytes(root, block_rsv, blocksize, 0); if (!ret) { - spin_lock(&block_rsv->lock); - block_rsv->size += blocksize; - spin_unlock(&block_rsv->lock); return block_rsv; } else if (ret && block_rsv != global_rsv) { ret = block_rsv_use_bytes(global_rsv, blocksize); @@ -5918,7 +6208,7 @@ static noinline int walk_down_proc(struct btrfs_trans_handle *trans, return 1; if (path->locks[level] && !wc->keep_locks) { - btrfs_tree_unlock(eb); + btrfs_tree_unlock_rw(eb, path->locks[level]); path->locks[level] = 0; } return 0; @@ -5942,7 +6232,7 @@ static noinline int walk_down_proc(struct btrfs_trans_handle *trans, * keep the tree lock */ if (path->locks[level] && level > 0) { - btrfs_tree_unlock(eb); + btrfs_tree_unlock_rw(eb, path->locks[level]); path->locks[level] = 0; } return 0; @@ -6055,7 +6345,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans, BUG_ON(level != btrfs_header_level(next)); path->nodes[level] = next; path->slots[level] = 0; - path->locks[level] = 1; + path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; wc->level = level; if (wc->level == 1) wc->reada_slot = 0; @@ -6126,7 +6416,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans, BUG_ON(level == 0); btrfs_tree_lock(eb); btrfs_set_lock_blocking(eb); - path->locks[level] = 1; + path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; ret = btrfs_lookup_extent_info(trans, root, eb->start, eb->len, @@ -6135,8 +6425,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans, BUG_ON(ret); BUG_ON(wc->refs[level] == 0); if (wc->refs[level] == 1) { - btrfs_tree_unlock(eb); - path->locks[level] = 0; + btrfs_tree_unlock_rw(eb, path->locks[level]); return 1; } } @@ -6158,7 +6447,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans, btrfs_header_generation(eb) == trans->transid) { btrfs_tree_lock(eb); btrfs_set_lock_blocking(eb); - path->locks[level] = 1; + path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; } clean_tree_block(trans, root, eb); } @@ -6237,7 +6526,8 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans, return 0; if (path->locks[level]) { - btrfs_tree_unlock(path->nodes[level]); + btrfs_tree_unlock_rw(path->nodes[level], + path->locks[level]); path->locks[level] = 0; } free_extent_buffer(path->nodes[level]); @@ -6259,8 +6549,8 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans, * also make sure backrefs for the shared block and all lower level * blocks are properly updated. */ -int btrfs_drop_snapshot(struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, int update_ref) +void btrfs_drop_snapshot(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, int update_ref) { struct btrfs_path *path; struct btrfs_trans_handle *trans; @@ -6271,12 +6561,20 @@ int btrfs_drop_snapshot(struct btrfs_root *root, int err = 0; int ret; int level; + bool root_dropped = false; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + err = -ENOMEM; + goto out; + } wc = kzalloc(sizeof(*wc), GFP_NOFS); - BUG_ON(!wc); + if (!wc) { + btrfs_free_path(path); + err = -ENOMEM; + goto out; + } trans = btrfs_start_transaction(tree_root, 0); BUG_ON(IS_ERR(trans)); @@ -6289,7 +6587,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, path->nodes[level] = btrfs_lock_root_node(root); btrfs_set_lock_blocking(path->nodes[level]); path->slots[level] = 0; - path->locks[level] = 1; + path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; memset(&wc->update_progress, 0, sizeof(wc->update_progress)); } else { @@ -6304,7 +6602,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, path->lowest_level = 0; if (ret < 0) { err = ret; - goto out; + goto out_free; } WARN_ON(ret > 0); @@ -6318,6 +6616,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, while (1) { btrfs_tree_lock(path->nodes[level]); btrfs_set_lock_blocking(path->nodes[level]); + path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; ret = btrfs_lookup_extent_info(trans, root, path->nodes[level]->start, @@ -6331,6 +6630,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, break; btrfs_tree_unlock(path->nodes[level]); + path->locks[level] = 0; WARN_ON(wc->refs[level] != 1); level--; } @@ -6411,11 +6711,24 @@ int btrfs_drop_snapshot(struct btrfs_root *root, free_extent_buffer(root->commit_root); kfree(root); } -out: + root_dropped = true; +out_free: btrfs_end_transaction_throttle(trans, tree_root); kfree(wc); btrfs_free_path(path); - return err; +out: + /* + * So if we need to stop dropping the snapshot for whatever reason we + * need to make sure to add it back to the dead root list so that we + * keep trying to do the work later. This also cleans up roots if we + * don't have it in the radix (like when we recover after a power fail + * or unmount) so we don't leak memory. + */ + if (root_dropped == false) + btrfs_add_dead_root(root); + if (err && err != -EAGAIN) + btrfs_std_error(root->fs_info, err); + return; } /* @@ -6457,7 +6770,7 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, level = btrfs_header_level(node); path->nodes[level] = node; path->slots[level] = 0; - path->locks[level] = 1; + path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; wc->refs[parent_level] = 1; wc->flags[parent_level] = BTRFS_BLOCK_FLAG_FULL_BACKREF; @@ -6532,30 +6845,45 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) return flags; } -static int set_block_group_ro(struct btrfs_block_group_cache *cache) +static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force) { struct btrfs_space_info *sinfo = cache->space_info; u64 num_bytes; + u64 min_allocable_bytes; int ret = -ENOSPC; - if (cache->ro) - return 0; + + /* + * We need some metadata space and system metadata space for + * allocating chunks in some corner cases until we force to set + * it to be readonly. + */ + if ((sinfo->flags & + (BTRFS_BLOCK_GROUP_SYSTEM | BTRFS_BLOCK_GROUP_METADATA)) && + !force) + min_allocable_bytes = 1 * 1024 * 1024; + else + min_allocable_bytes = 0; spin_lock(&sinfo->lock); spin_lock(&cache->lock); + + if (cache->ro) { + ret = 0; + goto out; + } + 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 + - cache->reserved_pinned + num_bytes <= sinfo->total_bytes) { + sinfo->bytes_may_use + sinfo->bytes_readonly + num_bytes + + min_allocable_bytes <= sinfo->total_bytes) { sinfo->bytes_readonly += num_bytes; - sinfo->bytes_reserved += cache->reserved_pinned; - cache->reserved_pinned = 0; cache->ro = 1; ret = 0; } - +out: spin_unlock(&cache->lock); spin_unlock(&sinfo->lock); return ret; @@ -6579,7 +6907,7 @@ int btrfs_set_block_group_ro(struct btrfs_root *root, do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, CHUNK_ALLOC_FORCE); - ret = set_block_group_ro(cache); + ret = set_block_group_ro(cache, 0); if (!ret) goto out; alloc_flags = get_alloc_profile(root, cache->space_info->flags); @@ -6587,7 +6915,7 @@ int btrfs_set_block_group_ro(struct btrfs_root *root, CHUNK_ALLOC_FORCE); if (ret < 0) goto out; - ret = set_block_group_ro(cache); + ret = set_block_group_ro(cache, 0); out: btrfs_end_transaction(trans, root); return ret; @@ -6688,6 +7016,10 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) struct btrfs_space_info *space_info; struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; struct btrfs_device *device; + u64 min_free; + u64 dev_min = 1; + u64 dev_nr = 0; + int index; int full = 0; int ret = 0; @@ -6697,8 +7029,10 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) if (!block_group) return -1; + min_free = btrfs_block_group_used(&block_group->item); + /* no bytes used, we're good */ - if (!btrfs_block_group_used(&block_group->item)) + if (!min_free) goto out; space_info = block_group->space_info; @@ -6714,10 +7048,9 @@ int btrfs_can_relocate(struct btrfs_root *root, 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 + - btrfs_block_group_used(&block_group->item) < - space_info->total_bytes)) { + (space_info->bytes_used + space_info->bytes_reserved + + space_info->bytes_pinned + space_info->bytes_readonly + + min_free < space_info->total_bytes)) { spin_unlock(&space_info->lock); goto out; } @@ -6734,9 +7067,31 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) if (full) goto out; + /* + * index: + * 0: raid10 + * 1: raid1 + * 2: dup + * 3: raid0 + * 4: single + */ + index = get_block_group_index(block_group); + if (index == 0) { + dev_min = 4; + /* Divide by 2 */ + min_free >>= 1; + } else if (index == 1) { + dev_min = 2; + } else if (index == 2) { + /* Multiply by 2 */ + min_free <<= 1; + } else if (index == 3) { + dev_min = fs_devices->rw_devices; + do_div(min_free, dev_min); + } + mutex_lock(&root->fs_info->chunk_mutex); list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) { - u64 min_free = btrfs_block_group_used(&block_group->item); u64 dev_offset; /* @@ -6747,7 +7102,11 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) ret = find_free_dev_extent(NULL, device, min_free, &dev_offset, NULL); if (!ret) + dev_nr++; + + if (dev_nr >= dev_min) break; + ret = -1; } } @@ -6887,7 +7246,8 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info) struct btrfs_space_info, list); if (space_info->bytes_pinned > 0 || - space_info->bytes_reserved > 0) { + space_info->bytes_reserved > 0 || + space_info->bytes_may_use > 0) { WARN_ON(1); dump_space_info(space_info, 0, 0); } @@ -6929,14 +7289,12 @@ int btrfs_read_block_groups(struct btrfs_root *root) return -ENOMEM; path->reada = 1; - cache_gen = btrfs_super_cache_generation(&root->fs_info->super_copy); - if (cache_gen != 0 && - btrfs_super_generation(&root->fs_info->super_copy) != cache_gen) + cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy); + if (btrfs_test_opt(root, SPACE_CACHE) && + btrfs_super_generation(root->fs_info->super_copy) != cache_gen) need_clear = 1; if (btrfs_test_opt(root, CLEAR_CACHE)) need_clear = 1; - if (!btrfs_test_opt(root, SPACE_CACHE) && cache_gen) - printk(KERN_INFO "btrfs: disk space caching is enabled\n"); while (1) { ret = find_first_block_group(root, path, &key); @@ -7024,7 +7382,7 @@ int btrfs_read_block_groups(struct btrfs_root *root) set_avail_alloc_bits(root->fs_info, cache->flags); if (btrfs_chunk_readonly(root, cache->key.objectid)) - set_block_group_ro(cache); + set_block_group_ro(cache, 1); } list_for_each_entry_rcu(space_info, &root->fs_info->space_info, list) { @@ -7038,9 +7396,9 @@ int btrfs_read_block_groups(struct btrfs_root *root) * mirrored block groups. */ list_for_each_entry(cache, &space_info->block_groups[3], list) - set_block_group_ro(cache); + set_block_group_ro(cache, 1); list_for_each_entry(cache, &space_info->block_groups[4], list) - set_block_group_ro(cache); + set_block_group_ro(cache, 1); } init_global_block_rsv(info); @@ -7170,11 +7528,15 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, spin_unlock(&cluster->refill_lock); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + ret = -ENOMEM; + goto out; + } - inode = lookup_free_space_inode(root, block_group, path); + inode = lookup_free_space_inode(tree_root, block_group, path); if (!IS_ERR(inode)) { - btrfs_orphan_add(trans, inode); + ret = btrfs_orphan_add(trans, inode); + BUG_ON(ret); clear_nlink(inode); /* One for the block groups ref */ spin_lock(&block_group->lock); @@ -7187,7 +7549,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, spin_unlock(&block_group->lock); } /* One for our lookup ref */ - iput(inode); + btrfs_add_delayed_iput(inode); } key.objectid = BTRFS_FREE_SPACE_OBJECTID; @@ -7258,7 +7620,7 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info) int mixed = 0; int ret; - disk_super = &fs_info->super_copy; + disk_super = fs_info->super_copy; if (!btrfs_super_root(disk_super)) return 1; @@ -7289,7 +7651,7 @@ out: int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) { - return unpin_extent_range(root, start, end); + return unpin_extent_range(root, start, end, false); } int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr, diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 7055d11..9a837a8 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -17,6 +17,7 @@ #include "compat.h" #include "ctree.h" #include "btrfs_inode.h" +#include "volumes.h" static struct kmem_cache *extent_state_cache; static struct kmem_cache *extent_buffer_cache; @@ -254,14 +255,14 @@ static void merge_cb(struct extent_io_tree *tree, struct extent_state *new, * * This should be called with the tree lock held. */ -static int merge_state(struct extent_io_tree *tree, - struct extent_state *state) +static void merge_state(struct extent_io_tree *tree, + struct extent_state *state) { struct extent_state *other; struct rb_node *other_node; if (state->state & (EXTENT_IOBITS | EXTENT_BOUNDARY)) - return 0; + return; other_node = rb_prev(&state->rb_node); if (other_node) { @@ -281,26 +282,19 @@ static int merge_state(struct extent_io_tree *tree, if (other->start == state->end + 1 && other->state == state->state) { merge_cb(tree, state, other); - other->start = state->start; - state->tree = NULL; - rb_erase(&state->rb_node, &tree->state); - free_extent_state(state); - state = NULL; + state->end = other->end; + other->tree = NULL; + rb_erase(&other->rb_node, &tree->state); + free_extent_state(other); } } - - return 0; } -static int set_state_cb(struct extent_io_tree *tree, +static void set_state_cb(struct extent_io_tree *tree, struct extent_state *state, int *bits) { - if (tree->ops && tree->ops->set_bit_hook) { - return tree->ops->set_bit_hook(tree->mapping->host, - state, bits); - } - - return 0; + if (tree->ops && tree->ops->set_bit_hook) + tree->ops->set_bit_hook(tree->mapping->host, state, bits); } static void clear_state_cb(struct extent_io_tree *tree, @@ -310,6 +304,9 @@ static void clear_state_cb(struct extent_io_tree *tree, tree->ops->clear_bit_hook(tree->mapping->host, state, bits); } +static void set_state_bits(struct extent_io_tree *tree, + struct extent_state *state, int *bits); + /* * insert an extent_state struct into the tree. 'bits' are set on the * struct before it is inserted. @@ -325,8 +322,6 @@ static int insert_state(struct extent_io_tree *tree, int *bits) { struct rb_node *node; - int bits_to_set = *bits & ~EXTENT_CTLBITS; - int ret; if (end < start) { printk(KERN_ERR "btrfs end < start %llu %llu\n", @@ -336,13 +331,9 @@ static int insert_state(struct extent_io_tree *tree, } state->start = start; state->end = end; - ret = set_state_cb(tree, state, bits); - if (ret) - return ret; - if (bits_to_set & EXTENT_DIRTY) - tree->dirty_bytes += end - start + 1; - state->state |= bits_to_set; + set_state_bits(tree, state, bits); + node = tree_insert(&tree->state, end, &state->rb_node); if (node) { struct extent_state *found; @@ -351,7 +342,6 @@ static int insert_state(struct extent_io_tree *tree, "%llu %llu\n", (unsigned long long)found->start, (unsigned long long)found->end, (unsigned long long)start, (unsigned long long)end); - free_extent_state(state); return -EEXIST; } state->tree = tree; @@ -359,13 +349,11 @@ static int insert_state(struct extent_io_tree *tree, return 0; } -static int split_cb(struct extent_io_tree *tree, struct extent_state *orig, +static void split_cb(struct extent_io_tree *tree, struct extent_state *orig, u64 split) { if (tree->ops && tree->ops->split_extent_hook) - return tree->ops->split_extent_hook(tree->mapping->host, - orig, split); - return 0; + tree->ops->split_extent_hook(tree->mapping->host, orig, split); } /* @@ -500,7 +488,8 @@ again: cached_state = NULL; } - if (cached && cached->tree && cached->start == start) { + if (cached && cached->tree && cached->start <= start && + cached->end > start) { if (clear) atomic_dec(&cached->refs); state = cached; @@ -660,34 +649,25 @@ again: if (start > end) break; - if (need_resched()) { - spin_unlock(&tree->lock); - cond_resched(); - spin_lock(&tree->lock); - } + cond_resched_lock(&tree->lock); } out: spin_unlock(&tree->lock); return 0; } -static int set_state_bits(struct extent_io_tree *tree, +static void set_state_bits(struct extent_io_tree *tree, struct extent_state *state, int *bits) { - int ret; int bits_to_set = *bits & ~EXTENT_CTLBITS; - ret = set_state_cb(tree, state, bits); - if (ret) - return ret; + set_state_cb(tree, state, bits); if ((bits_to_set & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) { u64 range = state->end - state->start + 1; tree->dirty_bytes += range; } state->state |= bits_to_set; - - return 0; } static void cache_state(struct extent_state *state, @@ -742,7 +722,8 @@ again: spin_lock(&tree->lock); if (cached_state && *cached_state) { state = *cached_state; - if (state->start == start && state->tree) { + if (state->start <= start && state->end > start && + state->tree) { node = &state->rb_node; goto hit_next; } @@ -779,17 +760,15 @@ hit_next: goto out; } - err = set_state_bits(tree, state, &bits); - if (err) - goto out; + set_state_bits(tree, state, &bits); - next_node = rb_next(node); cache_state(state, cached_state); merge_state(tree, state); if (last_end == (u64)-1) goto out; start = last_end + 1; + next_node = rb_next(&state->rb_node); if (next_node && start < end && prealloc && !need_resched()) { state = rb_entry(next_node, struct extent_state, rb_node); @@ -830,9 +809,7 @@ hit_next: if (err) goto out; if (state->end <= end) { - err = set_state_bits(tree, state, &bits); - if (err) - goto out; + set_state_bits(tree, state, &bits); cache_state(state, cached_state); merge_state(tree, state); if (last_end == (u64)-1) @@ -862,7 +839,6 @@ hit_next: * Avoid to free 'prealloc' if it can be merged with * the later extent. */ - atomic_inc(&prealloc->refs); err = insert_state(tree, prealloc, start, this_end, &bits); BUG_ON(err == -EEXIST); @@ -872,7 +848,6 @@ hit_next: goto out; } cache_state(prealloc, cached_state); - free_extent_state(prealloc); prealloc = NULL; start = this_end + 1; goto search_again; @@ -895,12 +870,204 @@ hit_next: err = split_state(tree, state, prealloc, end + 1); BUG_ON(err == -EEXIST); - err = set_state_bits(tree, prealloc, &bits); + set_state_bits(tree, prealloc, &bits); + cache_state(prealloc, cached_state); + merge_state(tree, prealloc); + prealloc = NULL; + goto out; + } + + goto search_again; + +out: + spin_unlock(&tree->lock); + if (prealloc) + free_extent_state(prealloc); + + return err; + +search_again: + if (start > end) + goto out; + spin_unlock(&tree->lock); + if (mask & __GFP_WAIT) + cond_resched(); + goto again; +} + +/** + * convert_extent - convert all bits in a given range from one bit to another + * @tree: the io tree to search + * @start: the start offset in bytes + * @end: the end offset in bytes (inclusive) + * @bits: the bits to set in this range + * @clear_bits: the bits to clear in this range + * @mask: the allocation mask + * + * This will go through and set bits for the given range. If any states exist + * already in this range they are set with the given bit and cleared of the + * clear_bits. This is only meant to be used by things that are mergeable, ie + * converting from say DELALLOC to DIRTY. This is not meant to be used with + * boundary bits like LOCK. + */ +int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, + int bits, int clear_bits, gfp_t mask) +{ + struct extent_state *state; + struct extent_state *prealloc = NULL; + struct rb_node *node; + int err = 0; + u64 last_start; + u64 last_end; + +again: + if (!prealloc && (mask & __GFP_WAIT)) { + prealloc = alloc_extent_state(mask); + if (!prealloc) + return -ENOMEM; + } + + spin_lock(&tree->lock); + /* + * this search will find all the extents that end after + * our range starts. + */ + node = tree_search(tree, start); + if (!node) { + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + err = insert_state(tree, prealloc, start, end, &bits); + prealloc = NULL; + BUG_ON(err == -EEXIST); + goto out; + } + state = rb_entry(node, struct extent_state, rb_node); +hit_next: + last_start = state->start; + last_end = state->end; + + /* + * | ---- desired range ---- | + * | state | + * + * Just lock what we found and keep going + */ + if (state->start == start && state->end <= end) { + struct rb_node *next_node; + + set_state_bits(tree, state, &bits); + clear_state_bit(tree, state, &clear_bits, 0); + + merge_state(tree, state); + if (last_end == (u64)-1) + goto out; + + start = last_end + 1; + next_node = rb_next(&state->rb_node); + if (next_node && start < end && prealloc && !need_resched()) { + state = rb_entry(next_node, struct extent_state, + rb_node); + if (state->start == start) + goto hit_next; + } + goto search_again; + } + + /* + * | ---- desired range ---- | + * | state | + * or + * | ------------- state -------------- | + * + * We need to split the extent we found, and may flip bits on + * second half. + * + * If the extent we found extends past our + * range, we just split and search again. It'll get split + * again the next time though. + * + * If the extent we found is inside our range, we set the + * desired bit on it. + */ + if (state->start < start) { + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + err = split_state(tree, state, prealloc, start); + BUG_ON(err == -EEXIST); + prealloc = NULL; + if (err) + goto out; + if (state->end <= end) { + set_state_bits(tree, state, &bits); + clear_state_bit(tree, state, &clear_bits, 0); + merge_state(tree, state); + if (last_end == (u64)-1) + goto out; + start = last_end + 1; + } + goto search_again; + } + /* + * | ---- desired range ---- | + * | state | or | state | + * + * There's a hole, we need to insert something in it and + * ignore the extent we found. + */ + if (state->start > start) { + u64 this_end; + if (end < last_start) + this_end = end; + else + this_end = last_start - 1; + + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + + /* + * Avoid to free 'prealloc' if it can be merged with + * the later extent. + */ + err = insert_state(tree, prealloc, start, this_end, + &bits); + BUG_ON(err == -EEXIST); if (err) { + free_extent_state(prealloc); prealloc = NULL; goto out; } - cache_state(prealloc, cached_state); + prealloc = NULL; + start = this_end + 1; + goto search_again; + } + /* + * | ---- desired range ---- | + * | state | + * We need to split the extent, and set the bit + * on the first half + */ + if (state->start <= end && state->end > end) { + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + + err = split_state(tree, state, prealloc, end + 1); + BUG_ON(err == -EEXIST); + + set_state_bits(tree, prealloc, &bits); + clear_state_bit(tree, prealloc, &clear_bits, 0); + merge_state(tree, prealloc); prealloc = NULL; goto out; @@ -949,7 +1116,7 @@ int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask) { return set_extent_bit(tree, start, end, - EXTENT_DELALLOC | EXTENT_DIRTY | EXTENT_UPTODATE, + EXTENT_DELALLOC | EXTENT_UPTODATE, 0, NULL, cached_state, mask); } @@ -1042,19 +1209,33 @@ int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask) mask); } -/* - * helper function to set both pages and extents in the tree writeback - */ -static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end) +int extent_range_clear_dirty_for_io(struct inode *inode, u64 start, u64 end) { unsigned long index = start >> PAGE_CACHE_SHIFT; unsigned long end_index = end >> PAGE_CACHE_SHIFT; struct page *page; while (index <= end_index) { - page = find_get_page(tree->mapping, index); - BUG_ON(!page); - set_page_writeback(page); + page = find_get_page(inode->i_mapping, index); + BUG_ON(!page); /* Pages should be in the extent_io_tree */ + clear_page_dirty_for_io(page); + page_cache_release(page); + index++; + } + return 0; +} + +int extent_range_redirty_for_io(struct inode *inode, u64 start, u64 end) +{ + unsigned long index = start >> PAGE_CACHE_SHIFT; + unsigned long end_index = end >> PAGE_CACHE_SHIFT; + struct page *page; + + while (index <= end_index) { + page = find_get_page(inode->i_mapping, index); + BUG_ON(!page); /* Pages should be in the extent_io_tree */ + account_page_redirty(page); + __set_page_dirty_nobuffers(page); page_cache_release(page); index++; } @@ -1062,43 +1243,22 @@ static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end) } /* - * find the first offset in the io tree with 'bits' set. zero is - * returned if we find something, and *start_ret and *end_ret are - * set to reflect the state struct that was found. - * - * If nothing was found, 1 is returned, < 0 on error + * helper function to set both pages and extents in the tree writeback */ -int find_first_extent_bit(struct extent_io_tree *tree, u64 start, - u64 *start_ret, u64 *end_ret, int bits) +static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end) { - struct rb_node *node; - struct extent_state *state; - int ret = 1; - - spin_lock(&tree->lock); - /* - * this search will find all the extents that end after - * our range starts. - */ - node = tree_search(tree, start); - if (!node) - goto out; + unsigned long index = start >> PAGE_CACHE_SHIFT; + unsigned long end_index = end >> PAGE_CACHE_SHIFT; + struct page *page; - while (1) { - state = rb_entry(node, struct extent_state, rb_node); - if (state->end >= start && (state->state & bits)) { - *start_ret = state->start; - *end_ret = state->end; - ret = 0; - break; - } - node = rb_next(node); - if (!node) - break; + while (index <= end_index) { + page = find_get_page(tree->mapping, index); + BUG_ON(!page); + set_page_writeback(page); + page_cache_release(page); + index++; } -out: - spin_unlock(&tree->lock); - return ret; + return 0; } /* find the first state struct with 'bits' set after 'start', and @@ -1133,6 +1293,30 @@ out: } /* + * find the first offset in the io tree with 'bits' set. zero is + * returned if we find something, and *start_ret and *end_ret are + * set to reflect the state struct that was found. + * + * If nothing was found, 1 is returned, < 0 on error + */ +int find_first_extent_bit(struct extent_io_tree *tree, u64 start, + u64 *start_ret, u64 *end_ret, int bits) +{ + struct extent_state *state; + int ret = 1; + + spin_lock(&tree->lock); + state = find_first_extent_bit_state(tree, start, bits); + if (state) { + *start_ret = state->start; + *end_ret = state->end; + ret = 0; + } + spin_unlock(&tree->lock); + return ret; +} + +/* * find a contiguous range of bytes in the file marked as delalloc, not * more than 'max_bytes'. start and end are used to return the range, * @@ -1339,6 +1523,7 @@ again: * shortening the size of the delalloc range we're searching */ free_extent_state(cached_state); + cached_state = NULL; if (!loops) { unsigned long offset = (*start) & (PAGE_CACHE_SIZE - 1); max_bytes = PAGE_CACHE_SIZE - offset; @@ -1564,7 +1749,8 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end, int bitset = 0; spin_lock(&tree->lock); - if (cached && cached->tree && cached->start == start) + if (cached && cached->tree && cached->start <= start && + cached->end > start) node = &cached->rb_node; else node = tree_search(tree, start); @@ -1644,6 +1830,368 @@ static int check_page_writeback(struct extent_io_tree *tree, return 0; } +/* + * When IO fails, either with EIO or csum verification fails, we + * try other mirrors that might have a good copy of the data. This + * io_failure_record is used to record state as we go through all the + * mirrors. If another mirror has good data, the page is set up to date + * and things continue. If a good mirror can't be found, the original + * bio end_io callback is called to indicate things have failed. + */ +struct io_failure_record { + struct page *page; + u64 start; + u64 len; + u64 logical; + unsigned long bio_flags; + int this_mirror; + int failed_mirror; + int in_validation; +}; + +static int free_io_failure(struct inode *inode, struct io_failure_record *rec, + int did_repair) +{ + int ret; + int err = 0; + struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree; + + set_state_private(failure_tree, rec->start, 0); + ret = clear_extent_bits(failure_tree, rec->start, + rec->start + rec->len - 1, + EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); + if (ret) + err = ret; + + if (did_repair) { + ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start, + rec->start + rec->len - 1, + EXTENT_DAMAGED, GFP_NOFS); + if (ret && !err) + err = ret; + } + + kfree(rec); + return err; +} + +static void repair_io_failure_callback(struct bio *bio, int err) +{ + complete(bio->bi_private); +} + +/* + * this bypasses the standard btrfs submit functions deliberately, as + * the standard behavior is to write all copies in a raid setup. here we only + * want to write the one bad copy. so we do the mapping for ourselves and issue + * submit_bio directly. + * to avoid any synchonization issues, wait for the data after writing, which + * actually prevents the read that triggered the error from finishing. + * currently, there can be no more than two copies of every data bit. thus, + * exactly one rewrite is required. + */ +int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start, + u64 length, u64 logical, struct page *page, + int mirror_num) +{ + struct bio *bio; + struct btrfs_device *dev; + DECLARE_COMPLETION_ONSTACK(compl); + u64 map_length = 0; + u64 sector; + struct btrfs_bio *bbio = NULL; + int ret; + + BUG_ON(!mirror_num); + + bio = bio_alloc(GFP_NOFS, 1); + if (!bio) + return -EIO; + bio->bi_private = &compl; + bio->bi_end_io = repair_io_failure_callback; + bio->bi_size = 0; + map_length = length; + + ret = btrfs_map_block(map_tree, WRITE, logical, + &map_length, &bbio, mirror_num); + if (ret) { + bio_put(bio); + return -EIO; + } + BUG_ON(mirror_num != bbio->mirror_num); + sector = bbio->stripes[mirror_num-1].physical >> 9; + bio->bi_sector = sector; + dev = bbio->stripes[mirror_num-1].dev; + kfree(bbio); + if (!dev || !dev->bdev || !dev->writeable) { + bio_put(bio); + return -EIO; + } + bio->bi_bdev = dev->bdev; + bio_add_page(bio, page, length, start-page_offset(page)); + submit_bio(WRITE_SYNC, bio); + wait_for_completion(&compl); + + if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) { + /* try to remap that extent elsewhere? */ + bio_put(bio); + return -EIO; + } + + printk(KERN_INFO "btrfs read error corrected: ino %lu off %llu (dev %s " + "sector %llu)\n", page->mapping->host->i_ino, start, + dev->name, sector); + + bio_put(bio); + return 0; +} + +/* + * each time an IO finishes, we do a fast check in the IO failure tree + * to see if we need to process or clean up an io_failure_record + */ +static int clean_io_failure(u64 start, struct page *page) +{ + u64 private; + u64 private_failure; + struct io_failure_record *failrec; + struct btrfs_mapping_tree *map_tree; + struct extent_state *state; + int num_copies; + int did_repair = 0; + int ret; + struct inode *inode = page->mapping->host; + + private = 0; + ret = count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private, + (u64)-1, 1, EXTENT_DIRTY, 0); + if (!ret) + return 0; + + ret = get_state_private(&BTRFS_I(inode)->io_failure_tree, start, + &private_failure); + if (ret) + return 0; + + failrec = (struct io_failure_record *)(unsigned long) private_failure; + BUG_ON(!failrec->this_mirror); + + if (failrec->in_validation) { + /* there was no real error, just free the record */ + pr_debug("clean_io_failure: freeing dummy error at %llu\n", + failrec->start); + did_repair = 1; + goto out; + } + + spin_lock(&BTRFS_I(inode)->io_tree.lock); + state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree, + failrec->start, + EXTENT_LOCKED); + spin_unlock(&BTRFS_I(inode)->io_tree.lock); + + if (state && state->start == failrec->start) { + map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree; + num_copies = btrfs_num_copies(map_tree, failrec->logical, + failrec->len); + if (num_copies > 1) { + ret = repair_io_failure(map_tree, start, failrec->len, + failrec->logical, page, + failrec->failed_mirror); + did_repair = !ret; + } + } + +out: + if (!ret) + ret = free_io_failure(inode, failrec, did_repair); + + return ret; +} + +/* + * this is a generic handler for readpage errors (default + * readpage_io_failed_hook). if other copies exist, read those and write back + * good data to the failed position. does not investigate in remapping the + * failed extent elsewhere, hoping the device will be smart enough to do this as + * needed + */ + +static int bio_readpage_error(struct bio *failed_bio, struct page *page, + u64 start, u64 end, int failed_mirror, + struct extent_state *state) +{ + struct io_failure_record *failrec = NULL; + u64 private; + struct extent_map *em; + struct inode *inode = page->mapping->host; + struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree; + struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree; + struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; + struct bio *bio; + int num_copies; + int ret; + int read_mode; + u64 logical; + + BUG_ON(failed_bio->bi_rw & REQ_WRITE); + + ret = get_state_private(failure_tree, start, &private); + if (ret) { + failrec = kzalloc(sizeof(*failrec), GFP_NOFS); + if (!failrec) + return -ENOMEM; + failrec->start = start; + failrec->len = end - start + 1; + failrec->this_mirror = 0; + failrec->bio_flags = 0; + failrec->in_validation = 0; + + read_lock(&em_tree->lock); + em = lookup_extent_mapping(em_tree, start, failrec->len); + if (!em) { + read_unlock(&em_tree->lock); + kfree(failrec); + return -EIO; + } + + if (em->start > start || em->start + em->len < start) { + free_extent_map(em); + em = NULL; + } + read_unlock(&em_tree->lock); + + if (!em || IS_ERR(em)) { + kfree(failrec); + return -EIO; + } + logical = start - em->start; + logical = em->block_start + logical; + if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) { + logical = em->block_start; + failrec->bio_flags = EXTENT_BIO_COMPRESSED; + extent_set_compress_type(&failrec->bio_flags, + em->compress_type); + } + pr_debug("bio_readpage_error: (new) logical=%llu, start=%llu, " + "len=%llu\n", logical, start, failrec->len); + failrec->logical = logical; + free_extent_map(em); + + /* set the bits in the private failure tree */ + ret = set_extent_bits(failure_tree, start, end, + EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); + if (ret >= 0) + ret = set_state_private(failure_tree, start, + (u64)(unsigned long)failrec); + /* set the bits in the inode's tree */ + if (ret >= 0) + ret = set_extent_bits(tree, start, end, EXTENT_DAMAGED, + GFP_NOFS); + if (ret < 0) { + kfree(failrec); + return ret; + } + } else { + failrec = (struct io_failure_record *)(unsigned long)private; + pr_debug("bio_readpage_error: (found) logical=%llu, " + "start=%llu, len=%llu, validation=%d\n", + failrec->logical, failrec->start, failrec->len, + failrec->in_validation); + /* + * when data can be on disk more than twice, add to failrec here + * (e.g. with a list for failed_mirror) to make + * clean_io_failure() clean all those errors at once. + */ + } + num_copies = btrfs_num_copies( + &BTRFS_I(inode)->root->fs_info->mapping_tree, + failrec->logical, failrec->len); + if (num_copies == 1) { + /* + * we only have a single copy of the data, so don't bother with + * all the retry and error correction code that follows. no + * matter what the error is, it is very likely to persist. + */ + pr_debug("bio_readpage_error: cannot repair, num_copies == 1. " + "state=%p, num_copies=%d, next_mirror %d, " + "failed_mirror %d\n", state, num_copies, + failrec->this_mirror, failed_mirror); + free_io_failure(inode, failrec, 0); + return -EIO; + } + + if (!state) { + spin_lock(&tree->lock); + state = find_first_extent_bit_state(tree, failrec->start, + EXTENT_LOCKED); + if (state && state->start != failrec->start) + state = NULL; + spin_unlock(&tree->lock); + } + + /* + * there are two premises: + * a) deliver good data to the caller + * b) correct the bad sectors on disk + */ + if (failed_bio->bi_vcnt > 1) { + /* + * to fulfill b), we need to know the exact failing sectors, as + * we don't want to rewrite any more than the failed ones. thus, + * we need separate read requests for the failed bio + * + * if the following BUG_ON triggers, our validation request got + * merged. we need separate requests for our algorithm to work. + */ + BUG_ON(failrec->in_validation); + failrec->in_validation = 1; + failrec->this_mirror = failed_mirror; + read_mode = READ_SYNC | REQ_FAILFAST_DEV; + } else { + /* + * we're ready to fulfill a) and b) alongside. get a good copy + * of the failed sector and if we succeed, we have setup + * everything for repair_io_failure to do the rest for us. + */ + if (failrec->in_validation) { + BUG_ON(failrec->this_mirror != failed_mirror); + failrec->in_validation = 0; + failrec->this_mirror = 0; + } + failrec->failed_mirror = failed_mirror; + failrec->this_mirror++; + if (failrec->this_mirror == failed_mirror) + failrec->this_mirror++; + read_mode = READ_SYNC; + } + + if (!state || failrec->this_mirror > num_copies) { + pr_debug("bio_readpage_error: (fail) state=%p, num_copies=%d, " + "next_mirror %d, failed_mirror %d\n", state, + num_copies, failrec->this_mirror, failed_mirror); + free_io_failure(inode, failrec, 0); + return -EIO; + } + + bio = bio_alloc(GFP_NOFS, 1); + bio->bi_private = state; + bio->bi_end_io = failed_bio->bi_end_io; + bio->bi_sector = failrec->logical >> 9; + bio->bi_bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev; + bio->bi_size = 0; + + bio_add_page(bio, page, failrec->len, start - page_offset(page)); + + pr_debug("bio_readpage_error: submitting new read[%#x] to " + "this_mirror=%d, num_copies=%d, in_validation=%d\n", read_mode, + failrec->this_mirror, num_copies, failrec->in_validation); + + tree->ops->submit_bio_hook(inode, read_mode, bio, failrec->this_mirror, + failrec->bio_flags, 0); + return 0; +} + /* lots and lots of room for performance fixes in the end_bio funcs */ /* @@ -1742,6 +2290,9 @@ static void end_bio_extent_readpage(struct bio *bio, int err) struct extent_state *cached = NULL; struct extent_state *state; + pr_debug("end_bio_extent_readpage: bi_vcnt=%d, idx=%d, err=%d, " + "mirror=%ld\n", bio->bi_vcnt, bio->bi_idx, err, + (long int)bio->bi_bdev); tree = &BTRFS_I(page->mapping->host)->io_tree; start = ((u64)page->index << PAGE_CACHE_SHIFT) + @@ -1772,12 +2323,26 @@ static void end_bio_extent_readpage(struct bio *bio, int err) state); if (ret) uptodate = 0; + else + clean_io_failure(start, page); } - if (!uptodate && tree->ops && - tree->ops->readpage_io_failed_hook) { - ret = tree->ops->readpage_io_failed_hook(bio, page, - start, end, NULL); + if (!uptodate) { + int failed_mirror; + failed_mirror = (int)(unsigned long)bio->bi_bdev; + /* + * The generic bio_readpage_error handles errors the + * following way: If possible, new read requests are + * created and submitted and will end up in + * end_bio_extent_readpage as well (if we're lucky, not + * in the !uptodate case). In that case it returns 0 and + * we just go on with the next page in our bio. If it + * can't handle the error it will return -EIO and we + * remain responsible for that page. + */ + ret = bio_readpage_error(bio, page, start, end, + failed_mirror, NULL); if (ret == 0) { +error_handled: uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); if (err) @@ -1785,6 +2350,13 @@ static void end_bio_extent_readpage(struct bio *bio, int err) uncache_state(&cached); continue; } + if (tree->ops && tree->ops->readpage_io_failed_hook) { + ret = tree->ops->readpage_io_failed_hook( + bio, page, start, end, + failed_mirror, state); + if (ret == 0) + goto error_handled; + } } if (uptodate) { @@ -1856,6 +2428,7 @@ static int submit_one_bio(int rw, struct bio *bio, int mirror_num, mirror_num, bio_flags, start); else submit_bio(rw, bio); + if (bio_flagged(bio, BIO_EOPNOTSUPP)) ret = -EOPNOTSUPP; bio_put(bio); @@ -1871,7 +2444,8 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, bio_end_io_t end_io_func, int mirror_num, unsigned long prev_bio_flags, - unsigned long bio_flags) + unsigned long bio_flags, + bool force_bio_submit) { int ret = 0; struct bio *bio; @@ -1890,6 +2464,7 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, sector; if (prev_bio_flags != bio_flags || !contig || + force_bio_submit || (tree->ops && tree->ops->merge_bio_hook && tree->ops->merge_bio_hook(page, offset, page_size, bio, bio_flags)) || @@ -1946,7 +2521,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree, struct page *page, get_extent_t *get_extent, struct bio **bio, int mirror_num, - unsigned long *bio_flags) + unsigned long *bio_flags, + u64 *prev_em_start) { struct inode *inode = page->mapping->host; u64 start = (u64)page->index << PAGE_CACHE_SHIFT; @@ -2002,6 +2578,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree, } } while (cur <= end) { + bool force_bio_submit = false; + if (cur >= last_byte) { char *userpage; struct extent_state *cached = NULL; @@ -2048,6 +2626,49 @@ static int __extent_read_full_page(struct extent_io_tree *tree, block_start = em->block_start; if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) block_start = EXTENT_MAP_HOLE; + + /* + * If we have a file range that points to a compressed extent + * and it's followed by a consecutive file range that points to + * to the same compressed extent (possibly with a different + * offset and/or length, so it either points to the whole extent + * or only part of it), we must make sure we do not submit a + * single bio to populate the pages for the 2 ranges because + * this makes the compressed extent read zero out the pages + * belonging to the 2nd range. Imagine the following scenario: + * + * File layout + * [0 - 8K] [8K - 24K] + * | | + * | | + * points to extent X, points to extent X, + * offset 4K, length of 8K offset 0, length 16K + * + * [extent X, compressed length = 4K uncompressed length = 16K] + * + * If the bio to read the compressed extent covers both ranges, + * it will decompress extent X into the pages belonging to the + * first range and then it will stop, zeroing out the remaining + * pages that belong to the other range that points to extent X. + * So here we make sure we submit 2 bios, one for the first + * range and another one for the third range. Both will target + * the same physical extent from disk, but we can't currently + * make the compressed bio endio callback populate the pages + * for both ranges because each compressed bio is tightly + * coupled with a single extent map, and each range can have + * an extent map with a different offset value relative to the + * uncompressed data of our extent and different lengths. This + * is a corner case so we prioritize correctness over + * non-optimal behavior (submitting 2 bios for the same extent). + */ + if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) && + prev_em_start && *prev_em_start != (u64)-1 && + *prev_em_start != em->orig_start) + force_bio_submit = true; + + if (prev_em_start) + *prev_em_start = em->orig_start; + free_extent_map(em); em = NULL; @@ -2102,7 +2723,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree, bdev, bio, pnr, end_bio_extent_readpage, mirror_num, *bio_flags, - this_bio_flag); + this_bio_flag, + force_bio_submit); nr++; *bio_flags = this_bio_flag; } @@ -2121,16 +2743,16 @@ out: } int extent_read_full_page(struct extent_io_tree *tree, struct page *page, - get_extent_t *get_extent) + get_extent_t *get_extent, int mirror_num) { struct bio *bio = NULL; unsigned long bio_flags = 0; int ret; - ret = __extent_read_full_page(tree, page, get_extent, &bio, 0, - &bio_flags); + ret = __extent_read_full_page(tree, page, get_extent, &bio, mirror_num, + &bio_flags, NULL); if (bio) - ret = submit_one_bio(READ, bio, 0, bio_flags); + ret = submit_one_bio(READ, bio, mirror_num, bio_flags); return ret; } @@ -2181,6 +2803,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, int compressed; int write_flags; unsigned long nr_written = 0; + bool fill_delalloc = true; if (wbc->sync_mode == WB_SYNC_ALL) write_flags = WRITE_SYNC; @@ -2190,6 +2813,9 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, trace___extent_writepage(page, inode, wbc); WARN_ON(!PageLocked(page)); + + ClearPageError(page); + pg_offset = i_size & (PAGE_CACHE_SIZE - 1); if (page->index > end_index || (page->index == end_index && !pg_offset)) { @@ -2211,10 +2837,13 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, set_page_extent_mapped(page); + if (!tree->ops || !tree->ops->fill_delalloc) + fill_delalloc = false; + delalloc_start = start; delalloc_end = 0; page_started = 0; - if (!epd->extent_locked) { + if (!epd->extent_locked && fill_delalloc) { u64 delalloc_to_write = 0; /* * make sure the wbc mapping index is at least updated @@ -2380,7 +3009,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, sector, iosize, pg_offset, bdev, &epd->bio, max_nr, end_bio_extent_writepage, - 0, 0, 0); + 0, 0, 0, false); if (ret) SetPageError(page); } @@ -2432,6 +3061,7 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, pgoff_t index; pgoff_t end; /* Inclusive */ int scanned = 0; + int tag; pagevec_init(&pvec, 0); if (wbc->range_cyclic) { @@ -2442,11 +3072,16 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, end = wbc->range_end >> PAGE_CACHE_SHIFT; scanned = 1; } + if (wbc->sync_mode == WB_SYNC_ALL) + tag = PAGECACHE_TAG_TOWRITE; + else + tag = PAGECACHE_TAG_DIRTY; retry: + if (wbc->sync_mode == WB_SYNC_ALL) + tag_pages_for_writeback(mapping, index, end); while (!done && !nr_to_write_done && (index <= end) && - (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, - PAGECACHE_TAG_DIRTY, min(end - index, - (pgoff_t)PAGEVEC_SIZE-1) + 1))) { + (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag, + min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) { unsigned i; scanned = 1; @@ -2460,10 +3095,16 @@ retry: * swizzled back from swapper_space to tmpfs file * mapping */ - if (tree->ops && tree->ops->write_cache_pages_lock_hook) - tree->ops->write_cache_pages_lock_hook(page); - else - lock_page(page); + if (tree->ops && + tree->ops->write_cache_pages_lock_hook) { + tree->ops->write_cache_pages_lock_hook(page, + data, flush_fn); + } else { + if (!trylock_page(page)) { + flush_fn(data); + lock_page(page); + } + } if (unlikely(page->mapping != mapping)) { unlock_page(page); @@ -2541,7 +3182,6 @@ int extent_write_full_page(struct extent_io_tree *tree, struct page *page, struct writeback_control *wbc) { int ret; - struct address_space *mapping = page->mapping; struct extent_page_data epd = { .bio = NULL, .tree = tree, @@ -2549,18 +3189,9 @@ int extent_write_full_page(struct extent_io_tree *tree, struct page *page, .extent_locked = 0, .sync_io = wbc->sync_mode == WB_SYNC_ALL, }; - struct writeback_control wbc_writepages = { - .sync_mode = wbc->sync_mode, - .older_than_this = NULL, - .nr_to_write = 64, - .range_start = page_offset(page) + PAGE_CACHE_SIZE, - .range_end = (loff_t)-1, - }; ret = __extent_writepage(page, wbc, &epd); - extent_write_cache_pages(tree, mapping, &wbc_writepages, - __extent_writepage, &epd, flush_write_bio); flush_epd_write_bio(&epd); return ret; } @@ -2584,7 +3215,6 @@ int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode, }; struct writeback_control wbc_writepages = { .sync_mode = mode, - .older_than_this = NULL, .nr_to_write = nr_pages * 2, .range_start = start, .range_end = end + 1, @@ -2638,6 +3268,7 @@ int extent_readpages(struct extent_io_tree *tree, struct bio *bio = NULL; unsigned page_idx; unsigned long bio_flags = 0; + u64 prev_em_start = (u64)-1; for (page_idx = 0; page_idx < nr_pages; page_idx++) { struct page *page = list_entry(pages->prev, struct page, lru); @@ -2647,7 +3278,8 @@ int extent_readpages(struct extent_io_tree *tree, if (!add_to_page_cache_lru(page, mapping, page->index, GFP_NOFS)) { __extent_read_full_page(tree, page, get_extent, - &bio, 0, &bio_flags); + &bio, 0, &bio_flags, + &prev_em_start); } page_cache_release(page); } @@ -2840,6 +3472,9 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, return -ENOMEM; path->leave_spinning = 1; + start = ALIGN(start, BTRFS_I(inode)->root->sectorsize); + len = ALIGN(len, BTRFS_I(inode)->root->sectorsize); + /* * lookup the last file extent. We're not using i_size here * because there might be preallocation past i_size @@ -2887,7 +3522,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0, &cached_state, GFP_NOFS); - em = get_extent_skip_holes(inode, off, last_for_get_extent, + em = get_extent_skip_holes(inode, start, last_for_get_extent, get_extent); if (!em) goto out; @@ -2976,7 +3611,7 @@ out: return ret; } -static inline struct page *extent_buffer_page(struct extent_buffer *eb, +inline struct page *extent_buffer_page(struct extent_buffer *eb, unsigned long i) { struct page *p; @@ -3001,7 +3636,7 @@ static inline struct page *extent_buffer_page(struct extent_buffer *eb, return p; } -static inline unsigned long num_extent_pages(u64 start, u64 len) +inline unsigned long num_extent_pages(u64 start, u64 len) { return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) - (start >> PAGE_CACHE_SHIFT); @@ -3022,8 +3657,15 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree, return NULL; eb->start = start; eb->len = len; - spin_lock_init(&eb->lock); - init_waitqueue_head(&eb->lock_wq); + rwlock_init(&eb->lock); + atomic_set(&eb->write_locks, 0); + atomic_set(&eb->read_locks, 0); + atomic_set(&eb->blocking_readers, 0); + atomic_set(&eb->blocking_writers, 0); + atomic_set(&eb->spinning_readers, 0); + atomic_set(&eb->spinning_writers, 0); + init_waitqueue_head(&eb->write_lock_wq); + init_waitqueue_head(&eb->read_lock_wq); #if LEAK_DEBUG spin_lock_irqsave(&leak_lock, flags); @@ -3119,7 +3761,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, i = 0; } for (; i < num_pages; i++, index++) { - p = find_or_create_page(mapping, index, GFP_NOFS | __GFP_HIGHMEM); + p = find_or_create_page(mapping, index, GFP_NOFS); if (!p) { WARN_ON(1); goto free_eb; @@ -3247,6 +3889,7 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree, PAGECACHE_TAG_DIRTY); } spin_unlock_irq(&page->mapping->tree_lock); + ClearPageError(page); unlock_page(page); } return 0; @@ -3266,6 +3909,22 @@ int set_extent_buffer_dirty(struct extent_io_tree *tree, return was_dirty; } +static int __eb_straddles_pages(u64 start, u64 len) +{ + if (len < PAGE_CACHE_SIZE) + return 1; + if (start & (PAGE_CACHE_SIZE - 1)) + return 1; + if ((start + len) & (PAGE_CACHE_SIZE - 1)) + return 1; + return 0; +} + +static int eb_straddles_pages(struct extent_buffer *eb) +{ + return __eb_straddles_pages(eb->start, eb->len); +} + int clear_extent_buffer_uptodate(struct extent_io_tree *tree, struct extent_buffer *eb, struct extent_state **cached_state) @@ -3277,8 +3936,10 @@ int clear_extent_buffer_uptodate(struct extent_io_tree *tree, num_pages = num_extent_pages(eb->start, eb->len); clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); - clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1, - cached_state, GFP_NOFS); + if (eb_straddles_pages(eb)) { + clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1, + cached_state, GFP_NOFS); + } for (i = 0; i < num_pages; i++) { page = extent_buffer_page(eb, i); if (page) @@ -3296,8 +3957,10 @@ int set_extent_buffer_uptodate(struct extent_io_tree *tree, num_pages = num_extent_pages(eb->start, eb->len); - set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1, - NULL, GFP_NOFS); + if (eb_straddles_pages(eb)) { + set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1, + NULL, GFP_NOFS); + } for (i = 0; i < num_pages; i++) { page = extent_buffer_page(eb, i); if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) || @@ -3320,9 +3983,12 @@ int extent_range_uptodate(struct extent_io_tree *tree, int uptodate; unsigned long index; - ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL); - if (ret) - return 1; + if (__eb_straddles_pages(start, end - start + 1)) { + ret = test_range_bit(tree, start, end, + EXTENT_UPTODATE, 1, NULL); + if (ret) + return 1; + } while (start <= end) { index = start >> PAGE_CACHE_SHIFT; page = find_get_page(tree->mapping, index); @@ -3350,10 +4016,12 @@ int extent_buffer_uptodate(struct extent_io_tree *tree, if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags)) return 1; - ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1, - EXTENT_UPTODATE, 1, cached_state); - if (ret) - return ret; + if (eb_straddles_pages(eb)) { + ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1, + EXTENT_UPTODATE, 1, cached_state); + if (ret) + return ret; + } num_pages = num_extent_pages(eb->start, eb->len); for (i = 0; i < num_pages; i++) { @@ -3367,8 +4035,7 @@ int extent_buffer_uptodate(struct extent_io_tree *tree, } int read_extent_buffer_pages(struct extent_io_tree *tree, - struct extent_buffer *eb, - u64 start, int wait, + struct extent_buffer *eb, u64 start, int wait, get_extent_t *get_extent, int mirror_num) { unsigned long i; @@ -3382,13 +4049,16 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, unsigned long num_pages; struct bio *bio = NULL; unsigned long bio_flags = 0; + u64 prev_em_start = (u64)-1; if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags)) return 0; - if (test_range_bit(tree, eb->start, eb->start + eb->len - 1, - EXTENT_UPTODATE, 1, NULL)) { - return 0; + if (eb_straddles_pages(eb)) { + if (test_range_bit(tree, eb->start, eb->start + eb->len - 1, + EXTENT_UPTODATE, 1, NULL)) { + return 0; + } } if (start) { @@ -3402,7 +4072,7 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, num_pages = num_extent_pages(eb->start, eb->len); for (i = start_i; i < num_pages; i++) { page = extent_buffer_page(eb, i); - if (!wait) { + if (wait == WAIT_NONE) { if (!trylock_page(page)) goto unlock_exit; } else { @@ -3435,7 +4105,8 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, ClearPageError(page); err = __extent_read_full_page(tree, page, get_extent, &bio, - mirror_num, &bio_flags); + mirror_num, &bio_flags, + &prev_em_start); if (err) ret = err; } else { @@ -3446,7 +4117,7 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, if (bio) submit_one_bio(READ, bio, mirror_num, bio_flags); - if (ret || !wait) + if (ret || wait != WAIT_COMPLETE) return ret; for (i = start_i; i < num_pages; i++) { @@ -3492,9 +4163,8 @@ void read_extent_buffer(struct extent_buffer *eb, void *dstv, page = extent_buffer_page(eb, i); cur = min(len, (PAGE_CACHE_SIZE - offset)); - kaddr = kmap_atomic(page, KM_USER1); + kaddr = page_address(page); memcpy(dst, kaddr + offset, cur); - kunmap_atomic(kaddr, KM_USER1); dst += cur; len -= cur; @@ -3504,9 +4174,9 @@ void read_extent_buffer(struct extent_buffer *eb, void *dstv, } int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, - unsigned long min_len, char **token, char **map, + unsigned long min_len, char **map, unsigned long *map_start, - unsigned long *map_len, int km) + unsigned long *map_len) { size_t offset = start & (PAGE_CACHE_SIZE - 1); char *kaddr; @@ -3536,42 +4206,12 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, } p = extent_buffer_page(eb, i); - kaddr = kmap_atomic(p, km); - *token = kaddr; + kaddr = page_address(p); *map = kaddr + offset; *map_len = PAGE_CACHE_SIZE - offset; return 0; } -int map_extent_buffer(struct extent_buffer *eb, unsigned long start, - unsigned long min_len, - char **token, char **map, - unsigned long *map_start, - unsigned long *map_len, int km) -{ - int err; - int save = 0; - if (eb->map_token) { - unmap_extent_buffer(eb, eb->map_token, km); - eb->map_token = NULL; - save = 1; - } - err = map_private_extent_buffer(eb, start, min_len, token, map, - map_start, map_len, km); - if (!err && save) { - eb->map_token = *token; - eb->kaddr = *map; - eb->map_start = *map_start; - eb->map_len = *map_len; - } - return err; -} - -void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km) -{ - kunmap_atomic(token, km); -} - int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv, unsigned long start, unsigned long len) @@ -3595,9 +4235,8 @@ int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv, cur = min(len, (PAGE_CACHE_SIZE - offset)); - kaddr = kmap_atomic(page, KM_USER0); + kaddr = page_address(page); ret = memcmp(ptr, kaddr + offset, cur); - kunmap_atomic(kaddr, KM_USER0); if (ret) break; @@ -3630,9 +4269,8 @@ void write_extent_buffer(struct extent_buffer *eb, const void *srcv, WARN_ON(!PageUptodate(page)); cur = min(len, PAGE_CACHE_SIZE - offset); - kaddr = kmap_atomic(page, KM_USER1); + kaddr = page_address(page); memcpy(kaddr + offset, src, cur); - kunmap_atomic(kaddr, KM_USER1); src += cur; len -= cur; @@ -3661,9 +4299,8 @@ void memset_extent_buffer(struct extent_buffer *eb, char c, WARN_ON(!PageUptodate(page)); cur = min(len, PAGE_CACHE_SIZE - offset); - kaddr = kmap_atomic(page, KM_USER0); + kaddr = page_address(page); memset(kaddr + offset, c, cur); - kunmap_atomic(kaddr, KM_USER0); len -= cur; offset = 0; @@ -3694,9 +4331,8 @@ void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src, cur = min(len, (unsigned long)(PAGE_CACHE_SIZE - offset)); - kaddr = kmap_atomic(page, KM_USER0); + kaddr = page_address(page); read_extent_buffer(src, kaddr + offset, src_offset, cur); - kunmap_atomic(kaddr, KM_USER0); src_offset += cur; len -= cur; @@ -3709,20 +4345,17 @@ static void move_pages(struct page *dst_page, struct page *src_page, unsigned long dst_off, unsigned long src_off, unsigned long len) { - char *dst_kaddr = kmap_atomic(dst_page, KM_USER0); + char *dst_kaddr = page_address(dst_page); if (dst_page == src_page) { memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len); } else { - char *src_kaddr = kmap_atomic(src_page, KM_USER1); + char *src_kaddr = page_address(src_page); char *p = dst_kaddr + dst_off + len; char *s = src_kaddr + src_off + len; while (len--) *--p = *--s; - - kunmap_atomic(src_kaddr, KM_USER1); } - kunmap_atomic(dst_kaddr, KM_USER0); } static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len) @@ -3735,20 +4368,17 @@ static void copy_pages(struct page *dst_page, struct page *src_page, unsigned long dst_off, unsigned long src_off, unsigned long len) { - char *dst_kaddr = kmap_atomic(dst_page, KM_USER0); + char *dst_kaddr = page_address(dst_page); char *src_kaddr; if (dst_page != src_page) { - src_kaddr = kmap_atomic(src_page, KM_USER1); + src_kaddr = page_address(src_page); } else { src_kaddr = dst_kaddr; BUG_ON(areas_overlap(src_off, dst_off, len)); } memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len); - kunmap_atomic(dst_kaddr, KM_USER0); - if (dst_page != src_page) - kunmap_atomic(src_kaddr, KM_USER1); } void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index a11a92e..2e32510 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -17,6 +17,8 @@ #define EXTENT_NODATASUM (1 << 10) #define EXTENT_DO_ACCOUNTING (1 << 11) #define EXTENT_FIRST_DELALLOC (1 << 12) +#define EXTENT_NEED_WAIT (1 << 13) +#define EXTENT_DAMAGED (1 << 14) #define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK) #define EXTENT_CTLBITS (EXTENT_DO_ACCOUNTING | EXTENT_FIRST_DELALLOC) @@ -32,6 +34,7 @@ #define EXTENT_BUFFER_BLOCKING 1 #define EXTENT_BUFFER_DIRTY 2 #define EXTENT_BUFFER_CORRUPT 3 +#define EXTENT_BUFFER_READAHEAD 4 /* this got triggered by readahead */ /* these are flags for extent_clear_unlock_delalloc */ #define EXTENT_CLEAR_UNLOCK_PAGE 0x1 @@ -67,7 +70,7 @@ struct extent_io_ops { unsigned long bio_flags); int (*readpage_io_hook)(struct page *page, u64 start, u64 end); int (*readpage_io_failed_hook)(struct bio *bio, struct page *page, - u64 start, u64 end, + u64 start, u64 end, int failed_mirror, struct extent_state *state); int (*writepage_io_failed_hook)(struct bio *bio, struct page *page, u64 start, u64 end, @@ -76,16 +79,17 @@ struct extent_io_ops { struct extent_state *state); int (*writepage_end_io_hook)(struct page *page, u64 start, u64 end, struct extent_state *state, int uptodate); - int (*set_bit_hook)(struct inode *inode, struct extent_state *state, - int *bits); - int (*clear_bit_hook)(struct inode *inode, struct extent_state *state, - int *bits); - int (*merge_extent_hook)(struct inode *inode, - struct extent_state *new, - struct extent_state *other); - int (*split_extent_hook)(struct inode *inode, - struct extent_state *orig, u64 split); - int (*write_cache_pages_lock_hook)(struct page *page); + void (*set_bit_hook)(struct inode *inode, struct extent_state *state, + int *bits); + void (*clear_bit_hook)(struct inode *inode, struct extent_state *state, + int *bits); + void (*merge_extent_hook)(struct inode *inode, + struct extent_state *new, + struct extent_state *other); + void (*split_extent_hook)(struct inode *inode, + struct extent_state *orig, u64 split); + int (*write_cache_pages_lock_hook)(struct page *page, void *data, + void (*flush_fn)(void *)); }; struct extent_io_tree { @@ -108,8 +112,6 @@ struct extent_state { wait_queue_head_t wq; atomic_t refs; unsigned long state; - u64 split_start; - u64 split_end; /* for use by the FS */ u64 private; @@ -120,8 +122,6 @@ struct extent_state { struct extent_buffer { u64 start; unsigned long len; - char *map_token; - char *kaddr; unsigned long map_start; unsigned long map_len; struct page *first_page; @@ -130,14 +130,26 @@ struct extent_buffer { struct rcu_head rcu_head; atomic_t refs; - /* the spinlock is used to protect most operations */ - spinlock_t lock; + /* count of read lock holders on the extent buffer */ + atomic_t write_locks; + atomic_t read_locks; + atomic_t blocking_writers; + atomic_t blocking_readers; + atomic_t spinning_readers; + atomic_t spinning_writers; + + /* protects write locks */ + rwlock_t lock; + + /* readers use lock_wq while they wait for the write + * lock holders to unlock + */ + wait_queue_head_t write_lock_wq; - /* - * when we keep the lock held while blocking, waiters go onto - * the wq + /* writers use read_lock_wq while they wait for readers + * to unlock */ - wait_queue_head_t lock_wq; + wait_queue_head_t read_lock_wq; }; static inline void extent_set_compress_type(unsigned long *bio_flags, @@ -177,7 +189,7 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end, int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); int extent_read_full_page(struct extent_io_tree *tree, struct page *page, - get_extent_t *get_extent); + get_extent_t *get_extent, int mirror_num); int __init extent_io_init(void); void extent_io_exit(void); @@ -206,6 +218,8 @@ int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); +int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, + int bits, int clear_bits, gfp_t mask); int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask); int find_first_extent_bit(struct extent_io_tree *tree, u64 start, @@ -240,9 +254,14 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, u64 start, unsigned long len); void free_extent_buffer(struct extent_buffer *eb); +#define WAIT_NONE 0 +#define WAIT_COMPLETE 1 +#define WAIT_PAGE_LOCK 2 int read_extent_buffer_pages(struct extent_io_tree *tree, struct extent_buffer *eb, u64 start, int wait, get_extent_t *get_extent, int mirror_num); +unsigned long num_extent_pages(u64 start, u64 len); +struct page *extent_buffer_page(struct extent_buffer *eb, unsigned long i); static inline void extent_buffer_get(struct extent_buffer *eb) { @@ -279,17 +298,14 @@ int clear_extent_buffer_uptodate(struct extent_io_tree *tree, int extent_buffer_uptodate(struct extent_io_tree *tree, struct extent_buffer *eb, struct extent_state *cached_state); -int map_extent_buffer(struct extent_buffer *eb, unsigned long offset, - unsigned long min_len, char **token, char **map, - unsigned long *map_start, - unsigned long *map_len, int km); int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset, - unsigned long min_len, char **token, char **map, + unsigned long min_len, char **map, unsigned long *map_start, - unsigned long *map_len, int km); -void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km); + unsigned long *map_len); int extent_range_uptodate(struct extent_io_tree *tree, u64 start, u64 end); +int extent_range_clear_dirty_for_io(struct inode *inode, u64 start, u64 end); +int extent_range_redirty_for_io(struct inode *inode, u64 start, u64 end); int extent_clear_unlock_delalloc(struct inode *inode, struct extent_io_tree *tree, u64 start, u64 end, struct page *locked_page, @@ -297,4 +313,10 @@ int extent_clear_unlock_delalloc(struct inode *inode, struct bio * btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs, gfp_t gfp_flags); + +struct btrfs_mapping_tree; + +int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start, + u64 length, u64 logical, struct page *page, + int mirror_num); #endif diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c index 2d04103..7c97b33 100644 --- a/fs/btrfs/extent_map.c +++ b/fs/btrfs/extent_map.c @@ -183,22 +183,10 @@ static int mergable_maps(struct extent_map *prev, struct extent_map *next) return 0; } -int unpin_extent_cache(struct extent_map_tree *tree, u64 start, u64 len) +static void try_merge_map(struct extent_map_tree *tree, struct extent_map *em) { - int ret = 0; struct extent_map *merge = NULL; struct rb_node *rb; - struct extent_map *em; - - write_lock(&tree->lock); - em = lookup_extent_mapping(tree, start, len); - - WARN_ON(!em || em->start != start); - - if (!em) - goto out; - - clear_bit(EXTENT_FLAG_PINNED, &em->flags); if (em->start != 0) { rb = rb_prev(&em->rb_node); @@ -225,6 +213,24 @@ int unpin_extent_cache(struct extent_map_tree *tree, u64 start, u64 len) merge->in_tree = 0; free_extent_map(merge); } +} + +int unpin_extent_cache(struct extent_map_tree *tree, u64 start, u64 len) +{ + int ret = 0; + struct extent_map *em; + + write_lock(&tree->lock); + em = lookup_extent_mapping(tree, start, len); + + WARN_ON(!em || em->start != start); + + if (!em) + goto out; + + clear_bit(EXTENT_FLAG_PINNED, &em->flags); + + try_merge_map(tree, em); free_extent_map(em); out: @@ -247,7 +253,6 @@ int add_extent_mapping(struct extent_map_tree *tree, struct extent_map *em) { int ret = 0; - struct extent_map *merge = NULL; struct rb_node *rb; struct extent_map *exist; @@ -263,30 +268,8 @@ int add_extent_mapping(struct extent_map_tree *tree, goto out; } atomic_inc(&em->refs); - if (em->start != 0) { - rb = rb_prev(&em->rb_node); - if (rb) - merge = rb_entry(rb, struct extent_map, rb_node); - if (rb && mergable_maps(merge, em)) { - em->start = merge->start; - em->len += merge->len; - em->block_len += merge->block_len; - em->block_start = merge->block_start; - merge->in_tree = 0; - rb_erase(&merge->rb_node, &tree->map); - free_extent_map(merge); - } - } - rb = rb_next(&em->rb_node); - if (rb) - merge = rb_entry(rb, struct extent_map, rb_node); - if (rb && mergable_maps(em, merge)) { - em->len += merge->len; - em->block_len += merge->len; - rb_erase(&merge->rb_node, &tree->map); - merge->in_tree = 0; - free_extent_map(merge); - } + + try_merge_map(tree, em); out: return ret; } @@ -299,19 +282,8 @@ static u64 range_end(u64 start, u64 len) return start + len; } -/** - * lookup_extent_mapping - lookup extent_map - * @tree: tree to lookup in - * @start: byte offset to start the search - * @len: length of the lookup range - * - * Find and return the first extent_map struct in @tree that intersects the - * [start, len] range. There may be additional objects in the tree that - * intersect, so check the object returned carefully to make sure that no - * additional lookups are needed. - */ -struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree, - u64 start, u64 len) +struct extent_map *__lookup_extent_mapping(struct extent_map_tree *tree, + u64 start, u64 len, int strict) { struct extent_map *em; struct rb_node *rb_node; @@ -320,38 +292,42 @@ struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree, u64 end = range_end(start, len); rb_node = __tree_search(&tree->map, start, &prev, &next); - if (!rb_node && prev) { - em = rb_entry(prev, struct extent_map, rb_node); - if (end > em->start && start < extent_map_end(em)) - goto found; - } - if (!rb_node && next) { - em = rb_entry(next, struct extent_map, rb_node); - if (end > em->start && start < extent_map_end(em)) - goto found; - } if (!rb_node) { - em = NULL; - goto out; - } - if (IS_ERR(rb_node)) { - em = ERR_CAST(rb_node); - goto out; + if (prev) + rb_node = prev; + else if (next) + rb_node = next; + else + return NULL; } + em = rb_entry(rb_node, struct extent_map, rb_node); - if (end > em->start && start < extent_map_end(em)) - goto found; - em = NULL; - goto out; + if (strict && !(end > em->start && start < extent_map_end(em))) + return NULL; -found: atomic_inc(&em->refs); -out: return em; } /** + * lookup_extent_mapping - lookup extent_map + * @tree: tree to lookup in + * @start: byte offset to start the search + * @len: length of the lookup range + * + * Find and return the first extent_map struct in @tree that intersects the + * [start, len] range. There may be additional objects in the tree that + * intersect, so check the object returned carefully to make sure that no + * additional lookups are needed. + */ +struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree, + u64 start, u64 len) +{ + return __lookup_extent_mapping(tree, start, len, 1); +} + +/** * search_extent_mapping - find a nearby extent map * @tree: tree to lookup in * @start: byte offset to start the search @@ -365,38 +341,7 @@ out: struct extent_map *search_extent_mapping(struct extent_map_tree *tree, u64 start, u64 len) { - struct extent_map *em; - struct rb_node *rb_node; - struct rb_node *prev = NULL; - struct rb_node *next = NULL; - - rb_node = __tree_search(&tree->map, start, &prev, &next); - if (!rb_node && prev) { - em = rb_entry(prev, struct extent_map, rb_node); - goto found; - } - if (!rb_node && next) { - em = rb_entry(next, struct extent_map, rb_node); - goto found; - } - if (!rb_node) { - em = NULL; - goto out; - } - if (IS_ERR(rb_node)) { - em = ERR_CAST(rb_node); - goto out; - } - em = rb_entry(rb_node, struct extent_map, rb_node); - goto found; - - em = NULL; - goto out; - -found: - atomic_inc(&em->refs); -out: - return em; + return __lookup_extent_mapping(tree, start, len, 0); } /** diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index 90d4ee5..d7c9b68 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -91,8 +91,7 @@ struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans, struct btrfs_csum_item *item; struct extent_buffer *leaf; u64 csum_offset = 0; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); int csums_in_item; file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID; @@ -162,8 +161,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, u64 item_last_offset = 0; u64 disk_bytenr; u32 diff; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); int ret; struct btrfs_path *path; struct btrfs_csum_item *item = NULL; @@ -177,6 +175,17 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, WARN_ON(bio->bi_vcnt <= 0); + /* + * the free space stuff is only read when it hasn't been + * updated in the current transaction. So, we can safely + * read from the commit root and sidestep a nasty deadlock + * between reading the free space cache and updating the csum tree. + */ + if (btrfs_is_free_space_inode(root, inode)) { + path->search_commit_root = 1; + path->skip_locking = 1; + } + disk_bytenr = (u64)bio->bi_sector << 9; if (dio) offset = logical_offset; @@ -279,10 +288,11 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, int ret; size_t size; u64 csum_end; - u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; if (search_commit) { path->skip_locking = 1; @@ -480,8 +490,7 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, u64 bytenr, u64 len) { struct extent_buffer *leaf; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); u64 csum_end; u64 end_byte = bytenr + len; u32 blocksize_bits = root->fs_info->sb->s_blocksize_bits; @@ -537,8 +546,7 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, u64 csum_end; struct extent_buffer *leaf; int ret; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); int blocksize_bits = root->fs_info->sb->s_blocksize_bits; root = root->fs_info->csum_root; @@ -664,15 +672,12 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans, struct btrfs_sector_sum *sector_sum; u32 nritems; u32 ins_size; - char *eb_map; - char *eb_token; - unsigned long map_len; - unsigned long map_start; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; + sector_sum = sums->sums; again: next_offset = (u64)-1; @@ -712,7 +717,7 @@ again: found_next = 1; if (ret != 0) goto insert; - slot = 0; + slot = path->slots[0]; } btrfs_item_key_to_cpu(path->nodes[0], &found_key, slot); if (found_key.objectid != BTRFS_EXTENT_CSUM_OBJECTID || @@ -814,30 +819,9 @@ found: item_end = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item); item_end = (struct btrfs_csum_item *)((unsigned char *)item_end + btrfs_item_size_nr(leaf, path->slots[0])); - eb_token = NULL; next_sector: - if (!eb_token || - (unsigned long)item + csum_size >= map_start + map_len) { - int err; - - if (eb_token) - unmap_extent_buffer(leaf, eb_token, KM_USER1); - eb_token = NULL; - err = map_private_extent_buffer(leaf, (unsigned long)item, - csum_size, - &eb_token, &eb_map, - &map_start, &map_len, KM_USER1); - if (err) - eb_token = NULL; - } - if (eb_token) { - memcpy(eb_token + ((unsigned long)item & (PAGE_CACHE_SIZE - 1)), - §or_sum->sum, csum_size); - } else { - write_extent_buffer(leaf, §or_sum->sum, - (unsigned long)item, csum_size); - } + write_extent_buffer(leaf, §or_sum->sum, (unsigned long)item, csum_size); total_bytes += root->sectorsize; sector_sum++; @@ -850,10 +834,7 @@ next_sector: goto next_sector; } } - if (eb_token) { - unmap_extent_buffer(leaf, eb_token, KM_USER1); - eb_token = NULL; - } + btrfs_mark_buffer_dirty(path->nodes[0]); if (total_bytes < sums->len) { btrfs_release_path(path); diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index fa4ef18..97fbe93 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -74,7 +74,7 @@ struct inode_defrag { * If an existing record is found the defrag item you * pass in is freed */ -static int __btrfs_add_inode_defrag(struct inode *inode, +static void __btrfs_add_inode_defrag(struct inode *inode, struct inode_defrag *defrag) { struct btrfs_root *root = BTRFS_I(inode)->root; @@ -106,11 +106,11 @@ static int __btrfs_add_inode_defrag(struct inode *inode, BTRFS_I(inode)->in_defrag = 1; rb_link_node(&defrag->rb_node, parent, p); rb_insert_color(&defrag->rb_node, &root->fs_info->defrag_inodes); - return 0; + return; exists: kfree(defrag); - return 0; + return; } @@ -123,7 +123,6 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, { struct btrfs_root *root = BTRFS_I(inode)->root; struct inode_defrag *defrag; - int ret = 0; u64 transid; if (!btrfs_test_opt(root, AUTO_DEFRAG)) @@ -150,9 +149,11 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, spin_lock(&root->fs_info->defrag_inodes_lock); if (!BTRFS_I(inode)->in_defrag) - ret = __btrfs_add_inode_defrag(inode, defrag); + __btrfs_add_inode_defrag(inode, defrag); + else + kfree(defrag); spin_unlock(&root->fs_info->defrag_inodes_lock); - return ret; + return 0; } /* @@ -855,7 +856,8 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans, btrfs_drop_extent_cache(inode, start, end - 1, 0); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; again: recow = 0; split = start; @@ -1034,11 +1036,13 @@ out: * on error we return an unlocked page and the error value * on success we return a locked page and 0 */ -static int prepare_uptodate_page(struct page *page, u64 pos) +static int prepare_uptodate_page(struct page *page, u64 pos, + bool force_uptodate) { int ret = 0; - if ((pos & (PAGE_CACHE_SIZE - 1)) && !PageUptodate(page)) { + if (((pos & (PAGE_CACHE_SIZE - 1)) || force_uptodate) && + !PageUptodate(page)) { ret = btrfs_readpage(NULL, page); if (ret) return ret; @@ -1059,12 +1063,13 @@ static int prepare_uptodate_page(struct page *page, u64 pos) static noinline int prepare_pages(struct btrfs_root *root, struct file *file, struct page **pages, size_t num_pages, loff_t pos, unsigned long first_index, - unsigned long last_index, size_t write_bytes) + size_t write_bytes, bool force_uptodate) { struct extent_state *cached_state = NULL; int i; unsigned long index = pos >> PAGE_CACHE_SHIFT; struct inode *inode = fdentry(file)->d_inode; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); int err = 0; int faili = 0; u64 start_pos; @@ -1073,15 +1078,10 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, start_pos = pos & ~((u64)root->sectorsize - 1); last_pos = ((u64)index + num_pages) << PAGE_CACHE_SHIFT; - if (start_pos > inode->i_size) { - err = btrfs_cont_expand(inode, i_size_read(inode), start_pos); - if (err) - return err; - } - again: for (i = 0; i < num_pages; i++) { - pages[i] = grab_cache_page(inode->i_mapping, index + i); + pages[i] = find_or_create_page(inode->i_mapping, index + i, + mask); if (!pages[i]) { faili = i - 1; err = -ENOMEM; @@ -1089,10 +1089,11 @@ again: } if (i == 0) - err = prepare_uptodate_page(pages[i], pos); + err = prepare_uptodate_page(pages[i], pos, + force_uptodate); if (i == num_pages - 1) err = prepare_uptodate_page(pages[i], - pos + write_bytes); + pos + write_bytes, false); if (err) { page_cache_release(pages[i]); faili = i - 1; @@ -1158,20 +1159,21 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, struct btrfs_root *root = BTRFS_I(inode)->root; struct page **pages = NULL; unsigned long first_index; - unsigned long last_index; size_t num_written = 0; int nrptrs; int ret = 0; + bool force_page_uptodate = false; nrptrs = min((iov_iter_count(i) + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE, PAGE_CACHE_SIZE / (sizeof(struct page *))); + nrptrs = min(nrptrs, current->nr_dirtied_pause - current->nr_dirtied); + nrptrs = max(nrptrs, 8); pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL); if (!pages) return -ENOMEM; first_index = pos >> PAGE_CACHE_SHIFT; - last_index = (pos + iov_iter_count(i)) >> PAGE_CACHE_SHIFT; while (iov_iter_count(i) > 0) { size_t offset = pos & (PAGE_CACHE_SIZE - 1); @@ -1205,8 +1207,8 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, * contents of pages from loop to loop */ ret = prepare_pages(root, file, pages, num_pages, - pos, first_index, last_index, - write_bytes); + pos, first_index, write_bytes, + force_page_uptodate); if (ret) { btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT); @@ -1223,12 +1225,15 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, if (copied < write_bytes) nrptrs = 1; - if (copied == 0) + if (copied == 0) { + force_page_uptodate = true; dirty_pages = 0; - else + } else { + force_page_uptodate = false; dirty_pages = (copied + offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + } /* * If we had a short copy we need to release the excess delaloc @@ -1238,9 +1243,11 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, * managed to copy. */ if (num_pages > dirty_pages) { - if (copied > 0) - atomic_inc( - &BTRFS_I(inode)->outstanding_extents); + if (copied > 0) { + spin_lock(&BTRFS_I(inode)->lock); + BTRFS_I(inode)->outstanding_extents++; + spin_unlock(&BTRFS_I(inode)->lock); + } btrfs_delalloc_release_space(inode, (num_pages - dirty_pages) << PAGE_CACHE_SHIFT); @@ -1336,6 +1343,7 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; loff_t *ppos = &iocb->ki_pos; + u64 start_pos; ssize_t num_written = 0; ssize_t err = 0; size_t count, ocount; @@ -1381,9 +1389,22 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, goto out; } - file_update_time(file); + err = btrfs_update_time(file); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } BTRFS_I(inode)->sequence++; + start_pos = round_down(pos, root->sectorsize); + if (start_pos > i_size_read(inode)) { + err = btrfs_cont_expand(inode, i_size_read(inode), start_pos); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + } + if (unlikely(file->f_flags & O_DIRECT)) { num_written = __btrfs_direct_write(iocb, iov, nr_segs, pos, ppos, count, ocount); @@ -1452,7 +1473,7 @@ int btrfs_release_file(struct inode *inode, struct file *filp) * important optimization for directories because holding the mutex prevents * new operations on the dir while we write to disk. */ -int btrfs_sync_file(struct file *file, int datasync) +int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) { struct dentry *dentry = file->f_path.dentry; struct inode *inode = dentry->d_inode; @@ -1462,9 +1483,13 @@ int btrfs_sync_file(struct file *file, int datasync) trace_btrfs_sync_file(file, datasync); + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + /* we wait first, since the writeback may change the inode */ root->log_batch++; - /* the VFS called filemap_fdatawrite for us */ btrfs_wait_ordered_range(inode, 0, (u64)-1); root->log_batch++; @@ -1472,8 +1497,10 @@ int btrfs_sync_file(struct file *file, int datasync) * check the transaction that last modified this inode * and see if its already been committed */ - if (!BTRFS_I(inode)->last_trans) + if (!BTRFS_I(inode)->last_trans) { + mutex_unlock(&inode->i_mutex); goto out; + } /* * if the last transaction that changed this file was before @@ -1484,6 +1511,7 @@ int btrfs_sync_file(struct file *file, int datasync) if (BTRFS_I(inode)->last_trans <= root->fs_info->last_trans_committed) { BTRFS_I(inode)->last_trans = 0; + mutex_unlock(&inode->i_mutex); goto out; } @@ -1496,12 +1524,15 @@ int btrfs_sync_file(struct file *file, int datasync) trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) { ret = PTR_ERR(trans); + mutex_unlock(&inode->i_mutex); goto out; } ret = btrfs_log_dentry_safe(trans, root, dentry); - if (ret < 0) + if (ret < 0) { + mutex_unlock(&inode->i_mutex); goto out; + } /* we've logged all the items and now have a consistent * version of the file in the log. It is possible that @@ -1513,7 +1544,7 @@ int btrfs_sync_file(struct file *file, int datasync) * file again, but that will end up using the synchronization * inside btrfs_sync_log to keep things safe. */ - mutex_unlock(&dentry->d_inode->i_mutex); + mutex_unlock(&inode->i_mutex); if (ret != BTRFS_NO_LOG_SYNC) { if (ret > 0) { @@ -1528,7 +1559,6 @@ int btrfs_sync_file(struct file *file, int datasync) } else { ret = btrfs_end_transaction(trans, root); } - mutex_lock(&dentry->d_inode->i_mutex); out: return ret > 0 ? -EIO : ret; } @@ -1592,10 +1622,6 @@ static long btrfs_fallocate(struct file *file, int mode, goto out; } - ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start); - if (ret) - goto out; - locked_end = alloc_end - 1; while (1) { struct btrfs_ordered_extent *ordered; @@ -1629,23 +1655,53 @@ static long btrfs_fallocate(struct file *file, int mode, cur_offset = alloc_start; while (1) { + u64 actual_end; + em = btrfs_get_extent(inode, NULL, 0, cur_offset, alloc_end - cur_offset, 0); BUG_ON(IS_ERR_OR_NULL(em)); last_byte = min(extent_map_end(em), alloc_end); + actual_end = min_t(u64, extent_map_end(em), offset + len); last_byte = (last_byte + mask) & ~mask; + if (em->block_start == EXTENT_MAP_HOLE || (cur_offset >= inode->i_size && !test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) { + + /* + * Make sure we have enough space before we do the + * allocation. + */ + ret = btrfs_check_data_free_space(inode, last_byte - + cur_offset); + if (ret) { + free_extent_map(em); + break; + } + ret = btrfs_prealloc_file_range(inode, mode, cur_offset, last_byte - cur_offset, 1 << inode->i_blkbits, offset + len, &alloc_hint); + + /* Let go of our reservation. */ + btrfs_free_reserved_data_space(inode, last_byte - + cur_offset); if (ret < 0) { free_extent_map(em); break; } + } else if (actual_end > inode->i_size && + !(mode & FALLOC_FL_KEEP_SIZE)) { + /* + * We didn't need to allocate any more space, but we + * still extended the size of the file so we need to + * update i_size. + */ + inode->i_ctime = CURRENT_TIME; + i_size_write(inode, actual_end); + btrfs_ordered_update_i_size(inode, actual_end, NULL); } free_extent_map(em); @@ -1657,15 +1713,168 @@ static long btrfs_fallocate(struct file *file, int mode, } unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end, &cached_state, GFP_NOFS); - - btrfs_free_reserved_data_space(inode, alloc_end - alloc_start); out: mutex_unlock(&inode->i_mutex); return ret; } +static int find_desired_extent(struct inode *inode, loff_t *offset, int origin) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + struct extent_map *em; + struct extent_state *cached_state = NULL; + u64 lockstart = *offset; + u64 lockend = i_size_read(inode); + u64 start = *offset; + u64 orig_start = *offset; + u64 len = i_size_read(inode); + u64 last_end = 0; + int ret = 0; + + lockend = max_t(u64, root->sectorsize, lockend); + if (lockend <= lockstart) + lockend = lockstart + root->sectorsize; + + len = lockend - lockstart + 1; + + len = max_t(u64, len, root->sectorsize); + if (inode->i_size == 0) + return -ENXIO; + + lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend, 0, + &cached_state, GFP_NOFS); + + /* + * Delalloc is such a pain. If we have a hole and we have pending + * delalloc for a portion of the hole we will get back a hole that + * exists for the entire range since it hasn't been actually written + * yet. So to take care of this case we need to look for an extent just + * before the position we want in case there is outstanding delalloc + * going on here. + */ + if (origin == SEEK_HOLE && start != 0) { + if (start <= root->sectorsize) + em = btrfs_get_extent_fiemap(inode, NULL, 0, 0, + root->sectorsize, 0); + else + em = btrfs_get_extent_fiemap(inode, NULL, 0, + start - root->sectorsize, + root->sectorsize, 0); + if (IS_ERR(em)) { + ret = -ENXIO; + goto out; + } + last_end = em->start + em->len; + if (em->block_start == EXTENT_MAP_DELALLOC) + last_end = min_t(u64, last_end, inode->i_size); + free_extent_map(em); + } + + while (1) { + em = btrfs_get_extent_fiemap(inode, NULL, 0, start, len, 0); + if (IS_ERR(em)) { + ret = -ENXIO; + break; + } + + if (em->block_start == EXTENT_MAP_HOLE) { + if (test_bit(EXTENT_FLAG_VACANCY, &em->flags)) { + if (last_end <= orig_start) { + free_extent_map(em); + ret = -ENXIO; + break; + } + } + + if (origin == SEEK_HOLE) { + *offset = start; + free_extent_map(em); + break; + } + } else { + if (origin == SEEK_DATA) { + if (em->block_start == EXTENT_MAP_DELALLOC) { + if (start >= inode->i_size) { + free_extent_map(em); + ret = -ENXIO; + break; + } + } + + *offset = start; + free_extent_map(em); + break; + } + } + + start = em->start + em->len; + last_end = em->start + em->len; + + if (em->block_start == EXTENT_MAP_DELALLOC) + last_end = min_t(u64, last_end, inode->i_size); + + if (test_bit(EXTENT_FLAG_VACANCY, &em->flags)) { + free_extent_map(em); + ret = -ENXIO; + break; + } + free_extent_map(em); + cond_resched(); + } + if (!ret) + *offset = min(*offset, inode->i_size); +out: + unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart, lockend, + &cached_state, GFP_NOFS); + return ret; +} + +static loff_t btrfs_file_llseek(struct file *file, loff_t offset, int origin) +{ + struct inode *inode = file->f_mapping->host; + int ret; + + mutex_lock(&inode->i_mutex); + switch (origin) { + case SEEK_END: + case SEEK_CUR: + offset = generic_file_llseek(file, offset, origin); + goto out; + case SEEK_DATA: + case SEEK_HOLE: + if (offset >= i_size_read(inode)) { + mutex_unlock(&inode->i_mutex); + return -ENXIO; + } + + ret = find_desired_extent(inode, &offset, origin); + if (ret) { + mutex_unlock(&inode->i_mutex); + return ret; + } + } + + if (offset < 0 && !(file->f_mode & FMODE_UNSIGNED_OFFSET)) { + offset = -EINVAL; + goto out; + } + if (offset > inode->i_sb->s_maxbytes) { + offset = -EINVAL; + goto out; + } + + /* Special lock needed here? */ + if (offset != file->f_pos) { + file->f_pos = offset; + file->f_version = 0; + } +out: + mutex_unlock(&inode->i_mutex); + return offset; +} + const struct file_operations btrfs_file_operations = { - .llseek = generic_file_llseek, + .llseek = btrfs_file_llseek, .read = do_sync_read, .write = do_sync_write, .aio_read = generic_file_aio_read, diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index bf0d615..ec23d43 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -20,6 +20,7 @@ #include <linux/sched.h> #include <linux/slab.h> #include <linux/math64.h> +#include <linux/ratelimit.h> #include "ctree.h" #include "free-space-cache.h" #include "transaction.h" @@ -84,6 +85,7 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, *block_group, struct btrfs_path *path) { struct inode *inode = NULL; + u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW; spin_lock(&block_group->lock); if (block_group->inode) @@ -98,7 +100,14 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, return inode; spin_lock(&block_group->lock); - if (!btrfs_fs_closing(root->fs_info)) { + if (!((BTRFS_I(inode)->flags & flags) == flags)) { + printk(KERN_INFO "Old style space inode found, converting.\n"); + BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM | + BTRFS_INODE_NODATACOW; + block_group->disk_cache_state = BTRFS_DC_CLEAR; + } + + if (!block_group->iref) { block_group->inode = igrab(inode); block_group->iref = 1; } @@ -116,12 +125,17 @@ int __create_free_space_inode(struct btrfs_root *root, struct btrfs_free_space_header *header; struct btrfs_inode_item *inode_item; struct extent_buffer *leaf; + u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC; int ret; ret = btrfs_insert_empty_inode(trans, root, path, ino); if (ret) return ret; + /* We inline crc's for the free disk space cache */ + if (ino != BTRFS_FREE_INO_OBJECTID) + flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW; + leaf = path->nodes[0]; inode_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item); @@ -134,8 +148,7 @@ int __create_free_space_inode(struct btrfs_root *root, btrfs_set_inode_uid(leaf, inode_item, 0); btrfs_set_inode_gid(leaf, inode_item, 0); btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600); - btrfs_set_inode_flags(leaf, inode_item, BTRFS_INODE_NOCOMPRESS | - BTRFS_INODE_PREALLOC | BTRFS_INODE_NODATASUM); + btrfs_set_inode_flags(leaf, inode_item, flags); btrfs_set_inode_nlink(leaf, inode_item, 1); btrfs_set_inode_transid(leaf, inode_item, trans->transid); btrfs_set_inode_block_group(leaf, inode_item, offset); @@ -184,15 +197,25 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, struct btrfs_path *path, struct inode *inode) { + struct btrfs_block_rsv *rsv; + u64 needed_bytes; loff_t oldsize; int ret = 0; - trans->block_rsv = root->orphan_block_rsv; - ret = btrfs_block_rsv_check(trans, root, - root->orphan_block_rsv, - 0, 5); - if (ret) - return ret; + rsv = trans->block_rsv; + trans->block_rsv = &root->fs_info->global_block_rsv; + + /* 1 for slack space, 1 for updating the inode */ + needed_bytes = btrfs_calc_trunc_metadata_size(root, 1) + + btrfs_calc_trans_metadata_size(root, 1); + + spin_lock(&trans->block_rsv->lock); + if (trans->block_rsv->reserved < needed_bytes) { + spin_unlock(&trans->block_rsv->lock); + trans->block_rsv = rsv; + return -ENOSPC; + } + spin_unlock(&trans->block_rsv->lock); oldsize = i_size_read(inode); btrfs_i_size_write(inode, 0); @@ -204,12 +227,16 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, */ ret = btrfs_truncate_inode_items(trans, root, inode, 0, BTRFS_EXTENT_DATA_KEY); + if (ret) { + trans->block_rsv = rsv; WARN_ON(1); return ret; } ret = btrfs_update_inode(trans, root, inode); + trans->block_rsv = rsv; + return ret; } @@ -232,31 +259,348 @@ static int readahead_cache(struct inode *inode) return 0; } +struct io_ctl { + void *cur, *orig; + struct page *page; + struct page **pages; + struct btrfs_root *root; + unsigned long size; + int index; + int num_pages; + unsigned check_crcs:1; +}; + +static int io_ctl_init(struct io_ctl *io_ctl, struct inode *inode, + struct btrfs_root *root) +{ + memset(io_ctl, 0, sizeof(struct io_ctl)); + io_ctl->num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; + io_ctl->pages = kzalloc(sizeof(struct page *) * io_ctl->num_pages, + GFP_NOFS); + if (!io_ctl->pages) + return -ENOMEM; + io_ctl->root = root; + if (btrfs_ino(inode) != BTRFS_FREE_INO_OBJECTID) + io_ctl->check_crcs = 1; + return 0; +} + +static void io_ctl_free(struct io_ctl *io_ctl) +{ + kfree(io_ctl->pages); +} + +static void io_ctl_unmap_page(struct io_ctl *io_ctl) +{ + if (io_ctl->cur) { + kunmap(io_ctl->page); + io_ctl->cur = NULL; + io_ctl->orig = NULL; + } +} + +static void io_ctl_map_page(struct io_ctl *io_ctl, int clear) +{ + WARN_ON(io_ctl->cur); + BUG_ON(io_ctl->index >= io_ctl->num_pages); + io_ctl->page = io_ctl->pages[io_ctl->index++]; + io_ctl->cur = kmap(io_ctl->page); + io_ctl->orig = io_ctl->cur; + io_ctl->size = PAGE_CACHE_SIZE; + if (clear) + memset(io_ctl->cur, 0, PAGE_CACHE_SIZE); +} + +static void io_ctl_drop_pages(struct io_ctl *io_ctl) +{ + int i; + + io_ctl_unmap_page(io_ctl); + + for (i = 0; i < io_ctl->num_pages; i++) { + ClearPageChecked(io_ctl->pages[i]); + unlock_page(io_ctl->pages[i]); + page_cache_release(io_ctl->pages[i]); + } +} + +static int io_ctl_prepare_pages(struct io_ctl *io_ctl, struct inode *inode, + int uptodate) +{ + struct page *page; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); + int i; + + for (i = 0; i < io_ctl->num_pages; i++) { + page = find_or_create_page(inode->i_mapping, i, mask); + if (!page) { + io_ctl_drop_pages(io_ctl); + return -ENOMEM; + } + io_ctl->pages[i] = page; + if (uptodate && !PageUptodate(page)) { + btrfs_readpage(NULL, page); + lock_page(page); + if (!PageUptodate(page)) { + printk(KERN_ERR "btrfs: error reading free " + "space cache\n"); + io_ctl_drop_pages(io_ctl); + return -EIO; + } + } + } + + for (i = 0; i < io_ctl->num_pages; i++) { + clear_page_dirty_for_io(io_ctl->pages[i]); + set_page_extent_mapped(io_ctl->pages[i]); + } + + return 0; +} + +static void io_ctl_set_generation(struct io_ctl *io_ctl, u64 generation) +{ + u64 *val; + + io_ctl_map_page(io_ctl, 1); + + /* + * Skip the csum areas. If we don't check crcs then we just have a + * 64bit chunk at the front of the first page. + */ + if (io_ctl->check_crcs) { + io_ctl->cur += (sizeof(u32) * io_ctl->num_pages); + io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages); + } else { + io_ctl->cur += sizeof(u64); + io_ctl->size -= sizeof(u64) * 2; + } + + val = io_ctl->cur; + *val = cpu_to_le64(generation); + io_ctl->cur += sizeof(u64); +} + +static int io_ctl_check_generation(struct io_ctl *io_ctl, u64 generation) +{ + u64 *gen; + + /* + * Skip the crc area. If we don't check crcs then we just have a 64bit + * chunk at the front of the first page. + */ + if (io_ctl->check_crcs) { + io_ctl->cur += sizeof(u32) * io_ctl->num_pages; + io_ctl->size -= sizeof(u64) + + (sizeof(u32) * io_ctl->num_pages); + } else { + io_ctl->cur += sizeof(u64); + io_ctl->size -= sizeof(u64) * 2; + } + + gen = io_ctl->cur; + if (le64_to_cpu(*gen) != generation) { + printk_ratelimited(KERN_ERR "btrfs: space cache generation " + "(%Lu) does not match inode (%Lu)\n", *gen, + generation); + io_ctl_unmap_page(io_ctl); + return -EIO; + } + io_ctl->cur += sizeof(u64); + return 0; +} + +static void io_ctl_set_crc(struct io_ctl *io_ctl, int index) +{ + u32 *tmp; + u32 crc = ~(u32)0; + unsigned offset = 0; + + if (!io_ctl->check_crcs) { + io_ctl_unmap_page(io_ctl); + return; + } + + if (index == 0) + offset = sizeof(u32) * io_ctl->num_pages;; + + crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc, + PAGE_CACHE_SIZE - offset); + btrfs_csum_final(crc, (char *)&crc); + io_ctl_unmap_page(io_ctl); + tmp = kmap(io_ctl->pages[0]); + tmp += index; + *tmp = crc; + kunmap(io_ctl->pages[0]); +} + +static int io_ctl_check_crc(struct io_ctl *io_ctl, int index) +{ + u32 *tmp, val; + u32 crc = ~(u32)0; + unsigned offset = 0; + + if (!io_ctl->check_crcs) { + io_ctl_map_page(io_ctl, 0); + return 0; + } + + if (index == 0) + offset = sizeof(u32) * io_ctl->num_pages; + + tmp = kmap(io_ctl->pages[0]); + tmp += index; + val = *tmp; + kunmap(io_ctl->pages[0]); + + io_ctl_map_page(io_ctl, 0); + crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc, + PAGE_CACHE_SIZE - offset); + btrfs_csum_final(crc, (char *)&crc); + if (val != crc) { + printk_ratelimited(KERN_ERR "btrfs: csum mismatch on free " + "space cache\n"); + io_ctl_unmap_page(io_ctl); + return -EIO; + } + + return 0; +} + +static int io_ctl_add_entry(struct io_ctl *io_ctl, u64 offset, u64 bytes, + void *bitmap) +{ + struct btrfs_free_space_entry *entry; + + if (!io_ctl->cur) + return -ENOSPC; + + entry = io_ctl->cur; + entry->offset = cpu_to_le64(offset); + entry->bytes = cpu_to_le64(bytes); + entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP : + BTRFS_FREE_SPACE_EXTENT; + io_ctl->cur += sizeof(struct btrfs_free_space_entry); + io_ctl->size -= sizeof(struct btrfs_free_space_entry); + + if (io_ctl->size >= sizeof(struct btrfs_free_space_entry)) + return 0; + + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + + /* No more pages to map */ + if (io_ctl->index >= io_ctl->num_pages) + return 0; + + /* map the next page */ + io_ctl_map_page(io_ctl, 1); + return 0; +} + +static int io_ctl_add_bitmap(struct io_ctl *io_ctl, void *bitmap) +{ + if (!io_ctl->cur) + return -ENOSPC; + + /* + * If we aren't at the start of the current page, unmap this one and + * map the next one if there is any left. + */ + if (io_ctl->cur != io_ctl->orig) { + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + if (io_ctl->index >= io_ctl->num_pages) + return -ENOSPC; + io_ctl_map_page(io_ctl, 0); + } + + memcpy(io_ctl->cur, bitmap, PAGE_CACHE_SIZE); + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + if (io_ctl->index < io_ctl->num_pages) + io_ctl_map_page(io_ctl, 0); + return 0; +} + +static void io_ctl_zero_remaining_pages(struct io_ctl *io_ctl) +{ + /* + * If we're not on the boundary we know we've modified the page and we + * need to crc the page. + */ + if (io_ctl->cur != io_ctl->orig) + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + else + io_ctl_unmap_page(io_ctl); + + while (io_ctl->index < io_ctl->num_pages) { + io_ctl_map_page(io_ctl, 1); + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + } +} + +static int io_ctl_read_entry(struct io_ctl *io_ctl, + struct btrfs_free_space *entry, u8 *type) +{ + struct btrfs_free_space_entry *e; + int ret; + + if (!io_ctl->cur) { + ret = io_ctl_check_crc(io_ctl, io_ctl->index); + if (ret) + return ret; + } + + e = io_ctl->cur; + entry->offset = le64_to_cpu(e->offset); + entry->bytes = le64_to_cpu(e->bytes); + *type = e->type; + io_ctl->cur += sizeof(struct btrfs_free_space_entry); + io_ctl->size -= sizeof(struct btrfs_free_space_entry); + + if (io_ctl->size >= sizeof(struct btrfs_free_space_entry)) + return 0; + + io_ctl_unmap_page(io_ctl); + + return 0; +} + +static int io_ctl_read_bitmap(struct io_ctl *io_ctl, + struct btrfs_free_space *entry) +{ + int ret; + + ret = io_ctl_check_crc(io_ctl, io_ctl->index); + if (ret) + return ret; + + memcpy(entry->bitmap, io_ctl->cur, PAGE_CACHE_SIZE); + io_ctl_unmap_page(io_ctl); + + return 0; +} + int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, struct btrfs_free_space_ctl *ctl, struct btrfs_path *path, u64 offset) { struct btrfs_free_space_header *header; struct extent_buffer *leaf; - struct page *page; - u32 *checksums = NULL, *crc; - char *disk_crcs = NULL; + struct io_ctl io_ctl; struct btrfs_key key; + struct btrfs_free_space *e, *n; struct list_head bitmaps; u64 num_entries; u64 num_bitmaps; u64 generation; - u32 cur_crc = ~(u32)0; - pgoff_t index = 0; - unsigned long first_page_offset; - int num_checksums; + u8 type; int ret = 0; INIT_LIST_HEAD(&bitmaps); /* Nothing in the space cache, goodbye */ if (!i_size_read(inode)) - goto out; + return 0; key.objectid = BTRFS_FREE_SPACE_OBJECTID; key.offset = offset; @@ -264,11 +608,10 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) - goto out; + return 0; else if (ret > 0) { btrfs_release_path(path); - ret = 0; - goto out; + return 0; } ret = -1; @@ -286,194 +629,102 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, " not match free space cache generation (%llu)\n", (unsigned long long)BTRFS_I(inode)->generation, (unsigned long long)generation); - goto out; + return 0; } if (!num_entries) - goto out; - - /* Setup everything for doing checksumming */ - num_checksums = i_size_read(inode) / PAGE_CACHE_SIZE; - checksums = crc = kzalloc(sizeof(u32) * num_checksums, GFP_NOFS); - if (!checksums) - goto out; - first_page_offset = (sizeof(u32) * num_checksums) + sizeof(u64); - disk_crcs = kzalloc(first_page_offset, GFP_NOFS); - if (!disk_crcs) - goto out; + return 0; + io_ctl_init(&io_ctl, inode, root); ret = readahead_cache(inode); if (ret) goto out; - while (1) { - struct btrfs_free_space_entry *entry; - struct btrfs_free_space *e; - void *addr; - unsigned long offset = 0; - unsigned long start_offset = 0; - int need_loop = 0; + ret = io_ctl_prepare_pages(&io_ctl, inode, 1); + if (ret) + goto out; - if (!num_entries && !num_bitmaps) - break; + ret = io_ctl_check_crc(&io_ctl, 0); + if (ret) + goto free_cache; - if (index == 0) { - start_offset = first_page_offset; - offset = start_offset; - } + ret = io_ctl_check_generation(&io_ctl, generation); + if (ret) + goto free_cache; - page = grab_cache_page(inode->i_mapping, index); - if (!page) + while (num_entries) { + e = kmem_cache_zalloc(btrfs_free_space_cachep, + GFP_NOFS); + if (!e) goto free_cache; - if (!PageUptodate(page)) { - btrfs_readpage(NULL, page); - lock_page(page); - if (!PageUptodate(page)) { - unlock_page(page); - page_cache_release(page); - printk(KERN_ERR "btrfs: error reading free " - "space cache\n"); - goto free_cache; - } - } - addr = kmap(page); - - if (index == 0) { - u64 *gen; - - memcpy(disk_crcs, addr, first_page_offset); - gen = addr + (sizeof(u32) * num_checksums); - if (*gen != BTRFS_I(inode)->generation) { - printk(KERN_ERR "btrfs: space cache generation" - " (%llu) does not match inode (%llu)\n", - (unsigned long long)*gen, - (unsigned long long) - BTRFS_I(inode)->generation); - kunmap(page); - unlock_page(page); - page_cache_release(page); - goto free_cache; - } - crc = (u32 *)disk_crcs; - } - entry = addr + start_offset; - - /* First lets check our crc before we do anything fun */ - cur_crc = ~(u32)0; - cur_crc = btrfs_csum_data(root, addr + start_offset, cur_crc, - PAGE_CACHE_SIZE - start_offset); - btrfs_csum_final(cur_crc, (char *)&cur_crc); - if (cur_crc != *crc) { - printk(KERN_ERR "btrfs: crc mismatch for page %lu\n", - index); - kunmap(page); - unlock_page(page); - page_cache_release(page); + ret = io_ctl_read_entry(&io_ctl, e, &type); + if (ret) { + kmem_cache_free(btrfs_free_space_cachep, e); goto free_cache; } - crc++; - while (1) { - if (!num_entries) - break; + if (!e->bytes) { + kmem_cache_free(btrfs_free_space_cachep, e); + goto free_cache; + } - need_loop = 1; - e = kmem_cache_zalloc(btrfs_free_space_cachep, - GFP_NOFS); - if (!e) { - kunmap(page); - unlock_page(page); - page_cache_release(page); + if (type == BTRFS_FREE_SPACE_EXTENT) { + spin_lock(&ctl->tree_lock); + ret = link_free_space(ctl, e); + spin_unlock(&ctl->tree_lock); + if (ret) { + printk(KERN_ERR "Duplicate entries in " + "free space cache, dumping\n"); + kmem_cache_free(btrfs_free_space_cachep, e); goto free_cache; } - - e->offset = le64_to_cpu(entry->offset); - e->bytes = le64_to_cpu(entry->bytes); - if (!e->bytes) { - kunmap(page); - kmem_cache_free(btrfs_free_space_cachep, e); - unlock_page(page); - page_cache_release(page); + } else { + BUG_ON(!num_bitmaps); + num_bitmaps--; + e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); + if (!e->bitmap) { + kmem_cache_free( + btrfs_free_space_cachep, e); goto free_cache; } - - if (entry->type == BTRFS_FREE_SPACE_EXTENT) { - spin_lock(&ctl->tree_lock); - ret = link_free_space(ctl, e); - spin_unlock(&ctl->tree_lock); - if (ret) { - printk(KERN_ERR "Duplicate entries in " - "free space cache, dumping\n"); - kunmap(page); - unlock_page(page); - page_cache_release(page); - goto free_cache; - } - } else { - e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); - if (!e->bitmap) { - kunmap(page); - kmem_cache_free( - btrfs_free_space_cachep, e); - unlock_page(page); - page_cache_release(page); - goto free_cache; - } - spin_lock(&ctl->tree_lock); - ret = link_free_space(ctl, e); - ctl->total_bitmaps++; - ctl->op->recalc_thresholds(ctl); - spin_unlock(&ctl->tree_lock); - if (ret) { - printk(KERN_ERR "Duplicate entries in " - "free space cache, dumping\n"); - kunmap(page); - unlock_page(page); - page_cache_release(page); - goto free_cache; - } - list_add_tail(&e->list, &bitmaps); + spin_lock(&ctl->tree_lock); + ret = link_free_space(ctl, e); + ctl->total_bitmaps++; + ctl->op->recalc_thresholds(ctl); + spin_unlock(&ctl->tree_lock); + if (ret) { + printk(KERN_ERR "Duplicate entries in " + "free space cache, dumping\n"); + kmem_cache_free(btrfs_free_space_cachep, e); + goto free_cache; } - - num_entries--; - offset += sizeof(struct btrfs_free_space_entry); - if (offset + sizeof(struct btrfs_free_space_entry) >= - PAGE_CACHE_SIZE) - break; - entry++; + list_add_tail(&e->list, &bitmaps); } - /* - * We read an entry out of this page, we need to move on to the - * next page. - */ - if (need_loop) { - kunmap(page); - goto next; - } + num_entries--; + } - /* - * We add the bitmaps at the end of the entries in order that - * the bitmap entries are added to the cache. - */ - e = list_entry(bitmaps.next, struct btrfs_free_space, list); + io_ctl_unmap_page(&io_ctl); + + /* + * We add the bitmaps at the end of the entries in order that + * the bitmap entries are added to the cache. + */ + list_for_each_entry_safe(e, n, &bitmaps, list) { list_del_init(&e->list); - memcpy(e->bitmap, addr, PAGE_CACHE_SIZE); - kunmap(page); - num_bitmaps--; -next: - unlock_page(page); - page_cache_release(page); - index++; + ret = io_ctl_read_bitmap(&io_ctl, e); + if (ret) + goto free_cache; } + io_ctl_drop_pages(&io_ctl); ret = 1; out: - kfree(checksums); - kfree(disk_crcs); + io_ctl_free(&io_ctl); return ret; free_cache: + io_ctl_drop_pages(&io_ctl); __btrfs_remove_free_space_cache(ctl); goto out; } @@ -485,7 +736,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root = fs_info->tree_root; struct inode *inode; struct btrfs_path *path; - int ret; + int ret = 0; bool matched; u64 used = btrfs_block_group_used(&block_group->item); @@ -517,6 +768,14 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, return 0; } + /* We may have converted the inode and made the cache invalid. */ + spin_lock(&block_group->lock); + if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { + spin_unlock(&block_group->lock); + goto out; + } + spin_unlock(&block_group->lock); + ret = __load_free_space_cache(fs_info->tree_root, inode, ctl, path, block_group->key.objectid); btrfs_free_path(path); @@ -550,6 +809,19 @@ out: return ret; } +/** + * __btrfs_write_out_cache - write out cached info to an inode + * @root - the root the inode belongs to + * @ctl - the free space cache we are going to write out + * @block_group - the block_group for this cache if it belongs to a block_group + * @trans - the trans handle + * @path - the path to use + * @offset - the offset for the key we'll insert + * + * This function writes out a free space cache struct to disk for quick recovery + * on mount. This will return 0 if it was successfull in writing the cache out, + * and -1 if it was not. + */ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, struct btrfs_free_space_ctl *ctl, struct btrfs_block_group_cache *block_group, @@ -560,64 +832,24 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, struct extent_buffer *leaf; struct rb_node *node; struct list_head *pos, *n; - struct page **pages; - struct page *page; struct extent_state *cached_state = NULL; struct btrfs_free_cluster *cluster = NULL; struct extent_io_tree *unpin = NULL; + struct io_ctl io_ctl; struct list_head bitmap_list; struct btrfs_key key; u64 start, end, len; - u64 bytes = 0; - u32 *crc, *checksums; - unsigned long first_page_offset; - int index = 0, num_pages = 0; int entries = 0; int bitmaps = 0; - int ret = -1; - bool next_page = false; - bool out_of_space = false; + int ret; + int err = -1; INIT_LIST_HEAD(&bitmap_list); - node = rb_first(&ctl->free_space_offset); - if (!node) - return 0; - if (!i_size_read(inode)) return -1; - num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> - PAGE_CACHE_SHIFT; - - /* Since the first page has all of our checksums and our generation we - * need to calculate the offset into the page that we can start writing - * our entries. - */ - first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64); - - filemap_write_and_wait(inode->i_mapping); - btrfs_wait_ordered_range(inode, inode->i_size & - ~(root->sectorsize - 1), (u64)-1); - - /* make sure we don't overflow that first page */ - if (first_page_offset + sizeof(struct btrfs_free_space_entry) >= PAGE_CACHE_SIZE) { - /* this is really the same as running out of space, where we also return 0 */ - printk(KERN_CRIT "Btrfs: free space cache was too big for the crc page\n"); - ret = 0; - goto out_update; - } - - /* We need a checksum per page. */ - crc = checksums = kzalloc(sizeof(u32) * num_pages, GFP_NOFS); - if (!crc) - return -1; - - pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS); - if (!pages) { - kfree(crc); - return -1; - } + io_ctl_init(&io_ctl, inode, root); /* Get the cluster for this block_group if it exists */ if (block_group && !list_empty(&block_group->cluster_list)) @@ -631,30 +863,9 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, */ unpin = root->fs_info->pinned_extents; - /* - * Lock all pages first so we can lock the extent safely. - * - * NOTE: Because we hold the ref the entire time we're going to write to - * the page find_get_page should never fail, so we don't do a check - * after find_get_page at this point. Just putting this here so people - * know and don't freak out. - */ - while (index < num_pages) { - page = grab_cache_page(inode->i_mapping, index); - if (!page) { - int i; - - for (i = 0; i < num_pages; i++) { - unlock_page(pages[i]); - page_cache_release(pages[i]); - } - goto out_free; - } - pages[index] = page; - index++; - } + /* Lock all pages first so we can lock the extent safely. */ + io_ctl_prepare_pages(&io_ctl, inode, 0); - index = 0; lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, 0, &cached_state, GFP_NOFS); @@ -665,192 +876,112 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, if (block_group) start = block_group->key.objectid; - /* Write out the extent entries */ - do { - struct btrfs_free_space_entry *entry; - void *addr; - unsigned long offset = 0; - unsigned long start_offset = 0; - - next_page = false; - - if (index == 0) { - start_offset = first_page_offset; - offset = start_offset; - } + node = rb_first(&ctl->free_space_offset); + if (!node && cluster) { + node = rb_first(&cluster->root); + cluster = NULL; + } - if (index >= num_pages) { - out_of_space = true; - break; - } + /* Make sure we can fit our crcs into the first page */ + if (io_ctl.check_crcs && + (io_ctl.num_pages * sizeof(u32)) >= PAGE_CACHE_SIZE) { + WARN_ON(1); + goto out_nospc; + } - page = pages[index]; + io_ctl_set_generation(&io_ctl, trans->transid); - addr = kmap(page); - entry = addr + start_offset; + /* Write out the extent entries */ + while (node) { + struct btrfs_free_space *e; - memset(addr, 0, PAGE_CACHE_SIZE); - while (node && !next_page) { - struct btrfs_free_space *e; + e = rb_entry(node, struct btrfs_free_space, offset_index); + entries++; - e = rb_entry(node, struct btrfs_free_space, offset_index); - entries++; + ret = io_ctl_add_entry(&io_ctl, e->offset, e->bytes, + e->bitmap); + if (ret) + goto out_nospc; - entry->offset = cpu_to_le64(e->offset); - entry->bytes = cpu_to_le64(e->bytes); - if (e->bitmap) { - entry->type = BTRFS_FREE_SPACE_BITMAP; - list_add_tail(&e->list, &bitmap_list); - bitmaps++; - } else { - entry->type = BTRFS_FREE_SPACE_EXTENT; - } - node = rb_next(node); - if (!node && cluster) { - node = rb_first(&cluster->root); - cluster = NULL; - } - offset += sizeof(struct btrfs_free_space_entry); - if (offset + sizeof(struct btrfs_free_space_entry) >= - PAGE_CACHE_SIZE) - next_page = true; - entry++; + if (e->bitmap) { + list_add_tail(&e->list, &bitmap_list); + bitmaps++; } + node = rb_next(node); + if (!node && cluster) { + node = rb_first(&cluster->root); + cluster = NULL; + } + } - /* - * We want to add any pinned extents to our free space cache - * so we don't leak the space - */ - while (block_group && !next_page && - (start < block_group->key.objectid + - block_group->key.offset)) { - ret = find_first_extent_bit(unpin, start, &start, &end, - EXTENT_DIRTY); - if (ret) { - ret = 0; - break; - } - - /* This pinned extent is out of our range */ - if (start >= block_group->key.objectid + - block_group->key.offset) - break; - - len = block_group->key.objectid + - block_group->key.offset - start; - len = min(len, end + 1 - start); - - entries++; - entry->offset = cpu_to_le64(start); - entry->bytes = cpu_to_le64(len); - entry->type = BTRFS_FREE_SPACE_EXTENT; - - start = end + 1; - offset += sizeof(struct btrfs_free_space_entry); - if (offset + sizeof(struct btrfs_free_space_entry) >= - PAGE_CACHE_SIZE) - next_page = true; - entry++; + /* + * We want to add any pinned extents to our free space cache + * so we don't leak the space + */ + while (block_group && (start < block_group->key.objectid + + block_group->key.offset)) { + ret = find_first_extent_bit(unpin, start, &start, &end, + EXTENT_DIRTY); + if (ret) { + ret = 0; + break; } - *crc = ~(u32)0; - *crc = btrfs_csum_data(root, addr + start_offset, *crc, - PAGE_CACHE_SIZE - start_offset); - kunmap(page); - btrfs_csum_final(*crc, (char *)crc); - crc++; + /* This pinned extent is out of our range */ + if (start >= block_group->key.objectid + + block_group->key.offset) + break; + + len = block_group->key.objectid + + block_group->key.offset - start; + len = min(len, end + 1 - start); - bytes += PAGE_CACHE_SIZE; + entries++; + ret = io_ctl_add_entry(&io_ctl, start, len, NULL); + if (ret) + goto out_nospc; - index++; - } while (node || next_page); + start = end + 1; + } /* Write out the bitmaps */ list_for_each_safe(pos, n, &bitmap_list) { - void *addr; struct btrfs_free_space *entry = list_entry(pos, struct btrfs_free_space, list); - if (index >= num_pages) { - out_of_space = true; - break; - } - page = pages[index]; - - addr = kmap(page); - memcpy(addr, entry->bitmap, PAGE_CACHE_SIZE); - *crc = ~(u32)0; - *crc = btrfs_csum_data(root, addr, *crc, PAGE_CACHE_SIZE); - kunmap(page); - btrfs_csum_final(*crc, (char *)crc); - crc++; - bytes += PAGE_CACHE_SIZE; - + ret = io_ctl_add_bitmap(&io_ctl, entry->bitmap); + if (ret) + goto out_nospc; list_del_init(&entry->list); - index++; - } - - if (out_of_space) { - btrfs_drop_pages(pages, num_pages); - unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, - i_size_read(inode) - 1, &cached_state, - GFP_NOFS); - ret = 0; - goto out_free; } /* Zero out the rest of the pages just to make sure */ - while (index < num_pages) { - void *addr; - - page = pages[index]; - addr = kmap(page); - memset(addr, 0, PAGE_CACHE_SIZE); - kunmap(page); - bytes += PAGE_CACHE_SIZE; - index++; - } - - /* Write the checksums and trans id to the first page */ - { - void *addr; - u64 *gen; - - page = pages[0]; + io_ctl_zero_remaining_pages(&io_ctl); - addr = kmap(page); - memcpy(addr, checksums, sizeof(u32) * num_pages); - gen = addr + (sizeof(u32) * num_pages); - *gen = trans->transid; - kunmap(page); - } - - ret = btrfs_dirty_pages(root, inode, pages, num_pages, 0, - bytes, &cached_state); - btrfs_drop_pages(pages, num_pages); + ret = btrfs_dirty_pages(root, inode, io_ctl.pages, io_ctl.num_pages, + 0, i_size_read(inode), &cached_state); + io_ctl_drop_pages(&io_ctl); unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, &cached_state, GFP_NOFS); - if (ret) { - ret = 0; - goto out_free; - } + if (ret) + goto out; - BTRFS_I(inode)->generation = trans->transid; - filemap_write_and_wait(inode->i_mapping); + ret = filemap_write_and_wait(inode->i_mapping); + if (ret) + goto out; key.objectid = BTRFS_FREE_SPACE_OBJECTID; key.offset = offset; key.type = 0; - ret = btrfs_search_slot(trans, root, &key, path, 1, 1); + ret = btrfs_search_slot(trans, root, &key, path, 0, 1); if (ret < 0) { - ret = -1; - clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, - EXTENT_DIRTY | EXTENT_DELALLOC | - EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS); - goto out_free; + clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1, + EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL, + GFP_NOFS); + goto out; } leaf = path->nodes[0]; if (ret > 0) { @@ -860,15 +991,16 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID || found_key.offset != offset) { - ret = -1; - clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, - EXTENT_DIRTY | EXTENT_DELALLOC | - EXTENT_DO_ACCOUNTING, 0, 0, NULL, - GFP_NOFS); + clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, + inode->i_size - 1, + EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, + NULL, GFP_NOFS); btrfs_release_path(path); - goto out_free; + goto out; } } + + BTRFS_I(inode)->generation = trans->transid; header = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_free_space_header); btrfs_set_free_space_entries(leaf, header, entries); @@ -877,19 +1009,26 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, btrfs_mark_buffer_dirty(leaf); btrfs_release_path(path); - ret = 1; - -out_free: - kfree(checksums); - kfree(pages); - -out_update: - if (ret != 1) { - invalidate_inode_pages2_range(inode->i_mapping, 0, index); + err = 0; +out: + io_ctl_free(&io_ctl); + if (err) { + invalidate_inode_pages2(inode->i_mapping); BTRFS_I(inode)->generation = 0; } btrfs_update_inode(trans, root, inode); - return ret; + return err; + +out_nospc: + list_for_each_safe(pos, n, &bitmap_list) { + struct btrfs_free_space *entry = + list_entry(pos, struct btrfs_free_space, list); + list_del_init(&entry->list); + } + io_ctl_drop_pages(&io_ctl); + unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, + i_size_read(inode) - 1, &cached_state, GFP_NOFS); + goto out; } int btrfs_write_out_cache(struct btrfs_root *root, @@ -916,14 +1055,15 @@ int btrfs_write_out_cache(struct btrfs_root *root, ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans, path, block_group->key.objectid); - if (ret < 0) { + if (ret) { spin_lock(&block_group->lock); block_group->disk_cache_state = BTRFS_DC_ERROR; spin_unlock(&block_group->lock); ret = 0; - +#ifdef DEBUG printk(KERN_ERR "btrfs: failed to write free space cace " "for block group %llu\n", block_group->key.objectid); +#endif } iput(inode); @@ -1219,9 +1359,9 @@ static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) div64_u64(extent_bytes, (sizeof(struct btrfs_free_space))); } -static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, - struct btrfs_free_space *info, u64 offset, - u64 bytes) +static inline void __bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info, + u64 offset, u64 bytes) { unsigned long start, count; @@ -1232,6 +1372,13 @@ static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, bitmap_clear(info->bitmap, start, count); info->bytes -= bytes; +} + +static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info, u64 offset, + u64 bytes) +{ + __bitmap_clear_bits(ctl, info, offset, bytes); ctl->free_space -= bytes; } @@ -1323,6 +1470,7 @@ static void add_new_bitmap(struct btrfs_free_space_ctl *ctl, { info->offset = offset_to_bitmap(ctl, offset); info->bytes = 0; + INIT_LIST_HEAD(&info->list); link_free_space(ctl, info); ctl->total_bitmaps++; @@ -1702,7 +1850,13 @@ again: info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 1, 0); if (!info) { - WARN_ON(1); + /* the tree logging code might be calling us before we + * have fully loaded the free space rbtree for this + * block group. So it is possible the entry won't + * be in the rbtree yet at all. The caching code + * will make sure not to put it in the rbtree if + * the logging code has pinned it. + */ goto out_lock; } } @@ -1741,6 +1895,7 @@ again: ctl->total_bitmaps--; } kmem_cache_free(btrfs_free_space_cachep, info); + ret = 0; goto out_lock; } @@ -1748,7 +1903,8 @@ again: unlink_free_space(ctl, info); info->offset += bytes; info->bytes -= bytes; - link_free_space(ctl, info); + ret = link_free_space(ctl, info); + WARN_ON(ret); goto out_lock; } @@ -2035,7 +2191,7 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group, return 0; ret = search_start; - bitmap_clear_bits(ctl, entry, ret, bytes); + __bitmap_clear_bits(ctl, entry, ret, bytes); return ret; } @@ -2090,7 +2246,6 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, continue; } } else { - ret = entry->offset; entry->offset += bytes; @@ -2165,6 +2320,7 @@ again: if (!found) { start = i; + cluster->max_size = 0; found = true; } @@ -2308,16 +2464,23 @@ setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, { struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry; - struct rb_node *node; int ret = -ENOSPC; + u64 bitmap_offset = offset_to_bitmap(ctl, offset); if (ctl->total_bitmaps == 0) return -ENOSPC; /* - * First check our cached list of bitmaps and see if there is an entry - * here that will work. + * The bitmap that covers offset won't be in the list unless offset + * is just its start offset. */ + entry = list_first_entry(bitmaps, struct btrfs_free_space, list); + if (entry->offset != bitmap_offset) { + entry = tree_search_offset(ctl, bitmap_offset, 1, 0); + if (entry && list_empty(&entry->list)) + list_add(&entry->list, bitmaps); + } + list_for_each_entry(entry, bitmaps, list) { if (entry->bytes < min_bytes) continue; @@ -2328,38 +2491,10 @@ setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, } /* - * If we do have entries on our list and we are here then we didn't find - * anything, so go ahead and get the next entry after the last entry in - * this list and start the search from there. + * The bitmaps list has all the bitmaps that record free space + * starting after offset, so no more search is required. */ - if (!list_empty(bitmaps)) { - entry = list_entry(bitmaps->prev, struct btrfs_free_space, - list); - node = rb_next(&entry->offset_index); - if (!node) - return -ENOSPC; - entry = rb_entry(node, struct btrfs_free_space, offset_index); - goto search; - } - - entry = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 0, 1); - if (!entry) - return -ENOSPC; - -search: - node = &entry->offset_index; - do { - entry = rb_entry(node, struct btrfs_free_space, offset_index); - node = rb_next(&entry->offset_index); - if (!entry->bitmap) - continue; - if (entry->bytes < min_bytes) - continue; - ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset, - bytes, min_bytes); - } while (ret && node); - - return ret; + return -ENOSPC; } /* @@ -2377,8 +2512,8 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, u64 offset, u64 bytes, u64 empty_size) { struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; - struct list_head bitmaps; struct btrfs_free_space *entry, *tmp; + LIST_HEAD(bitmaps); u64 min_bytes; int ret; @@ -2417,7 +2552,6 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, goto out; } - INIT_LIST_HEAD(&bitmaps); ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset, bytes, min_bytes); if (ret) @@ -2513,9 +2647,19 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, spin_unlock(&ctl->tree_lock); if (bytes >= minlen) { - int update_ret; - update_ret = btrfs_update_reserved_bytes(block_group, - bytes, 1, 1); + struct btrfs_space_info *space_info; + int update = 0; + + space_info = block_group->space_info; + spin_lock(&space_info->lock); + spin_lock(&block_group->lock); + if (!block_group->ro) { + block_group->reserved += bytes; + space_info->bytes_reserved += bytes; + update = 1; + } + spin_unlock(&block_group->lock); + spin_unlock(&space_info->lock); ret = btrfs_error_discard_extent(fs_info->extent_root, start, @@ -2523,9 +2667,16 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, &actually_trimmed); btrfs_add_free_space(block_group, start, bytes); - if (!update_ret) - btrfs_update_reserved_bytes(block_group, - bytes, 0, 1); + if (update) { + spin_lock(&space_info->lock); + spin_lock(&block_group->lock); + if (block_group->ro) + space_info->bytes_readonly += bytes; + block_group->reserved -= bytes; + space_info->bytes_reserved -= bytes; + spin_unlock(&space_info->lock); + spin_unlock(&block_group->lock); + } if (ret) break; @@ -2684,9 +2835,13 @@ int btrfs_write_out_ino_cache(struct btrfs_root *root, return 0; ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0); - if (ret < 0) + if (ret) { + btrfs_delalloc_release_metadata(inode, inode->i_size); +#ifdef DEBUG printk(KERN_ERR "btrfs: failed to write free ino cache " "for root %llu\n", root->root_key.objectid); +#endif + } iput(inode); return ret; diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c index b4087e0..b3d1efe 100644 --- a/fs/btrfs/inode-map.c +++ b/fs/btrfs/inode-map.c @@ -207,24 +207,14 @@ again: void btrfs_return_ino(struct btrfs_root *root, u64 objectid) { - struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; struct btrfs_free_space_ctl *pinned = root->free_ino_pinned; if (!btrfs_test_opt(root, INODE_MAP_CACHE)) return; - again: if (root->cached == BTRFS_CACHE_FINISHED) { - __btrfs_add_free_space(ctl, objectid, 1); + __btrfs_add_free_space(pinned, objectid, 1); } else { - /* - * If we are in the process of caching free ino chunks, - * to avoid adding the same inode number to the free_ino - * tree twice due to cross transaction, we'll leave it - * in the pinned tree until a transaction is committed - * or the caching work is done. - */ - mutex_lock(&root->fs_commit_mutex); spin_lock(&root->cache_lock); if (root->cached == BTRFS_CACHE_FINISHED) { @@ -236,11 +226,7 @@ again: start_caching(root); - if (objectid <= root->cache_progress || - objectid > root->highest_objectid) - __btrfs_add_free_space(ctl, objectid, 1); - else - __btrfs_add_free_space(pinned, objectid, 1); + __btrfs_add_free_space(pinned, objectid, 1); mutex_unlock(&root->fs_commit_mutex); } @@ -258,6 +244,7 @@ void btrfs_unpin_free_ino(struct btrfs_root *root) { struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; struct rb_root *rbroot = &root->free_ino_pinned->free_space_offset; + spinlock_t *rbroot_lock = &root->free_ino_pinned->tree_lock; struct btrfs_free_space *info; struct rb_node *n; u64 count; @@ -266,24 +253,30 @@ void btrfs_unpin_free_ino(struct btrfs_root *root) return; while (1) { + bool add_to_ctl = true; + + spin_lock(rbroot_lock); n = rb_first(rbroot); - if (!n) + if (!n) { + spin_unlock(rbroot_lock); break; + } info = rb_entry(n, struct btrfs_free_space, offset_index); BUG_ON(info->bitmap); if (info->offset > root->cache_progress) - goto free; + add_to_ctl = false; else if (info->offset + info->bytes > root->cache_progress) count = root->cache_progress - info->offset + 1; else count = info->bytes; - __btrfs_add_free_space(ctl, info->offset, count); -free: rb_erase(&info->offset_index, rbroot); - kfree(info); + spin_unlock(rbroot_lock); + if (add_to_ctl) + __btrfs_add_free_space(ctl, info->offset, count); + kmem_cache_free(btrfs_free_space_cachep, info); } } @@ -398,6 +391,8 @@ int btrfs_save_ino_cache(struct btrfs_root *root, struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; struct btrfs_path *path; struct inode *inode; + struct btrfs_block_rsv *rsv; + u64 num_bytes; u64 alloc_hint = 0; int ret; int prealloc; @@ -421,11 +416,26 @@ int btrfs_save_ino_cache(struct btrfs_root *root, if (!path) return -ENOMEM; + rsv = trans->block_rsv; + trans->block_rsv = &root->fs_info->trans_block_rsv; + + num_bytes = trans->bytes_reserved; + /* + * 1 item for inode item insertion if need + * 3 items for inode item update (in the worst case) + * 1 item for free space object + * 3 items for pre-allocation + */ + trans->bytes_reserved = btrfs_calc_trans_metadata_size(root, 8); + ret = btrfs_block_rsv_add_noflush(root, trans->block_rsv, + trans->bytes_reserved); + if (ret) + goto out; again: inode = lookup_free_ino_inode(root, path); if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { ret = PTR_ERR(inode); - goto out; + goto out_release; } if (IS_ERR(inode)) { @@ -434,7 +444,7 @@ again: ret = create_free_ino_inode(root, trans, path); if (ret) - goto out; + goto out_release; goto again; } @@ -465,21 +475,26 @@ again: /* Just to make sure we have enough space */ prealloc += 8 * PAGE_CACHE_SIZE; - ret = btrfs_check_data_free_space(inode, prealloc); + ret = btrfs_delalloc_reserve_space(inode, prealloc); if (ret) goto out_put; ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc, prealloc, prealloc, &alloc_hint); - if (ret) + if (ret) { + btrfs_delalloc_release_space(inode, prealloc); goto out_put; + } btrfs_free_reserved_data_space(inode, prealloc); + ret = btrfs_write_out_ino_cache(root, trans, path); out_put: iput(inode); +out_release: + btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved); out: - if (ret == 0) - ret = btrfs_write_out_ino_cache(root, trans, path); + trans->block_rsv = rsv; + trans->bytes_reserved = num_bytes; btrfs_free_path(path); return ret; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d42e6bf..cb10cb9 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -38,6 +38,7 @@ #include <linux/falloc.h> #include <linux/slab.h> #include <linux/ratelimit.h> +#include <linux/mount.h> #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -45,17 +46,17 @@ #include "btrfs_inode.h" #include "ioctl.h" #include "print-tree.h" -#include "volumes.h" #include "ordered-data.h" #include "xattr.h" #include "tree-log.h" +#include "volumes.h" #include "compression.h" #include "locking.h" #include "free-space-cache.h" #include "inode-map.h" struct btrfs_iget_args { - u64 ino; + struct btrfs_key *location; struct btrfs_root *root; }; @@ -93,6 +94,8 @@ static noinline int cow_file_range(struct inode *inode, struct page *locked_page, u64 start, u64 end, int *page_started, unsigned long *nr_written, int unlock); +static noinline int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode); static int btrfs_init_inode_security(struct btrfs_trans_handle *trans, struct inode *inode, struct inode *dir, @@ -340,6 +343,7 @@ static noinline int compress_file_range(struct inode *inode, int i; int will_compress; int compress_type = root->fs_info->compress_type; + int redirty = 0; /* if this is a small write inside eof, kick off a defragbot */ if (end <= BTRFS_I(inode)->disk_i_size && (end - start + 1) < 16 * 1024) @@ -393,11 +397,25 @@ again: (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) { WARN_ON(pages); pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); - BUG_ON(!pages); + if (!pages) { + /* just bail out to the uncompressed code */ + goto cont; + } if (BTRFS_I(inode)->force_compress) compress_type = BTRFS_I(inode)->force_compress; + /* + * we need to call clear_page_dirty_for_io on each + * page in the range. Otherwise applications with the file + * mmap'd can wander in and change the page contents while + * we are compressing them. + * + * If the compression fails for any reason, we set the pages + * dirty again later on. + */ + extent_range_clear_dirty_for_io(inode, start, end); + redirty = 1; ret = btrfs_compress_pages(compress_type, inode->i_mapping, start, total_compressed, pages, @@ -424,6 +442,7 @@ again: will_compress = 1; } } +cont: if (start == 0) { trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); @@ -534,6 +553,8 @@ cleanup_and_bail_uncompressed: __set_page_dirty_nobuffers(locked_page); /* unlocked later on in the async handlers */ } + if (redirty) + extent_range_redirty_for_io(inode, start, end); add_async_extent(async_cow, start, end - start + 1, 0, NULL, 0, BTRFS_COMPRESS_NONE); *num_added += 1; @@ -750,15 +771,6 @@ static u64 get_extent_allocation_hint(struct inode *inode, u64 start, return alloc_hint; } -static inline bool is_free_space_inode(struct btrfs_root *root, - struct inode *inode) -{ - if (root == root->fs_info->tree_root || - BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) - return true; - return false; -} - /* * when extent_io.c finds a delayed allocation range in the file, * the call backs end up in this code. The basic idea is to @@ -791,7 +803,7 @@ static noinline int cow_file_range(struct inode *inode, struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; int ret = 0; - BUG_ON(is_free_space_inode(root, inode)); + BUG_ON(btrfs_is_free_space_inode(root, inode)); trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); trans->block_rsv = &root->fs_info->delalloc_block_rsv; @@ -829,7 +841,7 @@ static noinline int cow_file_range(struct inode *inode, } BUG_ON(disk_num_bytes > - btrfs_super_total_bytes(&root->fs_info->super_copy)); + btrfs_super_total_bytes(root->fs_info->super_copy)); alloc_hint = get_extent_allocation_hint(inode, start, num_bytes); btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0); @@ -1070,9 +1082,10 @@ static noinline int run_delalloc_nocow(struct inode *inode, u64 ino = btrfs_ino(inode); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; - nolock = is_free_space_inode(root, inode); + nolock = btrfs_is_free_space_inode(root, inode); if (nolock) trans = btrfs_join_transaction_nolock(root); @@ -1291,15 +1304,16 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page, return ret; } -static int btrfs_split_extent_hook(struct inode *inode, - struct extent_state *orig, u64 split) +static void btrfs_split_extent_hook(struct inode *inode, + struct extent_state *orig, u64 split) { /* not delalloc, ignore it */ if (!(orig->state & EXTENT_DELALLOC)) - return 0; + return; - atomic_inc(&BTRFS_I(inode)->outstanding_extents); - return 0; + spin_lock(&BTRFS_I(inode)->lock); + BTRFS_I(inode)->outstanding_extents++; + spin_unlock(&BTRFS_I(inode)->lock); } /* @@ -1308,16 +1322,17 @@ static int btrfs_split_extent_hook(struct inode *inode, * extents, such as when we are doing sequential writes, so we can properly * account for the metadata space we'll need. */ -static int btrfs_merge_extent_hook(struct inode *inode, - struct extent_state *new, - struct extent_state *other) +static void btrfs_merge_extent_hook(struct inode *inode, + struct extent_state *new, + struct extent_state *other) { /* not delalloc, ignore it */ if (!(other->state & EXTENT_DELALLOC)) - return 0; + return; - atomic_dec(&BTRFS_I(inode)->outstanding_extents); - return 0; + spin_lock(&BTRFS_I(inode)->lock); + BTRFS_I(inode)->outstanding_extents--; + spin_unlock(&BTRFS_I(inode)->lock); } /* @@ -1325,8 +1340,8 @@ static int btrfs_merge_extent_hook(struct inode *inode, * bytes in this file, and to maintain the list of inodes that * have pending delalloc work to be done. */ -static int btrfs_set_bit_hook(struct inode *inode, - struct extent_state *state, int *bits) +static void btrfs_set_bit_hook(struct inode *inode, + struct extent_state *state, int *bits) { /* @@ -1337,12 +1352,15 @@ static int btrfs_set_bit_hook(struct inode *inode, if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - bool do_list = !is_free_space_inode(root, inode); + bool do_list = !btrfs_is_free_space_inode(root, inode); - if (*bits & EXTENT_FIRST_DELALLOC) + if (*bits & EXTENT_FIRST_DELALLOC) { *bits &= ~EXTENT_FIRST_DELALLOC; - else - atomic_inc(&BTRFS_I(inode)->outstanding_extents); + } else { + spin_lock(&BTRFS_I(inode)->lock); + BTRFS_I(inode)->outstanding_extents++; + spin_unlock(&BTRFS_I(inode)->lock); + } spin_lock(&root->fs_info->delalloc_lock); BTRFS_I(inode)->delalloc_bytes += len; @@ -1353,14 +1371,13 @@ static int btrfs_set_bit_hook(struct inode *inode, } spin_unlock(&root->fs_info->delalloc_lock); } - return 0; } /* * extent_io.c clear_bit_hook, see set_bit_hook for why */ -static int btrfs_clear_bit_hook(struct inode *inode, - struct extent_state *state, int *bits) +static void btrfs_clear_bit_hook(struct inode *inode, + struct extent_state *state, int *bits) { /* * set_bit and clear bit hooks normally require _irqsave/restore @@ -1370,12 +1387,15 @@ static int btrfs_clear_bit_hook(struct inode *inode, if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - bool do_list = !is_free_space_inode(root, inode); + bool do_list = !btrfs_is_free_space_inode(root, inode); - if (*bits & EXTENT_FIRST_DELALLOC) + if (*bits & EXTENT_FIRST_DELALLOC) { *bits &= ~EXTENT_FIRST_DELALLOC; - else if (!(*bits & EXTENT_DO_ACCOUNTING)) - atomic_dec(&BTRFS_I(inode)->outstanding_extents); + } else if (!(*bits & EXTENT_DO_ACCOUNTING)) { + spin_lock(&BTRFS_I(inode)->lock); + BTRFS_I(inode)->outstanding_extents--; + spin_unlock(&BTRFS_I(inode)->lock); + } if (*bits & EXTENT_DO_ACCOUNTING) btrfs_delalloc_release_metadata(inode, len); @@ -1394,7 +1414,6 @@ static int btrfs_clear_bit_hook(struct inode *inode, } spin_unlock(&root->fs_info->delalloc_lock); } - return 0; } /* @@ -1477,7 +1496,7 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; - if (is_free_space_inode(root, inode)) + if (btrfs_is_free_space_inode(root, inode)) ret = btrfs_bio_wq_end_io(root->fs_info, bio, 2); else ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); @@ -1644,7 +1663,8 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, int ret; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; path->leave_spinning = 1; @@ -1726,7 +1746,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) return 0; BUG_ON(!ordered_extent); - nolock = is_free_space_inode(root, inode); + nolock = btrfs_is_free_space_inode(root, inode); if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) { BUG_ON(!list_empty(&ordered_extent->list)); @@ -1738,7 +1758,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); trans->block_rsv = &root->fs_info->delalloc_block_rsv; - ret = btrfs_update_inode(trans, root, inode); + ret = btrfs_update_inode_fallback(trans, root, inode); BUG_ON(ret); } goto out; @@ -1787,18 +1807,18 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) &ordered_extent->list); ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent); - if (!ret) { - ret = btrfs_update_inode(trans, root, inode); + if (!ret || !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) { + ret = btrfs_update_inode_fallback(trans, root, inode); BUG_ON(ret); } ret = 0; out: - if (nolock) { - if (trans) - btrfs_end_transaction_nolock(trans, root); - } else { + if (root != root->fs_info->tree_root) btrfs_delalloc_release_metadata(inode, ordered_extent->len); - if (trans) + if (trans) { + if (nolock) + btrfs_end_transaction_nolock(trans, root); + else btrfs_end_transaction(trans, root); } @@ -1820,153 +1840,9 @@ static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, } /* - * When IO fails, either with EIO or csum verification fails, we - * try other mirrors that might have a good copy of the data. This - * io_failure_record is used to record state as we go through all the - * mirrors. If another mirror has good data, the page is set up to date - * and things continue. If a good mirror can't be found, the original - * bio end_io callback is called to indicate things have failed. - */ -struct io_failure_record { - struct page *page; - u64 start; - u64 len; - u64 logical; - unsigned long bio_flags; - int last_mirror; -}; - -static int btrfs_io_failed_hook(struct bio *failed_bio, - struct page *page, u64 start, u64 end, - struct extent_state *state) -{ - struct io_failure_record *failrec = NULL; - u64 private; - struct extent_map *em; - struct inode *inode = page->mapping->host; - struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree; - struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; - struct bio *bio; - int num_copies; - int ret; - int rw; - u64 logical; - - ret = get_state_private(failure_tree, start, &private); - if (ret) { - failrec = kmalloc(sizeof(*failrec), GFP_NOFS); - if (!failrec) - return -ENOMEM; - failrec->start = start; - failrec->len = end - start + 1; - failrec->last_mirror = 0; - failrec->bio_flags = 0; - - read_lock(&em_tree->lock); - em = lookup_extent_mapping(em_tree, start, failrec->len); - if (em->start > start || em->start + em->len < start) { - free_extent_map(em); - em = NULL; - } - read_unlock(&em_tree->lock); - - if (IS_ERR_OR_NULL(em)) { - kfree(failrec); - return -EIO; - } - logical = start - em->start; - logical = em->block_start + logical; - if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) { - logical = em->block_start; - failrec->bio_flags = EXTENT_BIO_COMPRESSED; - extent_set_compress_type(&failrec->bio_flags, - em->compress_type); - } - failrec->logical = logical; - free_extent_map(em); - set_extent_bits(failure_tree, start, end, EXTENT_LOCKED | - EXTENT_DIRTY, GFP_NOFS); - set_state_private(failure_tree, start, - (u64)(unsigned long)failrec); - } else { - failrec = (struct io_failure_record *)(unsigned long)private; - } - num_copies = btrfs_num_copies( - &BTRFS_I(inode)->root->fs_info->mapping_tree, - failrec->logical, failrec->len); - failrec->last_mirror++; - if (!state) { - spin_lock(&BTRFS_I(inode)->io_tree.lock); - state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree, - failrec->start, - EXTENT_LOCKED); - if (state && state->start != failrec->start) - state = NULL; - spin_unlock(&BTRFS_I(inode)->io_tree.lock); - } - if (!state || failrec->last_mirror > num_copies) { - set_state_private(failure_tree, failrec->start, 0); - clear_extent_bits(failure_tree, failrec->start, - failrec->start + failrec->len - 1, - EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); - kfree(failrec); - return -EIO; - } - bio = bio_alloc(GFP_NOFS, 1); - bio->bi_private = state; - bio->bi_end_io = failed_bio->bi_end_io; - bio->bi_sector = failrec->logical >> 9; - bio->bi_bdev = failed_bio->bi_bdev; - bio->bi_size = 0; - - bio_add_page(bio, page, failrec->len, start - page_offset(page)); - if (failed_bio->bi_rw & REQ_WRITE) - rw = WRITE; - else - rw = READ; - - ret = BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, - failrec->last_mirror, - failrec->bio_flags, 0); - return ret; -} - -/* - * each time an IO finishes, we do a fast check in the IO failure tree - * to see if we need to process or clean up an io_failure_record - */ -static int btrfs_clean_io_failures(struct inode *inode, u64 start) -{ - u64 private; - u64 private_failure; - struct io_failure_record *failure; - int ret; - - private = 0; - if (count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private, - (u64)-1, 1, EXTENT_DIRTY, 0)) { - ret = get_state_private(&BTRFS_I(inode)->io_failure_tree, - start, &private_failure); - if (ret == 0) { - failure = (struct io_failure_record *)(unsigned long) - private_failure; - set_state_private(&BTRFS_I(inode)->io_failure_tree, - failure->start, 0); - clear_extent_bits(&BTRFS_I(inode)->io_failure_tree, - failure->start, - failure->start + failure->len - 1, - EXTENT_DIRTY | EXTENT_LOCKED, - GFP_NOFS); - kfree(failure); - } - } - return 0; -} - -/* * when reads are done, we need to check csums to verify the data is correct - * if there's a match, we allow the bio to finish. If not, we go through - * the io_failure_record routines to find good copies + * if there's a match, we allow the bio to finish. If not, the code in + * extent_io.c will try to find good copies for us. */ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end, struct extent_state *state) @@ -2012,10 +1888,6 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end, kunmap_atomic(kaddr, KM_USER0); good: - /* if the io failure tree for this inode is non-empty, - * check to see if we've recovered from a failed IO - */ - btrfs_clean_io_failures(inode, start); return 0; zeroit: @@ -2080,89 +1952,6 @@ void btrfs_run_delayed_iputs(struct btrfs_root *root) up_read(&root->fs_info->cleanup_work_sem); } -/* - * calculate extra metadata reservation when snapshotting a subvolume - * contains orphan files. - */ -void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending, - u64 *bytes_to_reserve) -{ - struct btrfs_root *root; - struct btrfs_block_rsv *block_rsv; - u64 num_bytes; - int index; - - root = pending->root; - if (!root->orphan_block_rsv || list_empty(&root->orphan_list)) - return; - - block_rsv = root->orphan_block_rsv; - - /* orphan block reservation for the snapshot */ - num_bytes = block_rsv->size; - - /* - * after the snapshot is created, COWing tree blocks may use more - * space than it frees. So we should make sure there is enough - * reserved space. - */ - index = trans->transid & 0x1; - if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) { - num_bytes += block_rsv->size - - (block_rsv->reserved + block_rsv->freed[index]); - } - - *bytes_to_reserve += num_bytes; -} - -void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending) -{ - struct btrfs_root *root = pending->root; - struct btrfs_root *snap = pending->snap; - struct btrfs_block_rsv *block_rsv; - u64 num_bytes; - int index; - int ret; - - if (!root->orphan_block_rsv || list_empty(&root->orphan_list)) - return; - - /* refill source subvolume's orphan block reservation */ - block_rsv = root->orphan_block_rsv; - index = trans->transid & 0x1; - if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) { - num_bytes = block_rsv->size - - (block_rsv->reserved + block_rsv->freed[index]); - ret = btrfs_block_rsv_migrate(&pending->block_rsv, - root->orphan_block_rsv, - num_bytes); - BUG_ON(ret); - } - - /* setup orphan block reservation for the snapshot */ - block_rsv = btrfs_alloc_block_rsv(snap); - BUG_ON(!block_rsv); - - btrfs_add_durable_block_rsv(root->fs_info, block_rsv); - snap->orphan_block_rsv = block_rsv; - - num_bytes = root->orphan_block_rsv->size; - ret = btrfs_block_rsv_migrate(&pending->block_rsv, - block_rsv, num_bytes); - BUG_ON(ret); - -#if 0 - /* insert orphan item for the snapshot */ - WARN_ON(!root->orphan_item_inserted); - ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root, - snap->root_key.objectid); - BUG_ON(ret); - snap->orphan_item_inserted = 1; -#endif -} - enum btrfs_orphan_cleanup_state { ORPHAN_CLEANUP_STARTED = 1, ORPHAN_CLEANUP_DONE = 2, @@ -2214,7 +2003,8 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) if (!root->orphan_block_rsv) { block_rsv = btrfs_alloc_block_rsv(root); - BUG_ON(!block_rsv); + if (!block_rsv) + return -ENOMEM; } spin_lock(&root->orphan_lock); @@ -2247,9 +2037,6 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) } spin_unlock(&root->orphan_lock); - if (block_rsv) - btrfs_add_durable_block_rsv(root->fs_info, block_rsv); - /* grab metadata reservation from transaction handle */ if (reserve) { ret = btrfs_orphan_reserve_metadata(trans, inode); @@ -2259,7 +2046,7 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) /* insert an orphan item to track this unlinked/truncated file */ if (insert >= 1) { ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); - BUG_ON(ret); + BUG_ON(ret && ret != -EEXIST); } /* insert an orphan item to track subvolume contains orphan files */ @@ -2316,6 +2103,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) struct btrfs_key key, found_key; struct btrfs_trans_handle *trans; struct inode *inode; + u64 last_objectid = 0; int ret = 0, nr_unlink = 0, nr_truncate = 0; if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) @@ -2367,41 +2155,81 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) * crossing root thing. we store the inode number in the * offset of the orphan item. */ + + if (found_key.offset == last_objectid) { + printk(KERN_ERR "btrfs: Error removing orphan entry, " + "stopping orphan cleanup\n"); + ret = -EINVAL; + goto out; + } + + last_objectid = found_key.offset; + found_key.objectid = found_key.offset; found_key.type = BTRFS_INODE_ITEM_KEY; found_key.offset = 0; inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL); - if (IS_ERR(inode)) { - ret = PTR_ERR(inode); + ret = PTR_RET(inode); + if (ret && ret != -ESTALE) goto out; - } - /* - * add this inode to the orphan list so btrfs_orphan_del does - * the proper thing when we hit it - */ - spin_lock(&root->orphan_lock); - list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list); - spin_unlock(&root->orphan_lock); + if (ret == -ESTALE && root == root->fs_info->tree_root) { + struct btrfs_root *dead_root; + struct btrfs_fs_info *fs_info = root->fs_info; + int is_dead_root = 0; + /* + * this is an orphan in the tree root. Currently these + * could come from 2 sources: + * a) a snapshot deletion in progress + * b) a free space cache inode + * We need to distinguish those two, as the snapshot + * orphan must not get deleted. + * find_dead_roots already ran before us, so if this + * is a snapshot deletion, we should find the root + * in the dead_roots list + */ + spin_lock(&fs_info->trans_lock); + list_for_each_entry(dead_root, &fs_info->dead_roots, + root_list) { + if (dead_root->root_key.objectid == + found_key.objectid) { + is_dead_root = 1; + break; + } + } + spin_unlock(&fs_info->trans_lock); + if (is_dead_root) { + /* prevent this orphan from being found again */ + key.offset = found_key.objectid - 1; + continue; + } + } /* - * if this is a bad inode, means we actually succeeded in - * removing the inode, but not the orphan record, which means - * we need to manually delete the orphan since iput will just - * do a destroy_inode + * Inode is already gone but the orphan item is still there, + * kill the orphan item. */ - if (is_bad_inode(inode)) { - trans = btrfs_start_transaction(root, 0); + if (ret == -ESTALE) { + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { ret = PTR_ERR(trans); goto out; } - btrfs_orphan_del(trans, inode); + ret = btrfs_del_orphan_item(trans, root, + found_key.objectid); + BUG_ON(ret); btrfs_end_transaction(trans, root); - iput(inode); continue; } + /* + * add this inode to the orphan list so btrfs_orphan_del does + * the proper thing when we hit it + */ + spin_lock(&root->orphan_lock); + list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list); + spin_unlock(&root->orphan_lock); + /* if we have links, this was a truncate, lets do that */ if (inode->i_nlink) { if (!S_ISREG(inode->i_mode)) { @@ -2410,7 +2238,14 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) continue; } nr_truncate++; + /* + * Need to hold the imutex for reservation purposes, not + * a huge deal here but I have a WARN_ON in + * btrfs_delalloc_reserve_space to catch offenders. + */ + mutex_lock(&inode->i_mutex); ret = btrfs_truncate(inode); + mutex_unlock(&inode->i_mutex); } else { nr_unlink++; } @@ -2420,6 +2255,9 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) if (ret) goto out; } + /* release the path since we're done with it */ + btrfs_release_path(path); + root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; if (root->orphan_block_rsv) @@ -2516,7 +2354,9 @@ static void btrfs_read_locked_inode(struct inode *inode) filled = true; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + goto make_bad; + path->leave_spinning = 1; memcpy(&location, &BTRFS_I(inode)->location, sizeof(location)); @@ -2531,15 +2371,8 @@ static void btrfs_read_locked_inode(struct inode *inode) inode_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item); - if (!leaf->map_token) - map_private_extent_buffer(leaf, (unsigned long)inode_item, - sizeof(struct btrfs_inode_item), - &leaf->map_token, &leaf->kaddr, - &leaf->map_start, &leaf->map_len, - KM_USER1); - inode->i_mode = btrfs_inode_mode(leaf, inode_item); - inode->i_nlink = btrfs_inode_nlink(leaf, inode_item); + set_nlink(inode, btrfs_inode_nlink(leaf, inode_item)); inode->i_uid = btrfs_inode_uid(leaf, inode_item); inode->i_gid = btrfs_inode_gid(leaf, inode_item); btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item)); @@ -2575,11 +2408,6 @@ cache_acl: if (!maybe_acls) cache_no_acl(inode); - if (leaf->map_token) { - unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); - leaf->map_token = NULL; - } - btrfs_free_path(path); switch (inode->i_mode & S_IFMT) { @@ -2624,13 +2452,6 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, struct btrfs_inode_item *item, struct inode *inode) { - if (!leaf->map_token) - map_private_extent_buffer(leaf, (unsigned long)item, - sizeof(struct btrfs_inode_item), - &leaf->map_token, &leaf->kaddr, - &leaf->map_start, &leaf->map_len, - KM_USER1); - btrfs_set_inode_uid(leaf, item, inode->i_uid); btrfs_set_inode_gid(leaf, item, inode->i_gid); btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size); @@ -2659,17 +2480,12 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, btrfs_set_inode_rdev(leaf, item, inode->i_rdev); btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags); btrfs_set_inode_block_group(leaf, item, 0); - - if (leaf->map_token) { - unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); - leaf->map_token = NULL; - } } /* * copy everything in the in-memory inode into the btree. */ -noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, +static noinline int btrfs_update_inode_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode) { struct btrfs_inode_item *inode_item; @@ -2677,21 +2493,6 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, struct extent_buffer *leaf; int ret; - /* - * If the inode is a free space inode, we can deadlock during commit - * if we put it into the delayed code. - * - * The data relocation inode should also be directly updated - * without delay - */ - if (!is_free_space_inode(root, inode) - && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID) { - ret = btrfs_delayed_update_inode(trans, root, inode); - if (!ret) - btrfs_set_inode_last_trans(trans, inode); - return ret; - } - path = btrfs_alloc_path(); if (!path) return -ENOMEM; @@ -2720,6 +2521,43 @@ failed: } /* + * copy everything in the in-memory inode into the btree. + */ +noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode) +{ + int ret; + + /* + * If the inode is a free space inode, we can deadlock during commit + * if we put it into the delayed code. + * + * The data relocation inode should also be directly updated + * without delay + */ + if (!btrfs_is_free_space_inode(root, inode) + && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID) { + ret = btrfs_delayed_update_inode(trans, root, inode); + if (!ret) + btrfs_set_inode_last_trans(trans, inode); + return ret; + } + + return btrfs_update_inode_item(trans, root, inode); +} + +static noinline int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode) +{ + int ret; + + ret = btrfs_update_inode(trans, root, inode); + if (ret == -ENOSPC) + return btrfs_update_inode_item(trans, root, inode); + return ret; +} + +/* * unlink helper that gets used here in inode.c and in the tree logging * recovery code. It remove a link in a directory with a given name, and * also drops the back refs in the inode to the directory @@ -2857,7 +2695,16 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, u64 ino = btrfs_ino(inode); u64 dir_ino = btrfs_ino(dir); - trans = btrfs_start_transaction(root, 10); + /* + * 1 for the possible orphan item + * 1 for the dir item + * 1 for the dir index + * 1 for the inode ref + * 1 for the inode ref in the tree log + * 2 for the dir entries in the log + * 1 for the inode + */ + trans = btrfs_start_transaction(root, 8); if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC) return trans; @@ -2880,7 +2727,8 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, return ERR_PTR(-ENOMEM); } - trans = btrfs_start_transaction(root, 0); + /* 1 for the orphan item */ + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { btrfs_free_path(path); root->fs_info->enospc_unlink = 0; @@ -2985,6 +2833,12 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, err = 0; out: btrfs_free_path(path); + /* Migrate the orphan reservation over */ + if (!err) + err = btrfs_block_rsv_migrate(trans->block_rsv, + &root->fs_info->global_block_rsv, + trans->bytes_reserved); + if (err) { btrfs_end_transaction(trans, root); root->fs_info->enospc_unlink = 0; @@ -2999,6 +2853,9 @@ static void __unlink_end_trans(struct btrfs_trans_handle *trans, struct btrfs_root *root) { if (trans->block_rsv == &root->fs_info->global_block_rsv) { + btrfs_block_rsv_release(root, trans->block_rsv, + trans->bytes_reserved); + trans->block_rsv = &root->fs_info->trans_block_rsv; BUG_ON(!root->fs_info->enospc_unlink); root->fs_info->enospc_unlink = 0; } @@ -3021,13 +2878,16 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry) ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode, dentry->d_name.name, dentry->d_name.len); - BUG_ON(ret); + if (ret) + goto out; if (inode->i_nlink == 0) { ret = btrfs_orphan_add(trans, inode); - BUG_ON(ret); + if (ret) + goto out; } +out: nr = trans->blocks_used; __unlink_end_trans(trans, root); btrfs_btree_balance_dirty(root, nr); @@ -3170,6 +3030,11 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY); + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + path->reada = -1; + if (root->ref_cows || root == root->fs_info->tree_root) btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0); @@ -3182,10 +3047,6 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, if (min_type == 0 && root == BTRFS_I(inode)->root) btrfs_kill_delayed_inode_items(inode); - path = btrfs_alloc_path(); - BUG_ON(!path); - path->reada = -1; - key.objectid = ino; key.offset = (u64)-1; key.type = (u8)-1; @@ -3386,6 +3247,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from) pgoff_t index = from >> PAGE_CACHE_SHIFT; unsigned offset = from & (PAGE_CACHE_SIZE-1); struct page *page; + gfp_t mask = btrfs_alloc_write_mask(mapping); int ret = 0; u64 page_start; u64 page_end; @@ -3398,7 +3260,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from) ret = -ENOMEM; again: - page = grab_cache_page(mapping, index); + page = find_or_create_page(mapping, index, mask); if (!page) { btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); goto out; @@ -3519,7 +3381,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) u64 hint_byte = 0; hole_size = last_byte - cur_offset; - trans = btrfs_start_transaction(root, 2); + trans = btrfs_start_transaction(root, 3); if (IS_ERR(trans)) { err = PTR_ERR(trans); break; @@ -3528,19 +3390,26 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) err = btrfs_drop_extents(trans, inode, cur_offset, cur_offset + hole_size, &hint_byte, 1); - if (err) + if (err) { + btrfs_update_inode(trans, root, inode); + btrfs_end_transaction(trans, root); break; + } err = btrfs_insert_file_extent(trans, root, btrfs_ino(inode), cur_offset, 0, 0, hole_size, 0, hole_size, 0, 0, 0); - if (err) + if (err) { + btrfs_update_inode(trans, root, inode); + btrfs_end_transaction(trans, root); break; + } btrfs_drop_extent_cache(inode, hole_start, last_byte - 1, 0); + btrfs_update_inode(trans, root, inode); btrfs_end_transaction(trans, root); } free_extent_map(em); @@ -3558,6 +3427,8 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) static int btrfs_setsize(struct inode *inode, loff_t newsize) { + struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_trans_handle *trans; loff_t oldsize = i_size_read(inode); int ret; @@ -3565,16 +3436,19 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return 0; if (newsize > oldsize) { - i_size_write(inode, newsize); - btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); truncate_pagecache(inode, oldsize, newsize); ret = btrfs_cont_expand(inode, oldsize, newsize); - if (ret) { - btrfs_setsize(inode, oldsize); + if (ret) return ret; - } - mark_inode_dirty(inode); + trans = btrfs_start_transaction(root, 1); + if (IS_ERR(trans)) + return PTR_ERR(trans); + + i_size_write(inode, newsize); + btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); + ret = btrfs_update_inode(trans, root, inode); + btrfs_end_transaction_throttle(trans, root); } else { /* @@ -3614,9 +3488,9 @@ static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) if (attr->ia_valid) { setattr_copy(inode, attr); - mark_inode_dirty(inode); + err = btrfs_dirty_inode(inode); - if (attr->ia_valid & ATTR_MODE) + if (!err && attr->ia_valid & ATTR_MODE) err = btrfs_acl_chmod(inode); } @@ -3627,6 +3501,8 @@ void btrfs_evict_inode(struct inode *inode) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_block_rsv *rsv, *global_rsv; + u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); unsigned long nr; int ret; @@ -3634,7 +3510,7 @@ void btrfs_evict_inode(struct inode *inode) truncate_inode_pages(&inode->i_data, 0); if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 || - is_free_space_inode(root, inode))) + btrfs_is_free_space_inode(root, inode))) goto no_delete; if (is_bad_inode(inode)) { @@ -3642,7 +3518,8 @@ void btrfs_evict_inode(struct inode *inode) goto no_delete; } /* do we really want it for ->i_nlink > 0 and zero btrfs_root_refs? */ - btrfs_wait_ordered_range(inode, 0, (u64)-1); + if (!special_file(inode->i_mode)) + btrfs_wait_ordered_range(inode, 0, (u64)-1); if (root->fs_info->log_root_recovering) { BUG_ON(!list_empty(&BTRFS_I(inode)->i_orphan)); @@ -3654,22 +3531,55 @@ void btrfs_evict_inode(struct inode *inode) goto no_delete; } + rsv = btrfs_alloc_block_rsv(root); + if (!rsv) { + btrfs_orphan_del(NULL, inode); + goto no_delete; + } + rsv->size = min_size; + global_rsv = &root->fs_info->global_block_rsv; + btrfs_i_size_write(inode, 0); + /* + * This is a bit simpler than btrfs_truncate since + * + * 1) We've already reserved our space for our orphan item in the + * unlink. + * 2) We're going to delete the inode item, so we don't need to update + * it at all. + * + * So we just need to reserve some slack space in case we add bytes when + * doing the truncate. + */ while (1) { - trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); - trans->block_rsv = root->orphan_block_rsv; + ret = btrfs_block_rsv_refill_noflush(root, rsv, min_size); + + /* + * 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) + ret = btrfs_block_rsv_migrate(global_rsv, rsv, min_size); - ret = btrfs_block_rsv_check(trans, root, - root->orphan_block_rsv, 0, 5); if (ret) { - BUG_ON(ret != -EAGAIN); - ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); - continue; + printk(KERN_WARNING "Could not get space for a " + "delete, will truncate on mount %d\n", ret); + btrfs_orphan_del(NULL, inode); + btrfs_free_block_rsv(root, rsv); + goto no_delete; + } + + trans = btrfs_start_transaction(root, 0); + if (IS_ERR(trans)) { + btrfs_orphan_del(NULL, inode); + btrfs_free_block_rsv(root, rsv); + goto no_delete; } + trans->block_rsv = rsv; + ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0); if (ret != -EAGAIN) break; @@ -3678,14 +3588,17 @@ void btrfs_evict_inode(struct inode *inode) btrfs_end_transaction(trans, root); trans = NULL; btrfs_btree_balance_dirty(root, nr); - } + btrfs_free_block_rsv(root, rsv); + if (ret == 0) { + trans->block_rsv = root->orphan_block_rsv; ret = btrfs_orphan_del(trans, inode); BUG_ON(ret); } + trans->block_rsv = &root->fs_info->trans_block_rsv; if (!(root == root->fs_info->tree_root || root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) btrfs_return_ino(root, btrfs_ino(inode)); @@ -3713,7 +3626,8 @@ static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry, int ret = 0; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; di = btrfs_lookup_dir_item(NULL, root, path, btrfs_ino(dir), name, namelen, 0); @@ -3934,7 +3848,9 @@ again: static int btrfs_init_locked_inode(struct inode *inode, void *p) { struct btrfs_iget_args *args = p; - inode->i_ino = args->ino; + inode->i_ino = args->location->objectid; + memcpy(&BTRFS_I(inode)->location, args->location, + sizeof(*args->location)); BTRFS_I(inode)->root = args->root; btrfs_set_inode_space_info(args->root, inode); return 0; @@ -3943,20 +3859,20 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p) static int btrfs_find_actor(struct inode *inode, void *opaque) { struct btrfs_iget_args *args = opaque; - return args->ino == btrfs_ino(inode) && + return args->location->objectid == BTRFS_I(inode)->location.objectid && args->root == BTRFS_I(inode)->root; } static struct inode *btrfs_iget_locked(struct super_block *s, - u64 objectid, + struct btrfs_key *location, struct btrfs_root *root) { struct inode *inode; struct btrfs_iget_args args; - args.ino = objectid; + args.location = location; args.root = root; - inode = iget5_locked(s, objectid, btrfs_find_actor, + inode = iget5_locked(s, location->objectid, btrfs_find_actor, btrfs_init_locked_inode, (void *)&args); return inode; @@ -3970,18 +3886,22 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, { struct inode *inode; - inode = btrfs_iget_locked(s, location->objectid, root); + inode = btrfs_iget_locked(s, location, root); if (!inode) return ERR_PTR(-ENOMEM); if (inode->i_state & I_NEW) { - BTRFS_I(inode)->root = root; - memcpy(&BTRFS_I(inode)->location, location, sizeof(*location)); btrfs_read_locked_inode(inode); - inode_tree_add(inode); - unlock_new_inode(inode); - if (new) - *new = 1; + if (!is_bad_inode(inode)) { + inode_tree_add(inode); + unlock_new_inode(inode); + if (new) + *new = 1; + } else { + unlock_new_inode(inode); + iput(inode); + inode = ERR_PTR(-ESTALE); + } } return inode; @@ -4016,12 +3936,20 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry) struct btrfs_root *sub_root = root; struct btrfs_key location; int index; - int ret; + int ret = 0; if (dentry->d_name.len > BTRFS_NAME_LEN) return ERR_PTR(-ENAMETOOLONG); - ret = btrfs_inode_by_name(dir, dentry, &location); + if (unlikely(d_need_lookup(dentry))) { + memcpy(&location, dentry->d_fsdata, sizeof(struct btrfs_key)); + kfree(dentry->d_fsdata); + dentry->d_fsdata = NULL; + /* This thing is hashed, drop it for now */ + d_drop(dentry); + } else { + ret = btrfs_inode_by_name(dir, dentry, &location); + } if (ret < 0) return ERR_PTR(ret); @@ -4076,16 +4004,24 @@ static int btrfs_dentry_delete(const struct dentry *dentry) return 0; } +static void btrfs_dentry_release(struct dentry *dentry) +{ + if (dentry->d_fsdata) + kfree(dentry->d_fsdata); +} + static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd) { - struct inode *inode; - - inode = btrfs_lookup_dentry(dir, dentry); - if (IS_ERR(inode)) - return ERR_CAST(inode); + struct dentry *ret; - return d_splice_alias(inode, dentry); + ret = d_splice_alias(btrfs_lookup_dentry(dir, dentry), dentry); + if (unlikely(d_need_lookup(dentry))) { + spin_lock(&dentry->d_lock); + dentry->d_flags &= ~DCACHE_NEED_LOOKUP; + spin_unlock(&dentry->d_lock); + } + return ret; } unsigned char btrfs_filetype_table[] = { @@ -4104,6 +4040,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, struct btrfs_path *path; struct list_head ins_list; struct list_head del_list; + struct qstr q; int ret; struct extent_buffer *leaf; int slot; @@ -4194,6 +4131,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, while (di_cur < di_total) { struct btrfs_key location; + struct dentry *tmp; if (verify_dir_item(root, leaf, di)) break; @@ -4214,6 +4152,33 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)]; btrfs_dir_item_key_to_cpu(leaf, di, &location); + q.name = name_ptr; + q.len = name_len; + q.hash = full_name_hash(q.name, q.len); + tmp = d_lookup(filp->f_dentry, &q); + if (!tmp) { + struct btrfs_key *newkey; + + newkey = kzalloc(sizeof(struct btrfs_key), + GFP_NOFS); + if (!newkey) + goto no_dentry; + tmp = d_alloc(filp->f_dentry, &q); + if (!tmp) { + kfree(newkey); + dput(tmp); + goto no_dentry; + } + memcpy(newkey, &location, + sizeof(struct btrfs_key)); + tmp->d_fsdata = newkey; + tmp->d_flags |= DCACHE_NEED_LOOKUP; + d_rehash(tmp); + dput(tmp); + } else { + dput(tmp); + } +no_dentry: /* is this a reference to our own snapshot? If so * skip it */ @@ -4278,7 +4243,7 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) if (BTRFS_I(inode)->dummy_inode) return 0; - if (btrfs_fs_closing(root->fs_info) && is_free_space_inode(root, inode)) + if (btrfs_fs_closing(root->fs_info) && btrfs_is_free_space_inode(root, inode)) nolock = true; if (wbc->sync_mode == WB_SYNC_ALL) { @@ -4302,42 +4267,80 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) * FIXME, needs more benchmarking...there are no reasons other than performance * to keep or drop this code. */ -void btrfs_dirty_inode(struct inode *inode, int flags) +int btrfs_dirty_inode(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; int ret; if (BTRFS_I(inode)->dummy_inode) - return; + return 0; trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); ret = btrfs_update_inode(trans, root, inode); if (ret && ret == -ENOSPC) { /* whoops, lets try again with the full transaction */ btrfs_end_transaction(trans, root); trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) { - printk_ratelimited(KERN_ERR "btrfs: fail to " - "dirty inode %llu error %ld\n", - (unsigned long long)btrfs_ino(inode), - PTR_ERR(trans)); - return; - } + if (IS_ERR(trans)) + return PTR_ERR(trans); ret = btrfs_update_inode(trans, root, inode); - if (ret) { - printk_ratelimited(KERN_ERR "btrfs: fail to " - "dirty inode %llu error %d\n", - (unsigned long long)btrfs_ino(inode), - ret); - } } btrfs_end_transaction(trans, root); if (BTRFS_I(inode)->delayed_node) btrfs_balance_delayed_items(root); + + return ret; +} + +/* + * This is a copy of file_update_time. We need this so we can return error on + * ENOSPC for updating the inode in the case of file write and mmap writes. + */ +int btrfs_update_time(struct file *file) +{ + struct inode *inode = file->f_path.dentry->d_inode; + struct timespec now; + int ret; + enum { S_MTIME = 1, S_CTIME = 2, S_VERSION = 4 } sync_it = 0; + + /* First try to exhaust all avenues to not sync */ + if (IS_NOCMTIME(inode)) + return 0; + + now = current_fs_time(inode->i_sb); + if (!timespec_equal(&inode->i_mtime, &now)) + sync_it = S_MTIME; + + if (!timespec_equal(&inode->i_ctime, &now)) + sync_it |= S_CTIME; + + if (IS_I_VERSION(inode)) + sync_it |= S_VERSION; + + if (!sync_it) + return 0; + + /* Finally allowed to write? Takes lock. */ + if (mnt_want_write_file(file)) + return 0; + + /* Only change inode inside the lock region */ + if (sync_it & S_VERSION) + inode_inc_iversion(inode); + if (sync_it & S_CTIME) + inode->i_ctime = now; + if (sync_it & S_MTIME) + inode->i_mtime = now; + ret = btrfs_dirty_inode(inode); + if (!ret) + mark_inode_dirty_sync(inode); + mnt_drop_write(file->f_path.mnt); + return ret; } /* @@ -4439,7 +4442,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, int owner; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return ERR_PTR(-ENOMEM); inode = new_inode(root->fs_info->sb); if (!inode) { @@ -4474,7 +4478,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, inode->i_generation = BTRFS_I(inode)->generation; btrfs_set_inode_space_info(root, inode); - if (mode & S_IFDIR) + if (S_ISDIR(mode)) owner = 0; else owner = 1; @@ -4519,7 +4523,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, btrfs_inherit_iflags(inode, dir); - if ((mode & S_IFREG)) { + if (S_ISREG(mode)) { if (btrfs_test_opt(root, NODATASUM)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; if (btrfs_test_opt(root, NODATACOW) || @@ -4601,10 +4605,6 @@ static int btrfs_add_nondir(struct btrfs_trans_handle *trans, int err = btrfs_add_link(trans, dir, inode, dentry->d_name.name, dentry->d_name.len, backref, index); - if (!err) { - d_instantiate(dentry, inode); - return 0; - } if (err > 0) err = -EEXIST; return err; @@ -4652,13 +4652,21 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + + inode->i_op = &btrfs_special_inode_operations; err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { - inode->i_op = &btrfs_special_inode_operations; init_special_inode(inode, inode->i_mode, rdev); btrfs_update_inode(trans, root, inode); + d_instantiate(dentry, inode); } out_unlock: nr = trans->blocks_used; @@ -4710,15 +4718,23 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + inode->i_fop = &btrfs_file_operations; + inode->i_op = &btrfs_file_inode_operations; + err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { inode->i_mapping->a_ops = &btrfs_aops; inode->i_mapping->backing_dev_info = &root->fs_info->bdi; - inode->i_fop = &btrfs_file_operations; - inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; + d_instantiate(dentry, inode); } out_unlock: nr = trans->blocks_used; @@ -4773,11 +4789,11 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, if (err) { drop_inode = 1; } else { - struct dentry *parent = dget_parent(dentry); + struct dentry *parent = dentry->d_parent; err = btrfs_update_inode(trans, root, inode); BUG_ON(err); + d_instantiate(dentry, inode); btrfs_log_new_name(trans, inode, NULL, parent); - dput(parent); } nr = trans->blocks_used; @@ -5757,8 +5773,7 @@ again: if (test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags)) { ret = btrfs_ordered_update_i_size(inode, 0, ordered); if (!ret) - ret = btrfs_update_inode(trans, root, inode); - err = ret; + err = btrfs_update_inode_fallback(trans, root, inode); goto out; } @@ -5795,8 +5810,8 @@ again: add_pending_csums(trans, inode, ordered->file_offset, &ordered->list); ret = btrfs_ordered_update_i_size(inode, 0, ordered); - if (!ret) - btrfs_update_inode(trans, root, inode); + if (!ret || !test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags)) + btrfs_update_inode_fallback(trans, root, inode); ret = 0; out_unlock: unlock_extent_cached(&BTRFS_I(inode)->io_tree, ordered->file_offset, @@ -6251,7 +6266,7 @@ int btrfs_readpage(struct file *file, struct page *page) { struct extent_io_tree *tree; tree = &BTRFS_I(page->mapping->host)->io_tree; - return extent_read_full_page(tree, page, btrfs_get_extent); + return extent_read_full_page(tree, page, btrfs_get_extent, 0); } static int btrfs_writepage(struct page *page, struct writeback_control *wbc) @@ -6402,7 +6417,12 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) u64 page_start; u64 page_end; + /* Need this to keep space reservations serialized */ + mutex_lock(&inode->i_mutex); ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); + mutex_unlock(&inode->i_mutex); + if (!ret) + ret = btrfs_update_time(vma->vm_file); if (ret) { if (ret == -ENOMEM) ret = VM_FAULT_OOM; @@ -6503,6 +6523,7 @@ static int btrfs_truncate(struct inode *inode) struct btrfs_trans_handle *trans; unsigned long nr; u64 mask = root->sectorsize - 1; + u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); ret = btrfs_truncate_page(inode->i_mapping, inode->i_size); if (ret) @@ -6550,19 +6571,23 @@ static int btrfs_truncate(struct inode *inode) rsv = btrfs_alloc_block_rsv(root); if (!rsv) return -ENOMEM; - btrfs_add_durable_block_rsv(root->fs_info, rsv); + rsv->size = min_size; + /* + * 1 for the truncate slack space + * 1 for the orphan item we're going to add + * 1 for the orphan item deletion + * 1 for updating the inode. + */ trans = btrfs_start_transaction(root, 4); if (IS_ERR(trans)) { err = PTR_ERR(trans); goto out; } - /* - * Reserve space for the truncate process. Truncate should be adding - * space, but if there are snapshots it may end up using space. - */ - ret = btrfs_truncate_reserve_metadata(trans, root, rsv); + /* Migrate the slack space for the truncate to our reserve */ + ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, + min_size); BUG_ON(ret); ret = btrfs_orphan_add(trans, inode); @@ -6571,21 +6596,6 @@ static int btrfs_truncate(struct inode *inode) goto out; } - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); - - /* - * Ok so we've already migrated our bytes over for the truncate, so here - * just reserve the one slot we need for updating the inode. - */ - trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) { - err = PTR_ERR(trans); - goto out; - } - trans->block_rsv = rsv; - /* * setattr is responsible for setting the ordered_data_close flag, * but that is only tested during the last file release. That @@ -6607,20 +6617,31 @@ static int btrfs_truncate(struct inode *inode) btrfs_add_ordered_operation(trans, root, inode); while (1) { + ret = btrfs_block_rsv_refill(root, rsv, min_size); + if (ret) { + /* + * This can only happen with the original transaction we + * started above, every other time we shouldn't have a + * transaction started yet. + */ + if (ret == -EAGAIN) + goto end_trans; + err = ret; + break; + } + if (!trans) { - trans = btrfs_start_transaction(root, 3); + /* Just need the 1 for updating the inode */ + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { - err = PTR_ERR(trans); - goto out; + ret = err = PTR_ERR(trans); + trans = NULL; + break; } - - ret = btrfs_truncate_reserve_metadata(trans, root, - rsv); - BUG_ON(ret); - - trans->block_rsv = rsv; } + trans->block_rsv = rsv; + ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, BTRFS_EXTENT_DATA_KEY); @@ -6635,7 +6656,7 @@ static int btrfs_truncate(struct inode *inode) err = ret; break; } - +end_trans: nr = trans->blocks_used; btrfs_end_transaction(trans, root); trans = NULL; @@ -6655,14 +6676,16 @@ static int btrfs_truncate(struct inode *inode) ret = btrfs_orphan_del(NULL, inode); } - trans->block_rsv = &root->fs_info->trans_block_rsv; - ret = btrfs_update_inode(trans, root, inode); - if (ret && !err) - err = ret; + if (trans) { + trans->block_rsv = &root->fs_info->trans_block_rsv; + ret = btrfs_update_inode(trans, root, inode); + if (ret && !err) + err = ret; - nr = trans->blocks_used; - ret = btrfs_end_transaction_throttle(trans, root); - btrfs_btree_balance_dirty(root, nr); + nr = trans->blocks_used; + ret = btrfs_end_transaction_throttle(trans, root); + btrfs_btree_balance_dirty(root, nr); + } out: btrfs_free_block_rsv(root, rsv); @@ -6690,7 +6713,7 @@ int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, inode->i_op = &btrfs_dir_inode_operations; inode->i_fop = &btrfs_dir_file_operations; - inode->i_nlink = 1; + set_nlink(inode, 1); btrfs_i_size_write(inode, 0); err = btrfs_update_inode(trans, new_root, inode); @@ -6700,19 +6723,6 @@ int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, return 0; } -/* helper function for file defrag and space balancing. This - * forces readahead on a given range of bytes in an inode - */ -unsigned long btrfs_force_ra(struct address_space *mapping, - struct file_ra_state *ra, struct file *file, - pgoff_t offset, pgoff_t last_index) -{ - pgoff_t req_size = last_index - offset + 1; - - page_cache_sync_readahead(mapping, ra, file, offset, req_size); - return offset + req_size; -} - struct inode *btrfs_alloc_inode(struct super_block *sb) { struct btrfs_inode *ei; @@ -6730,19 +6740,21 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) ei->last_sub_trans = 0; ei->logged_trans = 0; ei->delalloc_bytes = 0; - ei->reserved_bytes = 0; ei->disk_i_size = 0; ei->flags = 0; + ei->csum_bytes = 0; ei->index_cnt = (u64)-1; ei->last_unlink_trans = 0; - atomic_set(&ei->outstanding_extents, 0); - atomic_set(&ei->reserved_extents, 0); + spin_lock_init(&ei->lock); + ei->outstanding_extents = 0; + ei->reserved_extents = 0; ei->ordered_data_close = 0; ei->orphan_meta_reserved = 0; ei->dummy_inode = 0; ei->in_defrag = 0; + ei->delalloc_meta_reserved = 0; ei->force_compress = BTRFS_COMPRESS_NONE; ei->delayed_node = NULL; @@ -6775,8 +6787,10 @@ void btrfs_destroy_inode(struct inode *inode) WARN_ON(!list_empty(&inode->i_dentry)); WARN_ON(inode->i_data.nrpages); - WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents)); - WARN_ON(atomic_read(&BTRFS_I(inode)->reserved_extents)); + WARN_ON(BTRFS_I(inode)->outstanding_extents); + WARN_ON(BTRFS_I(inode)->reserved_extents); + WARN_ON(BTRFS_I(inode)->delalloc_bytes); + WARN_ON(BTRFS_I(inode)->csum_bytes); /* * This can happen where we create an inode, but somebody else also @@ -6831,7 +6845,7 @@ int btrfs_drop_inode(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; if (btrfs_root_refs(&root->root_item) == 0 && - !is_free_space_inode(root, inode)) + !btrfs_is_free_space_inode(root, inode)) return 1; else return generic_drop_inode(inode); @@ -6900,11 +6914,13 @@ static int btrfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) { struct inode *inode = dentry->d_inode; + u32 blocksize = inode->i_sb->s_blocksize; + generic_fillattr(inode, stat); - stat->dev = BTRFS_I(inode)->root->anon_super.s_dev; + stat->dev = BTRFS_I(inode)->root->anon_dev; stat->blksize = PAGE_CACHE_SIZE; - stat->blocks = (inode_get_bytes(inode) + - BTRFS_I(inode)->delalloc_bytes) >> 9; + stat->blocks = (ALIGN(inode_get_bytes(inode), blocksize) + + ALIGN(BTRFS_I(inode)->delalloc_bytes, blocksize)) >> 9; return 0; } @@ -7069,9 +7085,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, BUG_ON(ret); if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { - struct dentry *parent = dget_parent(new_dentry); + struct dentry *parent = new_dentry->d_parent; btrfs_log_new_name(trans, old_inode, old_dir, parent); - dput(parent); btrfs_end_log_trans(root); } out_fail: @@ -7181,21 +7196,32 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + inode->i_fop = &btrfs_file_operations; + inode->i_op = &btrfs_file_inode_operations; + err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { inode->i_mapping->a_ops = &btrfs_aops; inode->i_mapping->backing_dev_info = &root->fs_info->bdi; - inode->i_fop = &btrfs_file_operations; - inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } if (drop_inode) goto out_unlock; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + err = -ENOMEM; + drop_inode = 1; + goto out_unlock; + } key.objectid = btrfs_ino(inode); key.offset = 0; btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY); @@ -7233,6 +7259,8 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, drop_inode = 1; out_unlock: + if (!err) + d_instantiate(dentry, inode); nr = trans->blocks_used; btrfs_end_transaction_throttle(trans, root); if (drop_inode) { @@ -7332,15 +7360,19 @@ static int btrfs_set_page_dirty(struct page *page) return __set_page_dirty_nobuffers(page); } -static int btrfs_permission(struct inode *inode, int mask, unsigned int flags) +static int btrfs_permission(struct inode *inode, int mask) { struct btrfs_root *root = BTRFS_I(inode)->root; + umode_t mode = inode->i_mode; - if (btrfs_root_readonly(root) && (mask & MAY_WRITE)) - return -EROFS; - if ((BTRFS_I(inode)->flags & BTRFS_INODE_READONLY) && (mask & MAY_WRITE)) - return -EACCES; - return generic_permission(inode, mask, flags, btrfs_check_acl); + if (mask & MAY_WRITE && + (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))) { + if (btrfs_root_readonly(root)) + return -EROFS; + if (BTRFS_I(inode)->flags & BTRFS_INODE_READONLY) + return -EACCES; + } + return generic_permission(inode, mask); } static const struct inode_operations btrfs_dir_inode_operations = { @@ -7360,10 +7392,12 @@ static const struct inode_operations btrfs_dir_inode_operations = { .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, .permission = btrfs_permission, + .get_acl = btrfs_get_acl, }; static const struct inode_operations btrfs_dir_ro_inode_operations = { .lookup = btrfs_lookup, .permission = btrfs_permission, + .get_acl = btrfs_get_acl, }; static const struct file_operations btrfs_dir_file_operations = { @@ -7385,7 +7419,6 @@ static struct extent_io_ops btrfs_extent_io_ops = { .readpage_end_io_hook = btrfs_readpage_end_io_hook, .writepage_end_io_hook = btrfs_writepage_end_io_hook, .writepage_start_hook = btrfs_writepage_start_hook, - .readpage_io_failed_hook = btrfs_io_failed_hook, .set_bit_hook = btrfs_set_bit_hook, .clear_bit_hook = btrfs_clear_bit_hook, .merge_extent_hook = btrfs_merge_extent_hook, @@ -7432,6 +7465,7 @@ static const struct inode_operations btrfs_file_inode_operations = { .removexattr = btrfs_removexattr, .permission = btrfs_permission, .fiemap = btrfs_fiemap, + .get_acl = btrfs_get_acl, }; static const struct inode_operations btrfs_special_inode_operations = { .getattr = btrfs_getattr, @@ -7441,19 +7475,23 @@ static const struct inode_operations btrfs_special_inode_operations = { .getxattr = btrfs_getxattr, .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, + .get_acl = btrfs_get_acl, }; static const struct inode_operations btrfs_symlink_inode_operations = { .readlink = generic_readlink, .follow_link = page_follow_link_light, .put_link = page_put_link, .getattr = btrfs_getattr, + .setattr = btrfs_setattr, .permission = btrfs_permission, .setxattr = btrfs_setxattr, .getxattr = btrfs_getxattr, .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, + .get_acl = btrfs_get_acl, }; const struct dentry_operations btrfs_dentry_operations = { .d_delete = btrfs_dentry_delete, + .d_release = btrfs_dentry_release, }; diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index a205027..ba26540 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -51,6 +51,7 @@ #include "volumes.h" #include "locking.h" #include "inode-map.h" +#include "backref.h" /* Mask out flags that are inappropriate for the given type of inode. */ static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags) @@ -117,7 +118,7 @@ void btrfs_update_iflags(struct inode *inode) /* * Inherit flags from the parent inode. * - * Unlike extN we don't have any flags we don't want to inherit currently. + * Currently only the compression flags and the cow flags are inherited. */ void btrfs_inherit_iflags(struct inode *inode, struct inode *dir) { @@ -128,12 +129,17 @@ void btrfs_inherit_iflags(struct inode *inode, struct inode *dir) flags = BTRFS_I(dir)->flags; - if (S_ISREG(inode->i_mode)) - flags &= ~BTRFS_INODE_DIRSYNC; - else if (!S_ISDIR(inode->i_mode)) - flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME); + if (flags & BTRFS_INODE_NOCOMPRESS) { + BTRFS_I(inode)->flags &= ~BTRFS_INODE_COMPRESS; + BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS; + } else if (flags & BTRFS_INODE_COMPRESS) { + BTRFS_I(inode)->flags &= ~BTRFS_INODE_NOCOMPRESS; + BTRFS_I(inode)->flags |= BTRFS_INODE_COMPRESS; + } + + if (flags & BTRFS_INODE_NODATACOW) + BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW; - BTRFS_I(inode)->flags = flags; btrfs_update_iflags(inode); } @@ -246,11 +252,11 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); + btrfs_update_iflags(inode); + inode->i_ctime = CURRENT_TIME; ret = btrfs_update_inode(trans, root, inode); BUG_ON(ret); - btrfs_update_iflags(inode); - inode->i_ctime = CURRENT_TIME; btrfs_end_transaction(trans, root); mnt_drop_write(file->f_path.mnt); @@ -277,6 +283,7 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) struct fstrim_range range; u64 minlen = ULLONG_MAX; u64 num_devices = 0; + u64 total_bytes = btrfs_super_total_bytes(root->fs_info->super_copy); int ret; if (!capable(CAP_SYS_ADMIN)) @@ -295,12 +302,15 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) } } rcu_read_unlock(); + if (!num_devices) return -EOPNOTSUPP; - if (copy_from_user(&range, arg, sizeof(range))) return -EFAULT; + if (range.start > total_bytes) + return -EINVAL; + range.len = min(range.len, total_bytes - range.start); range.minlen = max(range.minlen, minlen); ret = btrfs_trim_fs(root, &range); if (ret < 0) @@ -323,7 +333,7 @@ static noinline int create_subvol(struct btrfs_root *root, struct btrfs_inode_item *inode_item; struct extent_buffer *leaf; struct btrfs_root *new_root; - struct dentry *parent = dget_parent(dentry); + struct dentry *parent = dentry->d_parent; struct inode *dir; int ret; int err; @@ -332,10 +342,8 @@ static noinline int create_subvol(struct btrfs_root *root, u64 index = 0; ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid); - if (ret) { - dput(parent); + if (ret) return ret; - } dir = parent->d_inode; @@ -346,10 +354,8 @@ static noinline int create_subvol(struct btrfs_root *root, * 2 - dir items */ trans = btrfs_start_transaction(root, 6); - if (IS_ERR(trans)) { - dput(parent); + if (IS_ERR(trans)) return PTR_ERR(trans); - } leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0, objectid, NULL, 0, 0, 0); @@ -439,7 +445,6 @@ static noinline int create_subvol(struct btrfs_root *root, d_instantiate(dentry, btrfs_lookup_dentry(dir, dentry)); fail: - dput(parent); if (async_transid) { *async_transid = trans->transid; err = btrfs_commit_transaction_async(trans, root, 1); @@ -456,7 +461,6 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, bool readonly) { struct inode *inode; - struct dentry *parent; struct btrfs_pending_snapshot *pending_snapshot; struct btrfs_trans_handle *trans; int ret; @@ -504,9 +508,7 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, if (ret) goto fail; - parent = dget_parent(dentry); - inode = btrfs_lookup_dentry(parent->d_inode, dentry); - dput(parent); + inode = btrfs_lookup_dentry(dentry->d_parent->d_inode, dentry); if (IS_ERR(inode)) { ret = PTR_ERR(inode); goto fail; @@ -768,7 +770,7 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, int ret = 1; /* - * make sure that once we start defragging and extent, we keep on + * make sure that once we start defragging an extent, we keep on * defragging it */ if (start < *defrag_end) @@ -813,7 +815,6 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, * extent will force at least part of that big extent to be defragged. */ if (ret) { - *last_len += len; *defrag_end = extent_map_end(em); } else { *last_len = 0; @@ -851,13 +852,16 @@ static int cluster_pages_for_defrag(struct inode *inode, int i_done; struct btrfs_ordered_extent *ordered; struct extent_state *cached_state = NULL; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); if (isize == 0) return 0; file_end = (isize - 1) >> PAGE_CACHE_SHIFT; + mutex_lock(&inode->i_mutex); ret = btrfs_delalloc_reserve_space(inode, num_pages << PAGE_CACHE_SHIFT); + mutex_unlock(&inode->i_mutex); if (ret) return ret; again: @@ -867,8 +871,8 @@ again: /* step one, lock all the pages */ for (i = 0; i < num_pages; i++) { struct page *page; - page = grab_cache_page(inode->i_mapping, - start_index + i); + page = find_or_create_page(inode->i_mapping, + start_index + i, mask); if (!page) break; @@ -938,7 +942,9 @@ again: GFP_NOFS); if (i_done != num_pages) { - atomic_inc(&BTRFS_I(inode)->outstanding_extents); + spin_lock(&BTRFS_I(inode)->lock); + BTRFS_I(inode)->outstanding_extents++; + spin_unlock(&BTRFS_I(inode)->lock); btrfs_delalloc_release_space(inode, (num_pages - i_done) << PAGE_CACHE_SHIFT); } @@ -978,18 +984,20 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, struct btrfs_super_block *disk_super; struct file_ra_state *ra = NULL; unsigned long last_index; + u64 isize = i_size_read(inode); u64 features; u64 last_len = 0; u64 skip = 0; u64 defrag_end = 0; u64 newer_off = range->start; - int newer_left = 0; unsigned long i; + unsigned long ra_index = 0; int ret; int defrag_count = 0; int compress_type = BTRFS_COMPRESS_ZLIB; int extent_thresh = range->extent_thresh; - int newer_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT; + int max_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT; + int cluster = max_cluster; u64 new_align = ~((u64)128 * 1024 - 1); struct page **pages = NULL; @@ -1003,7 +1011,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, compress_type = range->compress_type; } - if (inode->i_size == 0) + if (isize == 0) return 0; /* @@ -1019,7 +1027,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, ra = &file->f_ra; } - pages = kmalloc(sizeof(struct page *) * newer_cluster, + pages = kmalloc(sizeof(struct page *) * max_cluster, GFP_NOFS); if (!pages) { ret = -ENOMEM; @@ -1028,10 +1036,10 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, /* find the last page to defrag */ if (range->start + range->len > range->start) { - last_index = min_t(u64, inode->i_size - 1, + last_index = min_t(u64, isize - 1, range->start + range->len - 1) >> PAGE_CACHE_SHIFT; } else { - last_index = (inode->i_size - 1) >> PAGE_CACHE_SHIFT; + last_index = (isize - 1) >> PAGE_CACHE_SHIFT; } if (newer_than) { @@ -1044,16 +1052,24 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, * the extents in the file evenly spaced */ i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; - newer_left = newer_cluster; } else goto out_ra; } else { i = range->start >> PAGE_CACHE_SHIFT; } if (!max_to_defrag) - max_to_defrag = last_index - 1; + max_to_defrag = last_index; - while (i <= last_index && defrag_count < max_to_defrag) { + /* + * make writeback starts from i, so the defrag range can be + * written sequentially. + */ + if (i < inode->i_mapping->writeback_index) + inode->i_mapping->writeback_index = i; + + while (i <= last_index && defrag_count < max_to_defrag && + (i < (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT)) { /* * make sure we stop running if someone unmounts * the FS @@ -1076,18 +1092,31 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, i = max(i + 1, next); continue; } + + if (!newer_than) { + cluster = (PAGE_CACHE_ALIGN(defrag_end) >> + PAGE_CACHE_SHIFT) - i; + cluster = min(cluster, max_cluster); + } else { + cluster = max_cluster; + } + if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) BTRFS_I(inode)->force_compress = compress_type; - btrfs_force_ra(inode->i_mapping, ra, file, i, newer_cluster); + if (i + cluster > ra_index) { + ra_index = max(i, ra_index); + btrfs_force_ra(inode->i_mapping, ra, file, ra_index, + cluster); + ra_index += max_cluster; + } - ret = cluster_pages_for_defrag(inode, pages, i, newer_cluster); + ret = cluster_pages_for_defrag(inode, pages, i, cluster); if (ret < 0) goto out_ra; defrag_count += ret; balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret); - i += ret; if (newer_than) { if (newer_off == (u64)-1) @@ -1102,12 +1131,17 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, if (!ret) { range->start = newer_off; i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; - newer_left = newer_cluster; } else { break; } } else { - i++; + if (ret > 0) { + i += ret; + last_len += ret << PAGE_CACHE_SHIFT; + } else { + i++; + last_len = 0; + } } } @@ -1133,16 +1167,14 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, mutex_unlock(&inode->i_mutex); } - disk_super = &root->fs_info->super_copy; + disk_super = root->fs_info->super_copy; features = btrfs_super_incompat_flags(disk_super); if (range->compress_type == BTRFS_COMPRESS_LZO) { features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO; btrfs_set_super_incompat_flags(disk_super, features); } - if (!file) - kfree(ra); - return defrag_count; + ret = defrag_count; out_ra: if (!file) @@ -1186,12 +1218,12 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, *devstr = '\0'; devstr = vol_args->name; devid = simple_strtoull(devstr, &end, 10); - printk(KERN_INFO "resizing devid %llu\n", + printk(KERN_INFO "btrfs: resizing devid %llu\n", (unsigned long long)devid); } device = btrfs_find_device(root, devid, NULL, NULL); if (!device) { - printk(KERN_INFO "resizer unable to find device %llu\n", + printk(KERN_INFO "btrfs: resizer unable to find device %llu\n", (unsigned long long)devid); ret = -EINVAL; goto out_unlock; @@ -1237,7 +1269,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, do_div(new_size, root->sectorsize); new_size *= root->sectorsize; - printk(KERN_INFO "new size for %s is %llu\n", + printk(KERN_INFO "btrfs: new size for %s is %llu\n", device->name, (unsigned long long)new_size); if (new_size > old_size) { @@ -1248,7 +1280,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, } ret = btrfs_grow_device(trans, device, new_size); btrfs_commit_transaction(trans, root); - } else { + } else if (new_size < old_size) { ret = btrfs_shrink_device(device, new_size); } @@ -1295,12 +1327,17 @@ static noinline int btrfs_ioctl_snap_create_transid(struct file *file, printk(KERN_INFO "btrfs: Snapshot src from " "another FS\n"); ret = -EINVAL; - fput(src_file); - goto out; + } else if (!inode_owner_or_capable(src_inode)) { + /* + * Subvolume creation is not restricted, but snapshots + * are limited to own subvolumes only + */ + ret = -EPERM; + } else { + ret = btrfs_mksubvol(&file->f_path, name, namelen, + BTRFS_I(src_inode)->root, + transid, readonly); } - ret = btrfs_mksubvol(&file->f_path, name, namelen, - BTRFS_I(src_inode)->root, - transid, readonly); fput(src_file); } out: @@ -1755,11 +1792,10 @@ static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info, key.objectid = key.offset; key.offset = (u64)-1; dirid = key.objectid; - } if (ptr < name) goto out; - memcpy(name, ptr, total_len); + memmove(name, ptr, total_len); name[total_len]='\0'; ret = 0; out: @@ -2184,6 +2220,11 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, if (!(src_file->f_mode & FMODE_READ)) goto out_fput; + /* don't make the dst file partly checksummed */ + if ((BTRFS_I(src)->flags & BTRFS_INODE_NODATASUM) != + (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) + goto out_fput; + ret = -EISDIR; if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode)) goto out_fput; @@ -2222,11 +2263,26 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, if (off + len == src->i_size) len = ALIGN(src->i_size, bs) - off; + if (len == 0) { + ret = 0; + goto out_unlock; + } + /* verify the end result is block aligned */ if (!IS_ALIGNED(off, bs) || !IS_ALIGNED(off + len, bs) || !IS_ALIGNED(destoff, bs)) goto out_unlock; + if (destoff > inode->i_size) { + ret = btrfs_cont_expand(inode, inode->i_size, destoff); + if (ret) + goto out_unlock; + } + + /* truncate page cache pages from target inode range */ + truncate_inode_pages_range(&inode->i_data, destoff, + PAGE_CACHE_ALIGN(destoff + len) - 1); + /* do any pending delalloc/csum calc on src, one way or another, and lock file content */ while (1) { @@ -2320,7 +2376,12 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, else new_key.offset = destoff; - trans = btrfs_start_transaction(root, 1); + /* + * 1 - adjusting old extent (we may have to split it) + * 1 - add new extent + * 1 - inode update + */ + trans = btrfs_start_transaction(root, 3); if (IS_ERR(trans)) { ret = PTR_ERR(trans); goto out; @@ -2328,14 +2389,21 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, if (type == BTRFS_FILE_EXTENT_REG || type == BTRFS_FILE_EXTENT_PREALLOC) { + /* + * a | --- range to clone ---| b + * | ------------- extent ------------- | + */ + + /* substract range b */ + if (key.offset + datal > off + len) + datal = off + len - key.offset; + + /* substract range a */ if (off > key.offset) { datao += off - key.offset; datal -= off - key.offset; } - if (key.offset + datal > off + len) - datal = off + len - key.offset; - ret = btrfs_drop_extents(trans, inode, new_key.offset, new_key.offset + datal, @@ -2380,6 +2448,20 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, new_key.offset += skip; } + /* + * Don't copy an inline extent into an offset + * greater than zero. Having an inline extent + * at such an offset results in chaos as btrfs + * isn't prepared for such cases. Just skip + * this case for the same reasons as commented + * at btrfs_ioctl_clone(). + */ + if (new_key.offset > 0) { + ret = -EOPNOTSUPP; + btrfs_end_transaction(trans, root); + goto out; + } + if (key.offset + datal > off+len) trim = key.offset + datal - (off+len); @@ -2432,7 +2514,6 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, if (endoff > inode->i_size) btrfs_i_size_write(inode, endoff); - BTRFS_I(inode)->flags = BTRFS_I(src)->flags; ret = btrfs_update_inode(trans, root, inode); BUG_ON(ret); btrfs_end_transaction(trans, root); @@ -2559,7 +2640,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp) return PTR_ERR(trans); } - dir_id = btrfs_super_root_dir(&root->fs_info->super_copy); + dir_id = btrfs_super_root_dir(root->fs_info->super_copy); di = btrfs_lookup_dir_item(trans, root->fs_info->tree_root, path, dir_id, "default", 7, 1); if (IS_ERR_OR_NULL(di)) { @@ -2575,7 +2656,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp) btrfs_mark_buffer_dirty(path->nodes[0]); btrfs_free_path(path); - disk_super = &root->fs_info->super_copy; + disk_super = root->fs_info->super_copy; features = btrfs_super_incompat_flags(disk_super); if (!(features & BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL)) { features |= BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL; @@ -2836,6 +2917,147 @@ static long btrfs_ioctl_scrub_progress(struct btrfs_root *root, return ret; } +static long btrfs_ioctl_ino_to_path(struct btrfs_root *root, void __user *arg) +{ + int ret = 0; + int i; + u64 rel_ptr; + int size; + struct btrfs_ioctl_ino_path_args *ipa = NULL; + struct inode_fs_paths *ipath = NULL; + struct btrfs_path *path; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + path = btrfs_alloc_path(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + ipa = memdup_user(arg, sizeof(*ipa)); + if (IS_ERR(ipa)) { + ret = PTR_ERR(ipa); + ipa = NULL; + goto out; + } + + size = min_t(u32, ipa->size, 4096); + ipath = init_ipath(size, root, path); + if (IS_ERR(ipath)) { + ret = PTR_ERR(ipath); + ipath = NULL; + goto out; + } + + ret = paths_from_inode(ipa->inum, ipath); + if (ret < 0) + goto out; + + for (i = 0; i < ipath->fspath->elem_cnt; ++i) { + rel_ptr = ipath->fspath->val[i] - + (u64)(unsigned long)ipath->fspath->val; + ipath->fspath->val[i] = rel_ptr; + } + + ret = copy_to_user((void *)(unsigned long)ipa->fspath, + (void *)(unsigned long)ipath->fspath, size); + if (ret) { + ret = -EFAULT; + goto out; + } + +out: + btrfs_free_path(path); + free_ipath(ipath); + kfree(ipa); + + return ret; +} + +static int build_ino_list(u64 inum, u64 offset, u64 root, void *ctx) +{ + struct btrfs_data_container *inodes = ctx; + const size_t c = 3 * sizeof(u64); + + if (inodes->bytes_left >= c) { + inodes->bytes_left -= c; + inodes->val[inodes->elem_cnt] = inum; + inodes->val[inodes->elem_cnt + 1] = offset; + inodes->val[inodes->elem_cnt + 2] = root; + inodes->elem_cnt += 3; + } else { + inodes->bytes_missing += c - inodes->bytes_left; + inodes->bytes_left = 0; + inodes->elem_missed += 3; + } + + return 0; +} + +static long btrfs_ioctl_logical_to_ino(struct btrfs_root *root, + void __user *arg) +{ + int ret = 0; + int size; + u64 extent_offset; + struct btrfs_ioctl_logical_ino_args *loi; + struct btrfs_data_container *inodes = NULL; + struct btrfs_path *path = NULL; + struct btrfs_key key; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + loi = memdup_user(arg, sizeof(*loi)); + if (IS_ERR(loi)) { + ret = PTR_ERR(loi); + loi = NULL; + goto out; + } + + path = btrfs_alloc_path(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + size = min_t(u32, loi->size, 4096); + inodes = init_data_container(size); + if (IS_ERR(inodes)) { + ret = PTR_ERR(inodes); + inodes = NULL; + goto out; + } + + ret = extent_from_logical(root->fs_info, loi->logical, path, &key); + + if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK) + ret = -ENOENT; + if (ret < 0) + goto out; + + extent_offset = loi->logical - key.objectid; + ret = iterate_extent_inodes(root->fs_info, path, key.objectid, + extent_offset, build_ino_list, inodes); + + if (ret < 0) + goto out; + + ret = copy_to_user((void *)(unsigned long)loi->inodes, + (void *)(unsigned long)inodes, size); + if (ret) + ret = -EFAULT; + +out: + btrfs_free_path(path); + kfree(inodes); + kfree(loi); + + return ret; +} + long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { @@ -2893,6 +3115,10 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_tree_search(file, argp); case BTRFS_IOC_INO_LOOKUP: return btrfs_ioctl_ino_lookup(file, argp); + case BTRFS_IOC_INO_PATHS: + return btrfs_ioctl_ino_to_path(root, argp); + case BTRFS_IOC_LOGICAL_INO: + return btrfs_ioctl_logical_to_ino(root, argp); case BTRFS_IOC_SPACE_INFO: return btrfs_ioctl_space_info(root, argp); case BTRFS_IOC_SYNC: diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h index ad1ea78..252ae99 100644 --- a/fs/btrfs/ioctl.h +++ b/fs/btrfs/ioctl.h @@ -193,6 +193,30 @@ struct btrfs_ioctl_space_args { struct btrfs_ioctl_space_info spaces[0]; }; +struct btrfs_data_container { + __u32 bytes_left; /* out -- bytes not needed to deliver output */ + __u32 bytes_missing; /* out -- additional bytes needed for result */ + __u32 elem_cnt; /* out */ + __u32 elem_missed; /* out */ + __u64 val[0]; /* out */ +}; + +struct btrfs_ioctl_ino_path_args { + __u64 inum; /* in */ + __u32 size; /* in */ + __u64 reserved[4]; + /* struct btrfs_data_container *fspath; out */ + __u64 fspath; /* out */ +}; + +struct btrfs_ioctl_logical_ino_args { + __u64 logical; /* in */ + __u32 size; /* in */ + __u64 reserved[4]; + /* struct btrfs_data_container *inodes; out */ + __u64 inodes; +}; + #define BTRFS_IOC_SNAP_CREATE _IOW(BTRFS_IOCTL_MAGIC, 1, \ struct btrfs_ioctl_vol_args) #define BTRFS_IOC_DEFRAG _IOW(BTRFS_IOCTL_MAGIC, 2, \ @@ -248,4 +272,9 @@ struct btrfs_ioctl_space_args { struct btrfs_ioctl_dev_info_args) #define BTRFS_IOC_FS_INFO _IOR(BTRFS_IOCTL_MAGIC, 31, \ struct btrfs_ioctl_fs_info_args) +#define BTRFS_IOC_INO_PATHS _IOWR(BTRFS_IOCTL_MAGIC, 35, \ + struct btrfs_ioctl_ino_path_args) +#define BTRFS_IOC_LOGICAL_INO _IOWR(BTRFS_IOCTL_MAGIC, 36, \ + struct btrfs_ioctl_ino_path_args) + #endif diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c index 66fa43d..d77b67c 100644 --- a/fs/btrfs/locking.c +++ b/fs/btrfs/locking.c @@ -24,185 +24,197 @@ #include "extent_io.h" #include "locking.h" -static inline void spin_nested(struct extent_buffer *eb) -{ - spin_lock(&eb->lock); -} +void btrfs_assert_tree_read_locked(struct extent_buffer *eb); /* - * Setting a lock to blocking will drop the spinlock and set the - * flag that forces other procs who want the lock to wait. After - * this you can safely schedule with the lock held. + * if we currently have a spinning reader or writer lock + * (indicated by the rw flag) this will bump the count + * of blocking holders and drop the spinlock. */ -void btrfs_set_lock_blocking(struct extent_buffer *eb) +void btrfs_set_lock_blocking_rw(struct extent_buffer *eb, int rw) { - if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) { - set_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags); - spin_unlock(&eb->lock); + if (rw == BTRFS_WRITE_LOCK) { + if (atomic_read(&eb->blocking_writers) == 0) { + WARN_ON(atomic_read(&eb->spinning_writers) != 1); + atomic_dec(&eb->spinning_writers); + btrfs_assert_tree_locked(eb); + atomic_inc(&eb->blocking_writers); + write_unlock(&eb->lock); + } + } else if (rw == BTRFS_READ_LOCK) { + btrfs_assert_tree_read_locked(eb); + atomic_inc(&eb->blocking_readers); + WARN_ON(atomic_read(&eb->spinning_readers) == 0); + atomic_dec(&eb->spinning_readers); + read_unlock(&eb->lock); } - /* exit with the spin lock released and the bit set */ + return; } /* - * clearing the blocking flag will take the spinlock again. - * After this you can't safely schedule + * if we currently have a blocking lock, take the spinlock + * and drop our blocking count */ -void btrfs_clear_lock_blocking(struct extent_buffer *eb) +void btrfs_clear_lock_blocking_rw(struct extent_buffer *eb, int rw) { - if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) { - spin_nested(eb); - clear_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags); - smp_mb__after_clear_bit(); + if (rw == BTRFS_WRITE_LOCK_BLOCKING) { + BUG_ON(atomic_read(&eb->blocking_writers) != 1); + write_lock(&eb->lock); + WARN_ON(atomic_read(&eb->spinning_writers)); + atomic_inc(&eb->spinning_writers); + if (atomic_dec_and_test(&eb->blocking_writers)) + wake_up(&eb->write_lock_wq); + } else if (rw == BTRFS_READ_LOCK_BLOCKING) { + BUG_ON(atomic_read(&eb->blocking_readers) == 0); + read_lock(&eb->lock); + atomic_inc(&eb->spinning_readers); + if (atomic_dec_and_test(&eb->blocking_readers)) + wake_up(&eb->read_lock_wq); } - /* exit with the spin lock held */ + return; } /* - * unfortunately, many of the places that currently set a lock to blocking - * don't end up blocking for very long, and often they don't block - * at all. For a dbench 50 run, if we don't spin on the blocking bit - * at all, the context switch rate can jump up to 400,000/sec or more. - * - * So, we're still stuck with this crummy spin on the blocking bit, - * at least until the most common causes of the short blocks - * can be dealt with. + * take a spinning read lock. This will wait for any blocking + * writers */ -static int btrfs_spin_on_block(struct extent_buffer *eb) +void btrfs_tree_read_lock(struct extent_buffer *eb) { - int i; - - for (i = 0; i < 512; i++) { - if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) - return 1; - if (need_resched()) - break; - cpu_relax(); +again: + wait_event(eb->write_lock_wq, atomic_read(&eb->blocking_writers) == 0); + read_lock(&eb->lock); + if (atomic_read(&eb->blocking_writers)) { + read_unlock(&eb->lock); + wait_event(eb->write_lock_wq, + atomic_read(&eb->blocking_writers) == 0); + goto again; } - return 0; + atomic_inc(&eb->read_locks); + atomic_inc(&eb->spinning_readers); } /* - * This is somewhat different from trylock. It will take the - * spinlock but if it finds the lock is set to blocking, it will - * return without the lock held. - * - * returns 1 if it was able to take the lock and zero otherwise - * - * After this call, scheduling is not safe without first calling - * btrfs_set_lock_blocking() + * returns 1 if we get the read lock and 0 if we don't + * this won't wait for blocking writers */ -int btrfs_try_spin_lock(struct extent_buffer *eb) +int btrfs_try_tree_read_lock(struct extent_buffer *eb) { - int i; + if (atomic_read(&eb->blocking_writers)) + return 0; - if (btrfs_spin_on_block(eb)) { - spin_nested(eb); - if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) - return 1; - spin_unlock(&eb->lock); + read_lock(&eb->lock); + if (atomic_read(&eb->blocking_writers)) { + read_unlock(&eb->lock); + return 0; } - /* spin for a bit on the BLOCKING flag */ - for (i = 0; i < 2; i++) { - cpu_relax(); - if (!btrfs_spin_on_block(eb)) - break; - - spin_nested(eb); - if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) - return 1; - spin_unlock(&eb->lock); - } - return 0; + atomic_inc(&eb->read_locks); + atomic_inc(&eb->spinning_readers); + return 1; } /* - * the autoremove wake function will return 0 if it tried to wake up - * a process that was already awake, which means that process won't - * count as an exclusive wakeup. The waitq code will continue waking - * procs until it finds one that was actually sleeping. - * - * For btrfs, this isn't quite what we want. We want a single proc - * to be notified that the lock is ready for taking. If that proc - * already happen to be awake, great, it will loop around and try for - * the lock. - * - * So, btrfs_wake_function always returns 1, even when the proc that we - * tried to wake up was already awake. + * returns 1 if we get the read lock and 0 if we don't + * this won't wait for blocking writers or readers */ -static int btrfs_wake_function(wait_queue_t *wait, unsigned mode, - int sync, void *key) +int btrfs_try_tree_write_lock(struct extent_buffer *eb) { - autoremove_wake_function(wait, mode, sync, key); + if (atomic_read(&eb->blocking_writers) || + atomic_read(&eb->blocking_readers)) + return 0; + write_lock(&eb->lock); + if (atomic_read(&eb->blocking_writers) || + atomic_read(&eb->blocking_readers)) { + write_unlock(&eb->lock); + return 0; + } + atomic_inc(&eb->write_locks); + atomic_inc(&eb->spinning_writers); return 1; } /* - * returns with the extent buffer spinlocked. - * - * This will spin and/or wait as required to take the lock, and then - * return with the spinlock held. - * - * After this call, scheduling is not safe without first calling - * btrfs_set_lock_blocking() + * drop a spinning read lock + */ +void btrfs_tree_read_unlock(struct extent_buffer *eb) +{ + btrfs_assert_tree_read_locked(eb); + WARN_ON(atomic_read(&eb->spinning_readers) == 0); + atomic_dec(&eb->spinning_readers); + atomic_dec(&eb->read_locks); + read_unlock(&eb->lock); +} + +/* + * drop a blocking read lock + */ +void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb) +{ + btrfs_assert_tree_read_locked(eb); + WARN_ON(atomic_read(&eb->blocking_readers) == 0); + if (atomic_dec_and_test(&eb->blocking_readers)) + wake_up(&eb->read_lock_wq); + atomic_dec(&eb->read_locks); +} + +/* + * take a spinning write lock. This will wait for both + * blocking readers or writers */ int btrfs_tree_lock(struct extent_buffer *eb) { - DEFINE_WAIT(wait); - wait.func = btrfs_wake_function; - - if (!btrfs_spin_on_block(eb)) - goto sleep; - - while(1) { - spin_nested(eb); - - /* nobody is blocking, exit with the spinlock held */ - if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) - return 0; - - /* - * we have the spinlock, but the real owner is blocking. - * wait for them - */ - spin_unlock(&eb->lock); - - /* - * spin for a bit, and if the blocking flag goes away, - * loop around - */ - cpu_relax(); - if (btrfs_spin_on_block(eb)) - continue; -sleep: - prepare_to_wait_exclusive(&eb->lock_wq, &wait, - TASK_UNINTERRUPTIBLE); - - if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) - schedule(); - - finish_wait(&eb->lock_wq, &wait); +again: + wait_event(eb->read_lock_wq, atomic_read(&eb->blocking_readers) == 0); + wait_event(eb->write_lock_wq, atomic_read(&eb->blocking_writers) == 0); + write_lock(&eb->lock); + if (atomic_read(&eb->blocking_readers)) { + write_unlock(&eb->lock); + wait_event(eb->read_lock_wq, + atomic_read(&eb->blocking_readers) == 0); + goto again; } + if (atomic_read(&eb->blocking_writers)) { + write_unlock(&eb->lock); + wait_event(eb->write_lock_wq, + atomic_read(&eb->blocking_writers) == 0); + goto again; + } + WARN_ON(atomic_read(&eb->spinning_writers)); + atomic_inc(&eb->spinning_writers); + atomic_inc(&eb->write_locks); return 0; } +/* + * drop a spinning or a blocking write lock. + */ int btrfs_tree_unlock(struct extent_buffer *eb) { - /* - * if we were a blocking owner, we don't have the spinlock held - * just clear the bit and look for waiters - */ - if (test_and_clear_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) - smp_mb__after_clear_bit(); - else - spin_unlock(&eb->lock); - - if (waitqueue_active(&eb->lock_wq)) - wake_up(&eb->lock_wq); + int blockers = atomic_read(&eb->blocking_writers); + + BUG_ON(blockers > 1); + + btrfs_assert_tree_locked(eb); + atomic_dec(&eb->write_locks); + + if (blockers) { + WARN_ON(atomic_read(&eb->spinning_writers)); + atomic_dec(&eb->blocking_writers); + smp_wmb(); + wake_up(&eb->write_lock_wq); + } else { + WARN_ON(atomic_read(&eb->spinning_writers) != 1); + atomic_dec(&eb->spinning_writers); + write_unlock(&eb->lock); + } return 0; } void btrfs_assert_tree_locked(struct extent_buffer *eb) { - if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) - assert_spin_locked(&eb->lock); + BUG_ON(!atomic_read(&eb->write_locks)); +} + +void btrfs_assert_tree_read_locked(struct extent_buffer *eb) +{ + BUG_ON(!atomic_read(&eb->read_locks)); } diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h index 5c33a56..17247dd 100644 --- a/fs/btrfs/locking.h +++ b/fs/btrfs/locking.h @@ -19,11 +19,43 @@ #ifndef __BTRFS_LOCKING_ #define __BTRFS_LOCKING_ +#define BTRFS_WRITE_LOCK 1 +#define BTRFS_READ_LOCK 2 +#define BTRFS_WRITE_LOCK_BLOCKING 3 +#define BTRFS_READ_LOCK_BLOCKING 4 + int btrfs_tree_lock(struct extent_buffer *eb); int btrfs_tree_unlock(struct extent_buffer *eb); int btrfs_try_spin_lock(struct extent_buffer *eb); -void btrfs_set_lock_blocking(struct extent_buffer *eb); -void btrfs_clear_lock_blocking(struct extent_buffer *eb); +void btrfs_tree_read_lock(struct extent_buffer *eb); +void btrfs_tree_read_unlock(struct extent_buffer *eb); +void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb); +void btrfs_set_lock_blocking_rw(struct extent_buffer *eb, int rw); +void btrfs_clear_lock_blocking_rw(struct extent_buffer *eb, int rw); void btrfs_assert_tree_locked(struct extent_buffer *eb); +int btrfs_try_tree_read_lock(struct extent_buffer *eb); +int btrfs_try_tree_write_lock(struct extent_buffer *eb); + +static inline void btrfs_tree_unlock_rw(struct extent_buffer *eb, int rw) +{ + if (rw == BTRFS_WRITE_LOCK || rw == BTRFS_WRITE_LOCK_BLOCKING) + btrfs_tree_unlock(eb); + else if (rw == BTRFS_READ_LOCK_BLOCKING) + btrfs_tree_read_unlock_blocking(eb); + else if (rw == BTRFS_READ_LOCK) + btrfs_tree_read_unlock(eb); + else + BUG(); +} + +static inline void btrfs_set_lock_blocking(struct extent_buffer *eb) +{ + btrfs_set_lock_blocking_rw(eb, BTRFS_WRITE_LOCK); +} + +static inline void btrfs_clear_lock_blocking(struct extent_buffer *eb) +{ + btrfs_clear_lock_blocking_rw(eb, BTRFS_WRITE_LOCK_BLOCKING); +} #endif diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c index fb2605d..f38e452 100644 --- a/fs/btrfs/print-tree.c +++ b/fs/btrfs/print-tree.c @@ -158,8 +158,7 @@ static void print_extent_ref_v0(struct extent_buffer *eb, int slot) void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l) { int i; - u32 type; - u32 nr = btrfs_header_nritems(l); + u32 type, nr; struct btrfs_item *item; struct btrfs_root_item *ri; struct btrfs_dir_item *di; @@ -172,6 +171,11 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l) struct btrfs_key key; struct btrfs_key found_key; + if (!l) + return; + + nr = btrfs_header_nritems(l); + printk(KERN_INFO "leaf %llu total ptrs %d free space %d\n", (unsigned long long)btrfs_header_bytenr(l), nr, btrfs_leaf_free_space(root, l)); diff --git a/fs/btrfs/ref-cache.c b/fs/btrfs/ref-cache.c deleted file mode 100644 index 82d569c..0000000 --- a/fs/btrfs/ref-cache.c +++ /dev/null @@ -1,68 +0,0 @@ -/* - * Copyright (C) 2008 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License v2 as published by the Free Software Foundation. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#include <linux/sched.h> -#include <linux/slab.h> -#include <linux/sort.h> -#include "ctree.h" -#include "ref-cache.h" -#include "transaction.h" - -static struct rb_node *tree_insert(struct rb_root *root, u64 bytenr, - struct rb_node *node) -{ - struct rb_node **p = &root->rb_node; - struct rb_node *parent = NULL; - struct btrfs_leaf_ref *entry; - - while (*p) { - parent = *p; - entry = rb_entry(parent, struct btrfs_leaf_ref, rb_node); - - if (bytenr < entry->bytenr) - p = &(*p)->rb_left; - else if (bytenr > entry->bytenr) - p = &(*p)->rb_right; - else - return parent; - } - - entry = rb_entry(node, struct btrfs_leaf_ref, rb_node); - rb_link_node(node, parent, p); - rb_insert_color(node, root); - return NULL; -} - -static struct rb_node *tree_search(struct rb_root *root, u64 bytenr) -{ - struct rb_node *n = root->rb_node; - struct btrfs_leaf_ref *entry; - - while (n) { - entry = rb_entry(n, struct btrfs_leaf_ref, rb_node); - WARN_ON(!entry->in_tree); - - if (bytenr < entry->bytenr) - n = n->rb_left; - else if (bytenr > entry->bytenr) - n = n->rb_right; - else - return n; - } - return NULL; -} diff --git a/fs/btrfs/ref-cache.h b/fs/btrfs/ref-cache.h deleted file mode 100644 index 24f7001..0000000 --- a/fs/btrfs/ref-cache.h +++ /dev/null @@ -1,52 +0,0 @@ -/* - * Copyright (C) 2008 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License v2 as published by the Free Software Foundation. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ -#ifndef __REFCACHE__ -#define __REFCACHE__ - -struct btrfs_extent_info { - /* bytenr and num_bytes find the extent in the extent allocation tree */ - u64 bytenr; - u64 num_bytes; - - /* objectid and offset find the back reference for the file */ - u64 objectid; - u64 offset; -}; - -struct btrfs_leaf_ref { - struct rb_node rb_node; - struct btrfs_leaf_ref_tree *tree; - int in_tree; - atomic_t usage; - - u64 root_gen; - u64 bytenr; - u64 owner; - u64 generation; - int nritems; - - struct list_head list; - struct btrfs_extent_info extents[]; -}; - -static inline size_t btrfs_leaf_ref_size(int nr_extents) -{ - return sizeof(struct btrfs_leaf_ref) + - sizeof(struct btrfs_extent_info) * nr_extents; -} -#endif diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 2ab5837..cfb5543 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -670,7 +670,6 @@ struct backref_node *build_backref_tree(struct reloc_control *rc, int cowonly; int ret; int err = 0; - bool need_check = true; path1 = btrfs_alloc_path(); path2 = btrfs_alloc_path(); @@ -893,7 +892,6 @@ again: cur->bytenr); lower = cur; - need_check = true; for (; level < BTRFS_MAX_LEVEL; level++) { if (!path2->nodes[level]) { BUG_ON(btrfs_root_bytenr(&root->root_item) != @@ -937,12 +935,14 @@ again: /* * add the block to pending list if we - * need check its backrefs, we only do this once - * while walking up a tree as we will catch - * anything else later on. + * need check its backrefs. only block + * at 'cur->level + 1' is added to the + * tail of pending list. this guarantees + * we check backrefs from lower level + * blocks to upper level blocks. */ - if (!upper->checked && need_check) { - need_check = false; + if (!upper->checked && + level == cur->level + 1) { list_add_tail(&edge->list[UPPER], &list); } else @@ -1174,6 +1174,8 @@ static int clone_backref_node(struct btrfs_trans_handle *trans, list_add_tail(&new_edge->list[UPPER], &new_node->lower); } + } else { + list_add_tail(&new_node->lower, &cache->leaves); } rb_node = tree_insert(&cache->rb_root, new_node->bytenr, @@ -2041,8 +2043,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc, BUG_ON(IS_ERR(trans)); trans->block_rsv = rc->block_rsv; - ret = btrfs_block_rsv_check(trans, root, rc->block_rsv, - min_reserved, 0); + ret = btrfs_block_rsv_refill(root, rc->block_rsv, min_reserved); if (ret) { BUG_ON(ret != -EAGAIN); ret = btrfs_commit_transaction(trans, root); @@ -2152,8 +2153,7 @@ int prepare_to_merge(struct reloc_control *rc, int err) again: if (!err) { num_bytes = rc->merging_rsv_size; - ret = btrfs_block_rsv_add(NULL, root, rc->block_rsv, - num_bytes); + ret = btrfs_block_rsv_add(root, rc->block_rsv, num_bytes); if (ret) err = ret; } @@ -2427,7 +2427,7 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans, num_bytes = calcu_metadata_size(rc, node, 1) * 2; trans->block_rsv = rc->block_rsv; - ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes); + ret = btrfs_block_rsv_add(root, rc->block_rsv, num_bytes); if (ret) { if (ret == -EAGAIN) rc->commit_transaction = 1; @@ -2922,6 +2922,7 @@ static int relocate_file_extent_cluster(struct inode *inode, unsigned long last_index; struct page *page; struct file_ra_state *ra; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); int nr = 0; int ret = 0; @@ -2946,7 +2947,9 @@ static int relocate_file_extent_cluster(struct inode *inode, index = (cluster->start - offset) >> PAGE_CACHE_SHIFT; last_index = (cluster->end - offset) >> PAGE_CACHE_SHIFT; while (index <= last_index) { + mutex_lock(&inode->i_mutex); ret = btrfs_delalloc_reserve_metadata(inode, PAGE_CACHE_SIZE); + mutex_unlock(&inode->i_mutex); if (ret) goto out; @@ -2955,7 +2958,8 @@ static int relocate_file_extent_cluster(struct inode *inode, page_cache_sync_readahead(inode->i_mapping, ra, NULL, index, last_index + 1 - index); - page = grab_cache_page(inode->i_mapping, index); + page = find_or_create_page(inode->i_mapping, index, + mask); if (!page) { btrfs_delalloc_release_metadata(inode, PAGE_CACHE_SIZE); @@ -3322,8 +3326,11 @@ static int find_data_references(struct reloc_control *rc, } key.objectid = ref_objectid; - key.offset = ref_offset; key.type = BTRFS_EXTENT_DATA_KEY; + if (ref_offset > ((u64)-1 << 32)) + key.offset = 0; + else + key.offset = ref_offset; path->search_commit_root = 1; path->skip_locking = 1; @@ -3644,14 +3651,11 @@ int prepare_to_relocate(struct reloc_control *rc) * btrfs_init_reloc_root will use them when there * is no reservation in transaction handle. */ - ret = btrfs_block_rsv_add(NULL, rc->extent_root, rc->block_rsv, + ret = btrfs_block_rsv_add(rc->extent_root, rc->block_rsv, rc->extent_root->nodesize * 256); if (ret) return ret; - rc->block_rsv->refill_used = 1; - btrfs_add_durable_block_rsv(rc->extent_root->fs_info, rc->block_rsv); - memset(&rc->cluster, 0, sizeof(rc->cluster)); rc->search_start = rc->block_group->key.objectid; rc->extents_found = 0; @@ -3776,8 +3780,7 @@ restart: } } - ret = btrfs_block_rsv_check(trans, rc->extent_root, - rc->block_rsv, 0, 5); + ret = btrfs_block_rsv_check(rc->extent_root, rc->block_rsv, 5); if (ret < 0) { if (ret != -EAGAIN) { err = ret; diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index ebe4544..f409990 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -71,13 +71,12 @@ out: return ret; } -int btrfs_set_root_node(struct btrfs_root_item *item, - struct extent_buffer *node) +void btrfs_set_root_node(struct btrfs_root_item *item, + struct extent_buffer *node) { btrfs_set_root_bytenr(item, node->start); btrfs_set_root_level(item, btrfs_header_level(node)); btrfs_set_root_generation(item, btrfs_header_generation(node)); - return 0; } /* diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index a8d03d5..ddf2c90 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -17,10 +17,14 @@ */ #include <linux/blkdev.h> +#include <linux/ratelimit.h> #include "ctree.h" #include "volumes.h" #include "disk-io.h" #include "ordered-data.h" +#include "transaction.h" +#include "backref.h" +#include "extent_io.h" /* * This is only the first step towards a full-features scrub. It reads all @@ -29,15 +33,12 @@ * any can be found. * * Future enhancements: - * - To enhance the performance, better read-ahead strategies for the - * extent-tree can be employed. * - In case an unrepairable extent is encountered, track which files are * affected and report them * - In case of a read error on files with nodatasum, map the file and read * the extent to trigger a writeback of the good copy * - track and record media errors, throw out bad devices * - add a mode to also read unallocated space - * - make the prefetch cancellable */ struct scrub_bio; @@ -63,7 +64,7 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix); struct scrub_page { u64 flags; /* extent flags */ u64 generation; - u64 mirror_num; + int mirror_num; int have_csum; u8 csum[BTRFS_CSUM_SIZE]; }; @@ -87,6 +88,7 @@ struct scrub_dev { int first_free; int curr; atomic_t in_flight; + atomic_t fixup_cnt; spinlock_t list_lock; wait_queue_head_t list_wait; u16 csum_size; @@ -100,6 +102,27 @@ struct scrub_dev { spinlock_t stat_lock; }; +struct scrub_fixup_nodatasum { + struct scrub_dev *sdev; + u64 logical; + struct btrfs_root *root; + struct btrfs_work work; + int mirror_num; +}; + +struct scrub_warning { + struct btrfs_path *path; + u64 extent_item_size; + char *scratch_buf; + char *msg_buf; + const char *errstr; + sector_t sector; + u64 logical; + struct btrfs_device *dev; + int msg_bufsize; + int scratch_bufsize; +}; + static void scrub_free_csums(struct scrub_dev *sdev) { while (!list_empty(&sdev->csum_list)) { @@ -175,14 +198,15 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev) if (i != SCRUB_BIOS_PER_DEV-1) sdev->bios[i]->next_free = i + 1; - else + else sdev->bios[i]->next_free = -1; } sdev->first_free = 0; sdev->curr = -1; atomic_set(&sdev->in_flight, 0); + atomic_set(&sdev->fixup_cnt, 0); atomic_set(&sdev->cancel_req, 0); - sdev->csum_size = btrfs_super_csum_size(&fs_info->super_copy); + sdev->csum_size = btrfs_super_csum_size(fs_info->super_copy); INIT_LIST_HEAD(&sdev->csum_list); spin_lock_init(&sdev->list_lock); @@ -195,24 +219,366 @@ nomem: return ERR_PTR(-ENOMEM); } +static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root, void *ctx) +{ + u64 isize; + u32 nlink; + int ret; + int i; + struct extent_buffer *eb; + struct btrfs_inode_item *inode_item; + struct scrub_warning *swarn = ctx; + struct btrfs_fs_info *fs_info = swarn->dev->dev_root->fs_info; + struct inode_fs_paths *ipath = NULL; + struct btrfs_root *local_root; + struct btrfs_key root_key; + + root_key.objectid = root; + root_key.type = BTRFS_ROOT_ITEM_KEY; + root_key.offset = (u64)-1; + local_root = btrfs_read_fs_root_no_name(fs_info, &root_key); + if (IS_ERR(local_root)) { + ret = PTR_ERR(local_root); + goto err; + } + + ret = inode_item_info(inum, 0, local_root, swarn->path); + if (ret) { + btrfs_release_path(swarn->path); + goto err; + } + + eb = swarn->path->nodes[0]; + inode_item = btrfs_item_ptr(eb, swarn->path->slots[0], + struct btrfs_inode_item); + isize = btrfs_inode_size(eb, inode_item); + nlink = btrfs_inode_nlink(eb, inode_item); + btrfs_release_path(swarn->path); + + ipath = init_ipath(4096, local_root, swarn->path); + if (IS_ERR(ipath)) { + ret = PTR_ERR(ipath); + ipath = NULL; + goto err; + } + ret = paths_from_inode(inum, ipath); + + if (ret < 0) + goto err; + + /* + * we deliberately ignore the bit ipath might have been too small to + * hold all of the paths here + */ + for (i = 0; i < ipath->fspath->elem_cnt; ++i) + printk(KERN_WARNING "btrfs: %s at logical %llu on dev " + "%s, sector %llu, root %llu, inode %llu, offset %llu, " + "length %llu, links %u (path: %s)\n", swarn->errstr, + swarn->logical, swarn->dev->name, + (unsigned long long)swarn->sector, root, inum, offset, + min(isize - offset, (u64)PAGE_SIZE), nlink, + (char *)(unsigned long)ipath->fspath->val[i]); + + free_ipath(ipath); + return 0; + +err: + printk(KERN_WARNING "btrfs: %s at logical %llu on dev " + "%s, sector %llu, root %llu, inode %llu, offset %llu: path " + "resolving failed with ret=%d\n", swarn->errstr, + swarn->logical, swarn->dev->name, + (unsigned long long)swarn->sector, root, inum, offset, ret); + + free_ipath(ipath); + return 0; +} + +static void scrub_print_warning(const char *errstr, struct scrub_bio *sbio, + int ix) +{ + struct btrfs_device *dev = sbio->sdev->dev; + struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; + struct btrfs_path *path; + struct btrfs_key found_key; + struct extent_buffer *eb; + struct btrfs_extent_item *ei; + struct scrub_warning swarn; + u32 item_size; + int ret; + u64 ref_root; + u8 ref_level; + unsigned long ptr = 0; + const int bufsize = 4096; + u64 extent_offset; + + path = btrfs_alloc_path(); + + swarn.scratch_buf = kmalloc(bufsize, GFP_NOFS); + swarn.msg_buf = kmalloc(bufsize, GFP_NOFS); + swarn.sector = (sbio->physical + ix * PAGE_SIZE) >> 9; + swarn.logical = sbio->logical + ix * PAGE_SIZE; + swarn.errstr = errstr; + swarn.dev = dev; + swarn.msg_bufsize = bufsize; + swarn.scratch_bufsize = bufsize; + + if (!path || !swarn.scratch_buf || !swarn.msg_buf) + goto out; + + ret = extent_from_logical(fs_info, swarn.logical, path, &found_key); + if (ret < 0) + goto out; + + extent_offset = swarn.logical - found_key.objectid; + swarn.extent_item_size = found_key.offset; + + eb = path->nodes[0]; + ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item); + item_size = btrfs_item_size_nr(eb, path->slots[0]); + + if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + do { + ret = tree_backref_for_extent(&ptr, eb, ei, item_size, + &ref_root, &ref_level); + printk(KERN_WARNING "%s at logical %llu on dev %s, " + "sector %llu: metadata %s (level %d) in tree " + "%llu\n", errstr, swarn.logical, dev->name, + (unsigned long long)swarn.sector, + ref_level ? "node" : "leaf", + ret < 0 ? -1 : ref_level, + ret < 0 ? -1 : ref_root); + } while (ret != 1); + } else { + swarn.path = path; + iterate_extent_inodes(fs_info, path, found_key.objectid, + extent_offset, + scrub_print_warning_inode, &swarn); + } + +out: + btrfs_free_path(path); + kfree(swarn.scratch_buf); + kfree(swarn.msg_buf); +} + +static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *ctx) +{ + struct page *page = NULL; + unsigned long index; + struct scrub_fixup_nodatasum *fixup = ctx; + int ret; + int corrected = 0; + struct btrfs_key key; + struct inode *inode = NULL; + u64 end = offset + PAGE_SIZE - 1; + struct btrfs_root *local_root; + + key.objectid = root; + key.type = BTRFS_ROOT_ITEM_KEY; + key.offset = (u64)-1; + local_root = btrfs_read_fs_root_no_name(fixup->root->fs_info, &key); + if (IS_ERR(local_root)) + return PTR_ERR(local_root); + + key.type = BTRFS_INODE_ITEM_KEY; + key.objectid = inum; + key.offset = 0; + inode = btrfs_iget(fixup->root->fs_info->sb, &key, local_root, NULL); + if (IS_ERR(inode)) + return PTR_ERR(inode); + + index = offset >> PAGE_CACHE_SHIFT; + + page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); + if (!page) { + ret = -ENOMEM; + goto out; + } + + if (PageUptodate(page)) { + struct btrfs_mapping_tree *map_tree; + if (PageDirty(page)) { + /* + * we need to write the data to the defect sector. the + * data that was in that sector is not in memory, + * because the page was modified. we must not write the + * modified page to that sector. + * + * TODO: what could be done here: wait for the delalloc + * runner to write out that page (might involve + * COW) and see whether the sector is still + * referenced afterwards. + * + * For the meantime, we'll treat this error + * incorrectable, although there is a chance that a + * later scrub will find the bad sector again and that + * there's no dirty page in memory, then. + */ + ret = -EIO; + goto out; + } + map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree; + ret = repair_io_failure(map_tree, offset, PAGE_SIZE, + fixup->logical, page, + fixup->mirror_num); + unlock_page(page); + corrected = !ret; + } else { + /* + * we need to get good data first. the general readpage path + * will call repair_io_failure for us, we just have to make + * sure we read the bad mirror. + */ + ret = set_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, + EXTENT_DAMAGED, GFP_NOFS); + if (ret) { + /* set_extent_bits should give proper error */ + WARN_ON(ret > 0); + if (ret > 0) + ret = -EFAULT; + goto out; + } + + ret = extent_read_full_page(&BTRFS_I(inode)->io_tree, page, + btrfs_get_extent, + fixup->mirror_num); + wait_on_page_locked(page); + + corrected = !test_range_bit(&BTRFS_I(inode)->io_tree, offset, + end, EXTENT_DAMAGED, 0, NULL); + if (!corrected) + clear_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, + EXTENT_DAMAGED, GFP_NOFS); + } + +out: + if (page) + put_page(page); + if (inode) + iput(inode); + + if (ret < 0) + return ret; + + if (ret == 0 && corrected) { + /* + * we only need to call readpage for one of the inodes belonging + * to this extent. so make iterate_extent_inodes stop + */ + return 1; + } + + return -EIO; +} + +static void scrub_fixup_nodatasum(struct btrfs_work *work) +{ + int ret; + struct scrub_fixup_nodatasum *fixup; + struct scrub_dev *sdev; + struct btrfs_trans_handle *trans = NULL; + struct btrfs_fs_info *fs_info; + struct btrfs_path *path; + int uncorrectable = 0; + + fixup = container_of(work, struct scrub_fixup_nodatasum, work); + sdev = fixup->sdev; + fs_info = fixup->root->fs_info; + + path = btrfs_alloc_path(); + if (!path) { + spin_lock(&sdev->stat_lock); + ++sdev->stat.malloc_errors; + spin_unlock(&sdev->stat_lock); + uncorrectable = 1; + goto out; + } + + trans = btrfs_join_transaction(fixup->root); + if (IS_ERR(trans)) { + uncorrectable = 1; + goto out; + } + + /* + * the idea is to trigger a regular read through the standard path. we + * read a page from the (failed) logical address by specifying the + * corresponding copynum of the failed sector. thus, that readpage is + * expected to fail. + * that is the point where on-the-fly error correction will kick in + * (once it's finished) and rewrite the failed sector if a good copy + * can be found. + */ + ret = iterate_inodes_from_logical(fixup->logical, fixup->root->fs_info, + path, scrub_fixup_readpage, + fixup); + if (ret < 0) { + uncorrectable = 1; + goto out; + } + WARN_ON(ret != 1); + + spin_lock(&sdev->stat_lock); + ++sdev->stat.corrected_errors; + spin_unlock(&sdev->stat_lock); + +out: + if (trans && !IS_ERR(trans)) + btrfs_end_transaction(trans, fixup->root); + if (uncorrectable) { + spin_lock(&sdev->stat_lock); + ++sdev->stat.uncorrectable_errors; + spin_unlock(&sdev->stat_lock); + printk_ratelimited(KERN_ERR "btrfs: unable to fixup " + "(nodatasum) error at logical %llu\n", + fixup->logical); + } + + btrfs_free_path(path); + kfree(fixup); + + /* see caller why we're pretending to be paused in the scrub counters */ + mutex_lock(&fs_info->scrub_lock); + atomic_dec(&fs_info->scrubs_running); + atomic_dec(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + atomic_dec(&sdev->fixup_cnt); + wake_up(&fs_info->scrub_pause_wait); + wake_up(&sdev->list_wait); +} + /* * scrub_recheck_error gets called when either verification of the page * failed or the bio failed to read, e.g. with EIO. In the latter case, * recheck_error gets called for every page in the bio, even though only * one may be bad */ -static void scrub_recheck_error(struct scrub_bio *sbio, int ix) +static int scrub_recheck_error(struct scrub_bio *sbio, int ix) { + struct scrub_dev *sdev = sbio->sdev; + u64 sector = (sbio->physical + ix * PAGE_SIZE) >> 9; + static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, + DEFAULT_RATELIMIT_BURST); + if (sbio->err) { - if (scrub_fixup_io(READ, sbio->sdev->dev->bdev, - (sbio->physical + ix * PAGE_SIZE) >> 9, + if (scrub_fixup_io(READ, sbio->sdev->dev->bdev, sector, sbio->bio->bi_io_vec[ix].bv_page) == 0) { if (scrub_fixup_check(sbio, ix) == 0) - return; + return 0; } + if (__ratelimit(&_rs)) + scrub_print_warning("i/o error", sbio, ix); + } else { + if (__ratelimit(&_rs)) + scrub_print_warning("checksum error", sbio, ix); } + spin_lock(&sdev->stat_lock); + ++sdev->stat.read_errors; + spin_unlock(&sdev->stat_lock); + scrub_fixup(sbio, ix); + return 1; } static int scrub_fixup_check(struct scrub_bio *sbio, int ix) @@ -250,7 +616,8 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix) struct scrub_dev *sdev = sbio->sdev; struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; - struct btrfs_multi_bio *multi = NULL; + struct btrfs_bio *bbio = NULL; + struct scrub_fixup_nodatasum *fixup; u64 logical = sbio->logical + ix * PAGE_SIZE; u64 length; int i; @@ -259,38 +626,57 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix) if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) && (sbio->spag[ix].have_csum == 0)) { + fixup = kzalloc(sizeof(*fixup), GFP_NOFS); + if (!fixup) + goto uncorrectable; + fixup->sdev = sdev; + fixup->logical = logical; + fixup->root = fs_info->extent_root; + fixup->mirror_num = sbio->spag[ix].mirror_num; /* - * nodatasum, don't try to fix anything - * FIXME: we can do better, open the inode and trigger a - * writeback + * increment scrubs_running to prevent cancel requests from + * completing as long as a fixup worker is running. we must also + * increment scrubs_paused to prevent deadlocking on pause + * requests used for transactions commits (as the worker uses a + * transaction context). it is safe to regard the fixup worker + * as paused for all matters practical. effectively, we only + * avoid cancellation requests from completing. */ - goto uncorrectable; + mutex_lock(&fs_info->scrub_lock); + atomic_inc(&fs_info->scrubs_running); + atomic_inc(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + atomic_inc(&sdev->fixup_cnt); + fixup->work.func = scrub_fixup_nodatasum; + btrfs_queue_worker(&fs_info->scrub_workers, &fixup->work); + return; } length = PAGE_SIZE; ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length, - &multi, 0); - if (ret || !multi || length < PAGE_SIZE) { + &bbio, 0); + if (ret || !bbio || length < PAGE_SIZE) { printk(KERN_ERR "scrub_fixup: btrfs_map_block failed us for %llu\n", (unsigned long long)logical); WARN_ON(1); + kfree(bbio); return; } - if (multi->num_stripes == 1) + if (bbio->num_stripes == 1) /* there aren't any replicas */ goto uncorrectable; /* * first find a good copy */ - for (i = 0; i < multi->num_stripes; ++i) { - if (i == sbio->spag[ix].mirror_num) + for (i = 0; i < bbio->num_stripes; ++i) { + if (i + 1 == sbio->spag[ix].mirror_num) continue; - if (scrub_fixup_io(READ, multi->stripes[i].dev->bdev, - multi->stripes[i].physical >> 9, + if (scrub_fixup_io(READ, bbio->stripes[i].dev->bdev, + bbio->stripes[i].physical >> 9, sbio->bio->bi_io_vec[ix].bv_page)) { /* I/O-error, this is not a good copy */ continue; @@ -299,7 +685,7 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix) if (scrub_fixup_check(sbio, ix) == 0) break; } - if (i == multi->num_stripes) + if (i == bbio->num_stripes) goto uncorrectable; if (!sdev->readonly) { @@ -314,25 +700,23 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix) } } - kfree(multi); + kfree(bbio); spin_lock(&sdev->stat_lock); ++sdev->stat.corrected_errors; spin_unlock(&sdev->stat_lock); - if (printk_ratelimit()) - printk(KERN_ERR "btrfs: fixed up at %llu\n", - (unsigned long long)logical); + printk_ratelimited(KERN_ERR "btrfs: fixed up error at logical %llu\n", + (unsigned long long)logical); return; uncorrectable: - kfree(multi); + kfree(bbio); spin_lock(&sdev->stat_lock); ++sdev->stat.uncorrectable_errors; spin_unlock(&sdev->stat_lock); - if (printk_ratelimit()) - printk(KERN_ERR "btrfs: unable to fixup at %llu\n", - (unsigned long long)logical); + printk_ratelimited(KERN_ERR "btrfs: unable to fixup (regular) error at " + "logical %llu\n", (unsigned long long)logical); } static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector, @@ -382,8 +766,14 @@ static void scrub_checksum(struct btrfs_work *work) int ret; if (sbio->err) { + ret = 0; for (i = 0; i < sbio->count; ++i) - scrub_recheck_error(sbio, i); + ret |= scrub_recheck_error(sbio, i); + if (!ret) { + spin_lock(&sdev->stat_lock); + ++sdev->stat.unverified_errors; + spin_unlock(&sdev->stat_lock); + } sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1); sbio->bio->bi_flags |= 1 << BIO_UPTODATE; @@ -396,10 +786,6 @@ static void scrub_checksum(struct btrfs_work *work) bi->bv_offset = 0; bi->bv_len = PAGE_SIZE; } - - spin_lock(&sdev->stat_lock); - ++sdev->stat.read_errors; - spin_unlock(&sdev->stat_lock); goto out; } for (i = 0; i < sbio->count; ++i) { @@ -420,8 +806,14 @@ static void scrub_checksum(struct btrfs_work *work) WARN_ON(1); } kunmap_atomic(buffer, KM_USER0); - if (ret) - scrub_recheck_error(sbio, i); + if (ret) { + ret = scrub_recheck_error(sbio, i); + if (!ret) { + spin_lock(&sdev->stat_lock); + ++sdev->stat.unverified_errors; + spin_unlock(&sdev->stat_lock); + } + } } out: @@ -557,57 +949,27 @@ static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer) static int scrub_submit(struct scrub_dev *sdev) { struct scrub_bio *sbio; - struct bio *bio; - int i; if (sdev->curr == -1) return 0; sbio = sdev->bios[sdev->curr]; - - bio = bio_alloc(GFP_NOFS, sbio->count); - if (!bio) - goto nomem; - - bio->bi_private = sbio; - bio->bi_end_io = scrub_bio_end_io; - bio->bi_bdev = sdev->dev->bdev; - bio->bi_sector = sbio->physical >> 9; - - for (i = 0; i < sbio->count; ++i) { - struct page *page; - int ret; - - page = alloc_page(GFP_NOFS); - if (!page) - goto nomem; - - ret = bio_add_page(bio, page, PAGE_SIZE, 0); - if (!ret) { - __free_page(page); - goto nomem; - } - } - sbio->err = 0; sdev->curr = -1; atomic_inc(&sdev->in_flight); - submit_bio(READ, bio); + submit_bio(READ, sbio->bio); return 0; - -nomem: - scrub_free_bio(bio); - - return -ENOMEM; } static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len, - u64 physical, u64 flags, u64 gen, u64 mirror_num, + u64 physical, u64 flags, u64 gen, int mirror_num, u8 *csum, int force) { struct scrub_bio *sbio; + struct page *page; + int ret; again: /* @@ -628,12 +990,22 @@ again: } sbio = sdev->bios[sdev->curr]; if (sbio->count == 0) { + struct bio *bio; + sbio->physical = physical; sbio->logical = logical; + bio = bio_alloc(GFP_NOFS, SCRUB_PAGES_PER_BIO); + if (!bio) + return -ENOMEM; + + bio->bi_private = sbio; + bio->bi_end_io = scrub_bio_end_io; + bio->bi_bdev = sdev->dev->bdev; + bio->bi_sector = sbio->physical >> 9; + sbio->err = 0; + sbio->bio = bio; } else if (sbio->physical + sbio->count * PAGE_SIZE != physical || sbio->logical + sbio->count * PAGE_SIZE != logical) { - int ret; - ret = scrub_submit(sdev); if (ret) return ret; @@ -643,6 +1015,20 @@ again: sbio->spag[sbio->count].generation = gen; sbio->spag[sbio->count].have_csum = 0; sbio->spag[sbio->count].mirror_num = mirror_num; + + page = alloc_page(GFP_NOFS); + if (!page) + return -ENOMEM; + + ret = bio_add_page(sbio->bio, page, PAGE_SIZE, 0); + if (!ret) { + __free_page(page); + ret = scrub_submit(sdev); + if (ret) + return ret; + goto again; + } + if (csum) { sbio->spag[sbio->count].have_csum = 1; memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size); @@ -701,7 +1087,7 @@ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len, /* scrub extent tries to collect up to 64 kB for each bio */ static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len, - u64 physical, u64 flags, u64 gen, u64 mirror_num) + u64 physical, u64 flags, u64 gen, int mirror_num) { int ret; u8 csum[BTRFS_CSUM_SIZE]; @@ -741,13 +1127,16 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev, int slot; int i; u64 nstripes; - int start_stripe; struct extent_buffer *l; struct btrfs_key key; u64 physical; u64 logical; u64 generation; - u64 mirror_num; + int mirror_num; + struct reada_control *reada1; + struct reada_control *reada2; + struct btrfs_key key_start; + struct btrfs_key key_end; u64 increment = map->stripe_len; u64 offset; @@ -758,102 +1147,88 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev, if (map->type & BTRFS_BLOCK_GROUP_RAID0) { offset = map->stripe_len * num; increment = map->stripe_len * map->num_stripes; - mirror_num = 0; + mirror_num = 1; } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { int factor = map->num_stripes / map->sub_stripes; offset = map->stripe_len * (num / map->sub_stripes); increment = map->stripe_len * factor; - mirror_num = num % map->sub_stripes; + mirror_num = num % map->sub_stripes + 1; } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { increment = map->stripe_len; - mirror_num = num % map->num_stripes; + mirror_num = num % map->num_stripes + 1; } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { increment = map->stripe_len; - mirror_num = num % map->num_stripes; + mirror_num = num % map->num_stripes + 1; } else { increment = map->stripe_len; - mirror_num = 0; + mirror_num = 1; } path = btrfs_alloc_path(); if (!path) return -ENOMEM; - path->reada = 2; path->search_commit_root = 1; path->skip_locking = 1; /* - * find all extents for each stripe and just read them to get - * them into the page cache - * FIXME: we can do better. build a more intelligent prefetching + * trigger the readahead for extent tree csum tree and wait for + * completion. During readahead, the scrub is officially paused + * to not hold off transaction commits */ logical = base + offset; - physical = map->stripes[num].physical; - ret = 0; - for (i = 0; i < nstripes; ++i) { - key.objectid = logical; - key.type = BTRFS_EXTENT_ITEM_KEY; - key.offset = (u64)0; - - ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); - if (ret < 0) - goto out_noplug; - /* - * we might miss half an extent here, but that doesn't matter, - * as it's only the prefetch - */ - while (1) { - l = path->nodes[0]; - slot = path->slots[0]; - if (slot >= btrfs_header_nritems(l)) { - ret = btrfs_next_leaf(root, path); - if (ret == 0) - continue; - if (ret < 0) - goto out_noplug; - - break; - } - btrfs_item_key_to_cpu(l, &key, slot); + wait_event(sdev->list_wait, + atomic_read(&sdev->in_flight) == 0); + atomic_inc(&fs_info->scrubs_paused); + wake_up(&fs_info->scrub_pause_wait); - if (key.objectid >= logical + map->stripe_len) - break; + /* FIXME it might be better to start readahead at commit root */ + key_start.objectid = logical; + key_start.type = BTRFS_EXTENT_ITEM_KEY; + key_start.offset = (u64)0; + key_end.objectid = base + offset + nstripes * increment; + key_end.type = BTRFS_EXTENT_ITEM_KEY; + key_end.offset = (u64)0; + reada1 = btrfs_reada_add(root, &key_start, &key_end); + + key_start.objectid = BTRFS_EXTENT_CSUM_OBJECTID; + key_start.type = BTRFS_EXTENT_CSUM_KEY; + key_start.offset = logical; + key_end.objectid = BTRFS_EXTENT_CSUM_OBJECTID; + key_end.type = BTRFS_EXTENT_CSUM_KEY; + key_end.offset = base + offset + nstripes * increment; + reada2 = btrfs_reada_add(csum_root, &key_start, &key_end); + + if (!IS_ERR(reada1)) + btrfs_reada_wait(reada1); + if (!IS_ERR(reada2)) + btrfs_reada_wait(reada2); - path->slots[0]++; - } - btrfs_release_path(path); - logical += increment; - physical += map->stripe_len; - cond_resched(); + mutex_lock(&fs_info->scrub_lock); + while (atomic_read(&fs_info->scrub_pause_req)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrub_pause_req) == 0); + mutex_lock(&fs_info->scrub_lock); } + atomic_dec(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + wake_up(&fs_info->scrub_pause_wait); /* * collect all data csums for the stripe to avoid seeking during * the scrub. This might currently (crc32) end up to be about 1MB */ - start_stripe = 0; blk_start_plug(&plug); -again: - logical = base + offset + start_stripe * increment; - for (i = start_stripe; i < nstripes; ++i) { - ret = btrfs_lookup_csums_range(csum_root, logical, - logical + map->stripe_len - 1, - &sdev->csum_list, 1); - if (ret) - goto out; - logical += increment; - cond_resched(); - } /* * now find all extents for each stripe and scrub them */ - logical = base + offset + start_stripe * increment; - physical = map->stripes[num].physical + start_stripe * map->stripe_len; + logical = base + offset; + physical = map->stripes[num].physical; ret = 0; - for (i = start_stripe; i < nstripes; ++i) { + for (i = 0; i < nstripes; ++i) { /* * canceled? */ @@ -882,11 +1257,14 @@ again: atomic_dec(&fs_info->scrubs_paused); mutex_unlock(&fs_info->scrub_lock); wake_up(&fs_info->scrub_pause_wait); - scrub_free_csums(sdev); - start_stripe = i; - goto again; } + ret = btrfs_lookup_csums_range(csum_root, logical, + logical + map->stripe_len - 1, + &sdev->csum_list, 1); + if (ret) + goto out; + key.objectid = logical; key.type = BTRFS_EXTENT_ITEM_KEY; key.offset = (u64)0; @@ -982,7 +1360,6 @@ next: out: blk_finish_plug(&plug); -out_noplug: btrfs_free_path(path); return ret < 0 ? ret : 0; } @@ -1158,18 +1535,22 @@ static noinline_for_stack int scrub_supers(struct scrub_dev *sdev) static noinline_for_stack int scrub_workers_get(struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; + int ret = 0; mutex_lock(&fs_info->scrub_lock); if (fs_info->scrub_workers_refcnt == 0) { btrfs_init_workers(&fs_info->scrub_workers, "scrub", fs_info->thread_pool_size, &fs_info->generic_worker); fs_info->scrub_workers.idle_thresh = 4; - btrfs_start_workers(&fs_info->scrub_workers, 1); + ret = btrfs_start_workers(&fs_info->scrub_workers); + if (ret) + goto out; } ++fs_info->scrub_workers_refcnt; +out: mutex_unlock(&fs_info->scrub_lock); - return 0; + return ret; } static noinline_for_stack void scrub_workers_put(struct btrfs_root *root) @@ -1253,10 +1634,11 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, ret = scrub_enumerate_chunks(sdev, start, end); wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0); - atomic_dec(&fs_info->scrubs_running); wake_up(&fs_info->scrub_pause_wait); + wait_event(sdev->list_wait, atomic_read(&sdev->fixup_cnt) == 0); + if (progress) memcpy(progress, &sdev->stat, sizeof(*progress)); diff --git a/fs/btrfs/struct-funcs.c b/fs/btrfs/struct-funcs.c index c0f7eca..bc1f6ad 100644 --- a/fs/btrfs/struct-funcs.c +++ b/fs/btrfs/struct-funcs.c @@ -50,36 +50,22 @@ u##bits btrfs_##name(struct extent_buffer *eb, \ unsigned long part_offset = (unsigned long)s; \ unsigned long offset = part_offset + offsetof(type, member); \ type *p; \ - /* ugly, but we want the fast path here */ \ - if (eb->map_token && offset >= eb->map_start && \ - offset + sizeof(((type *)0)->member) <= eb->map_start + \ - eb->map_len) { \ - p = (type *)(eb->kaddr + part_offset - eb->map_start); \ - return le##bits##_to_cpu(p->member); \ - } \ - { \ - int err; \ - char *map_token; \ - char *kaddr; \ - int unmap_on_exit = (eb->map_token == NULL); \ - unsigned long map_start; \ - unsigned long map_len; \ - u##bits res; \ - err = map_extent_buffer(eb, offset, \ - sizeof(((type *)0)->member), \ - &map_token, &kaddr, \ - &map_start, &map_len, KM_USER1); \ - if (err) { \ - __le##bits leres; \ - read_eb_member(eb, s, type, member, &leres); \ - return le##bits##_to_cpu(leres); \ - } \ - p = (type *)(kaddr + part_offset - map_start); \ - res = le##bits##_to_cpu(p->member); \ - if (unmap_on_exit) \ - unmap_extent_buffer(eb, map_token, KM_USER1); \ - return res; \ - } \ + int err; \ + char *kaddr; \ + unsigned long map_start; \ + unsigned long map_len; \ + u##bits res; \ + err = map_private_extent_buffer(eb, offset, \ + sizeof(((type *)0)->member), \ + &kaddr, &map_start, &map_len); \ + if (err) { \ + __le##bits leres; \ + read_eb_member(eb, s, type, member, &leres); \ + return le##bits##_to_cpu(leres); \ + } \ + p = (type *)(kaddr + part_offset - map_start); \ + res = le##bits##_to_cpu(p->member); \ + return res; \ } \ void btrfs_set_##name(struct extent_buffer *eb, \ type *s, u##bits val) \ @@ -87,36 +73,21 @@ void btrfs_set_##name(struct extent_buffer *eb, \ unsigned long part_offset = (unsigned long)s; \ unsigned long offset = part_offset + offsetof(type, member); \ type *p; \ - /* ugly, but we want the fast path here */ \ - if (eb->map_token && offset >= eb->map_start && \ - offset + sizeof(((type *)0)->member) <= eb->map_start + \ - eb->map_len) { \ - p = (type *)(eb->kaddr + part_offset - eb->map_start); \ - p->member = cpu_to_le##bits(val); \ - return; \ - } \ - { \ - int err; \ - char *map_token; \ - char *kaddr; \ - int unmap_on_exit = (eb->map_token == NULL); \ - unsigned long map_start; \ - unsigned long map_len; \ - err = map_extent_buffer(eb, offset, \ - sizeof(((type *)0)->member), \ - &map_token, &kaddr, \ - &map_start, &map_len, KM_USER1); \ - if (err) { \ - __le##bits val2; \ - val2 = cpu_to_le##bits(val); \ - write_eb_member(eb, s, type, member, &val2); \ - return; \ - } \ - p = (type *)(kaddr + part_offset - map_start); \ - p->member = cpu_to_le##bits(val); \ - if (unmap_on_exit) \ - unmap_extent_buffer(eb, map_token, KM_USER1); \ - } \ + int err; \ + char *kaddr; \ + unsigned long map_start; \ + unsigned long map_len; \ + err = map_private_extent_buffer(eb, offset, \ + sizeof(((type *)0)->member), \ + &kaddr, &map_start, &map_len); \ + if (err) { \ + __le##bits val2; \ + val2 = cpu_to_le##bits(val); \ + write_eb_member(eb, s, type, member, &val2); \ + return; \ + } \ + p = (type *)(kaddr + part_offset - map_start); \ + p->member = cpu_to_le##bits(val); \ } #include "ctree.h" @@ -125,15 +96,6 @@ void btrfs_node_key(struct extent_buffer *eb, struct btrfs_disk_key *disk_key, int nr) { unsigned long ptr = btrfs_node_key_ptr_offset(nr); - if (eb->map_token && ptr >= eb->map_start && - ptr + sizeof(*disk_key) <= eb->map_start + eb->map_len) { - memcpy(disk_key, eb->kaddr + ptr - eb->map_start, - sizeof(*disk_key)); - return; - } else if (eb->map_token) { - unmap_extent_buffer(eb, eb->map_token, KM_USER1); - eb->map_token = NULL; - } read_eb_member(eb, (struct btrfs_key_ptr *)ptr, struct btrfs_key_ptr, key, disk_key); } diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 15634d4..200f63b 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -40,6 +40,8 @@ #include <linux/magic.h> #include <linux/slab.h> #include <linux/cleancache.h> +#include <linux/mnt_namespace.h> +#include <linux/ratelimit.h> #include "compat.h" #include "delayed-inode.h" #include "ctree.h" @@ -58,6 +60,7 @@ #include <trace/events/btrfs.h> static const struct super_operations btrfs_super_ops; +static struct file_system_type btrfs_fs_type; static const char *btrfs_decode_error(struct btrfs_fs_info *fs_info, int errno, char nbuf[16]) @@ -162,7 +165,7 @@ enum { Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard, Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed, Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, - Opt_inode_cache, Opt_err, + Opt_inode_cache, Opt_no_space_cache, Opt_recovery, Opt_err, }; static match_table_t tokens = { @@ -195,6 +198,8 @@ static match_table_t tokens = { {Opt_subvolrootid, "subvolrootid=%d"}, {Opt_defrag, "autodefrag"}, {Opt_inode_cache, "inode_cache"}, + {Opt_no_space_cache, "nospace_cache"}, + {Opt_recovery, "recovery"}, {Opt_err, NULL}, }; @@ -206,14 +211,19 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) { struct btrfs_fs_info *info = root->fs_info; substring_t args[MAX_OPT_ARGS]; - char *p, *num, *orig; + char *p, *num, *orig = NULL; + u64 cache_gen; int intarg; int ret = 0; char *compress_type; bool compress_force = false; + cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy); + if (cache_gen) + btrfs_set_opt(info->mount_opt, SPACE_CACHE); + if (!options) - return 0; + goto out; /* * strsep changes the string, duplicate it because parse_options @@ -360,9 +370,12 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) btrfs_set_opt(info->mount_opt, DISCARD); break; case Opt_space_cache: - printk(KERN_INFO "btrfs: enabling disk space caching\n"); btrfs_set_opt(info->mount_opt, SPACE_CACHE); break; + case Opt_no_space_cache: + printk(KERN_INFO "btrfs: disabling disk space caching\n"); + btrfs_clear_opt(info->mount_opt, SPACE_CACHE); + break; case Opt_inode_cache: printk(KERN_INFO "btrfs: enabling inode map caching\n"); btrfs_set_opt(info->mount_opt, INODE_MAP_CACHE); @@ -381,6 +394,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) printk(KERN_INFO "btrfs: enabling auto defrag"); btrfs_set_opt(info->mount_opt, AUTO_DEFRAG); break; + case Opt_recovery: + printk(KERN_INFO "btrfs: enabling auto recovery"); + btrfs_set_opt(info->mount_opt, RECOVERY); + break; case Opt_err: printk(KERN_INFO "btrfs: unrecognized mount option " "'%s'\n", p); @@ -391,6 +408,8 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) } } out: + if (!ret && btrfs_test_opt(root, SPACE_CACHE)) + printk(KERN_INFO "btrfs: disk space caching is enabled\n"); kfree(orig); return ret; } @@ -406,12 +425,12 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags, u64 *subvol_rootid, struct btrfs_fs_devices **fs_devices) { substring_t args[MAX_OPT_ARGS]; - char *opts, *orig, *p; + char *device_name, *opts, *orig, *p; int error = 0; int intarg; if (!options) - goto out; + return 0; /* * strsep changes the string, duplicate it because parse_options @@ -430,6 +449,7 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags, token = match_token(p, tokens, args); switch (token) { case Opt_subvol: + kfree(*subvol_name); *subvol_name = match_strdup(&args[0]); break; case Opt_subvolid: @@ -457,29 +477,24 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags, } break; case Opt_device: - error = btrfs_scan_one_device(match_strdup(&args[0]), + device_name = match_strdup(&args[0]); + if (!device_name) { + error = -ENOMEM; + goto out; + } + error = btrfs_scan_one_device(device_name, flags, holder, fs_devices); + kfree(device_name); if (error) - goto out_free_opts; + goto out; break; default: break; } } - out_free_opts: +out: kfree(orig); - out: - /* - * If no subvolume name is specified we use the default one. Allocate - * a copy of the string "." here so that code later in the - * mount path doesn't care if it's the default volume or another one. - */ - if (!*subvol_name) { - *subvol_name = kstrdup(".", GFP_KERNEL); - if (!*subvol_name) - return -ENOMEM; - } return error; } @@ -492,7 +507,6 @@ static struct dentry *get_default_root(struct super_block *sb, struct btrfs_path *path; struct btrfs_key location; struct inode *inode; - struct dentry *dentry; u64 dir_id; int new = 0; @@ -517,7 +531,7 @@ static struct dentry *get_default_root(struct super_block *sb, * will mount by default if we haven't been given a specific subvolume * to mount. */ - dir_id = btrfs_super_root_dir(&root->fs_info->super_copy); + dir_id = btrfs_super_root_dir(root->fs_info->super_copy); di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0); if (IS_ERR(di)) { btrfs_free_path(path); @@ -566,29 +580,7 @@ setup_root: return dget(sb->s_root); } - if (new) { - const struct qstr name = { .name = "/", .len = 1 }; - - /* - * New inode, we need to make the dentry a sibling of s_root so - * everything gets cleaned up properly on unmount. - */ - dentry = d_alloc(sb->s_root, &name); - if (!dentry) { - iput(inode); - return ERR_PTR(-ENOMEM); - } - d_splice_alias(inode, dentry); - } else { - /* - * We found the inode in cache, just find a dentry for it and - * put the reference to the inode we just got. - */ - dentry = d_find_alias(inode); - iput(inode); - } - - return dentry; + return d_obtain_alias(inode); } static int btrfs_fill_super(struct super_block *sb, @@ -719,6 +711,8 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) seq_puts(seq, ",noacl"); if (btrfs_test_opt(root, SPACE_CACHE)) seq_puts(seq, ",space_cache"); + else + seq_puts(seq, ",nospace_cache"); if (btrfs_test_opt(root, CLEAR_CACHE)) seq_puts(seq, ",clear_cache"); if (btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED)) @@ -753,6 +747,111 @@ static int btrfs_set_super(struct super_block *s, void *data) return set_anon_super(s, data); } +/* + * subvolumes are identified by ino 256 + */ +static inline int is_subvolume_inode(struct inode *inode) +{ + if (inode && inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) + return 1; + return 0; +} + +/* + * This will strip out the subvol=%s argument for an argument string and add + * subvolid=0 to make sure we get the actual tree root for path walking to the + * subvol we want. + */ +static char *setup_root_args(char *args) +{ + unsigned copied = 0; + unsigned len = strlen(args) + 2; + char *pos; + char *ret; + + /* + * We need the same args as before, but minus + * + * subvol=a + * + * and add + * + * subvolid=0 + * + * which is a difference of 2 characters, so we allocate strlen(args) + + * 2 characters. + */ + ret = kzalloc(len * sizeof(char), GFP_NOFS); + if (!ret) + return NULL; + pos = strstr(args, "subvol="); + + /* This shouldn't happen, but just in case.. */ + if (!pos) { + kfree(ret); + return NULL; + } + + /* + * The subvol=<> arg is not at the front of the string, copy everybody + * up to that into ret. + */ + if (pos != args) { + *pos = '\0'; + strcpy(ret, args); + copied += strlen(args); + pos++; + } + + strncpy(ret + copied, "subvolid=0", len - copied); + + /* Length of subvolid=0 */ + copied += 10; + + /* + * If there is no , after the subvol= option then we know there's no + * other options and we can just return. + */ + pos = strchr(pos, ','); + if (!pos) + return ret; + + /* Copy the rest of the arguments into our buffer */ + strncpy(ret + copied, pos, len - copied); + copied += strlen(pos); + + return ret; +} + +static struct dentry *mount_subvol(const char *subvol_name, int flags, + const char *device_name, char *data) +{ + struct dentry *root; + struct vfsmount *mnt; + char *newargs; + + newargs = setup_root_args(data); + if (!newargs) + return ERR_PTR(-ENOMEM); + mnt = vfs_kern_mount(&btrfs_fs_type, flags, device_name, + newargs); + kfree(newargs); + if (IS_ERR(mnt)) + return ERR_CAST(mnt); + + root = mount_subtree(mnt, subvol_name); + + if (!IS_ERR(root) && !is_subvolume_inode(root->d_inode)) { + struct super_block *s = root->d_sb; + dput(root); + root = ERR_PTR(-EINVAL); + deactivate_locked_super(s); + printk(KERN_ERR "btrfs: '%s' is not a valid subvolume\n", + subvol_name); + } + + return root; +} /* * Find a superblock for the given device / mount point. @@ -767,7 +866,6 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, struct super_block *s; struct dentry *root; struct btrfs_fs_devices *fs_devices = NULL; - struct btrfs_root *tree_root = NULL; struct btrfs_fs_info *fs_info = NULL; fmode_t mode = FMODE_READ; char *subvol_name = NULL; @@ -781,21 +879,20 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, error = btrfs_parse_early_options(data, mode, fs_type, &subvol_name, &subvol_objectid, &subvol_rootid, &fs_devices); - if (error) + if (error) { + kfree(subvol_name); return ERR_PTR(error); + } - error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices); - if (error) - goto error_free_subvol_name; + if (subvol_name) { + root = mount_subvol(subvol_name, flags, device_name, data); + kfree(subvol_name); + return root; + } - error = btrfs_open_devices(fs_devices, mode, fs_type); + error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices); if (error) - goto error_free_subvol_name; - - if (!(flags & MS_RDONLY) && fs_devices->rw_devices == 0) { - error = -EACCES; - goto error_close_devices; - } + return ERR_PTR(error); /* * Setup a dummy root and fs_info for test/set super. This is because @@ -804,19 +901,40 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, * then open_ctree will properly initialize everything later. */ fs_info = kzalloc(sizeof(struct btrfs_fs_info), GFP_NOFS); - tree_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); - if (!fs_info || !tree_root) { + if (!fs_info) + return ERR_PTR(-ENOMEM); + + fs_info->tree_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); + if (!fs_info->tree_root) { error = -ENOMEM; - goto error_close_devices; + goto error_fs_info; } - fs_info->tree_root = tree_root; + fs_info->tree_root->fs_info = fs_info; fs_info->fs_devices = fs_devices; - tree_root->fs_info = fs_info; + + fs_info->super_copy = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS); + fs_info->super_for_commit = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS); + if (!fs_info->super_copy || !fs_info->super_for_commit) { + error = -ENOMEM; + goto error_fs_info; + } + + error = btrfs_open_devices(fs_devices, mode, fs_type); + if (error) + goto error_fs_info; + + if (!(flags & MS_RDONLY) && fs_devices->rw_devices == 0) { + error = -EACCES; + goto error_close_devices; + } bdev = fs_devices->latest_bdev; - s = sget(fs_type, btrfs_test_super, btrfs_set_super, tree_root); - if (IS_ERR(s)) - goto error_s; + s = sget(fs_type, btrfs_test_super, btrfs_set_super, + fs_info->tree_root); + if (IS_ERR(s)) { + error = PTR_ERR(s); + goto error_close_devices; + } if (s->s_root) { if ((flags ^ s->s_flags) & MS_RDONLY) { @@ -826,75 +944,35 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, } btrfs_close_devices(fs_devices); - kfree(fs_info); - kfree(tree_root); + free_fs_info(fs_info); } else { char b[BDEVNAME_SIZE]; s->s_flags = flags | MS_NOSEC; strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id)); + btrfs_sb(s)->fs_info->bdev_holder = fs_type; error = btrfs_fill_super(s, fs_devices, data, flags & MS_SILENT ? 1 : 0); if (error) { deactivate_locked_super(s); - goto error_free_subvol_name; + return ERR_PTR(error); } - btrfs_sb(s)->fs_info->bdev_holder = fs_type; s->s_flags |= MS_ACTIVE; } - /* if they gave us a subvolume name bind mount into that */ - if (strcmp(subvol_name, ".")) { - struct dentry *new_root; - - root = get_default_root(s, subvol_rootid); - if (IS_ERR(root)) { - error = PTR_ERR(root); - deactivate_locked_super(s); - goto error_free_subvol_name; - } - - mutex_lock(&root->d_inode->i_mutex); - new_root = lookup_one_len(subvol_name, root, - strlen(subvol_name)); - mutex_unlock(&root->d_inode->i_mutex); - - if (IS_ERR(new_root)) { - dput(root); - deactivate_locked_super(s); - error = PTR_ERR(new_root); - goto error_free_subvol_name; - } - if (!new_root->d_inode) { - dput(root); - dput(new_root); - deactivate_locked_super(s); - error = -ENXIO; - goto error_free_subvol_name; - } - dput(root); - root = new_root; - } else { - root = get_default_root(s, subvol_objectid); - if (IS_ERR(root)) { - error = PTR_ERR(root); - deactivate_locked_super(s); - goto error_free_subvol_name; - } + root = get_default_root(s, subvol_objectid); + if (IS_ERR(root)) { + deactivate_locked_super(s); + return root; } - kfree(subvol_name); return root; -error_s: - error = PTR_ERR(s); error_close_devices: btrfs_close_devices(fs_devices); - kfree(fs_info); - kfree(tree_root); -error_free_subvol_name: - kfree(subvol_name); +error_fs_info: + free_fs_info(fs_info); return ERR_PTR(error); } @@ -919,7 +997,7 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) if (root->fs_info->fs_devices->rw_devices == 0) return -EACCES; - if (btrfs_super_log_root(&root->fs_info->super_copy) != 0) + if (btrfs_super_log_root(root->fs_info->super_copy) != 0) return -EINVAL; ret = btrfs_cleanup_fs_roots(root->fs_info); @@ -976,11 +1054,11 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) u64 avail_space; u64 used_space; u64 min_stripe_size; - int min_stripes = 1; + int min_stripes = 1, num_stripes = 1; int i = 0, nr_devices; int ret; - nr_devices = fs_info->fs_devices->rw_devices; + nr_devices = fs_info->fs_devices->open_devices; BUG_ON(!nr_devices); devices_info = kmalloc(sizeof(*devices_info) * nr_devices, @@ -990,20 +1068,24 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) /* calc min stripe number for data space alloction */ type = btrfs_get_alloc_profile(root, 1); - if (type & BTRFS_BLOCK_GROUP_RAID0) + if (type & BTRFS_BLOCK_GROUP_RAID0) { min_stripes = 2; - else if (type & BTRFS_BLOCK_GROUP_RAID1) + num_stripes = nr_devices; + } else if (type & BTRFS_BLOCK_GROUP_RAID1) { min_stripes = 2; - else if (type & BTRFS_BLOCK_GROUP_RAID10) + num_stripes = 2; + } else if (type & BTRFS_BLOCK_GROUP_RAID10) { min_stripes = 4; + num_stripes = 4; + } if (type & BTRFS_BLOCK_GROUP_DUP) min_stripe_size = 2 * BTRFS_STRIPE_LEN; else min_stripe_size = BTRFS_STRIPE_LEN; - list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) { - if (!device->in_fs_metadata) + list_for_each_entry(device, &fs_devices->devices, dev_list) { + if (!device->in_fs_metadata || !device->bdev) continue; avail_space = device->total_bytes - device->bytes_used; @@ -1064,13 +1146,16 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) i = nr_devices - 1; avail_space = 0; while (nr_devices >= min_stripes) { + if (num_stripes > nr_devices) + num_stripes = nr_devices; + if (devices_info[i].max_avail >= min_stripe_size) { int j; u64 alloc_size; - avail_space += devices_info[i].max_avail * min_stripes; + avail_space += devices_info[i].max_avail * num_stripes; alloc_size = devices_info[i].max_avail; - for (j = i + 1 - min_stripes; j <= i; j++) + for (j = i + 1 - num_stripes; j <= i; j++) devices_info[j].max_avail -= alloc_size; } i--; @@ -1085,7 +1170,7 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) { struct btrfs_root *root = btrfs_sb(dentry->d_sb); - struct btrfs_super_block *disk_super = &root->fs_info->super_copy; + struct btrfs_super_block *disk_super = root->fs_info->super_copy; struct list_head *head = &root->fs_info->space_info; struct btrfs_space_info *found; u64 total_used = 0; @@ -1187,6 +1272,16 @@ static int btrfs_unfreeze(struct super_block *sb) return 0; } +static void btrfs_fs_dirty_inode(struct inode *inode, int flags) +{ + int ret; + + ret = btrfs_dirty_inode(inode); + if (ret) + printk_ratelimited(KERN_ERR "btrfs: fail to dirty inode %Lu " + "error %d\n", btrfs_ino(inode), ret); +} + static const struct super_operations btrfs_super_ops = { .drop_inode = btrfs_drop_inode, .evict_inode = btrfs_evict_inode, @@ -1194,7 +1289,7 @@ static const struct super_operations btrfs_super_ops = { .sync_fs = btrfs_sync_fs, .show_options = btrfs_show_options, .write_inode = btrfs_write_inode, - .dirty_inode = btrfs_dirty_inode, + .dirty_inode = btrfs_fs_dirty_inode, .alloc_inode = btrfs_alloc_inode, .destroy_inode = btrfs_destroy_inode, .statfs = btrfs_statfs, diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 51dcec8..292e847 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -55,6 +55,7 @@ static noinline int join_transaction(struct btrfs_root *root, int nofail) struct btrfs_transaction *cur_trans; spin_lock(&root->fs_info->trans_lock); +loop: if (root->fs_info->trans_no_join) { if (!nofail) { spin_unlock(&root->fs_info->trans_lock); @@ -75,16 +76,18 @@ static noinline int join_transaction(struct btrfs_root *root, int nofail) cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS); if (!cur_trans) return -ENOMEM; + spin_lock(&root->fs_info->trans_lock); if (root->fs_info->running_transaction) { + /* + * someone started a transaction after we unlocked. Make sure + * to redo the trans_no_join checks above + */ kmem_cache_free(btrfs_transaction_cachep, cur_trans); cur_trans = root->fs_info->running_transaction; - atomic_inc(&cur_trans->use_count); - atomic_inc(&cur_trans->num_writers); - cur_trans->num_joined++; - spin_unlock(&root->fs_info->trans_lock); - return 0; + goto loop; } + atomic_set(&cur_trans->num_writers, 1); cur_trans->num_joined = 0; init_waitqueue_head(&cur_trans->writer_wait); @@ -216,17 +219,11 @@ static void wait_current_trans(struct btrfs_root *root) spin_lock(&root->fs_info->trans_lock); cur_trans = root->fs_info->running_transaction; if (cur_trans && cur_trans->blocked) { - DEFINE_WAIT(wait); atomic_inc(&cur_trans->use_count); spin_unlock(&root->fs_info->trans_lock); - while (1) { - prepare_to_wait(&root->fs_info->transaction_wait, &wait, - TASK_UNINTERRUPTIBLE); - if (!cur_trans->blocked) - break; - schedule(); - } - finish_wait(&root->fs_info->transaction_wait, &wait); + + wait_event(root->fs_info->transaction_wait, + !cur_trans->blocked); put_transaction(cur_trans); } else { spin_unlock(&root->fs_info->trans_lock); @@ -260,7 +257,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root, { struct btrfs_trans_handle *h; struct btrfs_transaction *cur_trans; - int retries = 0; + u64 num_bytes = 0; int ret; if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) @@ -274,6 +271,19 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root, h->block_rsv = NULL; goto got_it; } + + /* + * Do the reservation before we join the transaction so we can do all + * the appropriate flushing if need be. + */ + if (num_items > 0 && root != root->fs_info->chunk_root) { + num_bytes = btrfs_calc_trans_metadata_size(root, num_items); + ret = btrfs_block_rsv_add(root, + &root->fs_info->trans_block_rsv, + num_bytes); + if (ret) + return ERR_PTR(ret); + } again: h = kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS); if (!h) @@ -310,24 +320,9 @@ again: goto again; } - if (num_items > 0) { - ret = btrfs_trans_reserve_metadata(h, root, num_items); - if (ret == -EAGAIN && !retries) { - retries++; - btrfs_commit_transaction(h, root); - goto again; - } else if (ret == -EAGAIN) { - /* - * We have already retried and got EAGAIN, so really we - * don't have space, so set ret to -ENOSPC. - */ - ret = -ENOSPC; - } - - if (ret < 0) { - btrfs_end_transaction(h, root); - return ERR_PTR(ret); - } + if (num_bytes) { + h->block_rsv = &root->fs_info->trans_block_rsv; + h->bytes_reserved = num_bytes; } got_it: @@ -359,19 +354,10 @@ struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *root } /* wait for a transaction commit to be fully complete */ -static noinline int wait_for_commit(struct btrfs_root *root, +static noinline void wait_for_commit(struct btrfs_root *root, struct btrfs_transaction *commit) { - DEFINE_WAIT(wait); - while (!commit->commit_done) { - prepare_to_wait(&commit->commit_wait, &wait, - TASK_UNINTERRUPTIBLE); - if (commit->commit_done) - break; - schedule(); - } - finish_wait(&commit->commit_wait, &wait); - return 0; + wait_event(commit->commit_wait, commit->commit_done); } int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid) @@ -435,8 +421,8 @@ static int should_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root) { int ret; - ret = btrfs_block_rsv_check(trans, root, - &root->fs_info->global_block_rsv, 0, 5); + + ret = btrfs_block_rsv_check(root, &root->fs_info->global_block_rsv, 5); return ret ? 1 : 0; } @@ -444,17 +430,26 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root) { struct btrfs_transaction *cur_trans = trans->transaction; + struct btrfs_block_rsv *rsv = trans->block_rsv; int updates; smp_mb(); if (cur_trans->blocked || cur_trans->delayed_refs.flushing) return 1; + /* + * We need to do this in case we're deleting csums so the global block + * rsv get's used instead of the csum block rsv. + */ + trans->block_rsv = NULL; + updates = trans->delayed_ref_updates; trans->delayed_ref_updates = 0; if (updates) btrfs_run_delayed_refs(trans, root, updates); + trans->block_rsv = rsv; + return should_end_transaction(trans, root); } @@ -465,11 +460,14 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_fs_info *info = root->fs_info; int count = 0; - if (--trans->use_count) { + if (trans->use_count > 1) { + trans->use_count--; trans->block_rsv = trans->orig_rsv; return 0; } + btrfs_trans_release_metadata(trans, root); + trans->block_rsv = NULL; while (count < 4) { unsigned long cur = trans->delayed_ref_updates; trans->delayed_ref_updates = 0; @@ -490,8 +488,6 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, count++; } - btrfs_trans_release_metadata(trans, root); - if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) && should_end_transaction(trans, root)) { trans->transaction->blocked = 1; @@ -572,50 +568,21 @@ int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans, int btrfs_write_marked_extents(struct btrfs_root *root, struct extent_io_tree *dirty_pages, int mark) { - int ret; int err = 0; int werr = 0; - struct page *page; - struct inode *btree_inode = root->fs_info->btree_inode; + struct address_space *mapping = root->fs_info->btree_inode->i_mapping; u64 start = 0; u64 end; - unsigned long index; - while (1) { - ret = find_first_extent_bit(dirty_pages, start, &start, &end, - mark); - if (ret) - break; - while (start <= end) { - cond_resched(); - - index = start >> PAGE_CACHE_SHIFT; - start = (u64)(index + 1) << PAGE_CACHE_SHIFT; - page = find_get_page(btree_inode->i_mapping, index); - if (!page) - continue; - - btree_lock_page_hook(page); - if (!page->mapping) { - unlock_page(page); - page_cache_release(page); - continue; - } - - if (PageWriteback(page)) { - if (PageDirty(page)) - wait_on_page_writeback(page); - else { - unlock_page(page); - page_cache_release(page); - continue; - } - } - err = write_one_page(page, 0); - if (err) - werr = err; - page_cache_release(page); - } + while (!find_first_extent_bit(dirty_pages, start, &start, &end, + mark)) { + convert_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT, mark, + GFP_NOFS); + err = filemap_fdatawrite_range(mapping, start, end); + if (err) + werr = err; + cond_resched(); + start = end + 1; } if (err) werr = err; @@ -631,39 +598,20 @@ int btrfs_write_marked_extents(struct btrfs_root *root, int btrfs_wait_marked_extents(struct btrfs_root *root, struct extent_io_tree *dirty_pages, int mark) { - int ret; int err = 0; int werr = 0; - struct page *page; - struct inode *btree_inode = root->fs_info->btree_inode; + struct address_space *mapping = root->fs_info->btree_inode->i_mapping; u64 start = 0; u64 end; - unsigned long index; - while (1) { - ret = find_first_extent_bit(dirty_pages, start, &start, &end, - mark); - if (ret) - break; - - clear_extent_bits(dirty_pages, start, end, mark, GFP_NOFS); - while (start <= end) { - index = start >> PAGE_CACHE_SHIFT; - start = (u64)(index + 1) << PAGE_CACHE_SHIFT; - page = find_get_page(btree_inode->i_mapping, index); - if (!page) - continue; - if (PageDirty(page)) { - btree_lock_page_hook(page); - wait_on_page_writeback(page); - err = write_one_page(page, 0); - if (err) - werr = err; - } - wait_on_page_writeback(page); - page_cache_release(page); - cond_resched(); - } + while (!find_first_extent_bit(dirty_pages, start, &start, &end, + EXTENT_NEED_WAIT)) { + clear_extent_bits(dirty_pages, start, end, EXTENT_NEED_WAIT, GFP_NOFS); + err = filemap_fdatawait_range(mapping, start, end); + if (err) + werr = err; + cond_resched(); + start = end + 1; } if (err) werr = err; @@ -683,7 +631,12 @@ int btrfs_write_and_wait_marked_extents(struct btrfs_root *root, ret = btrfs_write_marked_extents(root, dirty_pages, mark); ret2 = btrfs_wait_marked_extents(root, dirty_pages, mark); - return ret || ret2; + + if (ret) + return ret; + if (ret2) + return ret2; + return 0; } int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans, @@ -826,6 +779,10 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, btrfs_save_ino_cache(root, trans); + /* see comments in should_cow_block() */ + root->force_cow = 0; + smp_wmb(); + if (root->commit_root != root->node) { mutex_lock(&root->fs_commit_mutex); switch_commit_root(root); @@ -894,6 +851,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, struct btrfs_root *tree_root = fs_info->tree_root; struct btrfs_root *root = pending->root; struct btrfs_root *parent_root; + struct btrfs_block_rsv *rsv; struct inode *parent_inode; struct dentry *parent; struct dentry *dentry; @@ -905,6 +863,8 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, u64 objectid; u64 root_flags; + rsv = trans->block_rsv; + new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS); if (!new_root_item) { pending->error = -ENOMEM; @@ -918,11 +878,10 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, } btrfs_reloc_pre_snapshot(trans, pending, &to_reserve); - btrfs_orphan_pre_snapshot(trans, pending, &to_reserve); if (to_reserve > 0) { - ret = btrfs_block_rsv_add(trans, root, &pending->block_rsv, - to_reserve); + ret = btrfs_block_rsv_add_noflush(root, &pending->block_rsv, + to_reserve); if (ret) { pending->error = ret; goto fail; @@ -986,6 +945,10 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_tree_unlock(old); free_extent_buffer(old); + /* see comments in should_cow_block() */ + root->force_cow = 1; + smp_wmb(); + btrfs_set_root_node(new_root_item, tmp); /* record when the snapshot was created in key.offset */ key.offset = trans->transid; @@ -1009,9 +972,9 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, BUG_ON(IS_ERR(pending->snap)); btrfs_reloc_post_snapshot(trans, pending); - btrfs_orphan_post_snapshot(trans, pending); fail: kfree(new_root_item); + trans->block_rsv = rsv; btrfs_block_rsv_release(root, &pending->block_rsv, (u64)-1); return 0; } @@ -1038,7 +1001,7 @@ static void update_super_roots(struct btrfs_root *root) struct btrfs_root_item *root_item; struct btrfs_super_block *super; - super = &root->fs_info->super_copy; + super = root->fs_info->super_copy; root_item = &root->fs_info->chunk_root->root_item; super->chunk_root = root_item->bytenr; @@ -1049,7 +1012,7 @@ static void update_super_roots(struct btrfs_root *root) super->root = root_item->bytenr; super->generation = root_item->generation; super->root_level = root_item->level; - if (super->cache_generation != 0 || btrfs_test_opt(root, SPACE_CACHE)) + if (btrfs_test_opt(root, SPACE_CACHE)) super->cache_generation = root_item->generation; } @@ -1080,22 +1043,7 @@ int btrfs_transaction_blocked(struct btrfs_fs_info *info) static void wait_current_trans_commit_start(struct btrfs_root *root, struct btrfs_transaction *trans) { - DEFINE_WAIT(wait); - - if (trans->in_commit) - return; - - while (1) { - prepare_to_wait(&root->fs_info->transaction_blocked_wait, &wait, - TASK_UNINTERRUPTIBLE); - if (trans->in_commit) { - finish_wait(&root->fs_info->transaction_blocked_wait, - &wait); - break; - } - schedule(); - finish_wait(&root->fs_info->transaction_blocked_wait, &wait); - } + wait_event(root->fs_info->transaction_blocked_wait, trans->in_commit); } /* @@ -1105,24 +1053,8 @@ static void wait_current_trans_commit_start(struct btrfs_root *root, static void wait_current_trans_commit_start_and_unblock(struct btrfs_root *root, struct btrfs_transaction *trans) { - DEFINE_WAIT(wait); - - if (trans->commit_done || (trans->in_commit && !trans->blocked)) - return; - - while (1) { - prepare_to_wait(&root->fs_info->transaction_wait, &wait, - TASK_UNINTERRUPTIBLE); - if (trans->commit_done || - (trans->in_commit && !trans->blocked)) { - finish_wait(&root->fs_info->transaction_wait, - &wait); - break; - } - schedule(); - finish_wait(&root->fs_info->transaction_wait, - &wait); - } + wait_event(root->fs_info->transaction_wait, + trans->commit_done || (trans->in_commit && !trans->blocked)); } /* @@ -1205,14 +1137,15 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, btrfs_run_ordered_operations(root, 0); + btrfs_trans_release_metadata(trans, root); + trans->block_rsv = NULL; + /* make a pass through all the delayed refs we have so far * any runnings procs may add more while we are here */ ret = btrfs_run_delayed_refs(trans, root, 0); BUG_ON(ret); - btrfs_trans_release_metadata(trans, root); - cur_trans = trans->transaction; /* * set the flushing flag so procs in this transaction have to @@ -1229,8 +1162,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, atomic_inc(&cur_trans->use_count); btrfs_end_transaction(trans, root); - ret = wait_for_commit(root, cur_trans); - BUG_ON(ret); + wait_for_commit(root, cur_trans); put_transaction(cur_trans); @@ -1379,12 +1311,12 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, update_super_roots(root); if (!root->fs_info->log_root_recovering) { - btrfs_set_super_log_root(&root->fs_info->super_copy, 0); - btrfs_set_super_log_root_level(&root->fs_info->super_copy, 0); + btrfs_set_super_log_root(root->fs_info->super_copy, 0); + btrfs_set_super_log_root_level(root->fs_info->super_copy, 0); } - memcpy(&root->fs_info->super_for_commit, &root->fs_info->super_copy, - sizeof(root->fs_info->super_copy)); + memcpy(root->fs_info->super_for_commit, root->fs_info->super_copy, + sizeof(*root->fs_info->super_copy)); trans->transaction->blocked = 0; spin_lock(&root->fs_info->trans_lock); diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 88dec16..21faa12 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -276,8 +276,9 @@ static int process_one_buffer(struct btrfs_root *log, struct walk_control *wc, u64 gen) { if (wc->pin) - btrfs_pin_extent(log->fs_info->extent_root, - eb->start, eb->len, 0); + btrfs_pin_extent_for_log_replay(wc->trans, + log->fs_info->extent_root, + eb->start, eb->len); if (btrfs_buffer_uptodate(eb, gen)) { if (wc->write) @@ -1069,7 +1070,7 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, } btrfs_release_path(path); if (nlink != inode->i_nlink) { - inode->i_nlink = nlink; + set_nlink(inode, nlink); btrfs_update_inode(trans, root, inode); } BTRFS_I(inode)->index_cnt = (u64)-1; @@ -1679,7 +1680,8 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb, return 0; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; nritems = btrfs_header_nritems(eb); for (i = 0; i < nritems; i++) { @@ -1785,21 +1787,23 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans, return -ENOMEM; if (*level == 1) { - wc->process_func(root, next, wc, ptr_gen); + ret = wc->process_func(root, next, wc, ptr_gen); + if (ret) + return ret; path->slots[*level]++; if (wc->free) { btrfs_read_buffer(next, ptr_gen); btrfs_tree_lock(next); - clean_tree_block(trans, root, next); btrfs_set_lock_blocking(next); + clean_tree_block(trans, root, next); btrfs_wait_tree_block_writeback(next); btrfs_tree_unlock(next); WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID); - ret = btrfs_free_reserved_extent(root, + ret = btrfs_free_and_pin_reserved_extent(root, bytenr, blocksize); BUG_ON(ret); } @@ -1850,21 +1854,24 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans, parent = path->nodes[*level + 1]; root_owner = btrfs_header_owner(parent); - wc->process_func(root, path->nodes[*level], wc, + ret = wc->process_func(root, path->nodes[*level], wc, btrfs_header_generation(path->nodes[*level])); + if (ret) + return ret; + if (wc->free) { struct extent_buffer *next; next = path->nodes[*level]; btrfs_tree_lock(next); - clean_tree_block(trans, root, next); btrfs_set_lock_blocking(next); + clean_tree_block(trans, root, next); btrfs_wait_tree_block_writeback(next); btrfs_tree_unlock(next); WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID); - ret = btrfs_free_reserved_extent(root, + ret = btrfs_free_and_pin_reserved_extent(root, path->nodes[*level]->start, path->nodes[*level]->len); BUG_ON(ret); @@ -1926,14 +1933,14 @@ static int walk_log_tree(struct btrfs_trans_handle *trans, next = path->nodes[orig_level]; btrfs_tree_lock(next); - clean_tree_block(trans, log, next); btrfs_set_lock_blocking(next); + clean_tree_block(trans, log, next); btrfs_wait_tree_block_writeback(next); btrfs_tree_unlock(next); WARN_ON(log->root_key.objectid != BTRFS_TREE_LOG_OBJECTID); - ret = btrfs_free_reserved_extent(log, next->start, + ret = btrfs_free_and_pin_reserved_extent(log, next->start, next->len); BUG_ON(ret); } @@ -2049,10 +2056,10 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, /* wait for previous tree log sync to complete */ if (atomic_read(&root->log_commit[(index1 + 1) % 2])) wait_log_commit(trans, root, root->log_transid - 1); - while (1) { unsigned long batch = root->log_batch; - if (root->log_multiple_pids) { + /* when we're on an ssd, just kick the log commit out */ + if (!btrfs_test_opt(root, SSD) && root->log_multiple_pids) { mutex_unlock(&root->log_mutex); schedule_timeout_uninterruptible(1); mutex_lock(&root->log_mutex); @@ -2153,9 +2160,9 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, BUG_ON(ret); btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark); - btrfs_set_super_log_root(&root->fs_info->super_for_commit, + btrfs_set_super_log_root(root->fs_info->super_for_commit, log_root_tree->node->start); - btrfs_set_super_log_root_level(&root->fs_info->super_for_commit, + btrfs_set_super_log_root_level(root->fs_info->super_for_commit, btrfs_header_level(log_root_tree->node)); log_root_tree->log_batch = 0; diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 7745ad5..9899205 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -142,6 +142,7 @@ static noinline int run_scheduled_bios(struct btrfs_device *device) unsigned long limit; unsigned long last_waited = 0; int force_reg = 0; + int sync_pending = 0; struct blk_plug plug; /* @@ -229,6 +230,22 @@ loop_lock: BUG_ON(atomic_read(&cur->bi_cnt) == 0); + /* + * if we're doing the sync list, record that our + * plug has some sync requests on it + * + * If we're doing the regular list and there are + * sync requests sitting around, unplug before + * we add more + */ + if (pending_bios == &device->pending_sync_bios) { + sync_pending = 1; + } else if (sync_pending) { + blk_finish_plug(&plug); + blk_start_plug(&plug); + sync_pending = 0; + } + submit_bio(cur->bi_rw, cur); num_run++; batch_run++; @@ -278,6 +295,12 @@ loop_lock: btrfs_requeue_work(&device->work); goto done; } + /* unplug every 64 requests just for good measure */ + if (batch_run % 64 == 0) { + blk_finish_plug(&plug); + blk_start_plug(&plug); + sync_pending = 0; + } } cond_resched(); @@ -349,6 +372,14 @@ static noinline int device_list_add(const char *path, } INIT_LIST_HEAD(&device->dev_alloc_list); + /* init readahead state */ + spin_lock_init(&device->reada_lock); + device->reada_curr_zone = NULL; + atomic_set(&device->reada_in_flight, 0); + device->reada_next = 0; + INIT_RADIX_TREE(&device->reada_zones, GFP_NOFS & ~__GFP_WAIT); + INIT_RADIX_TREE(&device->reada_extents, GFP_NOFS & ~__GFP_WAIT); + mutex_lock(&fs_devices->device_list_mutex); list_add_rcu(&device->dev_list, &fs_devices->devices); mutex_unlock(&fs_devices->device_list_mutex); @@ -587,10 +618,8 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, set_blocksize(bdev, 4096); bh = btrfs_read_dev_super(bdev); - if (!bh) { - ret = -EINVAL; + if (!bh) goto error_close; - } disk_super = (struct btrfs_super_block *)bh->b_data; devid = btrfs_stack_device_id(&disk_super->dev_item); @@ -645,7 +674,7 @@ error: continue; } if (fs_devices->open_devices == 0) { - ret = -EIO; + ret = -EINVAL; goto out; } fs_devices->seeding = seeding; @@ -853,6 +882,7 @@ int find_free_dev_extent(struct btrfs_trans_handle *trans, max_hole_start = search_start; max_hole_size = 0; + hole_size = 0; if (search_start >= search_end) { ret = -ENOSPC; @@ -935,7 +965,14 @@ next: cond_resched(); } - hole_size = search_end- search_start; + /* + * At this point, search_start should be the end of + * allocated dev extents, and when shrinking the device, + * search_end may be smaller than search_start. + */ + if (search_end > search_start) + hole_size = search_end - search_start; + if (hole_size > max_hole_size) { max_hole_start = search_start; max_hole_size = hole_size; @@ -975,7 +1012,7 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, key.objectid = device->devid; key.offset = start; key.type = BTRFS_DEV_EXTENT_KEY; - +again: ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret > 0) { ret = btrfs_previous_item(root, path, key.objectid, @@ -988,6 +1025,9 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, struct btrfs_dev_extent); BUG_ON(found_key.offset > start || found_key.offset + btrfs_dev_extent_length(leaf, extent) < start); + key = found_key; + btrfs_release_path(path); + goto again; } else if (ret == 0) { leaf = path->nodes[0]; extent = btrfs_item_ptr(leaf, path->slots[0], @@ -995,8 +1035,13 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, } BUG_ON(ret); - if (device->bytes_used > 0) - device->bytes_used -= btrfs_dev_extent_length(leaf, extent); + if (device->bytes_used > 0) { + u64 len = btrfs_dev_extent_length(leaf, extent); + device->bytes_used -= len; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += len; + spin_unlock(&root->fs_info->free_chunk_lock); + } ret = btrfs_del_item(trans, root, path); out: @@ -1055,7 +1100,8 @@ static noinline int find_next_chunk(struct btrfs_root *root, struct btrfs_key found_key; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; key.objectid = objectid; key.offset = (u64)-1; @@ -1337,6 +1383,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (ret) goto error_undo; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space = device->total_bytes - + device->bytes_used; + spin_unlock(&root->fs_info->free_chunk_lock); + device->in_fs_metadata = 0; btrfs_scrub_cancel_dev(root, device); @@ -1368,8 +1419,8 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) call_rcu(&device->rcu, free_device); mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); - num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1; - btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices); + num_devices = btrfs_super_num_devices(root->fs_info->super_copy) - 1; + btrfs_set_super_num_devices(root->fs_info->super_copy, num_devices); if (cur_devices->open_devices == 0) { struct btrfs_fs_devices *fs_devices; @@ -1431,7 +1482,7 @@ static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans, struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; struct btrfs_fs_devices *old_devices; struct btrfs_fs_devices *seed_devices; - struct btrfs_super_block *disk_super = &root->fs_info->super_copy; + struct btrfs_super_block *disk_super = root->fs_info->super_copy; struct btrfs_device *device; u64 super_flags; @@ -1573,7 +1624,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) if ((sb->s_flags & MS_RDONLY) && !root->fs_info->fs_devices->seeding) return -EINVAL; - bdev = blkdev_get_by_path(device_path, FMODE_EXCL, + bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL, root->fs_info->bdev_holder); if (IS_ERR(bdev)) return PTR_ERR(bdev); @@ -1672,15 +1723,19 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) root->fs_info->fs_devices->num_can_discard++; root->fs_info->fs_devices->total_rw_bytes += device->total_bytes; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += device->total_bytes; + spin_unlock(&root->fs_info->free_chunk_lock); + if (!blk_queue_nonrot(bdev_get_queue(bdev))) root->fs_info->fs_devices->rotating = 1; - total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy); - btrfs_set_super_total_bytes(&root->fs_info->super_copy, + total_bytes = btrfs_super_total_bytes(root->fs_info->super_copy); + btrfs_set_super_total_bytes(root->fs_info->super_copy, total_bytes + device->total_bytes); - total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy); - btrfs_set_super_num_devices(&root->fs_info->super_copy, + total_bytes = btrfs_super_num_devices(root->fs_info->super_copy); + btrfs_set_super_num_devices(root->fs_info->super_copy, total_bytes + 1); mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); @@ -1771,7 +1826,7 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 new_size) { struct btrfs_super_block *super_copy = - &device->dev_root->fs_info->super_copy; + device->dev_root->fs_info->super_copy; u64 old_total = btrfs_super_total_bytes(super_copy); u64 diff = new_size - device->total_bytes; @@ -1830,7 +1885,7 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans, static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64 chunk_offset) { - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; struct btrfs_disk_key *disk_key; struct btrfs_chunk *chunk; u8 *ptr; @@ -2085,8 +2140,10 @@ int btrfs_balance(struct btrfs_root *dev_root) /* step two, relocate all the chunks */ path = btrfs_alloc_path(); - BUG_ON(!path); - + if (!path) { + ret = -ENOMEM; + goto error; + } key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID; key.offset = (u64)-1; key.type = BTRFS_CHUNK_ITEM_KEY; @@ -2154,7 +2211,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size) bool retried = false; struct extent_buffer *l; struct btrfs_key key; - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; u64 old_total = btrfs_super_total_bytes(super_copy); u64 old_size = device->total_bytes; u64 diff = device->total_bytes - new_size; @@ -2171,8 +2228,12 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size) lock_chunks(root); device->total_bytes = new_size; - if (device->writeable) + if (device->writeable) { device->fs_devices->total_rw_bytes -= diff; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space -= diff; + spin_unlock(&root->fs_info->free_chunk_lock); + } unlock_chunks(root); again: @@ -2236,6 +2297,9 @@ again: device->total_bytes = old_size; if (device->writeable) device->fs_devices->total_rw_bytes += diff; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += diff; + spin_unlock(&root->fs_info->free_chunk_lock); unlock_chunks(root); goto done; } @@ -2271,7 +2335,7 @@ static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans, struct btrfs_key *key, struct btrfs_chunk *chunk, int item_size) { - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; struct btrfs_disk_key disk_key; u32 array_size; u8 *ptr; @@ -2434,9 +2498,10 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, total_avail = device->total_bytes - device->bytes_used; else total_avail = 0; - /* avail is off by max(alloc_start, 1MB), but that is the same - * for all devices, so it doesn't hurt the sorting later on - */ + + /* If there is no space on this device, skip it. */ + if (total_avail == 0) + continue; ret = find_free_dev_extent(trans, device, max_stripe_size * dev_stripes, @@ -2593,6 +2658,11 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans, index++; } + spin_lock(&extent_root->fs_info->free_chunk_lock); + extent_root->fs_info->free_chunk_space -= (stripe_size * + map->num_stripes); + spin_unlock(&extent_root->fs_info->free_chunk_lock); + index = 0; stripe = &chunk->stripe; while (index < map->num_stripes) { @@ -2685,7 +2755,8 @@ static noinline int init_first_rw_device(struct btrfs_trans_handle *trans, ret = find_next_chunk(fs_info->chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID, &chunk_offset); - BUG_ON(ret); + if (ret) + return ret; alloc_profile = BTRFS_BLOCK_GROUP_METADATA | (fs_info->metadata_alloc_profile & @@ -2825,7 +2896,7 @@ static int find_live_mirror(struct map_lookup *map, int first, int num, static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 logical, u64 *length, - struct btrfs_multi_bio **multi_ret, + struct btrfs_bio **bbio_ret, int mirror_num) { struct extent_map *em; @@ -2843,18 +2914,18 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, int i; int num_stripes; int max_errors = 0; - struct btrfs_multi_bio *multi = NULL; + struct btrfs_bio *bbio = NULL; - if (multi_ret && !(rw & (REQ_WRITE | REQ_DISCARD))) + if (bbio_ret && !(rw & (REQ_WRITE | REQ_DISCARD))) stripes_allocated = 1; again: - if (multi_ret) { - multi = kzalloc(btrfs_multi_bio_size(stripes_allocated), + if (bbio_ret) { + bbio = kzalloc(btrfs_bio_size(stripes_allocated), GFP_NOFS); - if (!multi) + if (!bbio) return -ENOMEM; - atomic_set(&multi->error, 0); + atomic_set(&bbio->error, 0); } read_lock(&em_tree->lock); @@ -2875,7 +2946,7 @@ again: if (mirror_num > map->num_stripes) mirror_num = 0; - /* if our multi bio struct is too small, back off and try again */ + /* if our btrfs_bio struct is too small, back off and try again */ if (rw & REQ_WRITE) { if (map->type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP)) { @@ -2894,11 +2965,11 @@ again: stripes_required = map->num_stripes; } } - if (multi_ret && (rw & (REQ_WRITE | REQ_DISCARD)) && + if (bbio_ret && (rw & (REQ_WRITE | REQ_DISCARD)) && stripes_allocated < stripes_required) { stripes_allocated = map->num_stripes; free_extent_map(em); - kfree(multi); + kfree(bbio); goto again; } stripe_nr = offset; @@ -2927,7 +2998,7 @@ again: *length = em->len - offset; } - if (!multi_ret) + if (!bbio_ret) goto out; num_stripes = 1; @@ -2952,13 +3023,17 @@ again: stripe_index = find_live_mirror(map, 0, map->num_stripes, current->pid % map->num_stripes); + mirror_num = stripe_index + 1; } } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { - if (rw & (REQ_WRITE | REQ_DISCARD)) + if (rw & (REQ_WRITE | REQ_DISCARD)) { num_stripes = map->num_stripes; - else if (mirror_num) + } else if (mirror_num) { stripe_index = mirror_num - 1; + } else { + mirror_num = 1; + } } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { int factor = map->num_stripes / map->sub_stripes; @@ -2978,6 +3053,7 @@ again: stripe_index = find_live_mirror(map, stripe_index, map->sub_stripes, stripe_index + current->pid % map->sub_stripes); + mirror_num = stripe_index + 1; } } else { /* @@ -2986,15 +3062,16 @@ again: * stripe_index is the number of our device in the stripe array */ stripe_index = do_div(stripe_nr, map->num_stripes); + mirror_num = stripe_index + 1; } BUG_ON(stripe_index >= map->num_stripes); if (rw & REQ_DISCARD) { for (i = 0; i < num_stripes; i++) { - multi->stripes[i].physical = + bbio->stripes[i].physical = map->stripes[stripe_index].physical + stripe_offset + stripe_nr * map->stripe_len; - multi->stripes[i].dev = map->stripes[stripe_index].dev; + bbio->stripes[i].dev = map->stripes[stripe_index].dev; if (map->type & BTRFS_BLOCK_GROUP_RAID0) { u64 stripes; @@ -3015,16 +3092,16 @@ again: } stripes = stripe_nr_end - 1 - j; do_div(stripes, map->num_stripes); - multi->stripes[i].length = map->stripe_len * + bbio->stripes[i].length = map->stripe_len * (stripes - stripe_nr + 1); if (i == 0) { - multi->stripes[i].length -= + bbio->stripes[i].length -= stripe_offset; stripe_offset = 0; } if (stripe_index == last_stripe) - multi->stripes[i].length -= + bbio->stripes[i].length -= stripe_end_offset; } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { u64 stripes; @@ -3049,11 +3126,11 @@ again: } stripes = stripe_nr_end - 1 - j; do_div(stripes, factor); - multi->stripes[i].length = map->stripe_len * + bbio->stripes[i].length = map->stripe_len * (stripes - stripe_nr + 1); if (i < map->sub_stripes) { - multi->stripes[i].length -= + bbio->stripes[i].length -= stripe_offset; if (i == map->sub_stripes - 1) stripe_offset = 0; @@ -3061,11 +3138,11 @@ again: if (stripe_index >= last_stripe && stripe_index <= (last_stripe + map->sub_stripes - 1)) { - multi->stripes[i].length -= + bbio->stripes[i].length -= stripe_end_offset; } } else - multi->stripes[i].length = *length; + bbio->stripes[i].length = *length; stripe_index++; if (stripe_index == map->num_stripes) { @@ -3076,19 +3153,20 @@ again: } } else { for (i = 0; i < num_stripes; i++) { - multi->stripes[i].physical = + bbio->stripes[i].physical = map->stripes[stripe_index].physical + stripe_offset + stripe_nr * map->stripe_len; - multi->stripes[i].dev = + bbio->stripes[i].dev = map->stripes[stripe_index].dev; stripe_index++; } } - if (multi_ret) { - *multi_ret = multi; - multi->num_stripes = num_stripes; - multi->max_errors = max_errors; + if (bbio_ret) { + *bbio_ret = bbio; + bbio->num_stripes = num_stripes; + bbio->max_errors = max_errors; + bbio->mirror_num = mirror_num; } out: free_extent_map(em); @@ -3097,9 +3175,9 @@ out: int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 logical, u64 *length, - struct btrfs_multi_bio **multi_ret, int mirror_num) + struct btrfs_bio **bbio_ret, int mirror_num) { - return __btrfs_map_block(map_tree, rw, logical, length, multi_ret, + return __btrfs_map_block(map_tree, rw, logical, length, bbio_ret, mirror_num); } @@ -3168,30 +3246,32 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, return 0; } -static void end_bio_multi_stripe(struct bio *bio, int err) +static void btrfs_end_bio(struct bio *bio, int err) { - struct btrfs_multi_bio *multi = bio->bi_private; + struct btrfs_bio *bbio = bio->bi_private; int is_orig_bio = 0; if (err) - atomic_inc(&multi->error); + atomic_inc(&bbio->error); - if (bio == multi->orig_bio) + if (bio == bbio->orig_bio) is_orig_bio = 1; - if (atomic_dec_and_test(&multi->stripes_pending)) { + if (atomic_dec_and_test(&bbio->stripes_pending)) { if (!is_orig_bio) { bio_put(bio); - bio = multi->orig_bio; + bio = bbio->orig_bio; } - bio->bi_private = multi->private; - bio->bi_end_io = multi->end_io; + bio->bi_private = bbio->private; + bio->bi_end_io = bbio->end_io; + bio->bi_bdev = (struct block_device *) + (unsigned long)bbio->mirror_num; /* only send an error to the higher layers if it is * beyond the tolerance of the multi-bio */ - if (atomic_read(&multi->error) > multi->max_errors) { + if (atomic_read(&bbio->error) > bbio->max_errors) { err = -EIO; - } else if (err) { + } else { /* * this bio is actually up to date, we didn't * go over the max number of errors @@ -3199,7 +3279,7 @@ static void end_bio_multi_stripe(struct bio *bio, int err) set_bit(BIO_UPTODATE, &bio->bi_flags); err = 0; } - kfree(multi); + kfree(bbio); bio_endio(bio, err); } else if (!is_orig_bio) { @@ -3279,20 +3359,20 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, u64 logical = (u64)bio->bi_sector << 9; u64 length = 0; u64 map_length; - struct btrfs_multi_bio *multi = NULL; int ret; int dev_nr = 0; int total_devs = 1; + struct btrfs_bio *bbio = NULL; length = bio->bi_size; map_tree = &root->fs_info->mapping_tree; map_length = length; - ret = btrfs_map_block(map_tree, rw, logical, &map_length, &multi, + ret = btrfs_map_block(map_tree, rw, logical, &map_length, &bbio, mirror_num); BUG_ON(ret); - total_devs = multi->num_stripes; + total_devs = bbio->num_stripes; if (map_length < length) { printk(KERN_CRIT "mapping failed logical %llu bio len %llu " "len %llu\n", (unsigned long long)logical, @@ -3300,25 +3380,28 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, (unsigned long long)map_length); BUG(); } - multi->end_io = first_bio->bi_end_io; - multi->private = first_bio->bi_private; - multi->orig_bio = first_bio; - atomic_set(&multi->stripes_pending, multi->num_stripes); + + bbio->orig_bio = first_bio; + bbio->private = first_bio->bi_private; + bbio->end_io = first_bio->bi_end_io; + atomic_set(&bbio->stripes_pending, bbio->num_stripes); while (dev_nr < total_devs) { - if (total_devs > 1) { - if (dev_nr < total_devs - 1) { - bio = bio_clone(first_bio, GFP_NOFS); - BUG_ON(!bio); - } else { - bio = first_bio; - } - bio->bi_private = multi; - bio->bi_end_io = end_bio_multi_stripe; + if (dev_nr < total_devs - 1) { + bio = bio_clone(first_bio, GFP_NOFS); + BUG_ON(!bio); + } else { + bio = first_bio; } - bio->bi_sector = multi->stripes[dev_nr].physical >> 9; - dev = multi->stripes[dev_nr].dev; + bio->bi_private = bbio; + bio->bi_end_io = btrfs_end_bio; + bio->bi_sector = bbio->stripes[dev_nr].physical >> 9; + dev = bbio->stripes[dev_nr].dev; if (dev && dev->bdev && (rw != WRITE || dev->writeable)) { + pr_debug("btrfs_map_bio: rw %d, secor=%llu, dev=%lu " + "(%s id %llu), size=%u\n", rw, + (u64)bio->bi_sector, (u_long)dev->bdev->bd_dev, + dev->name, dev->devid, bio->bi_size); bio->bi_bdev = dev->bdev; if (async_submit) schedule_bio(root, dev, rw, bio); @@ -3331,8 +3414,6 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, } dev_nr++; } - if (total_devs == 1) - kfree(multi); return 0; } @@ -3593,15 +3674,20 @@ static int read_one_dev(struct btrfs_root *root, fill_device_from_item(leaf, dev_item, device); device->dev_root = root->fs_info->dev_root; device->in_fs_metadata = 1; - if (device->writeable) + if (device->writeable) { device->fs_devices->total_rw_bytes += device->total_bytes; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += device->total_bytes - + device->bytes_used; + spin_unlock(&root->fs_info->free_chunk_lock); + } ret = 0; return ret; } int btrfs_read_sys_array(struct btrfs_root *root) { - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; struct extent_buffer *sb; struct btrfs_disk_key *disk_key; struct btrfs_chunk *chunk; @@ -3619,7 +3705,7 @@ int btrfs_read_sys_array(struct btrfs_root *root) if (!sb) return -ENOMEM; btrfs_set_buffer_uptodate(sb); - btrfs_set_buffer_lockdep_class(sb, 0); + btrfs_set_buffer_lockdep_class(root->root_key.objectid, sb, 0); write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE); array_size = btrfs_super_sys_array_size(super_copy); diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 6d866db..78f2d4d 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -92,6 +92,20 @@ struct btrfs_device { struct btrfs_work work; struct rcu_head rcu; struct work_struct rcu_work; + + /* readahead state */ + spinlock_t reada_lock; + atomic_t reada_in_flight; + u64 reada_next; + struct reada_zone *reada_curr_zone; + struct radix_tree_root reada_zones; + struct radix_tree_root reada_extents; + + /* for sending down flush barriers */ + struct bio *flush_bio; + struct completion flush_wait; + int nobarriers; + }; struct btrfs_fs_devices { @@ -136,7 +150,10 @@ struct btrfs_bio_stripe { u64 length; /* only used for discard mappings */ }; -struct btrfs_multi_bio { +struct btrfs_bio; +typedef void (btrfs_bio_end_io_t) (struct btrfs_bio *bio, int err); + +struct btrfs_bio { atomic_t stripes_pending; bio_end_io_t *end_io; struct bio *orig_bio; @@ -144,6 +161,7 @@ struct btrfs_multi_bio { atomic_t error; int max_errors; int num_stripes; + int mirror_num; struct btrfs_bio_stripe stripes[]; }; @@ -171,7 +189,7 @@ struct map_lookup { int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start, u64 end, u64 *length); -#define btrfs_multi_bio_size(n) (sizeof(struct btrfs_multi_bio) + \ +#define btrfs_bio_size(n) (sizeof(struct btrfs_bio) + \ (sizeof(struct btrfs_bio_stripe) * (n))) int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans, @@ -180,7 +198,7 @@ int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans, u64 chunk_offset, u64 start, u64 num_bytes); int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 logical, u64 *length, - struct btrfs_multi_bio **multi_ret, int mirror_num); + struct btrfs_bio **bbio_ret, int mirror_num); int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, u64 chunk_start, u64 physical, u64 devid, u64 **logical, int *naddrs, int *stripe_len); diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index 5366fe4..60ff45e 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -102,48 +102,82 @@ static int do_setxattr(struct btrfs_trans_handle *trans, if (!path) return -ENOMEM; - /* first lets see if we already have this xattr */ - di = btrfs_lookup_xattr(trans, root, path, btrfs_ino(inode), name, - strlen(name), -1); - if (IS_ERR(di)) { - ret = PTR_ERR(di); - goto out; - } - - /* ok we already have this xattr, lets remove it */ - if (di) { - /* if we want create only exit */ - if (flags & XATTR_CREATE) { - ret = -EEXIST; + if (flags & XATTR_REPLACE) { + di = btrfs_lookup_xattr(trans, root, path, btrfs_ino(inode), name, + name_len, -1); + if (IS_ERR(di)) { + ret = PTR_ERR(di); + goto out; + } else if (!di) { + ret = -ENODATA; goto out; } - ret = btrfs_delete_one_dir_name(trans, root, path, di); - BUG_ON(ret); + if (ret) + goto out; btrfs_release_path(path); - /* if we don't have a value then we are removing the xattr */ + /* + * remove the attribute + */ if (!value) goto out; - } else { + } + +again: + ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode), + name, name_len, value, size); + /* + * If we're setting an xattr to a new value but the new value is say + * exactly BTRFS_MAX_XATTR_SIZE, we could end up with EOVERFLOW getting + * back from split_leaf. This is because it thinks we'll be extending + * the existing item size, but we're asking for enough space to add the + * item itself. So if we get EOVERFLOW just set ret to EEXIST and let + * the rest of the function figure it out. + */ + if (ret == -EOVERFLOW) + ret = -EEXIST; + + if (ret == -EEXIST) { + if (flags & XATTR_CREATE) + goto out; + /* + * We can't use the path we already have since we won't have the + * proper locking for a delete, so release the path and + * re-lookup to delete the thing. + */ btrfs_release_path(path); + di = btrfs_lookup_xattr(trans, root, path, btrfs_ino(inode), + name, name_len, -1); + if (IS_ERR(di)) { + ret = PTR_ERR(di); + goto out; + } else if (!di) { + /* Shouldn't happen but just in case... */ + btrfs_release_path(path); + goto again; + } - if (flags & XATTR_REPLACE) { - /* we couldn't find the attr to replace */ - ret = -ENODATA; + ret = btrfs_delete_one_dir_name(trans, root, path, di); + if (ret) goto out; + + /* + * We have a value to set, so go back and try to insert it now. + */ + if (value) { + btrfs_release_path(path); + goto again; } } - - /* ok we have to create a completely new xattr */ - ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode), - name, name_len, value, size); - BUG_ON(ret); out: btrfs_free_path(path); return ret; } +/* + * @value: "" makes the attribute to empty, NULL removes it + */ int __btrfs_setxattr(struct btrfs_trans_handle *trans, struct inode *inode, const char *name, const void *value, size_t size, int flags) @@ -276,21 +310,40 @@ const struct xattr_handler *btrfs_xattr_handlers[] = { /* * Check if the attribute is in a supported namespace. * - * This applied after the check for the synthetic attributes in the system + * This is applied after the check for the synthetic attributes in the system * namespace. */ -static bool btrfs_is_valid_xattr(const char *name) +static int btrfs_is_valid_xattr(const char *name) { - return !strncmp(name, XATTR_SECURITY_PREFIX, - XATTR_SECURITY_PREFIX_LEN) || - !strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN) || - !strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) || - !strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN); + int len = strlen(name); + int prefixlen = 0; + + if (!strncmp(name, XATTR_SECURITY_PREFIX, + XATTR_SECURITY_PREFIX_LEN)) + prefixlen = XATTR_SECURITY_PREFIX_LEN; + else if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) + prefixlen = XATTR_SYSTEM_PREFIX_LEN; + else if (!strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN)) + prefixlen = XATTR_TRUSTED_PREFIX_LEN; + else if (!strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN)) + prefixlen = XATTR_USER_PREFIX_LEN; + else + return -EOPNOTSUPP; + + /* + * The name cannot consist of just prefix + */ + if (len <= prefixlen) + return -EINVAL; + + return 0; } ssize_t btrfs_getxattr(struct dentry *dentry, const char *name, void *buffer, size_t size) { + int ret; + /* * If this is a request for a synthetic attribute in the system.* * namespace use the generic infrastructure to resolve a handler @@ -299,8 +352,9 @@ ssize_t btrfs_getxattr(struct dentry *dentry, const char *name, if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) return generic_getxattr(dentry, name, buffer, size); - if (!btrfs_is_valid_xattr(name)) - return -EOPNOTSUPP; + ret = btrfs_is_valid_xattr(name); + if (ret) + return ret; return __btrfs_getxattr(dentry->d_inode, name, buffer, size); } @@ -308,6 +362,7 @@ int btrfs_setxattr(struct dentry *dentry, const char *name, const void *value, size_t size, int flags) { struct btrfs_root *root = BTRFS_I(dentry->d_inode)->root; + int ret; /* * The permission on security.* and system.* is not checked @@ -324,8 +379,9 @@ int btrfs_setxattr(struct dentry *dentry, const char *name, const void *value, if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) return generic_setxattr(dentry, name, value, size, flags); - if (!btrfs_is_valid_xattr(name)) - return -EOPNOTSUPP; + ret = btrfs_is_valid_xattr(name); + if (ret) + return ret; if (size == 0) value = ""; /* empty EA, do not remove */ @@ -337,6 +393,7 @@ int btrfs_setxattr(struct dentry *dentry, const char *name, const void *value, int btrfs_removexattr(struct dentry *dentry, const char *name) { struct btrfs_root *root = BTRFS_I(dentry->d_inode)->root; + int ret; /* * The permission on security.* and system.* is not checked @@ -353,43 +410,44 @@ int btrfs_removexattr(struct dentry *dentry, const char *name) if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) return generic_removexattr(dentry, name); - if (!btrfs_is_valid_xattr(name)) - return -EOPNOTSUPP; + ret = btrfs_is_valid_xattr(name); + if (ret) + return ret; return __btrfs_setxattr(NULL, dentry->d_inode, name, NULL, 0, XATTR_REPLACE); } -int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, - struct inode *inode, struct inode *dir, - const struct qstr *qstr) +int btrfs_initxattrs(struct inode *inode, const struct xattr *xattr_array, + void *fs_info) { - int err; - size_t len; - void *value; - char *suffix; + const struct xattr *xattr; + struct btrfs_trans_handle *trans = fs_info; char *name; + int err = 0; - err = security_inode_init_security(inode, dir, qstr, &suffix, &value, - &len); - if (err) { - if (err == -EOPNOTSUPP) - return 0; - return err; - } - - name = kmalloc(XATTR_SECURITY_PREFIX_LEN + strlen(suffix) + 1, - GFP_NOFS); - if (!name) { - err = -ENOMEM; - } else { + for (xattr = xattr_array; xattr->name != NULL; xattr++) { + name = kmalloc(XATTR_SECURITY_PREFIX_LEN + + strlen(xattr->name) + 1, GFP_NOFS); + if (!name) { + err = -ENOMEM; + break; + } strcpy(name, XATTR_SECURITY_PREFIX); - strcpy(name + XATTR_SECURITY_PREFIX_LEN, suffix); - err = __btrfs_setxattr(trans, inode, name, value, len, 0); + strcpy(name + XATTR_SECURITY_PREFIX_LEN, xattr->name); + err = __btrfs_setxattr(trans, inode, name, + xattr->value, xattr->value_len, 0); kfree(name); + if (err < 0) + break; } - - kfree(suffix); - kfree(value); return err; } + +int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, + struct inode *inode, struct inode *dir, + const struct qstr *qstr) +{ + return security_inode_init_security(inode, dir, qstr, + &btrfs_initxattrs, trans); +} diff --git a/fs/hfs/bnode.c b/fs/hfs/bnode.c index cdb41a1..8daea16 100644 --- a/fs/hfs/bnode.c +++ b/fs/hfs/bnode.c @@ -287,7 +287,6 @@ static struct hfs_bnode *__hfs_bnode_create(struct hfs_btree *tree, u32 cnid) page_cache_release(page); goto fail; } - page_cache_release(page); node->page[i] = page; } @@ -397,11 +396,11 @@ node_error: void hfs_bnode_free(struct hfs_bnode *node) { - //int i; + int i; - //for (i = 0; i < node->tree->pages_per_bnode; i++) - // if (node->page[i]) - // page_cache_release(node->page[i]); + for (i = 0; i < node->tree->pages_per_bnode; i++) + if (node->page[i]) + page_cache_release(node->page[i]); kfree(node); } diff --git a/fs/hfs/brec.c b/fs/hfs/brec.c index 92fb358..db240c5 100644 --- a/fs/hfs/brec.c +++ b/fs/hfs/brec.c @@ -132,13 +132,16 @@ skip: hfs_bnode_write(node, entry, data_off + key_len, entry_len); hfs_bnode_dump(node); - if (new_node) { - /* update parent key if we inserted a key - * at the start of the first node - */ - if (!rec && new_node != node) - hfs_brec_update_parent(fd); + /* + * update parent key if we inserted a key + * at the start of the node and it is not the new node + */ + if (!rec && new_node != node) { + hfs_bnode_read_key(node, fd->search_key, data_off + size); + hfs_brec_update_parent(fd); + } + if (new_node) { hfs_bnode_put(fd->bnode); if (!new_node->parent) { hfs_btree_inc_height(tree); @@ -167,9 +170,6 @@ skip: goto again; } - if (!rec) - hfs_brec_update_parent(fd); - return 0; } @@ -366,6 +366,8 @@ again: if (IS_ERR(parent)) return PTR_ERR(parent); __hfs_brec_find(parent, fd); + if (fd->record < 0) + return -ENOENT; hfs_bnode_dump(parent); rec = fd->record; diff --git a/fs/hfs/dir.c b/fs/hfs/dir.c index b4d70b1..bce4eef 100644 --- a/fs/hfs/dir.c +++ b/fs/hfs/dir.c @@ -198,7 +198,7 @@ static int hfs_create(struct inode *dir, struct dentry *dentry, int mode, res = hfs_cat_create(inode->i_ino, dir, &dentry->d_name, inode); if (res) { - inode->i_nlink = 0; + clear_nlink(inode); hfs_delete_inode(inode); iput(inode); return res; @@ -227,7 +227,7 @@ static int hfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) res = hfs_cat_create(inode->i_ino, dir, &dentry->d_name, inode); if (res) { - inode->i_nlink = 0; + clear_nlink(inode); hfs_delete_inode(inode); iput(inode); return res; diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index fff16c9..a1a9fdc 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -123,8 +123,8 @@ static ssize_t hfs_direct_IO(int rw, struct kiocb *iocb, struct inode *inode = file->f_path.dentry->d_inode->i_mapping->host; ssize_t ret; - ret = blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov, - offset, nr_segs, hfs_get_block, NULL); + ret = blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs, + hfs_get_block); /* * In case of error extending write may have instantiated a few @@ -183,7 +183,7 @@ struct inode *hfs_new_inode(struct inode *dir, struct qstr *name, int mode) inode->i_mode = mode; inode->i_uid = current_fsuid(); inode->i_gid = current_fsgid(); - inode->i_nlink = 1; + set_nlink(inode, 1); inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; HFS_I(inode)->flags = 0; HFS_I(inode)->rsrc_inode = NULL; @@ -313,7 +313,7 @@ static int hfs_read_inode(struct inode *inode, void *data) /* Initialize the inode */ inode->i_uid = hsb->s_uid; inode->i_gid = hsb->s_gid; - inode->i_nlink = 1; + set_nlink(inode, 1); if (idata->key) HFS_I(inode)->cat_key = *idata->key; @@ -615,6 +615,8 @@ int hfs_inode_setattr(struct dentry *dentry, struct iattr * attr) if ((attr->ia_valid & ATTR_SIZE) && attr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + error = vmtruncate(inode, attr->ia_size); if (error) return error; @@ -625,12 +627,18 @@ int hfs_inode_setattr(struct dentry *dentry, struct iattr * attr) return 0; } -static int hfs_file_fsync(struct file *filp, int datasync) +static int hfs_file_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *inode = filp->f_mapping->host; struct super_block * sb; int ret, err; + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + /* sync the inode to buffers */ ret = write_inode_now(inode, 0); @@ -647,6 +655,7 @@ static int hfs_file_fsync(struct file *filp, int datasync) err = sync_blockdev(sb->s_bdev); if (!ret) ret = err; + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/hfs/super.c b/fs/hfs/super.c index 1b55f70..cac813d 100644 --- a/fs/hfs/super.c +++ b/fs/hfs/super.c @@ -138,9 +138,9 @@ static int hfs_show_options(struct seq_file *seq, struct vfsmount *mnt) struct hfs_sb_info *sbi = HFS_SB(mnt->mnt_sb); if (sbi->s_creator != cpu_to_be32(0x3f3f3f3f)) - seq_printf(seq, ",creator=%.4s", (char *)&sbi->s_creator); + seq_show_option_n(seq, "creator", (char *)&sbi->s_creator, 4); if (sbi->s_type != cpu_to_be32(0x3f3f3f3f)) - seq_printf(seq, ",type=%.4s", (char *)&sbi->s_type); + seq_show_option_n(seq, "type", (char *)&sbi->s_type, 4); seq_printf(seq, ",uid=%u,gid=%u", sbi->s_uid, sbi->s_gid); if (sbi->s_file_umask != 0133) seq_printf(seq, ",file_umask=%o", sbi->s_file_umask); diff --git a/fs/hfsplus/bnode.c b/fs/hfsplus/bnode.c index 1c42cc5..a1e9109 100644 --- a/fs/hfsplus/bnode.c +++ b/fs/hfsplus/bnode.c @@ -454,7 +454,6 @@ static struct hfs_bnode *__hfs_bnode_create(struct hfs_btree *tree, u32 cnid) page_cache_release(page); goto fail; } - page_cache_release(page); node->page[i] = page; } @@ -566,13 +565,11 @@ node_error: void hfs_bnode_free(struct hfs_bnode *node) { -#if 0 int i; for (i = 0; i < node->tree->pages_per_bnode; i++) if (node->page[i]) page_cache_release(node->page[i]); -#endif kfree(node); } diff --git a/fs/hfsplus/brec.c b/fs/hfsplus/brec.c index 2312de3..7429c40 100644 --- a/fs/hfsplus/brec.c +++ b/fs/hfsplus/brec.c @@ -43,6 +43,10 @@ u16 hfs_brec_keylen(struct hfs_bnode *node, u16 rec) node->tree->node_size - (rec + 1) * 2); if (!recoff) return 0; + if (recoff > node->tree->node_size - 2) { + printk(KERN_ERR "hfs: recoff %d too large\n", recoff); + return 0; + } retval = hfs_bnode_read_u16(node, recoff) + 2; if (retval > node->tree->max_key_len + 2) { @@ -126,13 +130,16 @@ skip: hfs_bnode_write(node, entry, data_off + key_len, entry_len); hfs_bnode_dump(node); - if (new_node) { - /* update parent key if we inserted a key - * at the start of the first node - */ - if (!rec && new_node != node) - hfs_brec_update_parent(fd); + /* + * update parent key if we inserted a key + * at the start of the node and it is not the new node + */ + if (!rec && new_node != node) { + hfs_bnode_read_key(node, fd->search_key, data_off + size); + hfs_brec_update_parent(fd); + } + if (new_node) { hfs_bnode_put(fd->bnode); if (!new_node->parent) { hfs_btree_inc_height(tree); @@ -162,9 +169,6 @@ skip: goto again; } - if (!rec) - hfs_brec_update_parent(fd); - return 0; } @@ -364,6 +368,8 @@ again: if (IS_ERR(parent)) return PTR_ERR(parent); __hfs_brec_find(parent, fd); + if (fd->record < 0) + return -ENOENT; hfs_bnode_dump(parent); rec = fd->record; diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index 408073a..ec2a9c2 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -212,7 +212,9 @@ int hfsplus_create_cat(u32 cnid, struct inode *dir, dprint(DBG_CAT_MOD, "create_cat: %s,%u(%d)\n", str->name, cnid, inode->i_nlink); - hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + err = hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + if (err) + return err; hfsplus_cat_build_key(sb, fd.search_key, cnid, NULL); entry_size = hfsplus_fill_cat_thread(sb, &entry, @@ -269,7 +271,9 @@ int hfsplus_delete_cat(u32 cnid, struct inode *dir, struct qstr *str) dprint(DBG_CAT_MOD, "delete_cat: %s,%u\n", str ? str->name : NULL, cnid); - hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + err = hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + if (err) + return err; if (!str) { int len; @@ -347,12 +351,14 @@ int hfsplus_rename_cat(u32 cnid, struct hfs_find_data src_fd, dst_fd; hfsplus_cat_entry entry; int entry_size, type; - int err = 0; + int err; dprint(DBG_CAT_MOD, "rename_cat: %u - %lu,%s - %lu,%s\n", cnid, src_dir->i_ino, src_name->name, dst_dir->i_ino, dst_name->name); - hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &src_fd); + err = hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &src_fd); + if (err) + return err; dst_fd = src_fd; /* find the old dir entry and read the data */ diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 159f5eb..5adb740 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -38,7 +38,9 @@ static struct dentry *hfsplus_lookup(struct inode *dir, struct dentry *dentry, sb = dir->i_sb; dentry->d_fsdata = NULL; - hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + err = hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + if (err) + return ERR_PTR(err); hfsplus_cat_build_key(sb, fd.search_key, dir->i_ino, &dentry->d_name); again: err = hfs_brec_read(&fd, &entry, sizeof(entry)); @@ -132,7 +134,9 @@ static int hfsplus_readdir(struct file *filp, void *dirent, filldir_t filldir) if (filp->f_pos >= inode->i_size) return 0; - hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + err = hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + if (err) + return err; hfsplus_cat_build_key(sb, fd.search_key, inode->i_ino, NULL); err = hfs_brec_find(&fd); if (err) @@ -422,7 +426,7 @@ static int hfsplus_symlink(struct inode *dir, struct dentry *dentry, goto out; out_err: - inode->i_nlink = 0; + clear_nlink(inode); hfsplus_delete_inode(inode); iput(inode); out: @@ -447,7 +451,7 @@ static int hfsplus_mknod(struct inode *dir, struct dentry *dentry, res = hfsplus_create_cat(inode->i_ino, dir, &dentry->d_name, inode); if (res) { - inode->i_nlink = 0; + clear_nlink(inode); hfsplus_delete_inode(inode); iput(inode); goto out; diff --git a/fs/hfsplus/extents.c b/fs/hfsplus/extents.c index 9d8c087..32b12e5 100644 --- a/fs/hfsplus/extents.c +++ b/fs/hfsplus/extents.c @@ -119,22 +119,31 @@ static void __hfsplus_ext_write_extent(struct inode *inode, set_bit(HFSPLUS_I_EXT_DIRTY, &hip->flags); } -static void hfsplus_ext_write_extent_locked(struct inode *inode) +static int hfsplus_ext_write_extent_locked(struct inode *inode) { + int res; + if (HFSPLUS_I(inode)->extent_state & HFSPLUS_EXT_DIRTY) { struct hfs_find_data fd; - hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); + res = hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); + if (res) + return res; __hfsplus_ext_write_extent(inode, &fd); hfs_find_exit(&fd); } + return 0; } -void hfsplus_ext_write_extent(struct inode *inode) +int hfsplus_ext_write_extent(struct inode *inode) { + int res; + mutex_lock(&HFSPLUS_I(inode)->extents_lock); - hfsplus_ext_write_extent_locked(inode); + res = hfsplus_ext_write_extent_locked(inode); mutex_unlock(&HFSPLUS_I(inode)->extents_lock); + + return res; } static inline int __hfsplus_ext_read_extent(struct hfs_find_data *fd, @@ -194,9 +203,11 @@ static int hfsplus_ext_read_extent(struct inode *inode, u32 block) block < hip->cached_start + hip->cached_blocks) return 0; - hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); - res = __hfsplus_ext_cache_extent(&fd, inode, block); - hfs_find_exit(&fd); + res = hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); + if (!res) { + res = __hfsplus_ext_cache_extent(&fd, inode, block); + hfs_find_exit(&fd); + } return res; } @@ -209,6 +220,7 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock, struct hfsplus_inode_info *hip = HFSPLUS_I(inode); int res = -EIO; u32 ablock, dblock, mask; + sector_t sector; int was_dirty = 0; int shift; @@ -255,10 +267,12 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock, done: dprint(DBG_EXTENT, "get_block(%lu): %llu - %u\n", inode->i_ino, (long long)iblock, dblock); + mask = (1 << sbi->fs_shift) - 1; - map_bh(bh_result, sb, - (dblock << sbi->fs_shift) + sbi->blockoffset + - (iblock & mask)); + sector = ((sector_t)dblock << sbi->fs_shift) + + sbi->blockoffset + (iblock & mask); + map_bh(bh_result, sb, sector); + if (create) { set_buffer_new(bh_result); hip->phys_size += sb->s_blocksize; @@ -371,7 +385,9 @@ int hfsplus_free_fork(struct super_block *sb, u32 cnid, if (total_blocks == blocks) return 0; - hfs_find_init(HFSPLUS_SB(sb)->ext_tree, &fd); + res = hfs_find_init(HFSPLUS_SB(sb)->ext_tree, &fd); + if (res) + return res; do { res = __hfsplus_ext_read_extent(&fd, ext_entry, cnid, total_blocks, type); @@ -469,7 +485,9 @@ out: insert_extent: dprint(DBG_EXTENT, "insert new extent\n"); - hfsplus_ext_write_extent_locked(inode); + res = hfsplus_ext_write_extent_locked(inode); + if (res) + goto out; memset(hip->cached_extents, 0, sizeof(hfsplus_extent_rec)); hip->cached_extents[0].start_block = cpu_to_be32(start); @@ -500,7 +518,6 @@ void hfsplus_file_truncate(struct inode *inode) struct page *page; void *fsdata; loff_t size = inode->i_size; - int res; res = pagecache_write_begin(NULL, mapping, size, 0, AOP_FLAG_UNINTERRUPTIBLE, @@ -523,7 +540,12 @@ void hfsplus_file_truncate(struct inode *inode) goto out; mutex_lock(&hip->extents_lock); - hfs_find_init(HFSPLUS_SB(sb)->ext_tree, &fd); + res = hfs_find_init(HFSPLUS_SB(sb)->ext_tree, &fd); + if (res) { + mutex_unlock(&hip->extents_lock); + /* XXX: We lack error handling of hfsplus_file_truncate() */ + return; + } while (1) { if (alloc_cnt == hip->first_blocks) { hfsplus_free_extents(sb, hip->first_extents, diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 4e7f64b..d7674d0 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -386,7 +386,7 @@ extern const struct file_operations hfsplus_dir_operations; /* extents.c */ int hfsplus_ext_cmp_key(const hfsplus_btree_key *, const hfsplus_btree_key *); -void hfsplus_ext_write_extent(struct inode *); +int hfsplus_ext_write_extent(struct inode *); int hfsplus_get_block(struct inode *, sector_t, struct buffer_head *, int); int hfsplus_free_fork(struct super_block *, u32, struct hfsplus_fork_raw *, int); @@ -404,7 +404,8 @@ int hfsplus_cat_read_inode(struct inode *, struct hfs_find_data *); int hfsplus_cat_write_inode(struct inode *); struct inode *hfsplus_new_inode(struct super_block *, int); void hfsplus_delete_inode(struct inode *); -int hfsplus_file_fsync(struct file *file, int datasync); +int hfsplus_file_fsync(struct file *file, loff_t start, loff_t end, + int datasync); /* ioctl.c */ long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index b248a6cf..40e1413 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -119,8 +119,8 @@ static ssize_t hfsplus_direct_IO(int rw, struct kiocb *iocb, struct inode *inode = file->f_path.dentry->d_inode->i_mapping->host; ssize_t ret; - ret = blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov, - offset, nr_segs, hfsplus_get_block, NULL); + ret = blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs, + hfsplus_get_block); /* * In case of error extending write may have instantiated a few @@ -195,11 +195,13 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, hip->flags = 0; set_bit(HFSPLUS_I_RSRC, &hip->flags); - hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); - err = hfsplus_find_cat(sb, dir->i_ino, &fd); - if (!err) - err = hfsplus_cat_read_inode(inode, &fd); - hfs_find_exit(&fd); + err = hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); + if (!err) { + err = hfsplus_find_cat(sb, dir->i_ino, &fd); + if (!err) + err = hfsplus_cat_read_inode(inode, &fd); + hfs_find_exit(&fd); + } if (err) { iput(inode); return ERR_PTR(err); @@ -296,6 +298,8 @@ static int hfsplus_setattr(struct dentry *dentry, struct iattr *attr) if ((attr->ia_valid & ATTR_SIZE) && attr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + error = vmtruncate(inode, attr->ia_size); if (error) return error; @@ -306,13 +310,19 @@ static int hfsplus_setattr(struct dentry *dentry, struct iattr *attr) return 0; } -int hfsplus_file_fsync(struct file *file, int datasync) +int hfsplus_file_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { struct inode *inode = file->f_mapping->host; struct hfsplus_inode_info *hip = HFSPLUS_I(inode); struct hfsplus_sb_info *sbi = HFSPLUS_SB(inode->i_sb); int error = 0, error2; + error = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (error) + return error; + mutex_lock(&inode->i_mutex); + /* * Sync inode metadata into the catalog and extent trees. */ @@ -340,6 +350,8 @@ int hfsplus_file_fsync(struct file *file, int datasync) if (!test_bit(HFSPLUS_SB_NOBARRIER, &sbi->flags)) blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); + mutex_unlock(&inode->i_mutex); + return error; } @@ -379,7 +391,7 @@ struct inode *hfsplus_new_inode(struct super_block *sb, int mode) inode->i_mode = mode; inode->i_uid = current_fsuid(); inode->i_gid = current_fsgid(); - inode->i_nlink = 1; + set_nlink(inode, 1); inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; hip = HFSPLUS_I(inode); @@ -500,7 +512,7 @@ int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) hfs_bnode_read(fd->bnode, &entry, fd->entryoffset, sizeof(struct hfsplus_cat_folder)); hfsplus_get_perms(inode, &folder->permissions, 1); - inode->i_nlink = 1; + set_nlink(inode, 1); inode->i_size = 2 + be32_to_cpu(folder->valence); inode->i_atime = hfsp_mt2ut(folder->access_date); inode->i_mtime = hfsp_mt2ut(folder->content_mod_date); @@ -520,11 +532,11 @@ int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) hfsplus_inode_read_fork(inode, HFSPLUS_IS_RSRC(inode) ? &file->rsrc_fork : &file->data_fork); hfsplus_get_perms(inode, &file->permissions, 0); - inode->i_nlink = 1; + set_nlink(inode, 1); if (S_ISREG(inode->i_mode)) { if (file->permissions.dev) - inode->i_nlink = - be32_to_cpu(file->permissions.dev); + set_nlink(inode, + be32_to_cpu(file->permissions.dev)); inode->i_op = &hfsplus_file_inode_operations; inode->i_fop = &hfsplus_file_operations; inode->i_mapping->a_ops = &hfsplus_aops; diff --git a/fs/hfsplus/options.c b/fs/hfsplus/options.c index bb62a5882..c8d6b4f 100644 --- a/fs/hfsplus/options.c +++ b/fs/hfsplus/options.c @@ -211,9 +211,9 @@ int hfsplus_show_options(struct seq_file *seq, struct vfsmount *mnt) struct hfsplus_sb_info *sbi = HFSPLUS_SB(mnt->mnt_sb); if (sbi->creator != HFSPLUS_DEF_CR_TYPE) - seq_printf(seq, ",creator=%.4s", (char *)&sbi->creator); + seq_show_option_n(seq, "creator", (char *)&sbi->creator, 4); if (sbi->type != HFSPLUS_DEF_CR_TYPE) - seq_printf(seq, ",type=%.4s", (char *)&sbi->type); + seq_show_option_n(seq, "type", (char *)&sbi->type, 4); seq_printf(seq, ",umask=%o,uid=%u,gid=%u", sbi->umask, sbi->uid, sbi->gid); if (sbi->part >= 0) diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index c3a76fd..d24a9b6 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -73,11 +73,13 @@ struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID || inode->i_ino == HFSPLUS_ROOT_CNID) { - hfs_find_init(HFSPLUS_SB(inode->i_sb)->cat_tree, &fd); - err = hfsplus_find_cat(inode->i_sb, inode->i_ino, &fd); - if (!err) - err = hfsplus_cat_read_inode(inode, &fd); - hfs_find_exit(&fd); + err = hfs_find_init(HFSPLUS_SB(inode->i_sb)->cat_tree, &fd); + if (!err) { + err = hfsplus_find_cat(inode->i_sb, inode->i_ino, &fd); + if (!err) + err = hfsplus_cat_read_inode(inode, &fd); + hfs_find_exit(&fd); + } } else { err = hfsplus_system_read_inode(inode); } @@ -133,9 +135,13 @@ static int hfsplus_system_write_inode(struct inode *inode) static int hfsplus_write_inode(struct inode *inode, struct writeback_control *wbc) { + int err; + dprint(DBG_INODE, "hfsplus_write_inode: %lu\n", inode->i_ino); - hfsplus_ext_write_extent(inode); + err = hfsplus_ext_write_extent(inode); + if (err) + return err; if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID || inode->i_ino == HFSPLUS_ROOT_CNID) @@ -338,6 +344,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) struct inode *root, *inode; struct qstr str; struct nls_table *nls = NULL; + u64 last_fs_block, last_fs_page; int err; err = -EINVAL; @@ -393,6 +400,17 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) if (!sbi->rsrc_clump_blocks) sbi->rsrc_clump_blocks = 1; + err = -EFBIG; + last_fs_block = sbi->total_blocks - 1; + last_fs_page = (last_fs_block << sbi->alloc_blksz_shift) >> + PAGE_CACHE_SHIFT; + + if ((last_fs_block > (sector_t)(~0ULL) >> (sbi->alloc_blksz_shift - 9)) || + (last_fs_page > (pgoff_t)(~0ULL))) { + printk(KERN_ERR "hfs: filesystem size too large.\n"); + goto out_free_vhdr; + } + /* Set up operations so we can load metadata */ sb->s_op = &hfsplus_sops; sb->s_maxbytes = MAX_LFS_FILESIZE; @@ -417,6 +435,8 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) sb->s_flags |= MS_RDONLY; } + err = -EINVAL; + /* Load metadata objects (B*Trees) */ sbi->ext_tree = hfs_btree_open(sb, HFSPLUS_EXT_CNID); if (!sbi->ext_tree) { @@ -447,7 +467,9 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) str.len = sizeof(HFSP_HIDDENDIR_NAME) - 1; str.name = HFSP_HIDDENDIR_NAME; - hfs_find_init(sbi->cat_tree, &fd); + err = hfs_find_init(sbi->cat_tree, &fd); + if (err) + goto out_put_root; hfsplus_cat_build_key(sb, fd.search_key, HFSPLUS_ROOT_CNID, &str); if (!hfs_brec_read(&fd, &entry, sizeof(entry))) { hfs_find_exit(&fd); diff --git a/fs/hfsplus/unicode.c b/fs/hfsplus/unicode.c index a3f0bfc..a32998f 100644 --- a/fs/hfsplus/unicode.c +++ b/fs/hfsplus/unicode.c @@ -142,7 +142,11 @@ int hfsplus_uni2asc(struct super_block *sb, /* search for single decomposed char */ if (likely(compose)) ce1 = hfsplus_compose_lookup(hfsplus_compose_table, c0); - if (ce1 && (cc = ce1[0])) { + if (ce1) + cc = ce1[0]; + else + cc = 0; + if (cc) { /* start of a possibly decomposed Hangul char */ if (cc != 0xffff) goto done; @@ -209,7 +213,8 @@ int hfsplus_uni2asc(struct super_block *sb, i++; ce2 = ce1; } - if ((cc = ce2[0])) { + cc = ce2[0]; + if (cc) { ip += i; ustrlen -= i; goto done; @@ -301,7 +306,11 @@ int hfsplus_asc2uni(struct super_block *sb, struct hfsplus_unistr *ustr, while (outlen < HFSPLUS_MAX_STRLEN && len > 0) { size = asc2unichar(sb, astr, len, &c); - if (decompose && (dstr = decompose_unichar(c, &dsize))) { + if (decompose) + dstr = decompose_unichar(c, &dsize); + else + dstr = NULL; + if (dstr) { if (outlen + dsize > HFSPLUS_MAX_STRLEN) break; do { @@ -346,15 +355,23 @@ int hfsplus_hash_dentry(const struct dentry *dentry, const struct inode *inode, astr += size; len -= size; - if (decompose && (dstr = decompose_unichar(c, &dsize))) { + if (decompose) + dstr = decompose_unichar(c, &dsize); + else + dstr = NULL; + if (dstr) { do { c2 = *dstr++; - if (!casefold || (c2 = case_fold(c2))) + if (casefold) + c2 = case_fold(c2); + if (!casefold || c2) hash = partial_name_hash(c2, hash); } while (--dsize > 0); } else { c2 = c; - if (!casefold || (c2 = case_fold(c2))) + if (casefold) + c2 = case_fold(c2); + if (!casefold || c2) hash = partial_name_hash(c2, hash); } } @@ -422,12 +439,14 @@ int hfsplus_compare_dentry(const struct dentry *parent, c1 = *dstr1; c2 = *dstr2; if (casefold) { - if (!(c1 = case_fold(c1))) { + c1 = case_fold(c1); + if (!c1) { dstr1++; dsize1--; continue; } - if (!(c2 = case_fold(c2))) { + c2 = case_fold(c2); + if (!c2) { dstr2++; dsize2--; continue; diff --git a/fs/hfsplus/wrapper.c b/fs/hfsplus/wrapper.c index aac1563..90effcc 100644 --- a/fs/hfsplus/wrapper.c +++ b/fs/hfsplus/wrapper.c @@ -184,10 +184,6 @@ int hfsplus_read_wrapper(struct super_block *sb) if (hfsplus_get_last_session(sb, &part_start, &part_size)) goto out; - if ((u64)part_start + part_size > 0x100000000ULL) { - pr_err("hfs: volumes larger than 2TB are not supported yet\n"); - goto out; - } error = -ENOMEM; sbi->s_vhdr_buf = kmalloc(hfsplus_min_io_size(sb), GFP_KERNEL); @@ -215,8 +211,9 @@ reread: if (!hfsplus_read_mdb(sbi->s_vhdr, &wd)) goto out_free_backup_vhdr; wd.ablk_size >>= HFSPLUS_SECTOR_SHIFT; - part_start += wd.ablk_start + wd.embed_start * wd.ablk_size; - part_size = wd.embed_count * wd.ablk_size; + part_start += (sector_t)wd.ablk_start + + (sector_t)wd.embed_start * wd.ablk_size; + part_size = (sector_t)wd.embed_count * wd.ablk_size; goto reread; default: /* diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index 2638c834e..104e4d9 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -265,7 +265,7 @@ static int hostfs_show_options(struct seq_file *seq, struct vfsmount *vfs) size_t offset = strlen(root_ino) + 1; if (strlen(root_path) > offset) - seq_printf(seq, ",%s", root_path + offset); + seq_show_option(seq, root_path + offset, NULL); return 0; } @@ -362,9 +362,20 @@ retry: return 0; } -int hostfs_fsync(struct file *file, int datasync) +int hostfs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { - return fsync_file(HOSTFS_I(file->f_mapping->host)->fd, datasync); + struct inode *inode = file->f_mapping->host; + int ret; + + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + + mutex_lock(&inode->i_mutex); + ret = fsync_file(HOSTFS_I(inode)->fd, datasync); + mutex_unlock(&inode->i_mutex); + + return ret; } static const struct file_operations hostfs_file_fops = { @@ -530,7 +541,7 @@ static int read_name(struct inode *ino, char *name) ino->i_ino = st.ino; ino->i_mode = st.mode; - ino->i_nlink = st.nlink; + set_nlink(ino, st.nlink); ino->i_uid = st.uid; ino->i_gid = st.gid; ino->i_atime = st.atime; @@ -748,12 +759,12 @@ int hostfs_rename(struct inode *from_ino, struct dentry *from, return err; } -int hostfs_permission(struct inode *ino, int desired, unsigned int flags) +int hostfs_permission(struct inode *ino, int desired) { char *name; int r = 0, w = 0, x = 0, err; - if (flags & IPERM_FLAG_RCU) + if (desired & MAY_NOT_BLOCK) return -ECHILD; if (desired & MAY_READ) r = 1; @@ -770,7 +781,7 @@ int hostfs_permission(struct inode *ino, int desired, unsigned int flags) err = access_file(name, r, w, x); __putname(name); if (!err) - err = generic_permission(ino, desired, flags, NULL); + err = generic_permission(ino, desired); return err; } diff --git a/fs/hostfs/hostfs_user.c b/fs/hostfs/hostfs_user.c index d51a983..dd7bc38 100644 --- a/fs/hostfs/hostfs_user.c +++ b/fs/hostfs/hostfs_user.c @@ -16,7 +16,6 @@ #include <sys/vfs.h> #include "hostfs.h" #include "os.h" -#include "user.h" #include <utime.h> static void stat64_to_hostfs(const struct stat64 *buf, struct hostfs_stat *p) diff --git a/fs/hpfs/alloc.c b/fs/hpfs/alloc.c index 7a5eb2c..e0d57c0 100644 --- a/fs/hpfs/alloc.c +++ b/fs/hpfs/alloc.c @@ -8,6 +8,58 @@ #include "hpfs_fn.h" +static void hpfs_claim_alloc(struct super_block *s, secno sec) +{ + struct hpfs_sb_info *sbi = hpfs_sb(s); + if (sbi->sb_n_free != (unsigned)-1) { + if (unlikely(!sbi->sb_n_free)) { + hpfs_error(s, "free count underflow, allocating sector %08x", sec); + sbi->sb_n_free = -1; + return; + } + sbi->sb_n_free--; + } +} + +static void hpfs_claim_free(struct super_block *s, secno sec) +{ + struct hpfs_sb_info *sbi = hpfs_sb(s); + if (sbi->sb_n_free != (unsigned)-1) { + if (unlikely(sbi->sb_n_free >= sbi->sb_fs_size)) { + hpfs_error(s, "free count overflow, freeing sector %08x", sec); + sbi->sb_n_free = -1; + return; + } + sbi->sb_n_free++; + } +} + +static void hpfs_claim_dirband_alloc(struct super_block *s, secno sec) +{ + struct hpfs_sb_info *sbi = hpfs_sb(s); + if (sbi->sb_n_free_dnodes != (unsigned)-1) { + if (unlikely(!sbi->sb_n_free_dnodes)) { + hpfs_error(s, "dirband free count underflow, allocating sector %08x", sec); + sbi->sb_n_free_dnodes = -1; + return; + } + sbi->sb_n_free_dnodes--; + } +} + +static void hpfs_claim_dirband_free(struct super_block *s, secno sec) +{ + struct hpfs_sb_info *sbi = hpfs_sb(s); + if (sbi->sb_n_free_dnodes != (unsigned)-1) { + if (unlikely(sbi->sb_n_free_dnodes >= sbi->sb_dirband_size / 4)) { + hpfs_error(s, "dirband free count overflow, freeing sector %08x", sec); + sbi->sb_n_free_dnodes = -1; + return; + } + sbi->sb_n_free_dnodes++; + } +} + /* * Check if a sector is allocated in bitmap * This is really slow. Turned on only if chk==2 @@ -203,9 +255,15 @@ secno hpfs_alloc_sector(struct super_block *s, secno near, unsigned n, int forwa } sec = 0; ret: + if (sec) { + i = 0; + do + hpfs_claim_alloc(s, sec + i); + while (unlikely(++i < n)); + } if (sec && f_p) { for (i = 0; i < forward; i++) { - if (!hpfs_alloc_if_possible(s, sec + i + 1)) { + if (!hpfs_alloc_if_possible(s, sec + n + i)) { hpfs_error(s, "Prealloc doesn't work! Wanted %d, allocated at %08x, can't allocate %d", forward, sec, i); sec = 0; break; @@ -228,6 +286,7 @@ static secno alloc_in_dirband(struct super_block *s, secno near) nr >>= 2; sec = alloc_in_bmp(s, (~0x3fff) | nr, 1, 0); if (!sec) return 0; + hpfs_claim_dirband_alloc(s, sec); return ((sec & 0x3fff) << 2) + sbi->sb_dirband_start; } @@ -242,6 +301,7 @@ int hpfs_alloc_if_possible(struct super_block *s, secno sec) bmp[(sec & 0x3fff) >> 5] &= cpu_to_le32(~(1 << (sec & 0x1f))); hpfs_mark_4buffers_dirty(&qbh); hpfs_brelse4(&qbh); + hpfs_claim_alloc(s, sec); return 1; } hpfs_brelse4(&qbh); @@ -275,6 +335,7 @@ void hpfs_free_sectors(struct super_block *s, secno sec, unsigned n) return; } bmp[(sec & 0x3fff) >> 5] |= cpu_to_le32(1 << (sec & 0x1f)); + hpfs_claim_free(s, sec); if (!--n) { hpfs_mark_4buffers_dirty(&qbh); hpfs_brelse4(&qbh); @@ -359,6 +420,7 @@ void hpfs_free_dnode(struct super_block *s, dnode_secno dno) bmp[ssec >> 5] |= cpu_to_le32(1 << (ssec & 0x1f)); hpfs_mark_4buffers_dirty(&qbh); hpfs_brelse4(&qbh); + hpfs_claim_dirband_free(s, dno); } } @@ -366,7 +428,7 @@ struct dnode *hpfs_alloc_dnode(struct super_block *s, secno near, dnode_secno *dno, struct quad_buffer_head *qbh) { struct dnode *d; - if (hpfs_count_one_bitmap(s, hpfs_sb(s)->sb_dmap) > FREE_DNODES_ADD) { + if (hpfs_get_free_dnodes(s) > FREE_DNODES_ADD) { if (!(*dno = alloc_in_dirband(s, near))) if (!(*dno = hpfs_alloc_sector(s, near, 4, 0))) return NULL; } else { diff --git a/fs/hpfs/dir.c b/fs/hpfs/dir.c index f46ae02..46549c7 100644 --- a/fs/hpfs/dir.c +++ b/fs/hpfs/dir.c @@ -29,25 +29,31 @@ static loff_t hpfs_dir_lseek(struct file *filp, loff_t off, int whence) struct hpfs_inode_info *hpfs_inode = hpfs_i(i); struct super_block *s = i->i_sb; + /* Somebody else will have to figure out what to do here */ + if (whence == SEEK_DATA || whence == SEEK_HOLE) + return -EINVAL; + + mutex_lock(&i->i_mutex); hpfs_lock(s); /*printk("dir lseek\n");*/ if (new_off == 0 || new_off == 1 || new_off == 11 || new_off == 12 || new_off == 13) goto ok; - mutex_lock(&i->i_mutex); pos = ((loff_t) hpfs_de_as_down_as_possible(s, hpfs_inode->i_dno) << 4) + 1; while (pos != new_off) { if (map_pos_dirent(i, &pos, &qbh)) hpfs_brelse4(&qbh); else goto fail; if (pos == 12) goto fail; } - mutex_unlock(&i->i_mutex); + hpfs_add_pos(i, &filp->f_pos); ok: + filp->f_pos = new_off; hpfs_unlock(s); - return filp->f_pos = new_off; -fail: mutex_unlock(&i->i_mutex); + return new_off; +fail: /*printk("illegal lseek: %016llx\n", new_off);*/ hpfs_unlock(s); + mutex_unlock(&i->i_mutex); return -ESPIPE; } @@ -243,7 +249,7 @@ struct dentry *hpfs_lookup(struct inode *dir, struct dentry *dentry, struct name result->i_mode &= ~0111; result->i_op = &hpfs_file_iops; result->i_fop = &hpfs_file_ops; - result->i_nlink = 1; + set_nlink(result, 1); } unlock_new_inode(result); } diff --git a/fs/hpfs/file.c b/fs/hpfs/file.c index 89c500e..5ecfffe 100644 --- a/fs/hpfs/file.c +++ b/fs/hpfs/file.c @@ -18,9 +18,14 @@ static int hpfs_file_release(struct inode *inode, struct file *file) return 0; } -int hpfs_file_fsync(struct file *file, int datasync) +int hpfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; + int ret; + + ret = filemap_write_and_wait_range(file->f_mapping, start, end); + if (ret) + return ret; return sync_blockdev(inode->i_sb->s_bdev); } @@ -111,9 +116,12 @@ static int hpfs_write_begin(struct file *file, struct address_space *mapping, hpfs_get_block, &hpfs_i(mapping->host)->mmu_private); if (unlikely(ret)) { - loff_t isize = mapping->host->i_size; + loff_t isize; + hpfs_lock(mapping->host->i_sb); + isize = mapping->host->i_size; if (pos + len > isize) vmtruncate(mapping->host, isize); + hpfs_unlock(mapping->host->i_sb); } return ret; diff --git a/fs/hpfs/hpfs_fn.h b/fs/hpfs/hpfs_fn.h index dd552f8..f1f2ca7 100644 --- a/fs/hpfs/hpfs_fn.h +++ b/fs/hpfs/hpfs_fn.h @@ -258,7 +258,7 @@ void hpfs_set_ea(struct inode *, struct fnode *, const char *, /* file.c */ -int hpfs_file_fsync(struct file *, int); +int hpfs_file_fsync(struct file *, loff_t, loff_t, int); extern const struct file_operations hpfs_file_ops; extern const struct inode_operations hpfs_file_iops; extern const struct address_space_operations hpfs_aops; @@ -311,10 +311,10 @@ static inline struct hpfs_sb_info *hpfs_sb(struct super_block *sb) /* super.c */ -void hpfs_error(struct super_block *, const char *, ...) - __attribute__((format (printf, 2, 3))); +__printf(2, 3) +void hpfs_error(struct super_block *, const char *, ...); int hpfs_stop_cycles(struct super_block *, int, int *, int *, char *); -unsigned hpfs_count_one_bitmap(struct super_block *, secno); +unsigned hpfs_get_free_dnodes(struct super_block *); /* * local time (HPFS) to GMT (Unix) diff --git a/fs/hpfs/inode.c b/fs/hpfs/inode.c index 338cd83..3b2cec2 100644 --- a/fs/hpfs/inode.c +++ b/fs/hpfs/inode.c @@ -53,7 +53,7 @@ void hpfs_read_inode(struct inode *i) i->i_mode &= ~0111; i->i_op = &hpfs_file_iops; i->i_fop = &hpfs_file_ops; - i->i_nlink = 0;*/ + clear_nlink(i);*/ make_bad_inode(i); return; } @@ -77,7 +77,7 @@ void hpfs_read_inode(struct inode *i) i->i_mode = S_IFLNK | 0777; i->i_op = &page_symlink_inode_operations; i->i_data.a_ops = &hpfs_symlink_aops; - i->i_nlink = 1; + set_nlink(i, 1); i->i_size = ea_size; i->i_blocks = 1; brelse(bh); @@ -101,7 +101,7 @@ void hpfs_read_inode(struct inode *i) } if (S_ISBLK(mode) || S_ISCHR(mode) || S_ISFIFO(mode) || S_ISSOCK(mode)) { brelse(bh); - i->i_nlink = 1; + set_nlink(i, 1); i->i_size = 0; i->i_blocks = 1; init_special_inode(i, mode, @@ -125,13 +125,13 @@ void hpfs_read_inode(struct inode *i) hpfs_count_dnodes(i->i_sb, hpfs_inode->i_dno, &n_dnodes, &n_subdirs, NULL); i->i_blocks = 4 * n_dnodes; i->i_size = 2048 * n_dnodes; - i->i_nlink = 2 + n_subdirs; + set_nlink(i, 2 + n_subdirs); } else { i->i_mode |= S_IFREG; if (!hpfs_inode->i_ea_mode) i->i_mode &= ~0111; i->i_op = &hpfs_file_iops; i->i_fop = &hpfs_file_ops; - i->i_nlink = 1; + set_nlink(i, 1); i->i_size = le32_to_cpu(fnode->file_size); i->i_blocks = ((i->i_size + 511) >> 9) + 1; i->i_data.a_ops = &hpfs_aops; diff --git a/fs/hpfs/namei.c b/fs/hpfs/namei.c index acf95da..ea91fcb 100644 --- a/fs/hpfs/namei.c +++ b/fs/hpfs/namei.c @@ -56,7 +56,7 @@ static int hpfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) result->i_fop = &hpfs_dir_ops; result->i_blocks = 4; result->i_size = 2048; - result->i_nlink = 2; + set_nlink(result, 2); if (dee.read_only) result->i_mode &= ~0222; @@ -150,7 +150,7 @@ static int hpfs_create(struct inode *dir, struct dentry *dentry, int mode, struc result->i_mode &= ~0111; result->i_op = &hpfs_file_iops; result->i_fop = &hpfs_file_ops; - result->i_nlink = 1; + set_nlink(result, 1); hpfs_i(result)->i_parent_dir = dir->i_ino; result->i_ctime.tv_sec = result->i_mtime.tv_sec = result->i_atime.tv_sec = local_to_gmt(dir->i_sb, le32_to_cpu(dee.creation_date)); result->i_ctime.tv_nsec = 0; @@ -242,7 +242,7 @@ static int hpfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t hpfs_i(result)->i_ea_size = 0; result->i_uid = current_fsuid(); result->i_gid = current_fsgid(); - result->i_nlink = 1; + set_nlink(result, 1); result->i_size = 0; result->i_blocks = 1; init_special_inode(result, mode, rdev); @@ -318,7 +318,7 @@ static int hpfs_symlink(struct inode *dir, struct dentry *dentry, const char *sy result->i_uid = current_fsuid(); result->i_gid = current_fsgid(); result->i_blocks = 1; - result->i_nlink = 1; + set_nlink(result, 1); result->i_size = strlen(symlink); result->i_op = &page_symlink_inode_operations; result->i_data.a_ops = &hpfs_symlink_aops; @@ -398,7 +398,7 @@ again: hpfs_unlock(dir->i_sb); return -ENOSPC; } - if (generic_permission(inode, MAY_WRITE, 0, NULL) || + if (generic_permission(inode, MAY_WRITE) || !S_ISREG(inode->i_mode) || get_write_access(inode)) { d_rehash(dentry); diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index f760c15..b03e766 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -115,7 +115,7 @@ static void hpfs_put_super(struct super_block *s) kfree(sbi); } -unsigned hpfs_count_one_bitmap(struct super_block *s, secno secno) +static unsigned hpfs_count_one_bitmap(struct super_block *s, secno secno) { struct quad_buffer_head qbh; unsigned long *bits; @@ -123,7 +123,7 @@ unsigned hpfs_count_one_bitmap(struct super_block *s, secno secno) bits = hpfs_map_4sectors(s, secno, &qbh, 4); if (!bits) - return 0; + return (unsigned)-1; count = bitmap_weight(bits, 2048 * BITS_PER_BYTE); hpfs_brelse4(&qbh); return count; @@ -134,29 +134,45 @@ static unsigned count_bitmaps(struct super_block *s) unsigned n, count, n_bands; n_bands = (hpfs_sb(s)->sb_fs_size + 0x3fff) >> 14; count = 0; - for (n = 0; n < n_bands; n++) - count += hpfs_count_one_bitmap(s, le32_to_cpu(hpfs_sb(s)->sb_bmp_dir[n])); + for (n = 0; n < n_bands; n++) { + unsigned c; + c = hpfs_count_one_bitmap(s, le32_to_cpu(hpfs_sb(s)->sb_bmp_dir[n])); + if (c != (unsigned)-1) + count += c; + } return count; } +unsigned hpfs_get_free_dnodes(struct super_block *s) +{ + struct hpfs_sb_info *sbi = hpfs_sb(s); + if (sbi->sb_n_free_dnodes == (unsigned)-1) { + unsigned c = hpfs_count_one_bitmap(s, sbi->sb_dmap); + if (c == (unsigned)-1) + return 0; + sbi->sb_n_free_dnodes = c; + } + return sbi->sb_n_free_dnodes; +} + static int hpfs_statfs(struct dentry *dentry, struct kstatfs *buf) { struct super_block *s = dentry->d_sb; struct hpfs_sb_info *sbi = hpfs_sb(s); u64 id = huge_encode_dev(s->s_bdev->bd_dev); + hpfs_lock(s); - /*if (sbi->sb_n_free == -1) {*/ + if (sbi->sb_n_free == (unsigned)-1) sbi->sb_n_free = count_bitmaps(s); - sbi->sb_n_free_dnodes = hpfs_count_one_bitmap(s, sbi->sb_dmap); - /*}*/ + buf->f_type = s->s_magic; buf->f_bsize = 512; buf->f_blocks = sbi->sb_fs_size; buf->f_bfree = sbi->sb_n_free; buf->f_bavail = sbi->sb_n_free; buf->f_files = sbi->sb_dirband_size / 4; - buf->f_ffree = sbi->sb_n_free_dnodes; + buf->f_ffree = hpfs_get_free_dnodes(s); buf->f_fsid.val[0] = (u32)id; buf->f_fsid.val[1] = (u32)(id >> 32); buf->f_namelen = 254; diff --git a/fs/hppfs/hppfs.c b/fs/hppfs/hppfs.c index 9d71c95..f590b11 100644 --- a/fs/hppfs/hppfs.c +++ b/fs/hppfs/hppfs.c @@ -574,9 +574,10 @@ static int hppfs_readdir(struct file *file, void *ent, filldir_t filldir) return err; } -static int hppfs_fsync(struct file *file, int datasync) +static int hppfs_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { - return 0; + return filemap_write_and_wait_range(file->f_mapping, start, end); } static const struct file_operations hppfs_dir_fops = { @@ -701,7 +702,7 @@ static struct inode *get_inode(struct super_block *sb, struct dentry *dentry) inode->i_ctime = proc_ino->i_ctime; inode->i_ino = proc_ino->i_ino; inode->i_mode = proc_ino->i_mode; - inode->i_nlink = proc_ino->i_nlink; + set_nlink(inode, proc_ino->i_nlink); inode->i_size = proc_ino->i_size; inode->i_blocks = proc_ino->i_blocks; diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 6327a06..0aa424a 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -94,7 +94,7 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma) vma->vm_flags |= VM_HUGETLB | VM_RESERVED; vma->vm_ops = &hugetlb_vm_ops; - if (vma->vm_pgoff & ~(huge_page_mask(h) >> PAGE_SHIFT)) + if (vma->vm_pgoff & (~huge_page_mask(h) >> PAGE_SHIFT)) return -EINVAL; vma_len = (loff_t)(vma->vm_end - vma->vm_start); @@ -484,6 +484,7 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb, uid_t uid, inode->i_op = &page_symlink_inode_operations; break; } + lockdep_annotate_inode_mutex_key(inode); } return inode; } @@ -592,9 +593,15 @@ static int hugetlbfs_statfs(struct dentry *dentry, struct kstatfs *buf) spin_lock(&sbinfo->stat_lock); /* If no limits set, just report 0 for max/free/used * blocks, like simple_statfs() */ - if (sbinfo->max_blocks >= 0) { - buf->f_blocks = sbinfo->max_blocks; - buf->f_bavail = buf->f_bfree = sbinfo->free_blocks; + if (sbinfo->spool) { + long free_pages; + + spin_lock(&sbinfo->spool->lock); + buf->f_blocks = sbinfo->spool->max_hpages; + free_pages = sbinfo->spool->max_hpages + - sbinfo->spool->used_hpages; + buf->f_bavail = buf->f_bfree = free_pages; + spin_unlock(&sbinfo->spool->lock); buf->f_files = sbinfo->max_inodes; buf->f_ffree = sbinfo->free_inodes; } @@ -610,6 +617,10 @@ static void hugetlbfs_put_super(struct super_block *sb) if (sbi) { sb->s_fs_info = NULL; + + if (sbi->spool) + hugepage_put_subpool(sbi->spool); + kfree(sbi); } } @@ -841,10 +852,14 @@ hugetlbfs_fill_super(struct super_block *sb, void *data, int silent) sb->s_fs_info = sbinfo; sbinfo->hstate = config.hstate; spin_lock_init(&sbinfo->stat_lock); - sbinfo->max_blocks = config.nr_blocks; - sbinfo->free_blocks = config.nr_blocks; sbinfo->max_inodes = config.nr_inodes; sbinfo->free_inodes = config.nr_inodes; + sbinfo->spool = NULL; + if (config.nr_blocks != -1) { + sbinfo->spool = hugepage_new_subpool(config.nr_blocks); + if (!sbinfo->spool) + goto out_free; + } sb->s_maxbytes = MAX_LFS_FILESIZE; sb->s_blocksize = huge_page_size(config.hstate); sb->s_blocksize_bits = huge_page_shift(config.hstate); @@ -864,38 +879,12 @@ hugetlbfs_fill_super(struct super_block *sb, void *data, int silent) sb->s_root = root; return 0; out_free: + if (sbinfo->spool) + kfree(sbinfo->spool); kfree(sbinfo); return -ENOMEM; } -int hugetlb_get_quota(struct address_space *mapping, long delta) -{ - int ret = 0; - struct hugetlbfs_sb_info *sbinfo = HUGETLBFS_SB(mapping->host->i_sb); - - if (sbinfo->free_blocks > -1) { - spin_lock(&sbinfo->stat_lock); - if (sbinfo->free_blocks - delta >= 0) - sbinfo->free_blocks -= delta; - else - ret = -ENOMEM; - spin_unlock(&sbinfo->stat_lock); - } - - return ret; -} - -void hugetlb_put_quota(struct address_space *mapping, long delta) -{ - struct hugetlbfs_sb_info *sbinfo = HUGETLBFS_SB(mapping->host->i_sb); - - if (sbinfo->free_blocks > -1) { - spin_lock(&sbinfo->stat_lock); - sbinfo->free_blocks += delta; - spin_unlock(&sbinfo->stat_lock); - } -} - static struct dentry *hugetlbfs_mount(struct file_system_type *fs_type, int flags, const char *dev_name, void *data) { @@ -963,7 +952,7 @@ struct file *hugetlb_file_setup(const char *name, size_t size, d_instantiate(path.dentry, inode); inode->i_size = size; - inode->i_nlink = 0; + clear_nlink(inode); error = -ENFILE; file = alloc_file(&path, FMODE_WRITE | FMODE_READ, @@ -1024,6 +1013,7 @@ static int __init init_hugetlbfs_fs(void) static void __exit exit_hugetlbfs_fs(void) { kmem_cache_destroy(hugetlbfs_inode_cachep); + kern_unmount(hugetlbfs_vfsmount); unregister_filesystem(&hugetlbfs_fs_type); bdi_destroy(&hugetlbfs_backing_dev_info); } diff --git a/fs/isofs/dir.c b/fs/isofs/dir.c index 0542b6e..f20437c 100644 --- a/fs/isofs/dir.c +++ b/fs/isofs/dir.c @@ -254,19 +254,16 @@ static int isofs_readdir(struct file *filp, char *tmpname; struct iso_directory_record *tmpde; struct inode *inode = filp->f_path.dentry->d_inode; - struct isofs_sb_info *sbi = ISOFS_SB(inode->i_sb); tmpname = (char *)__get_free_page(GFP_KERNEL); if (tmpname == NULL) return -ENOMEM; - mutex_lock(&sbi->s_mutex); tmpde = (struct iso_directory_record *) (tmpname+1024); result = do_isofs_readdir(inode, filp, dirent, filldir, tmpname, tmpde); free_page((unsigned long) tmpname); - mutex_unlock(&sbi->s_mutex); return result; } diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index 26f6364..2f9197f 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -20,6 +20,7 @@ #include <linux/statfs.h> #include <linux/cdrom.h> #include <linux/parser.h> +#include <linux/mpage.h> #include "isofs.h" #include "zisofs.h" @@ -67,7 +68,7 @@ static void isofs_put_super(struct super_block *sb) return; } -static int isofs_read_inode(struct inode *); +static int isofs_read_inode(struct inode *, int relocated); static int isofs_statfs (struct dentry *, struct kstatfs *); static struct kmem_cache *isofs_inode_cachep; @@ -854,7 +855,6 @@ root_found: sbi->s_utf8 = opt.utf8; sbi->s_nocompress = opt.nocompress; sbi->s_overriderockperm = opt.overriderockperm; - mutex_init(&sbi->s_mutex); /* * It would be incredibly stupid to allow people to mark every file * on the disk as suid, so we merely allow them to set the default @@ -1140,7 +1140,13 @@ struct buffer_head *isofs_bread(struct inode *inode, sector_t block) static int isofs_readpage(struct file *file, struct page *page) { - return block_read_full_page(page,isofs_get_block); + return mpage_readpage(page, isofs_get_block); +} + +static int isofs_readpages(struct file *file, struct address_space *mapping, + struct list_head *pages, unsigned nr_pages) +{ + return mpage_readpages(mapping, pages, nr_pages, isofs_get_block); } static sector_t _isofs_bmap(struct address_space *mapping, sector_t block) @@ -1150,6 +1156,7 @@ static sector_t _isofs_bmap(struct address_space *mapping, sector_t block) static const struct address_space_operations isofs_aops = { .readpage = isofs_readpage, + .readpages = isofs_readpages, .bmap = _isofs_bmap }; @@ -1256,7 +1263,7 @@ out_toomany: goto out; } -static int isofs_read_inode(struct inode *inode) +static int isofs_read_inode(struct inode *inode, int relocated) { struct super_block *sb = inode->i_sb; struct isofs_sb_info *sbi = ISOFS_SB(sb); @@ -1311,7 +1318,7 @@ static int isofs_read_inode(struct inode *inode) inode->i_mode = S_IFDIR | sbi->s_dmode; else inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO; - inode->i_nlink = 1; /* + set_nlink(inode, 1); /* * Set to 1. We know there are 2, but * the find utility tries to optimize * if it is 2, and it screws up. It is @@ -1329,7 +1336,7 @@ static int isofs_read_inode(struct inode *inode) */ inode->i_mode = S_IFREG | S_IRUGO | S_IXUGO; } - inode->i_nlink = 1; + set_nlink(inode, 1); } inode->i_uid = sbi->s_uid; inode->i_gid = sbi->s_gid; @@ -1401,7 +1408,7 @@ static int isofs_read_inode(struct inode *inode) */ if (!high_sierra) { - parse_rock_ridge_inode(de, inode); + parse_rock_ridge_inode(de, inode, relocated); /* if we want uid/gid set, override the rock ridge setting */ if (sbi->s_uid_set) inode->i_uid = sbi->s_uid; @@ -1480,9 +1487,10 @@ static int isofs_iget5_set(struct inode *ino, void *data) * offset that point to the underlying meta-data for the inode. The * code below is otherwise similar to the iget() code in * include/linux/fs.h */ -struct inode *isofs_iget(struct super_block *sb, - unsigned long block, - unsigned long offset) +struct inode *__isofs_iget(struct super_block *sb, + unsigned long block, + unsigned long offset, + int relocated) { unsigned long hashval; struct inode *inode; @@ -1504,7 +1512,7 @@ struct inode *isofs_iget(struct super_block *sb, return ERR_PTR(-ENOMEM); if (inode->i_state & I_NEW) { - ret = isofs_read_inode(inode); + ret = isofs_read_inode(inode, relocated); if (ret < 0) { iget_failed(inode); inode = ERR_PTR(ret); diff --git a/fs/isofs/isofs.h b/fs/isofs/isofs.h index 2882dc0..f9c9793 100644 --- a/fs/isofs/isofs.h +++ b/fs/isofs/isofs.h @@ -55,7 +55,6 @@ struct isofs_sb_info { gid_t s_gid; uid_t s_uid; struct nls_table *s_nls_iocharset; /* Native language support table */ - struct mutex s_mutex; /* replaces BKL, please remove if possible */ }; #define ISOFS_INVALID_MODE ((mode_t) -1) @@ -108,7 +107,7 @@ extern int iso_date(char *, int); struct inode; /* To make gcc happy */ -extern int parse_rock_ridge_inode(struct iso_directory_record *, struct inode *); +extern int parse_rock_ridge_inode(struct iso_directory_record *, struct inode *, int relocated); extern int get_rock_ridge_filename(struct iso_directory_record *, char *, struct inode *); extern int isofs_name_translate(struct iso_directory_record *, char *, struct inode *); @@ -119,9 +118,24 @@ extern struct dentry *isofs_lookup(struct inode *, struct dentry *, struct namei extern struct buffer_head *isofs_bread(struct inode *, sector_t); extern int isofs_get_blocks(struct inode *, sector_t, struct buffer_head **, unsigned long); -extern struct inode *isofs_iget(struct super_block *sb, - unsigned long block, - unsigned long offset); +struct inode *__isofs_iget(struct super_block *sb, + unsigned long block, + unsigned long offset, + int relocated); + +static inline struct inode *isofs_iget(struct super_block *sb, + unsigned long block, + unsigned long offset) +{ + return __isofs_iget(sb, block, offset, 0); +} + +static inline struct inode *isofs_iget_reloc(struct super_block *sb, + unsigned long block, + unsigned long offset) +{ + return __isofs_iget(sb, block, offset, 1); +} /* Because the inode number is no longer relevant to finding the * underlying meta-data for an inode, we are free to choose a more diff --git a/fs/isofs/namei.c b/fs/isofs/namei.c index 4fb3e80..1e2946f 100644 --- a/fs/isofs/namei.c +++ b/fs/isofs/namei.c @@ -168,7 +168,6 @@ struct dentry *isofs_lookup(struct inode *dir, struct dentry *dentry, struct nam int found; unsigned long uninitialized_var(block); unsigned long uninitialized_var(offset); - struct isofs_sb_info *sbi = ISOFS_SB(dir->i_sb); struct inode *inode; struct page *page; @@ -176,21 +175,13 @@ struct dentry *isofs_lookup(struct inode *dir, struct dentry *dentry, struct nam if (!page) return ERR_PTR(-ENOMEM); - mutex_lock(&sbi->s_mutex); found = isofs_find_entry(dir, dentry, &block, &offset, page_address(page), 1024 + page_address(page)); __free_page(page); - inode = NULL; - if (found) { - inode = isofs_iget(dir->i_sb, block, offset); - if (IS_ERR(inode)) { - mutex_unlock(&sbi->s_mutex); - return ERR_CAST(inode); - } - } - mutex_unlock(&sbi->s_mutex); + inode = found ? isofs_iget(dir->i_sb, block, offset) : NULL; + return d_splice_alias(inode, dentry); } diff --git a/fs/isofs/rock.c b/fs/isofs/rock.c index f9cd04d..1780949 100644 --- a/fs/isofs/rock.c +++ b/fs/isofs/rock.c @@ -30,6 +30,7 @@ struct rock_state { int cont_size; int cont_extent; int cont_offset; + int cont_loops; struct inode *inode; }; @@ -73,6 +74,9 @@ static void init_rock_state(struct rock_state *rs, struct inode *inode) rs->inode = inode; } +/* Maximum number of Rock Ridge continuation entries */ +#define RR_MAX_CE_ENTRIES 32 + /* * Returns 0 if the caller should continue scanning, 1 if the scan must end * and -ve on error. @@ -105,6 +109,8 @@ static int rock_continue(struct rock_state *rs) goto out; } ret = -EIO; + if (++rs->cont_loops >= RR_MAX_CE_ENTRIES) + goto out; bh = sb_bread(rs->inode->i_sb, rs->cont_extent); if (bh) { memcpy(rs->buffer, bh->b_data + rs->cont_offset, @@ -288,12 +294,16 @@ eio: goto out; } +#define RR_REGARD_XA 1 +#define RR_RELOC_DE 2 + static int parse_rock_ridge_inode_internal(struct iso_directory_record *de, - struct inode *inode, int regard_xa) + struct inode *inode, int flags) { int symlink_len = 0; int cnt, sig; + unsigned int reloc_block; struct inode *reloc; struct rock_ridge *rr; int rootflag; @@ -305,7 +315,7 @@ parse_rock_ridge_inode_internal(struct iso_directory_record *de, init_rock_state(&rs, inode); setup_rock_ridge(de, inode, &rs); - if (regard_xa) { + if (flags & RR_REGARD_XA) { rs.chr += 14; rs.len -= 14; if (rs.len < 0) @@ -352,6 +362,9 @@ repeat: rs.cont_size = isonum_733(rr->u.CE.size); break; case SIG('E', 'R'): + /* Invalid length of ER tag id? */ + if (rr->u.ER.len_id + offsetof(struct rock_ridge, u.ER.data) > rr->len) + goto out; ISOFS_SB(inode->i_sb)->s_rock = 1; printk(KERN_DEBUG "ISO 9660 Extensions: "); { @@ -363,7 +376,7 @@ repeat: break; case SIG('P', 'X'): inode->i_mode = isonum_733(rr->u.PX.mode); - inode->i_nlink = isonum_733(rr->u.PX.n_links); + set_nlink(inode, isonum_733(rr->u.PX.n_links)); inode->i_uid = isonum_733(rr->u.PX.uid); inode->i_gid = isonum_733(rr->u.PX.gid); break; @@ -485,18 +498,28 @@ repeat: "relocated directory\n"); goto out; case SIG('C', 'L'): - ISOFS_I(inode)->i_first_extent = - isonum_733(rr->u.CL.location); - reloc = - isofs_iget(inode->i_sb, - ISOFS_I(inode)->i_first_extent, - 0); + if (flags & RR_RELOC_DE) { + printk(KERN_ERR + "ISOFS: Recursive directory relocation " + "is not supported\n"); + goto eio; + } + reloc_block = isonum_733(rr->u.CL.location); + if (reloc_block == ISOFS_I(inode)->i_iget5_block && + ISOFS_I(inode)->i_iget5_offset == 0) { + printk(KERN_ERR + "ISOFS: Directory relocation points to " + "itself\n"); + goto eio; + } + ISOFS_I(inode)->i_first_extent = reloc_block; + reloc = isofs_iget_reloc(inode->i_sb, reloc_block, 0); if (IS_ERR(reloc)) { ret = PTR_ERR(reloc); goto out; } inode->i_mode = reloc->i_mode; - inode->i_nlink = reloc->i_nlink; + set_nlink(inode, reloc->i_nlink); inode->i_uid = reloc->i_uid; inode->i_gid = reloc->i_gid; inode->i_rdev = reloc->i_rdev; @@ -637,9 +660,11 @@ static char *get_symlink_chunk(char *rpnt, struct rock_ridge *rr, char *plimit) return rpnt; } -int parse_rock_ridge_inode(struct iso_directory_record *de, struct inode *inode) +int parse_rock_ridge_inode(struct iso_directory_record *de, struct inode *inode, + int relocated) { - int result = parse_rock_ridge_inode_internal(de, inode, 0); + int flags = relocated ? RR_RELOC_DE : 0; + int result = parse_rock_ridge_inode_internal(de, inode, flags); /* * if rockridge flag was reset and we didn't look for attributes @@ -647,7 +672,8 @@ int parse_rock_ridge_inode(struct iso_directory_record *de, struct inode *inode) */ if ((ISOFS_SB(inode->i_sb)->s_rock_offset == -1) && (ISOFS_SB(inode->i_sb)->s_rock == 2)) { - result = parse_rock_ridge_inode_internal(de, inode, 14); + result = parse_rock_ridge_inode_internal(de, inode, + flags | RR_REGARD_XA); } return result; } @@ -678,7 +704,6 @@ static int rock_ridge_symlink_readpage(struct file *file, struct page *page) init_rock_state(&rs, inode); block = ei->i_iget5_block; - mutex_lock(&sbi->s_mutex); bh = sb_bread(inode->i_sb, block); if (!bh) goto out_noread; @@ -748,7 +773,6 @@ repeat: goto fail; brelse(bh); *rpnt = '\0'; - mutex_unlock(&sbi->s_mutex); SetPageUptodate(page); kunmap(page); unlock_page(page); @@ -765,7 +789,6 @@ out_bad_span: printk("symlink spans iso9660 blocks\n"); fail: brelse(bh); - mutex_unlock(&sbi->s_mutex); error: SetPageError(page); kunmap(page); diff --git a/fs/jbd/checkpoint.c b/fs/jbd/checkpoint.c index e4b87bc..5c93ffc 100644 --- a/fs/jbd/checkpoint.c +++ b/fs/jbd/checkpoint.c @@ -22,6 +22,8 @@ #include <linux/jbd.h> #include <linux/errno.h> #include <linux/slab.h> +#include <linux/blkdev.h> +#include <trace/events/jbd.h> /* * Unlink a buffer from a transaction checkpoint list. @@ -95,10 +97,14 @@ static int __try_to_free_cp_buf(struct journal_head *jh) if (jh->b_jlist == BJ_None && !buffer_locked(bh) && !buffer_dirty(bh) && !buffer_write_io_error(bh)) { + /* + * Get our reference so that bh cannot be freed before + * we unlock it + */ + get_bh(bh); JBUFFER_TRACE(jh, "remove from checkpoint list"); ret = __journal_remove_checkpoint(jh) + 1; jbd_unlock_bh_state(bh); - journal_remove_journal_head(bh); BUFFER_TRACE(bh, "release"); __brelse(bh); } else { @@ -220,8 +226,8 @@ restart: spin_lock(&journal->j_list_lock); goto restart; } + get_bh(bh); if (buffer_locked(bh)) { - get_bh(bh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); wait_on_buffer(bh); @@ -240,7 +246,6 @@ restart: */ released = __journal_remove_checkpoint(jh); jbd_unlock_bh_state(bh); - journal_remove_journal_head(bh); __brelse(bh); } @@ -253,9 +258,12 @@ static void __flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count) { int i; + struct blk_plug plug; + blk_start_plug(&plug); for (i = 0; i < *batch_count; i++) - write_dirty_buffer(bhs[i], WRITE); + write_dirty_buffer(bhs[i], WRITE_SYNC); + blk_finish_plug(&plug); for (i = 0; i < *batch_count; i++) { struct buffer_head *bh = bhs[i]; @@ -304,12 +312,12 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh, ret = 1; if (unlikely(buffer_write_io_error(bh))) ret = -EIO; + get_bh(bh); J_ASSERT_JH(jh, !buffer_jbddirty(bh)); BUFFER_TRACE(bh, "remove from checkpoint"); __journal_remove_checkpoint(jh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); - journal_remove_journal_head(bh); __brelse(bh); } else { /* @@ -358,6 +366,7 @@ int log_do_checkpoint(journal_t *journal) * journal straight away. */ result = cleanup_journal_tail(journal); + trace_jbd_checkpoint(journal, result); jbd_debug(1, "cleanup_journal_tail returned %d\n", result); if (result <= 0) return result; @@ -444,8 +453,6 @@ out: * * Return <0 on error, 0 on success, 1 if there was nothing to clean up. * - * Called with the journal lock held. - * * This is the only part of the journaling code which really needs to be * aware of transaction aborts. Checkpointing involves writing to the * main filesystem area rather than to the journal, so it can proceed @@ -463,13 +470,14 @@ int cleanup_journal_tail(journal_t *journal) if (is_journal_aborted(journal)) return 1; - /* OK, work out the oldest transaction remaining in the log, and + /* + * OK, work out the oldest transaction remaining in the log, and * the log block it starts at. * * If the log is now empty, we need to work out which is the * next transaction ID we will write, and where it will - * start. */ - + * start. + */ spin_lock(&journal->j_state_lock); spin_lock(&journal->j_list_lock); transaction = journal->j_checkpoint_transactions; @@ -495,7 +503,25 @@ int cleanup_journal_tail(journal_t *journal) spin_unlock(&journal->j_state_lock); return 1; } + spin_unlock(&journal->j_state_lock); + /* + * We need to make sure that any blocks that were recently written out + * --- perhaps by log_do_checkpoint() --- are flushed out before we + * drop the transactions from the journal. It's unlikely this will be + * necessary, especially with an appropriately sized journal, but we + * need this to guarantee correctness. Fortunately + * cleanup_journal_tail() doesn't get called all that often. + */ + if (journal->j_flags & JFS_BARRIER) + blkdev_issue_flush(journal->j_fs_dev, GFP_KERNEL, NULL); + + spin_lock(&journal->j_state_lock); + if (!tid_gt(first_tid, journal->j_tail_sequence)) { + spin_unlock(&journal->j_state_lock); + /* Someone else cleaned up journal so return 0 */ + return 0; + } /* OK, update the superblock to recover the freed space. * Physical blocks come first: have we wrapped beyond the end of * the log? */ @@ -503,6 +529,7 @@ int cleanup_journal_tail(journal_t *journal) if (blocknr < journal->j_tail) freed = freed + journal->j_last - journal->j_first; + trace_jbd_cleanup_journal_tail(journal, first_tid, blocknr, freed); jbd_debug(1, "Cleaning journal tail from %d to %d (offset %u), " "freeing %u\n", @@ -523,9 +550,9 @@ int cleanup_journal_tail(journal_t *journal) /* * journal_clean_one_cp_list * - * Find all the written-back checkpoint buffers in the given list and release them. + * Find all the written-back checkpoint buffers in the given list and release + * them. * - * Called with the journal locked. * Called with j_list_lock held. * Returns number of bufers reaped (for debug) */ @@ -632,8 +659,8 @@ out: * checkpoint lists. * * The function returns 1 if it frees the transaction, 0 otherwise. + * The function can free jh and bh. * - * This function is called with the journal locked. * This function is called with j_list_lock held. * This function is called with jbd_lock_bh_state(jh2bh(jh)) */ @@ -652,13 +679,14 @@ int __journal_remove_checkpoint(struct journal_head *jh) } journal = transaction->t_journal; + JBUFFER_TRACE(jh, "removing from transaction"); __buffer_unlink(jh); jh->b_cp_transaction = NULL; + journal_put_journal_head(jh); if (transaction->t_checkpoint_list != NULL || transaction->t_checkpoint_io_list != NULL) goto out; - JBUFFER_TRACE(jh, "transaction has no more buffers"); /* * There is one special case to worry about: if we have just pulled the @@ -669,10 +697,8 @@ int __journal_remove_checkpoint(struct journal_head *jh) * The locking here around t_state is a bit sleazy. * See the comment at the end of journal_commit_transaction(). */ - if (transaction->t_state != T_FINISHED) { - JBUFFER_TRACE(jh, "belongs to running/committing transaction"); + if (transaction->t_state != T_FINISHED) goto out; - } /* OK, that was the last buffer for the transaction: we can now safely remove this transaction from the log */ @@ -684,7 +710,6 @@ int __journal_remove_checkpoint(struct journal_head *jh) wake_up(&journal->j_wait_logspace); ret = 1; out: - JBUFFER_TRACE(jh, "exit"); return ret; } @@ -703,6 +728,8 @@ void __journal_insert_checkpoint(struct journal_head *jh, J_ASSERT_JH(jh, buffer_dirty(jh2bh(jh)) || buffer_jbddirty(jh2bh(jh))); J_ASSERT_JH(jh, jh->b_cp_transaction == NULL); + /* Get reference for checkpointing transaction */ + journal_grab_journal_head(jh2bh(jh)); jh->b_cp_transaction = transaction; if (!transaction->t_checkpoint_list) { @@ -752,6 +779,7 @@ void __journal_drop_transaction(journal_t *journal, transaction_t *transaction) J_ASSERT(journal->j_committing_transaction != transaction); J_ASSERT(journal->j_running_transaction != transaction); + trace_jbd_drop_transaction(journal, transaction); jbd_debug(1, "Dropping transaction %d, all done\n", transaction->t_tid); kfree(transaction); } diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c index dcd23f8..931bf95 100644 --- a/fs/jbd/commit.c +++ b/fs/jbd/commit.c @@ -21,6 +21,7 @@ #include <linux/pagemap.h> #include <linux/bio.h> #include <linux/blkdev.h> +#include <trace/events/jbd.h> /* * Default IO end handler for temporary BJ_IO buffer_heads. @@ -209,6 +210,8 @@ write_out_data: if (!trylock_buffer(bh)) { BUFFER_TRACE(bh, "needs blocking lock"); spin_unlock(&journal->j_list_lock); + trace_jbd_do_submit_data(journal, + commit_transaction); /* Write out all data to prevent deadlocks */ journal_do_submit_data(wbuf, bufs, write_op); bufs = 0; @@ -241,6 +244,8 @@ write_out_data: jbd_unlock_bh_state(bh); if (bufs == journal->j_wbufsize) { spin_unlock(&journal->j_list_lock); + trace_jbd_do_submit_data(journal, + commit_transaction); journal_do_submit_data(wbuf, bufs, write_op); bufs = 0; goto write_out_data; @@ -258,10 +263,6 @@ write_out_data: jbd_unlock_bh_state(bh); if (locked) unlock_buffer(bh); - journal_remove_journal_head(bh); - /* One for our safety reference, other for - * journal_remove_journal_head() */ - put_bh(bh); release_data_buffer(bh); } @@ -271,6 +272,7 @@ write_out_data: } } spin_unlock(&journal->j_list_lock); + trace_jbd_do_submit_data(journal, commit_transaction); journal_do_submit_data(wbuf, bufs, write_op); return err; @@ -321,12 +323,14 @@ void journal_commit_transaction(journal_t *journal) commit_transaction = journal->j_running_transaction; J_ASSERT(commit_transaction->t_state == T_RUNNING); + trace_jbd_start_commit(journal, commit_transaction); jbd_debug(1, "JBD: starting commit of transaction %d\n", commit_transaction->t_tid); spin_lock(&journal->j_state_lock); commit_transaction->t_state = T_LOCKED; + trace_jbd_commit_locking(journal, commit_transaction); spin_lock(&commit_transaction->t_handle_lock); while (commit_transaction->t_updates) { DEFINE_WAIT(wait); @@ -397,6 +401,7 @@ void journal_commit_transaction(journal_t *journal) */ journal_switch_revoke_table(journal); + trace_jbd_commit_flushing(journal, commit_transaction); commit_transaction->t_state = T_FLUSH; journal->j_committing_transaction = commit_transaction; journal->j_running_transaction = NULL; @@ -451,14 +456,9 @@ void journal_commit_transaction(journal_t *journal) } if (buffer_jbd(bh) && bh2jh(bh) == jh && jh->b_transaction == commit_transaction && - jh->b_jlist == BJ_Locked) { + jh->b_jlist == BJ_Locked) __journal_unfile_buffer(jh); - jbd_unlock_bh_state(bh); - journal_remove_journal_head(bh); - put_bh(bh); - } else { - jbd_unlock_bh_state(bh); - } + jbd_unlock_bh_state(bh); release_data_buffer(bh); cond_resched_lock(&journal->j_list_lock); } @@ -498,6 +498,7 @@ void journal_commit_transaction(journal_t *journal) commit_transaction->t_state = T_COMMIT; spin_unlock(&journal->j_state_lock); + trace_jbd_commit_logging(journal, commit_transaction); J_ASSERT(commit_transaction->t_nr_buffers <= commit_transaction->t_outstanding_credits); @@ -802,10 +803,16 @@ restart_loop: while (commit_transaction->t_forget) { transaction_t *cp_transaction; struct buffer_head *bh; + int try_to_free = 0; jh = commit_transaction->t_forget; spin_unlock(&journal->j_list_lock); bh = jh2bh(jh); + /* + * Get a reference so that bh cannot be freed before we are + * done with it. + */ + get_bh(bh); jbd_lock_bh_state(bh); J_ASSERT_JH(jh, jh->b_transaction == commit_transaction || jh->b_transaction == journal->j_running_transaction); @@ -881,28 +888,27 @@ restart_loop: __journal_insert_checkpoint(jh, commit_transaction); if (is_journal_aborted(journal)) clear_buffer_jbddirty(bh); - JBUFFER_TRACE(jh, "refile for checkpoint writeback"); - __journal_refile_buffer(jh); - jbd_unlock_bh_state(bh); } else { J_ASSERT_BH(bh, !buffer_dirty(bh)); - /* The buffer on BJ_Forget list and not jbddirty means + /* + * The buffer on BJ_Forget list and not jbddirty means * it has been freed by this transaction and hence it * could not have been reallocated until this * transaction has committed. *BUT* it could be * reallocated once we have written all the data to * disk and before we process the buffer on BJ_Forget - * list. */ - JBUFFER_TRACE(jh, "refile or unfile freed buffer"); - __journal_refile_buffer(jh); - if (!jh->b_transaction) { - jbd_unlock_bh_state(bh); - /* needs a brelse */ - journal_remove_journal_head(bh); - release_buffer_page(bh); - } else - jbd_unlock_bh_state(bh); + * list. + */ + if (!jh->b_next_transaction) + try_to_free = 1; } + JBUFFER_TRACE(jh, "refile or unfile freed buffer"); + __journal_refile_buffer(jh); + jbd_unlock_bh_state(bh); + if (try_to_free) + release_buffer_page(bh); + else + __brelse(bh); cond_resched_lock(&journal->j_list_lock); } spin_unlock(&journal->j_list_lock); @@ -969,6 +975,7 @@ restart_loop: } spin_unlock(&journal->j_list_lock); + trace_jbd_end_commit(journal, commit_transaction); jbd_debug(1, "JBD: commit %d complete, head %d\n", journal->j_commit_sequence, journal->j_tail_sequence); diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c index 9f36384..fea8dd6 100644 --- a/fs/jbd/journal.c +++ b/fs/jbd/journal.c @@ -38,6 +38,9 @@ #include <linux/debugfs.h> #include <linux/ratelimit.h> +#define CREATE_TRACE_POINTS +#include <trace/events/jbd.h> + #include <asm/uaccess.h> #include <asm/page.h> @@ -1065,6 +1068,7 @@ void journal_update_superblock(journal_t *journal, int wait) } else write_dirty_buffer(bh, WRITE); + trace_jbd_update_superblock_end(journal, wait); out: /* If we have just flushed the log (by marking s_start==0), then * any future commit will have to be careful to update the @@ -1807,10 +1811,9 @@ static void journal_free_journal_head(struct journal_head *jh) * When a buffer has its BH_JBD bit set it is immune from being released by * core kernel code, mainly via ->b_count. * - * A journal_head may be detached from its buffer_head when the journal_head's - * b_transaction, b_cp_transaction and b_next_transaction pointers are NULL. - * Various places in JBD call journal_remove_journal_head() to indicate that the - * journal_head can be dropped if needed. + * A journal_head is detached from its buffer_head when the journal_head's + * b_jcount reaches zero. Running transaction (b_transaction) and checkpoint + * transaction (b_cp_transaction) hold their references to b_jcount. * * Various places in the kernel want to attach a journal_head to a buffer_head * _before_ attaching the journal_head to a transaction. To protect the @@ -1823,17 +1826,16 @@ static void journal_free_journal_head(struct journal_head *jh) * (Attach a journal_head if needed. Increments b_jcount) * struct journal_head *jh = journal_add_journal_head(bh); * ... - * jh->b_transaction = xxx; - * journal_put_journal_head(jh); - * - * Now, the journal_head's b_jcount is zero, but it is safe from being released - * because it has a non-zero b_transaction. + * (Get another reference for transaction) + * journal_grab_journal_head(bh); + * jh->b_transaction = xxx; + * (Put original reference) + * journal_put_journal_head(jh); */ /* * Give a buffer_head a journal_head. * - * Doesn't need the journal lock. * May sleep. */ struct journal_head *journal_add_journal_head(struct buffer_head *bh) @@ -1897,61 +1899,29 @@ static void __journal_remove_journal_head(struct buffer_head *bh) struct journal_head *jh = bh2jh(bh); J_ASSERT_JH(jh, jh->b_jcount >= 0); - - get_bh(bh); - if (jh->b_jcount == 0) { - if (jh->b_transaction == NULL && - jh->b_next_transaction == NULL && - jh->b_cp_transaction == NULL) { - J_ASSERT_JH(jh, jh->b_jlist == BJ_None); - J_ASSERT_BH(bh, buffer_jbd(bh)); - J_ASSERT_BH(bh, jh2bh(jh) == bh); - BUFFER_TRACE(bh, "remove journal_head"); - if (jh->b_frozen_data) { - printk(KERN_WARNING "%s: freeing " - "b_frozen_data\n", - __func__); - jbd_free(jh->b_frozen_data, bh->b_size); - } - if (jh->b_committed_data) { - printk(KERN_WARNING "%s: freeing " - "b_committed_data\n", - __func__); - jbd_free(jh->b_committed_data, bh->b_size); - } - bh->b_private = NULL; - jh->b_bh = NULL; /* debug, really */ - clear_buffer_jbd(bh); - __brelse(bh); - journal_free_journal_head(jh); - } else { - BUFFER_TRACE(bh, "journal_head was locked"); - } + J_ASSERT_JH(jh, jh->b_transaction == NULL); + J_ASSERT_JH(jh, jh->b_next_transaction == NULL); + J_ASSERT_JH(jh, jh->b_cp_transaction == NULL); + J_ASSERT_JH(jh, jh->b_jlist == BJ_None); + J_ASSERT_BH(bh, buffer_jbd(bh)); + J_ASSERT_BH(bh, jh2bh(jh) == bh); + BUFFER_TRACE(bh, "remove journal_head"); + if (jh->b_frozen_data) { + printk(KERN_WARNING "%s: freeing b_frozen_data\n", __func__); + jbd_free(jh->b_frozen_data, bh->b_size); } + if (jh->b_committed_data) { + printk(KERN_WARNING "%s: freeing b_committed_data\n", __func__); + jbd_free(jh->b_committed_data, bh->b_size); + } + bh->b_private = NULL; + jh->b_bh = NULL; /* debug, really */ + clear_buffer_jbd(bh); + journal_free_journal_head(jh); } /* - * journal_remove_journal_head(): if the buffer isn't attached to a transaction - * and has a zero b_jcount then remove and release its journal_head. If we did - * see that the buffer is not used by any transaction we also "logically" - * decrement ->b_count. - * - * We in fact take an additional increment on ->b_count as a convenience, - * because the caller usually wants to do additional things with the bh - * after calling here. - * The caller of journal_remove_journal_head() *must* run __brelse(bh) at some - * time. Once the caller has run __brelse(), the buffer is eligible for - * reaping by try_to_free_buffers(). - */ -void journal_remove_journal_head(struct buffer_head *bh) -{ - jbd_lock_bh_journal_head(bh); - __journal_remove_journal_head(bh); - jbd_unlock_bh_journal_head(bh); -} - -/* - * Drop a reference on the passed journal_head. If it fell to zero then try to + * Drop a reference on the passed journal_head. If it fell to zero then * release the journal_head from the buffer_head. */ void journal_put_journal_head(struct journal_head *jh) @@ -1961,11 +1931,12 @@ void journal_put_journal_head(struct journal_head *jh) jbd_lock_bh_journal_head(bh); J_ASSERT_JH(jh, jh->b_jcount > 0); --jh->b_jcount; - if (!jh->b_jcount && !jh->b_transaction) { + if (!jh->b_jcount) { __journal_remove_journal_head(bh); + jbd_unlock_bh_journal_head(bh); __brelse(bh); - } - jbd_unlock_bh_journal_head(bh); + } else + jbd_unlock_bh_journal_head(bh); } /* diff --git a/fs/jbd/recovery.c b/fs/jbd/recovery.c index 5b43e96..008bf06 100644 --- a/fs/jbd/recovery.c +++ b/fs/jbd/recovery.c @@ -20,6 +20,7 @@ #include <linux/fs.h> #include <linux/jbd.h> #include <linux/errno.h> +#include <linux/blkdev.h> #endif /* @@ -263,6 +264,9 @@ int journal_recover(journal_t *journal) err2 = sync_blockdev(journal->j_fs_dev); if (!err) err = err2; + /* Flush disk caches to get replayed data on the permanent storage */ + if (journal->j_flags & JFS_BARRIER) + blkdev_issue_flush(journal->j_fs_dev, GFP_KERNEL, NULL); return err; } diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c index d7ab092..7c86b37 100644 --- a/fs/jbd/transaction.c +++ b/fs/jbd/transaction.c @@ -26,6 +26,7 @@ #include <linux/mm.h> #include <linux/highmem.h> #include <linux/hrtimer.h> +#include <linux/backing-dev.h> static void __journal_temp_unlink_buffer(struct journal_head *jh); @@ -99,11 +100,10 @@ static int start_this_handle(journal_t *journal, handle_t *handle) alloc_transaction: if (!journal->j_running_transaction) { - new_transaction = kzalloc(sizeof(*new_transaction), - GFP_NOFS|__GFP_NOFAIL); + new_transaction = kzalloc(sizeof(*new_transaction), GFP_NOFS); if (!new_transaction) { - ret = -ENOMEM; - goto out; + congestion_wait(BLK_RW_ASYNC, HZ/50); + goto alloc_transaction; } } @@ -696,7 +696,6 @@ repeat: if (!jh->b_transaction) { JBUFFER_TRACE(jh, "no transaction"); J_ASSERT_JH(jh, !jh->b_next_transaction); - jh->b_transaction = transaction; JBUFFER_TRACE(jh, "file as BJ_Reserved"); spin_lock(&journal->j_list_lock); __journal_file_buffer(jh, transaction, BJ_Reserved); @@ -818,7 +817,6 @@ int journal_get_create_access(handle_t *handle, struct buffer_head *bh) * committed and so it's safe to clear the dirty bit. */ clear_buffer_dirty(jh2bh(jh)); - jh->b_transaction = transaction; /* first access by this transaction */ jh->b_modified = 0; @@ -844,8 +842,8 @@ int journal_get_create_access(handle_t *handle, struct buffer_head *bh) */ JBUFFER_TRACE(jh, "cancelling revoke"); journal_cancel_revoke(handle, jh); - journal_put_journal_head(jh); out: + journal_put_journal_head(jh); return err; } @@ -1069,8 +1067,9 @@ int journal_dirty_data(handle_t *handle, struct buffer_head *bh) ret = -EIO; goto no_journal; } - - if (jh->b_transaction != NULL) { + /* We might have slept so buffer could be refiled now */ + if (jh->b_transaction != NULL && + jh->b_transaction != handle->h_transaction) { JBUFFER_TRACE(jh, "unfile from commit"); __journal_temp_unlink_buffer(jh); /* It still points to the committing @@ -1091,8 +1090,6 @@ int journal_dirty_data(handle_t *handle, struct buffer_head *bh) if (jh->b_jlist != BJ_SyncData && jh->b_jlist != BJ_Locked) { JBUFFER_TRACE(jh, "not on correct data list: unfile"); J_ASSERT_JH(jh, jh->b_jlist != BJ_Shadow); - __journal_temp_unlink_buffer(jh); - jh->b_transaction = handle->h_transaction; JBUFFER_TRACE(jh, "file as data"); __journal_file_buffer(jh, handle->h_transaction, BJ_SyncData); @@ -1300,8 +1297,6 @@ int journal_forget (handle_t *handle, struct buffer_head *bh) __journal_file_buffer(jh, transaction, BJ_Forget); } else { __journal_unfile_buffer(jh); - journal_remove_journal_head(bh); - __brelse(bh); if (!buffer_jbd(bh)) { spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); @@ -1622,19 +1617,32 @@ static void __journal_temp_unlink_buffer(struct journal_head *jh) mark_buffer_dirty(bh); /* Expose it to the VM */ } +/* + * Remove buffer from all transactions. + * + * Called with bh_state lock and j_list_lock + * + * jh and bh may be already freed when this function returns. + */ void __journal_unfile_buffer(struct journal_head *jh) { __journal_temp_unlink_buffer(jh); jh->b_transaction = NULL; + journal_put_journal_head(jh); } void journal_unfile_buffer(journal_t *journal, struct journal_head *jh) { - jbd_lock_bh_state(jh2bh(jh)); + struct buffer_head *bh = jh2bh(jh); + + /* Get reference so that buffer cannot be freed before we unlock it */ + get_bh(bh); + jbd_lock_bh_state(bh); spin_lock(&journal->j_list_lock); __journal_unfile_buffer(jh); spin_unlock(&journal->j_list_lock); - jbd_unlock_bh_state(jh2bh(jh)); + jbd_unlock_bh_state(bh); + __brelse(bh); } /* @@ -1661,16 +1669,12 @@ __journal_try_to_free_buffer(journal_t *journal, struct buffer_head *bh) /* A written-back ordered data buffer */ JBUFFER_TRACE(jh, "release data"); __journal_unfile_buffer(jh); - journal_remove_journal_head(bh); - __brelse(bh); } } else if (jh->b_cp_transaction != NULL && jh->b_transaction == NULL) { /* written-back checkpointed metadata buffer */ if (jh->b_jlist == BJ_None) { JBUFFER_TRACE(jh, "remove from checkpoint list"); __journal_remove_checkpoint(jh); - journal_remove_journal_head(bh); - __brelse(bh); } } spin_unlock(&journal->j_list_lock); @@ -1733,7 +1737,7 @@ int journal_try_to_free_buffers(journal_t *journal, /* * We take our own ref against the journal_head here to avoid * having to add tons of locking around each instance of - * journal_remove_journal_head() and journal_put_journal_head(). + * journal_put_journal_head(). */ jh = journal_grab_journal_head(bh); if (!jh) @@ -1770,10 +1774,9 @@ static int __dispose_buffer(struct journal_head *jh, transaction_t *transaction) int may_free = 1; struct buffer_head *bh = jh2bh(jh); - __journal_unfile_buffer(jh); - if (jh->b_cp_transaction) { JBUFFER_TRACE(jh, "on running+cp transaction"); + __journal_temp_unlink_buffer(jh); /* * We don't want to write the buffer anymore, clear the * bit so that we don't confuse checks in @@ -1784,8 +1787,7 @@ static int __dispose_buffer(struct journal_head *jh, transaction_t *transaction) may_free = 0; } else { JBUFFER_TRACE(jh, "on running transaction"); - journal_remove_journal_head(bh); - __brelse(bh); + __journal_unfile_buffer(jh); } return may_free; } @@ -2096,6 +2098,8 @@ void __journal_file_buffer(struct journal_head *jh, if (jh->b_transaction) __journal_temp_unlink_buffer(jh); + else + journal_grab_journal_head(bh); jh->b_transaction = transaction; switch (jlist) { @@ -2153,9 +2157,10 @@ void journal_file_buffer(struct journal_head *jh, * already started to be used by a subsequent transaction, refile the * buffer on that transaction's metadata list. * - * Called under journal->j_list_lock - * + * Called under j_list_lock * Called under jbd_lock_bh_state(jh2bh(jh)) + * + * jh and bh may be already free when this function returns */ void __journal_refile_buffer(struct journal_head *jh) { @@ -2179,6 +2184,11 @@ void __journal_refile_buffer(struct journal_head *jh) was_dirty = test_clear_buffer_jbddirty(bh); __journal_temp_unlink_buffer(jh); + /* + * We set b_transaction here because b_next_transaction will inherit + * our jh reference and thus __journal_file_buffer() must not take a + * new one. + */ jh->b_transaction = jh->b_next_transaction; jh->b_next_transaction = NULL; if (buffer_freed(bh)) @@ -2195,30 +2205,21 @@ void __journal_refile_buffer(struct journal_head *jh) } /* - * For the unlocked version of this call, also make sure that any - * hanging journal_head is cleaned up if necessary. - * - * __journal_refile_buffer is usually called as part of a single locked - * operation on a buffer_head, in which the caller is probably going to - * be hooking the journal_head onto other lists. In that case it is up - * to the caller to remove the journal_head if necessary. For the - * unlocked journal_refile_buffer call, the caller isn't going to be - * doing anything else to the buffer so we need to do the cleanup - * ourselves to avoid a jh leak. - * - * *** The journal_head may be freed by this call! *** + * __journal_refile_buffer() with necessary locking added. We take our bh + * reference so that we can safely unlock bh. + * + * The jh and bh may be freed by this call. */ void journal_refile_buffer(journal_t *journal, struct journal_head *jh) { struct buffer_head *bh = jh2bh(jh); + /* Get reference so that buffer cannot be freed before we unlock it */ + get_bh(bh); jbd_lock_bh_state(bh); spin_lock(&journal->j_list_lock); - __journal_refile_buffer(jh); jbd_unlock_bh_state(bh); - journal_remove_journal_head(bh); - spin_unlock(&journal->j_list_lock); __brelse(bh); } diff --git a/fs/jfs/acl.c b/fs/jfs/acl.c index e5de942..45559dc 100644 --- a/fs/jfs/acl.c +++ b/fs/jfs/acl.c @@ -27,7 +27,7 @@ #include "jfs_xattr.h" #include "jfs_acl.h" -static struct posix_acl *jfs_get_acl(struct inode *inode, int type) +struct posix_acl *jfs_get_acl(struct inode *inode, int type) { struct posix_acl *acl; char *ea_name; @@ -114,30 +114,9 @@ out: return rc; } -int jfs_check_acl(struct inode *inode, int mask, unsigned int flags) -{ - struct posix_acl *acl; - - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - - acl = jfs_get_acl(inode, ACL_TYPE_ACCESS); - if (IS_ERR(acl)) - return PTR_ERR(acl); - if (acl) { - int error = posix_acl_permission(inode, acl, mask); - posix_acl_release(acl); - return error; - } - - return -EAGAIN; -} - int jfs_init_acl(tid_t tid, struct inode *inode, struct inode *dir) { struct posix_acl *acl = NULL; - struct posix_acl *clone; - mode_t mode; int rc = 0; if (S_ISLNK(inode->i_mode)) @@ -153,20 +132,11 @@ int jfs_init_acl(tid_t tid, struct inode *inode, struct inode *dir) if (rc) goto cleanup; } - clone = posix_acl_clone(acl, GFP_KERNEL); - if (!clone) { - rc = -ENOMEM; - goto cleanup; - } - mode = inode->i_mode; - rc = posix_acl_create_masq(clone, &mode); - if (rc >= 0) { - inode->i_mode = mode; - if (rc > 0) - rc = jfs_set_acl(tid, inode, ACL_TYPE_ACCESS, - clone); - } - posix_acl_release(clone); + rc = posix_acl_create(&acl, GFP_KERNEL, &inode->i_mode); + if (rc < 0) + goto cleanup; /* posix_acl_release(NULL) is no-op */ + if (rc > 0) + rc = jfs_set_acl(tid, inode, ACL_TYPE_ACCESS, acl); cleanup: posix_acl_release(acl); } else @@ -180,8 +150,9 @@ cleanup: int jfs_acl_chmod(struct inode *inode) { - struct posix_acl *acl, *clone; + struct posix_acl *acl; int rc; + tid_t tid; if (S_ISLNK(inode->i_mode)) return -EOPNOTSUPP; @@ -190,22 +161,18 @@ int jfs_acl_chmod(struct inode *inode) if (IS_ERR(acl) || !acl) return PTR_ERR(acl); - clone = posix_acl_clone(acl, GFP_KERNEL); - posix_acl_release(acl); - if (!clone) - return -ENOMEM; - - rc = posix_acl_chmod_masq(clone, inode->i_mode); - if (!rc) { - tid_t tid = txBegin(inode->i_sb, 0); - mutex_lock(&JFS_IP(inode)->commit_mutex); - rc = jfs_set_acl(tid, inode, ACL_TYPE_ACCESS, clone); - if (!rc) - rc = txCommit(tid, 1, &inode, 0); - txEnd(tid); - mutex_unlock(&JFS_IP(inode)->commit_mutex); - } + rc = posix_acl_chmod(&acl, GFP_KERNEL, inode->i_mode); + if (rc) + return rc; - posix_acl_release(clone); + tid = txBegin(inode->i_sb, 0); + mutex_lock(&JFS_IP(inode)->commit_mutex); + rc = jfs_set_acl(tid, inode, ACL_TYPE_ACCESS, acl); + if (!rc) + rc = txCommit(tid, 1, &inode, 0); + txEnd(tid); + mutex_unlock(&JFS_IP(inode)->commit_mutex); + + posix_acl_release(acl); return rc; } diff --git a/fs/jfs/file.c b/fs/jfs/file.c index 2f3f531..844f946 100644 --- a/fs/jfs/file.c +++ b/fs/jfs/file.c @@ -28,19 +28,26 @@ #include "jfs_acl.h" #include "jfs_debug.h" -int jfs_fsync(struct file *file, int datasync) +int jfs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; int rc = 0; + rc = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (rc) + return rc; + + mutex_lock(&inode->i_mutex); if (!(inode->i_state & I_DIRTY) || (datasync && !(inode->i_state & I_DIRTY_DATASYNC))) { /* Make sure committed changes hit the disk */ jfs_flush_journal(JFS_SBI(inode->i_sb)->log, 1); + mutex_unlock(&inode->i_mutex); return rc; } rc |= jfs_commit_inode(inode, 1); + mutex_unlock(&inode->i_mutex); return rc ? -EIO : 0; } @@ -110,6 +117,8 @@ int jfs_setattr(struct dentry *dentry, struct iattr *iattr) if ((iattr->ia_valid & ATTR_SIZE) && iattr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + rc = vmtruncate(inode, iattr->ia_size); if (rc) return rc; @@ -131,7 +140,7 @@ const struct inode_operations jfs_file_inode_operations = { .removexattr = jfs_removexattr, .setattr = jfs_setattr, #ifdef CONFIG_JFS_POSIX_ACL - .check_acl = jfs_check_acl, + .get_acl = jfs_get_acl, #endif }; diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c index 09100b4..13fc885 100644 --- a/fs/jfs/inode.c +++ b/fs/jfs/inode.c @@ -329,8 +329,8 @@ static ssize_t jfs_direct_IO(int rw, struct kiocb *iocb, struct inode *inode = file->f_mapping->host; ssize_t ret; - ret = blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov, - offset, nr_segs, jfs_get_block, NULL); + ret = blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs, + jfs_get_block); /* * In case of error extending write may have instantiated a few diff --git a/fs/jfs/jfs_acl.h b/fs/jfs/jfs_acl.h index f9285c4..ad84fe5 100644 --- a/fs/jfs/jfs_acl.h +++ b/fs/jfs/jfs_acl.h @@ -20,7 +20,7 @@ #ifdef CONFIG_JFS_POSIX_ACL -int jfs_check_acl(struct inode *, int, unsigned int flags); +struct posix_acl *jfs_get_acl(struct inode *inode, int type); int jfs_init_acl(tid_t, struct inode *, struct inode *); int jfs_acl_chmod(struct inode *inode); diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c index 4496872..9cbd11a 100644 --- a/fs/jfs/jfs_dmap.c +++ b/fs/jfs/jfs_dmap.c @@ -3161,7 +3161,7 @@ static int dbAllocDmapBU(struct bmap * bmp, struct dmap * dp, s64 blkno, { int rc; int dbitno, word, rembits, nb, nwords, wbitno, agno; - s8 oldroot, *leaf; + s8 oldroot; struct dmaptree *tp = (struct dmaptree *) & dp->tree; /* save the current value of the root (i.e. maximum free string) @@ -3169,9 +3169,6 @@ static int dbAllocDmapBU(struct bmap * bmp, struct dmap * dp, s64 blkno, */ oldroot = tp->stree[ROOT]; - /* pick up a pointer to the leaves of the dmap tree */ - leaf = tp->stree + LEAFIND; - /* determine the bit number and word within the dmap of the * starting block. */ diff --git a/fs/jfs/jfs_dtree.c b/fs/jfs/jfs_dtree.c index b6f17c0..f6f32fa 100644 --- a/fs/jfs/jfs_dtree.c +++ b/fs/jfs/jfs_dtree.c @@ -3048,9 +3048,9 @@ int jfs_readdir(struct file *filp, void *dirent, filldir_t filldir) dir_index = (u32) filp->f_pos; /* - * NFSv4 reserves cookies 1 and 2 for . and .. so we add - * the value we return to the vfs is one greater than the - * one we use internally. + * NFSv4 reserves cookies 1 and 2 for . and .. so the value + * we return to the vfs is one greater than the one we use + * internally. */ if (dir_index) dir_index--; @@ -3103,7 +3103,7 @@ int jfs_readdir(struct file *filp, void *dirent, filldir_t filldir) * self "." */ filp->f_pos = 1; - if (filldir(dirent, ".", 1, 0, ip->i_ino, + if (filldir(dirent, ".", 1, 1, ip->i_ino, DT_DIR)) return 0; } @@ -3111,7 +3111,7 @@ int jfs_readdir(struct file *filp, void *dirent, filldir_t filldir) * parent ".." */ filp->f_pos = 2; - if (filldir(dirent, "..", 2, 1, PARENT(ip), DT_DIR)) + if (filldir(dirent, "..", 2, 2, PARENT(ip), DT_DIR)) return 0; /* diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c index b78b2f9..1b6f15f 100644 --- a/fs/jfs/jfs_imap.c +++ b/fs/jfs/jfs_imap.c @@ -457,7 +457,7 @@ struct inode *diReadSpecial(struct super_block *sb, ino_t inum, int secondary) /* read the page of fixed disk inode (AIT) in raw mode */ mp = read_metapage(ip, address << sbi->l2nbperpage, PSIZE, 1); if (mp == NULL) { - ip->i_nlink = 1; /* Don't want iput() deleting it */ + set_nlink(ip, 1); /* Don't want iput() deleting it */ iput(ip); return (NULL); } @@ -469,7 +469,7 @@ struct inode *diReadSpecial(struct super_block *sb, ino_t inum, int secondary) /* copy on-disk inode to in-memory inode */ if ((copy_from_dinode(dp, ip)) != 0) { /* handle bad return by returning NULL for ip */ - ip->i_nlink = 1; /* Don't want iput() deleting it */ + set_nlink(ip, 1); /* Don't want iput() deleting it */ iput(ip); /* release the page */ release_metapage(mp); @@ -3076,7 +3076,7 @@ static int copy_from_dinode(struct dinode * dip, struct inode *ip) ip->i_mode |= 0001; } } - ip->i_nlink = le32_to_cpu(dip->di_nlink); + set_nlink(ip, le32_to_cpu(dip->di_nlink)); jfs_ip->saved_uid = le32_to_cpu(dip->di_uid); if (sbi->uid == -1) diff --git a/fs/jfs/jfs_inode.c b/fs/jfs/jfs_inode.c index 2686531..7f464c5 100644 --- a/fs/jfs/jfs_inode.c +++ b/fs/jfs/jfs_inode.c @@ -95,7 +95,7 @@ struct inode *ialloc(struct inode *parent, umode_t mode) if (insert_inode_locked(inode) < 0) { rc = -EINVAL; - goto fail_unlock; + goto fail_put; } inode_init_owner(inode, parent, mode); @@ -156,8 +156,7 @@ struct inode *ialloc(struct inode *parent, umode_t mode) fail_drop: dquot_drop(inode); inode->i_flags |= S_NOQUOTA; -fail_unlock: - inode->i_nlink = 0; + clear_nlink(inode); unlock_new_inode(inode); fail_put: iput(inode); diff --git a/fs/jfs/jfs_inode.h b/fs/jfs/jfs_inode.h index ec2fb8b..9271cfe 100644 --- a/fs/jfs/jfs_inode.h +++ b/fs/jfs/jfs_inode.h @@ -21,7 +21,7 @@ struct fid; extern struct inode *ialloc(struct inode *, umode_t); -extern int jfs_fsync(struct file *, int); +extern int jfs_fsync(struct file *, loff_t, loff_t, int); extern long jfs_ioctl(struct file *, unsigned int, unsigned long); extern long jfs_compat_ioctl(struct file *, unsigned int, unsigned long); extern struct inode *jfs_iget(struct super_block *, unsigned long); diff --git a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c index ee55e45..bfb2a91 100644 --- a/fs/jfs/jfs_logmgr.c +++ b/fs/jfs/jfs_logmgr.c @@ -67,6 +67,7 @@ #include <linux/buffer_head.h> /* for sync_blockdev() */ #include <linux/bio.h> #include <linux/freezer.h> +#include <linux/export.h> #include <linux/delay.h> #include <linux/mutex.h> #include <linux/seq_file.h> diff --git a/fs/jfs/jfs_txnmgr.c b/fs/jfs/jfs_txnmgr.c index f6cc0c0..af96060 100644 --- a/fs/jfs/jfs_txnmgr.c +++ b/fs/jfs/jfs_txnmgr.c @@ -1143,7 +1143,6 @@ int txCommit(tid_t tid, /* transaction identifier */ struct jfs_log *log; struct tblock *tblk; struct lrd *lrd; - int lsn; struct inode *ip; struct jfs_inode_info *jfs_ip; int k, n; @@ -1310,7 +1309,7 @@ int txCommit(tid_t tid, /* transaction identifier */ */ lrd->type = cpu_to_le16(LOG_COMMIT); lrd->length = 0; - lsn = lmLog(log, tblk, lrd, NULL); + lmLog(log, tblk, lrd, NULL); lmGroupCommit(log, tblk); @@ -2935,7 +2934,6 @@ int jfs_sync(void *arg) { struct inode *ip; struct jfs_inode_info *jfs_ip; - int rc; tid_t tid; do { @@ -2961,7 +2959,7 @@ int jfs_sync(void *arg) */ TXN_UNLOCK(); tid = txBegin(ip->i_sb, COMMIT_INODE); - rc = txCommit(tid, 1, &ip, 0); + txCommit(tid, 1, &ip, 0); txEnd(tid); mutex_unlock(&jfs_ip->commit_mutex); diff --git a/fs/jfs/jfs_umount.c b/fs/jfs/jfs_umount.c index adcf92d..7971f37 100644 --- a/fs/jfs/jfs_umount.c +++ b/fs/jfs/jfs_umount.c @@ -68,7 +68,7 @@ int jfs_umount(struct super_block *sb) /* * Wait for outstanding transactions to be written to log: */ - jfs_flush_journal(log, 1); + jfs_flush_journal(log, 2); /* * close fileset inode allocation map (aka fileset inode) @@ -146,7 +146,7 @@ int jfs_umount_rw(struct super_block *sb) * * remove file system from log active file system list. */ - jfs_flush_journal(log, 1); + jfs_flush_journal(log, 2); /* * Make sure all metadata makes it to disk diff --git a/fs/jfs/namei.c b/fs/jfs/namei.c index eaaf2b5..a112ad9 100644 --- a/fs/jfs/namei.c +++ b/fs/jfs/namei.c @@ -172,7 +172,7 @@ static int jfs_create(struct inode *dip, struct dentry *dentry, int mode, mutex_unlock(&JFS_IP(dip)->commit_mutex); if (rc) { free_ea_wmap(ip); - ip->i_nlink = 0; + clear_nlink(ip); unlock_new_inode(ip); iput(ip); } else { @@ -292,7 +292,7 @@ static int jfs_mkdir(struct inode *dip, struct dentry *dentry, int mode) goto out3; } - ip->i_nlink = 2; /* for '.' */ + set_nlink(ip, 2); /* for '.' */ ip->i_op = &jfs_dir_inode_operations; ip->i_fop = &jfs_dir_operations; @@ -311,7 +311,7 @@ static int jfs_mkdir(struct inode *dip, struct dentry *dentry, int mode) mutex_unlock(&JFS_IP(dip)->commit_mutex); if (rc) { free_ea_wmap(ip); - ip->i_nlink = 0; + clear_nlink(ip); unlock_new_inode(ip); iput(ip); } else { @@ -844,7 +844,7 @@ static int jfs_link(struct dentry *old_dentry, rc = txCommit(tid, 2, &iplist[0], 0); if (rc) { - ip->i_nlink--; /* never instantiated */ + drop_nlink(ip); /* never instantiated */ iput(ip); } else d_instantiate(dentry, ip); @@ -893,7 +893,7 @@ static int jfs_symlink(struct inode *dip, struct dentry *dentry, unchar *i_fastsymlink; s64 xlen = 0; int bmask = 0, xsize; - s64 extent = 0, xaddr; + s64 xaddr; struct metapage *mp; struct super_block *sb; struct tblock *tblk; @@ -993,7 +993,6 @@ static int jfs_symlink(struct inode *dip, struct dentry *dentry, txAbort(tid, 0); goto out3; } - extent = xaddr; ip->i_size = ssize - 1; while (ssize) { /* This is kind of silly since PATH_MAX == 4K */ @@ -1049,7 +1048,7 @@ static int jfs_symlink(struct inode *dip, struct dentry *dentry, mutex_unlock(&JFS_IP(dip)->commit_mutex); if (rc) { free_ea_wmap(ip); - ip->i_nlink = 0; + clear_nlink(ip); unlock_new_inode(ip); iput(ip); } else { @@ -1434,7 +1433,7 @@ static int jfs_mknod(struct inode *dir, struct dentry *dentry, mutex_unlock(&JFS_IP(dir)->commit_mutex); if (rc) { free_ea_wmap(ip); - ip->i_nlink = 0; + clear_nlink(ip); unlock_new_inode(ip); iput(ip); } else { @@ -1456,34 +1455,23 @@ static struct dentry *jfs_lookup(struct inode *dip, struct dentry *dentry, struc ino_t inum; struct inode *ip; struct component_name key; - const char *name = dentry->d_name.name; - int len = dentry->d_name.len; int rc; - jfs_info("jfs_lookup: name = %s", name); - - if ((name[0] == '.') && (len == 1)) - inum = dip->i_ino; - else if (strcmp(name, "..") == 0) - inum = PARENT(dip); - else { - if ((rc = get_UCSname(&key, dentry))) - return ERR_PTR(rc); - rc = dtSearch(dip, &key, &inum, &btstack, JFS_LOOKUP); - free_UCSname(&key); - if (rc == -ENOENT) { - d_add(dentry, NULL); - return NULL; - } else if (rc) { - jfs_err("jfs_lookup: dtSearch returned %d", rc); - return ERR_PTR(rc); - } - } - - ip = jfs_iget(dip->i_sb, inum); - if (IS_ERR(ip)) { - jfs_err("jfs_lookup: iget failed on inum %d", (uint) inum); - return ERR_CAST(ip); + jfs_info("jfs_lookup: name = %s", dentry->d_name.name); + + if ((rc = get_UCSname(&key, dentry))) + return ERR_PTR(rc); + rc = dtSearch(dip, &key, &inum, &btstack, JFS_LOOKUP); + free_UCSname(&key); + if (rc == -ENOENT) { + ip = NULL; + } else if (rc) { + jfs_err("jfs_lookup: dtSearch returned %d", rc); + ip = ERR_PTR(rc); + } else { + ip = jfs_iget(dip->i_sb, inum); + if (IS_ERR(ip)) + jfs_err("jfs_lookup: iget failed on inum %d", (uint)inum); } return d_splice_alias(ip, dentry); @@ -1548,7 +1536,7 @@ const struct inode_operations jfs_dir_inode_operations = { .removexattr = jfs_removexattr, .setattr = jfs_setattr, #ifdef CONFIG_JFS_POSIX_ACL - .check_acl = jfs_check_acl, + .get_acl = jfs_get_acl, #endif }; @@ -1597,8 +1585,6 @@ out: static int jfs_ci_revalidate(struct dentry *dentry, struct nameidata *nd) { - if (nd && nd->flags & LOOKUP_RCU) - return -ECHILD; /* * This is not negative dentry. Always valid. * @@ -1624,10 +1610,8 @@ static int jfs_ci_revalidate(struct dentry *dentry, struct nameidata *nd) * case sensitive name which is specified by user if this is * for creation. */ - if (!(nd->flags & (LOOKUP_CONTINUE | LOOKUP_PARENT))) { - if (nd->flags & (LOOKUP_CREATE | LOOKUP_RENAME_TARGET)) - return 0; - } + if (nd->flags & (LOOKUP_CREATE | LOOKUP_RENAME_TARGET)) + return 0; return 1; } diff --git a/fs/jfs/super.c b/fs/jfs/super.c index 06c8a67..a44eff0 100644 --- a/fs/jfs/super.c +++ b/fs/jfs/super.c @@ -485,7 +485,6 @@ static int jfs_fill_super(struct super_block *sb, void *data, int silent) goto out_unload; } inode->i_ino = 0; - inode->i_nlink = 1; inode->i_size = sb->s_bdev->bd_inode->i_size; inode->i_mapping->a_ops = &jfs_metapage_aops; insert_inode_hash(inode); diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c index 24838f1..26683e1 100644 --- a/fs/jfs/xattr.c +++ b/fs/jfs/xattr.c @@ -693,8 +693,7 @@ static int can_set_system_xattr(struct inode *inode, const char *name, return rc; } if (acl) { - mode_t mode = inode->i_mode; - rc = posix_acl_equiv_mode(acl, &mode); + rc = posix_acl_equiv_mode(acl, &inode->i_mode); posix_acl_release(acl); if (rc < 0) { printk(KERN_ERR @@ -702,7 +701,6 @@ static int can_set_system_xattr(struct inode *inode, const char *name, rc); return rc; } - inode->i_mode = mode; mark_inode_dirty(inode); } /* @@ -1091,38 +1089,37 @@ int jfs_removexattr(struct dentry *dentry, const char *name) } #ifdef CONFIG_JFS_SECURITY -int jfs_init_security(tid_t tid, struct inode *inode, struct inode *dir, - const struct qstr *qstr) +int jfs_initxattrs(struct inode *inode, const struct xattr *xattr_array, + void *fs_info) { - int rc; - size_t len; - void *value; - char *suffix; + const struct xattr *xattr; + tid_t *tid = fs_info; char *name; - - rc = security_inode_init_security(inode, dir, qstr, &suffix, &value, - &len); - if (rc) { - if (rc == -EOPNOTSUPP) - return 0; - return rc; - } - name = kmalloc(XATTR_SECURITY_PREFIX_LEN + 1 + strlen(suffix), - GFP_NOFS); - if (!name) { - rc = -ENOMEM; - goto kmalloc_failed; + int err = 0; + + for (xattr = xattr_array; xattr->name != NULL; xattr++) { + name = kmalloc(XATTR_SECURITY_PREFIX_LEN + + strlen(xattr->name) + 1, GFP_NOFS); + if (!name) { + err = -ENOMEM; + break; + } + strcpy(name, XATTR_SECURITY_PREFIX); + strcpy(name + XATTR_SECURITY_PREFIX_LEN, xattr->name); + + err = __jfs_setxattr(*tid, inode, name, + xattr->value, xattr->value_len, 0); + kfree(name); + if (err < 0) + break; } - strcpy(name, XATTR_SECURITY_PREFIX); - strcpy(name + XATTR_SECURITY_PREFIX_LEN, suffix); - - rc = __jfs_setxattr(tid, inode, name, value, len, 0); - - kfree(name); -kmalloc_failed: - kfree(suffix); - kfree(value); + return err; +} - return rc; +int jfs_init_security(tid_t tid, struct inode *inode, struct inode *dir, + const struct qstr *qstr) +{ + return security_inode_init_security(inode, dir, qstr, + &jfs_initxattrs, &tid); } #endif diff --git a/fs/nfsd/Kconfig b/fs/nfsd/Kconfig index fbb2a5e..10e6366 100644 --- a/fs/nfsd/Kconfig +++ b/fs/nfsd/Kconfig @@ -28,18 +28,6 @@ config NFSD If unsure, say N. -config NFSD_DEPRECATED - bool "Include support for deprecated syscall interface to NFSD" - depends on NFSD - default y - help - The syscall interface to nfsd was obsoleted in 2.6.0 by a new - filesystem based interface. The old interface is due for removal - in 2.6.40. If you wish to remove the interface before then - say N. - - In unsure, say Y. - config NFSD_V2_ACL bool depends on NFSD diff --git a/fs/nfsd/cache.h b/fs/nfsd/cache.h index d892be6..93cc9d3 100644 --- a/fs/nfsd/cache.h +++ b/fs/nfsd/cache.h @@ -69,7 +69,7 @@ enum { int nfsd_reply_cache_init(void); void nfsd_reply_cache_shutdown(void); -int nfsd_cache_lookup(struct svc_rqst *, int); +int nfsd_cache_lookup(struct svc_rqst *); void nfsd_cache_update(struct svc_rqst *, int, __be32 *); #ifdef CONFIG_NFSD_V4 diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index d7c4f02..a0205fc 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -16,7 +16,6 @@ #include <linux/module.h> #include <linux/exportfs.h> -#include <linux/nfsd/syscall.h> #include <net/ipv6.h> #include "nfsd.h" @@ -318,7 +317,6 @@ static void svc_export_put(struct kref *ref) struct svc_export *exp = container_of(ref, struct svc_export, h.ref); path_put(&exp->ex_path); auth_domain_put(exp->ex_client); - kfree(exp->ex_pathname); nfsd4_fslocs_free(&exp->ex_fslocs); kfree(exp); } @@ -528,11 +526,6 @@ static int svc_export_parse(struct cache_detail *cd, char *mesg, int mlen) exp.ex_client = dom; - err = -ENOMEM; - exp.ex_pathname = kstrdup(buf, GFP_KERNEL); - if (!exp.ex_pathname) - goto out2; - /* expiry */ err = -EINVAL; exp.h.expiry_time = get_expiry(&mesg); @@ -613,8 +606,6 @@ out4: nfsd4_fslocs_free(&exp.ex_fslocs); kfree(exp.ex_uuid); out3: - kfree(exp.ex_pathname); -out2: path_put(&exp.ex_path); out1: auth_domain_put(dom); @@ -678,7 +669,6 @@ static void svc_export_init(struct cache_head *cnew, struct cache_head *citem) new->ex_client = item->ex_client; new->ex_path.dentry = dget(item->ex_path.dentry); new->ex_path.mnt = mntget(item->ex_path.mnt); - new->ex_pathname = NULL; new->ex_fslocs.locations = NULL; new->ex_fslocs.locations_count = 0; new->ex_fslocs.migrated = 0; @@ -696,8 +686,6 @@ static void export_update(struct cache_head *cnew, struct cache_head *citem) new->ex_fsid = item->ex_fsid; new->ex_uuid = item->ex_uuid; item->ex_uuid = NULL; - new->ex_pathname = item->ex_pathname; - item->ex_pathname = NULL; new->ex_fslocs.locations = item->ex_fslocs.locations; item->ex_fslocs.locations = NULL; new->ex_fslocs.locations_count = item->ex_fslocs.locations_count; @@ -797,58 +785,6 @@ exp_find_key(svc_client *clp, int fsid_type, u32 *fsidv, struct cache_req *reqp) return ek; } -#ifdef CONFIG_NFSD_DEPRECATED -static int exp_set_key(svc_client *clp, int fsid_type, u32 *fsidv, - struct svc_export *exp) -{ - struct svc_expkey key, *ek; - - key.ek_client = clp; - key.ek_fsidtype = fsid_type; - memcpy(key.ek_fsid, fsidv, key_len(fsid_type)); - key.ek_path = exp->ex_path; - key.h.expiry_time = NEVER; - key.h.flags = 0; - - ek = svc_expkey_lookup(&key); - if (ek) - ek = svc_expkey_update(&key,ek); - if (ek) { - cache_put(&ek->h, &svc_expkey_cache); - return 0; - } - return -ENOMEM; -} - -/* - * Find the client's export entry matching xdev/xino. - */ -static inline struct svc_expkey * -exp_get_key(svc_client *clp, dev_t dev, ino_t ino) -{ - u32 fsidv[3]; - - if (old_valid_dev(dev)) { - mk_fsid(FSID_DEV, fsidv, dev, ino, 0, NULL); - return exp_find_key(clp, FSID_DEV, fsidv, NULL); - } - mk_fsid(FSID_ENCODE_DEV, fsidv, dev, ino, 0, NULL); - return exp_find_key(clp, FSID_ENCODE_DEV, fsidv, NULL); -} - -/* - * Find the client's export entry matching fsid - */ -static inline struct svc_expkey * -exp_get_fsid_key(svc_client *clp, int fsid) -{ - u32 fsidv[2]; - - mk_fsid(FSID_NUM, fsidv, 0, 0, fsid, NULL); - - return exp_find_key(clp, FSID_NUM, fsidv, NULL); -} -#endif static svc_export *exp_get_by_name(svc_client *clp, const struct path *path, struct cache_req *reqp) @@ -890,275 +826,7 @@ static struct svc_export *exp_parent(svc_client *clp, struct path *path) return exp; } -#ifdef CONFIG_NFSD_DEPRECATED -/* - * Hashtable locking. Write locks are placed only by user processes - * wanting to modify export information. - * Write locking only done in this file. Read locking - * needed externally. - */ -static DECLARE_RWSEM(hash_sem); - -void -exp_readlock(void) -{ - down_read(&hash_sem); -} - -static inline void -exp_writelock(void) -{ - down_write(&hash_sem); -} - -void -exp_readunlock(void) -{ - up_read(&hash_sem); -} - -static inline void -exp_writeunlock(void) -{ - up_write(&hash_sem); -} -#else - -/* hash_sem not needed once deprecated interface is removed */ -void exp_readlock(void) {} -static inline void exp_writelock(void){} -void exp_readunlock(void) {} -static inline void exp_writeunlock(void){} - -#endif - -#ifdef CONFIG_NFSD_DEPRECATED -static void exp_do_unexport(svc_export *unexp); -static int exp_verify_string(char *cp, int max); - -static void exp_fsid_unhash(struct svc_export *exp) -{ - struct svc_expkey *ek; - - if ((exp->ex_flags & NFSEXP_FSID) == 0) - return; - - ek = exp_get_fsid_key(exp->ex_client, exp->ex_fsid); - if (!IS_ERR(ek)) { - sunrpc_invalidate(&ek->h, &svc_expkey_cache); - cache_put(&ek->h, &svc_expkey_cache); - } -} - -static int exp_fsid_hash(svc_client *clp, struct svc_export *exp) -{ - u32 fsid[2]; - - if ((exp->ex_flags & NFSEXP_FSID) == 0) - return 0; - - mk_fsid(FSID_NUM, fsid, 0, 0, exp->ex_fsid, NULL); - return exp_set_key(clp, FSID_NUM, fsid, exp); -} - -static int exp_hash(struct auth_domain *clp, struct svc_export *exp) -{ - u32 fsid[2]; - struct inode *inode = exp->ex_path.dentry->d_inode; - dev_t dev = inode->i_sb->s_dev; - - if (old_valid_dev(dev)) { - mk_fsid(FSID_DEV, fsid, dev, inode->i_ino, 0, NULL); - return exp_set_key(clp, FSID_DEV, fsid, exp); - } - mk_fsid(FSID_ENCODE_DEV, fsid, dev, inode->i_ino, 0, NULL); - return exp_set_key(clp, FSID_ENCODE_DEV, fsid, exp); -} - -static void exp_unhash(struct svc_export *exp) -{ - struct svc_expkey *ek; - struct inode *inode = exp->ex_path.dentry->d_inode; - - ek = exp_get_key(exp->ex_client, inode->i_sb->s_dev, inode->i_ino); - if (!IS_ERR(ek)) { - sunrpc_invalidate(&ek->h, &svc_expkey_cache); - cache_put(&ek->h, &svc_expkey_cache); - } -} - -/* - * Export a file system. - */ -int -exp_export(struct nfsctl_export *nxp) -{ - svc_client *clp; - struct svc_export *exp = NULL; - struct svc_export new; - struct svc_expkey *fsid_key = NULL; - struct path path; - int err; - - /* Consistency check */ - err = -EINVAL; - if (!exp_verify_string(nxp->ex_path, NFS_MAXPATHLEN) || - !exp_verify_string(nxp->ex_client, NFSCLNT_IDMAX)) - goto out; - - dprintk("exp_export called for %s:%s (%x/%ld fl %x).\n", - nxp->ex_client, nxp->ex_path, - (unsigned)nxp->ex_dev, (long)nxp->ex_ino, - nxp->ex_flags); - - /* Try to lock the export table for update */ - exp_writelock(); - - /* Look up client info */ - if (!(clp = auth_domain_find(nxp->ex_client))) - goto out_unlock; - - - /* Look up the dentry */ - err = kern_path(nxp->ex_path, 0, &path); - if (err) - goto out_put_clp; - err = -EINVAL; - - exp = exp_get_by_name(clp, &path, NULL); - - memset(&new, 0, sizeof(new)); - - /* must make sure there won't be an ex_fsid clash */ - if ((nxp->ex_flags & NFSEXP_FSID) && - (!IS_ERR(fsid_key = exp_get_fsid_key(clp, nxp->ex_dev))) && - fsid_key->ek_path.mnt && - (fsid_key->ek_path.mnt != path.mnt || - fsid_key->ek_path.dentry != path.dentry)) - goto finish; - - if (!IS_ERR(exp)) { - /* just a flags/id/fsid update */ - - exp_fsid_unhash(exp); - exp->ex_flags = nxp->ex_flags; - exp->ex_anon_uid = nxp->ex_anon_uid; - exp->ex_anon_gid = nxp->ex_anon_gid; - exp->ex_fsid = nxp->ex_dev; - - err = exp_fsid_hash(clp, exp); - goto finish; - } - - err = check_export(path.dentry->d_inode, &nxp->ex_flags, NULL); - if (err) goto finish; - - err = -ENOMEM; - - dprintk("nfsd: creating export entry %p for client %p\n", exp, clp); - - new.h.expiry_time = NEVER; - new.h.flags = 0; - new.ex_pathname = kstrdup(nxp->ex_path, GFP_KERNEL); - if (!new.ex_pathname) - goto finish; - new.ex_client = clp; - new.ex_path = path; - new.ex_flags = nxp->ex_flags; - new.ex_anon_uid = nxp->ex_anon_uid; - new.ex_anon_gid = nxp->ex_anon_gid; - new.ex_fsid = nxp->ex_dev; - - exp = svc_export_lookup(&new); - if (exp) - exp = svc_export_update(&new, exp); - - if (!exp) - goto finish; - - if (exp_hash(clp, exp) || - exp_fsid_hash(clp, exp)) { - /* failed to create at least one index */ - exp_do_unexport(exp); - cache_flush(); - } else - err = 0; -finish: - kfree(new.ex_pathname); - if (!IS_ERR_OR_NULL(exp)) - exp_put(exp); - if (!IS_ERR_OR_NULL(fsid_key)) - cache_put(&fsid_key->h, &svc_expkey_cache); - path_put(&path); -out_put_clp: - auth_domain_put(clp); -out_unlock: - exp_writeunlock(); -out: - return err; -} - -/* - * Unexport a file system. The export entry has already - * been removed from the client's list of exported fs's. - */ -static void -exp_do_unexport(svc_export *unexp) -{ - sunrpc_invalidate(&unexp->h, &svc_export_cache); - exp_unhash(unexp); - exp_fsid_unhash(unexp); -} - - -/* - * unexport syscall. - */ -int -exp_unexport(struct nfsctl_export *nxp) -{ - struct auth_domain *dom; - svc_export *exp; - struct path path; - int err; - - /* Consistency check */ - if (!exp_verify_string(nxp->ex_path, NFS_MAXPATHLEN) || - !exp_verify_string(nxp->ex_client, NFSCLNT_IDMAX)) - return -EINVAL; - - exp_writelock(); - - err = -EINVAL; - dom = auth_domain_find(nxp->ex_client); - if (!dom) { - dprintk("nfsd: unexport couldn't find %s\n", nxp->ex_client); - goto out_unlock; - } - - err = kern_path(nxp->ex_path, 0, &path); - if (err) - goto out_domain; - - err = -EINVAL; - exp = exp_get_by_name(dom, &path, NULL); - path_put(&path); - if (IS_ERR(exp)) - goto out_domain; - - exp_do_unexport(exp); - exp_put(exp); - err = 0; - -out_domain: - auth_domain_put(dom); - cache_flush(); -out_unlock: - exp_writeunlock(); - return err; -} -#endif /* CONFIG_NFSD_DEPRECATED */ /* * Obtain the root fh on behalf of a client. @@ -1330,7 +998,7 @@ rqst_exp_parent(struct svc_rqst *rqstp, struct path *path) return exp; } -static struct svc_export *find_fsidzero_export(struct svc_rqst *rqstp) +struct svc_export *rqst_find_fsidzero_export(struct svc_rqst *rqstp) { u32 fsidv[2]; @@ -1350,7 +1018,7 @@ exp_pseudoroot(struct svc_rqst *rqstp, struct svc_fh *fhp) struct svc_export *exp; __be32 rv; - exp = find_fsidzero_export(rqstp); + exp = rqst_find_fsidzero_export(rqstp); if (IS_ERR(exp)) return nfserrno(PTR_ERR(exp)); rv = fh_compose(fhp, exp, exp->ex_path.dentry, NULL); @@ -1367,7 +1035,6 @@ static void *e_start(struct seq_file *m, loff_t *pos) unsigned hash, export; struct cache_head *ch; - exp_readlock(); read_lock(&svc_export_cache.hash_lock); if (!n--) return SEQ_START_TOKEN; @@ -1418,7 +1085,6 @@ static void e_stop(struct seq_file *m, void *p) __releases(svc_export_cache.hash_lock) { read_unlock(&svc_export_cache.hash_lock); - exp_readunlock(); } static struct flags { @@ -1550,97 +1216,6 @@ const struct seq_operations nfs_exports_op = { .show = e_show, }; -#ifdef CONFIG_NFSD_DEPRECATED -/* - * Add or modify a client. - * Change requests may involve the list of host addresses. The list of - * exports and possibly existing uid maps are left untouched. - */ -int -exp_addclient(struct nfsctl_client *ncp) -{ - struct auth_domain *dom; - int i, err; - struct in6_addr addr6; - - /* First, consistency check. */ - err = -EINVAL; - if (! exp_verify_string(ncp->cl_ident, NFSCLNT_IDMAX)) - goto out; - if (ncp->cl_naddr > NFSCLNT_ADDRMAX) - goto out; - - /* Lock the hashtable */ - exp_writelock(); - - dom = unix_domain_find(ncp->cl_ident); - - err = -ENOMEM; - if (!dom) - goto out_unlock; - - /* Insert client into hashtable. */ - for (i = 0; i < ncp->cl_naddr; i++) { - ipv6_addr_set_v4mapped(ncp->cl_addrlist[i].s_addr, &addr6); - auth_unix_add_addr(&init_net, &addr6, dom); - } - auth_unix_forget_old(dom); - auth_domain_put(dom); - - err = 0; - -out_unlock: - exp_writeunlock(); -out: - return err; -} - -/* - * Delete a client given an identifier. - */ -int -exp_delclient(struct nfsctl_client *ncp) -{ - int err; - struct auth_domain *dom; - - err = -EINVAL; - if (!exp_verify_string(ncp->cl_ident, NFSCLNT_IDMAX)) - goto out; - - /* Lock the hashtable */ - exp_writelock(); - - dom = auth_domain_find(ncp->cl_ident); - /* just make sure that no addresses work - * and that it will expire soon - */ - if (dom) { - err = auth_unix_forget_old(dom); - auth_domain_put(dom); - } - - exp_writeunlock(); -out: - return err; -} - -/* - * Verify that string is non-empty and does not exceed max length. - */ -static int -exp_verify_string(char *cp, int max) -{ - int i; - - for (i = 0; i < max; i++) - if (!cp[i]) - return i; - cp[i] = 0; - printk(KERN_NOTICE "nfsd: couldn't validate string %s\n", cp); - return 0; -} -#endif /* CONFIG_NFSD_DEPRECATED */ /* * Initialize the exports module. @@ -1667,10 +1242,8 @@ nfsd_export_init(void) void nfsd_export_flush(void) { - exp_writelock(); cache_purge(&svc_expkey_cache); cache_purge(&svc_export_cache); - exp_writeunlock(); } /* @@ -1682,12 +1255,9 @@ nfsd_export_shutdown(void) dprintk("nfsd: shutting down export module.\n"); - exp_writelock(); - cache_unregister(&svc_expkey_cache); cache_unregister(&svc_export_cache); svcauth_unix_purge(); - exp_writeunlock(); dprintk("nfsd: export shutdown complete.\n"); } diff --git a/fs/nfsd/lockd.c b/fs/nfsd/lockd.c index 7c831a2..77e7a5c 100644 --- a/fs/nfsd/lockd.c +++ b/fs/nfsd/lockd.c @@ -35,10 +35,8 @@ nlm_fopen(struct svc_rqst *rqstp, struct nfs_fh *f, struct file **filp) memcpy((char*)&fh.fh_handle.fh_base, f->data, f->size); fh.fh_export = NULL; - exp_readlock(); nfserr = nfsd_open(rqstp, &fh, S_IFREG, NFSD_MAY_LOCK, filp); fh_put(&fh); - exp_readunlock(); /* We return nlm error codes as nlm doesn't know * about nfsd, but nfsd does know about nlm.. */ diff --git a/fs/nfsd/nfs4acl.c b/fs/nfsd/nfs4acl.c index ad88f1c..435a9be 100644 --- a/fs/nfsd/nfs4acl.c +++ b/fs/nfsd/nfs4acl.c @@ -36,6 +36,7 @@ #include <linux/slab.h> #include <linux/nfs_fs.h> +#include <linux/export.h> #include "acl.h" @@ -372,8 +373,10 @@ sort_pacl(struct posix_acl *pacl) * by uid/gid. */ int i, j; - if (pacl->a_count <= 4) - return; /* no users or groups */ + /* no users or groups */ + if (!pacl || pacl->a_count <= 4) + return; + i = 1; while (pacl->a_entries[i].e_tag == ACL_USER) i++; @@ -497,13 +500,12 @@ posix_state_to_acl(struct posix_acl_state *state, unsigned int flags) /* * ACLs with no ACEs are treated differently in the inheritable - * and effective cases: when there are no inheritable ACEs, we - * set a zero-length default posix acl: + * and effective cases: when there are no inheritable ACEs, + * calls ->set_acl with a NULL ACL structure. */ - if (state->empty && (flags & NFS4_ACL_TYPE_DEFAULT)) { - pacl = posix_acl_alloc(0, GFP_KERNEL); - return pacl ? pacl : ERR_PTR(-ENOMEM); - } + if (state->empty && (flags & NFS4_ACL_TYPE_DEFAULT)) + return NULL; + /* * When there are no effective ACEs, the following will end * up setting a 3-element effective posix ACL with all diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 02eb4ed..51b4f43 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -39,6 +39,8 @@ #define NFSDDBG_FACILITY NFSDDBG_PROC +static void nfsd4_mark_cb_fault(struct nfs4_client *, int reason); + #define NFSPROC4_CB_NULL 0 #define NFSPROC4_CB_COMPOUND 1 @@ -351,7 +353,7 @@ static void encode_cb_recall4args(struct xdr_stream *xdr, __be32 *p; encode_nfs_cb_opnum4(xdr, OP_CB_RECALL); - encode_stateid4(xdr, &dp->dl_stateid); + encode_stateid4(xdr, &dp->dl_stid.sc_stateid); p = xdr_reserve_space(xdr, 4); *p++ = xdr_zero; /* truncate */ @@ -460,6 +462,8 @@ static int decode_cb_sequence4resok(struct xdr_stream *xdr, */ status = 0; out: + if (status) + nfsd4_mark_cb_fault(cb->cb_clp, status); return status; out_overflow: print_overflow_msg(__func__, xdr); @@ -629,9 +633,11 @@ static int max_cb_time(void) static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *conn, struct nfsd4_session *ses) { + int maxtime = max_cb_time(); struct rpc_timeout timeparms = { - .to_initval = max_cb_time(), + .to_initval = maxtime, .to_retries = 0, + .to_maxval = maxtime, }; struct rpc_create_args args = { .net = &init_net, @@ -686,6 +692,12 @@ static void nfsd4_mark_cb_down(struct nfs4_client *clp, int reason) warn_no_callback_path(clp, reason); } +static void nfsd4_mark_cb_fault(struct nfs4_client *clp, int reason) +{ + clp->cl_cb_state = NFSD4_CB_FAULT; + warn_no_callback_path(clp, reason); +} + static void nfsd4_cb_probe_done(struct rpc_task *task, void *calldata) { struct nfs4_client *clp = container_of(calldata, struct nfs4_client, cl_cb_null); @@ -773,8 +785,12 @@ static bool nfsd41_cb_get_slot(struct nfs4_client *clp, struct rpc_task *task) { if (test_and_set_bit(0, &clp->cl_cb_slot_busy) != 0) { rpc_sleep_on(&clp->cl_cb_waitq, task, NULL); - dprintk("%s slot is busy\n", __func__); - return false; + /* Race breaker */ + if (test_and_set_bit(0, &clp->cl_cb_slot_busy) != 0) { + dprintk("%s slot is busy\n", __func__); + return false; + } + rpc_wake_up_queued_task(&clp->cl_cb_waitq, task); } return true; } @@ -787,7 +803,7 @@ static void nfsd4_cb_prepare(struct rpc_task *task, void *calldata) { struct nfsd4_callback *cb = calldata; struct nfs4_delegation *dp = container_of(cb, struct nfs4_delegation, dl_recall); - struct nfs4_client *clp = dp->dl_client; + struct nfs4_client *clp = dp->dl_stid.sc_client; u32 minorversion = clp->cl_minorversion; cb->cb_minorversion = minorversion; @@ -809,7 +825,7 @@ static void nfsd4_cb_done(struct rpc_task *task, void *calldata) { struct nfsd4_callback *cb = calldata; struct nfs4_delegation *dp = container_of(cb, struct nfs4_delegation, dl_recall); - struct nfs4_client *clp = dp->dl_client; + struct nfs4_client *clp = dp->dl_stid.sc_client; dprintk("%s: minorversion=%d\n", __func__, clp->cl_minorversion); @@ -832,7 +848,7 @@ static void nfsd4_cb_recall_done(struct rpc_task *task, void *calldata) { struct nfsd4_callback *cb = calldata; struct nfs4_delegation *dp = container_of(cb, struct nfs4_delegation, dl_recall); - struct nfs4_client *clp = dp->dl_client; + struct nfs4_client *clp = dp->dl_stid.sc_client; struct rpc_clnt *current_rpc_client = clp->cl_cb_client; nfsd4_cb_done(task, calldata); @@ -1006,7 +1022,7 @@ void nfsd4_do_callback_rpc(struct work_struct *w) void nfsd4_cb_recall(struct nfs4_delegation *dp) { struct nfsd4_callback *cb = &dp->dl_recall; - struct nfs4_client *clp = dp->dl_client; + struct nfs4_client *clp = dp->dl_stid.sc_client; dp->dl_retries = 1; cb->cb_op = dp; diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index d06a02c..9a959de 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -35,6 +35,7 @@ #include <linux/file.h> #include <linux/slab.h> +#include "idmap.h" #include "cache.h" #include "xdr4.h" #include "vfs.h" @@ -170,12 +171,30 @@ do_open_permission(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfs return status; } +static __be32 nfsd_check_obj_isreg(struct svc_fh *fh) +{ + umode_t mode = fh->fh_dentry->d_inode->i_mode; + + if (S_ISREG(mode)) + return nfs_ok; + if (S_ISDIR(mode)) + return nfserr_isdir; + /* + * Using err_symlink as our catch-all case may look odd; but + * there's no other obvious error for this case in 4.0, and we + * happen to know that it will cause the linux v4 client to do + * the right thing on attempts to open something other than a + * regular file. + */ + return nfserr_symlink; +} + static __be32 do_open_lookup(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_open *open) { struct svc_fh resfh; + int accmode; __be32 status; - int created = 0; fh_init(&resfh, NFS4_FHSIZE); open->op_truncate = 0; @@ -204,7 +223,7 @@ do_open_lookup(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_o open->op_fname.len, &open->op_iattr, &resfh, open->op_createmode, (u32 *)open->op_verf.data, - &open->op_truncate, &created); + &open->op_truncate, &open->op_created); /* * Following rfc 3530 14.2.16, use the returned bitmask @@ -213,7 +232,7 @@ do_open_lookup(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_o */ if (open->op_createmode == NFS4_CREATE_EXCLUSIVE && status == 0) open->op_bmval[1] = (FATTR4_WORD1_TIME_ACCESS | - FATTR4_WORD1_TIME_MODIFY); + FATTR4_WORD1_TIME_MODIFY); } else { status = nfsd_lookup(rqstp, current_fh, open->op_fname.data, open->op_fname.len, &resfh); @@ -221,6 +240,9 @@ do_open_lookup(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_o } if (status) goto out; + status = nfsd_check_obj_isreg(&resfh); + if (status) + goto out; if (is_create_with_attrs(open) && open->op_acl != NULL) do_set_nfs4_acl(rqstp, &resfh, open->op_acl, open->op_bmval); @@ -229,11 +251,12 @@ do_open_lookup(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_o fh_dup2(current_fh, &resfh); /* set reply cache */ - fh_copy_shallow(&open->op_stateowner->so_replay.rp_openfh, + fh_copy_shallow(&open->op_openowner->oo_owner.so_replay.rp_openfh, &resfh.fh_handle); - if (!created) - status = do_open_permission(rqstp, current_fh, open, - NFSD_MAY_NOP); + accmode = NFSD_MAY_NOP; + if (open->op_created) + accmode |= NFSD_MAY_OWNER_OVERRIDE; + status = do_open_permission(rqstp, current_fh, open, accmode); out: fh_put(&resfh); @@ -244,6 +267,7 @@ static __be32 do_open_fhandle(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_open *open) { __be32 status; + int accmode = 0; /* Only reclaims from previously confirmed clients are valid */ if ((status = nfs4_check_open_reclaim(&open->op_clientid))) @@ -256,14 +280,24 @@ do_open_fhandle(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_ memset(&open->op_cinfo, 0, sizeof(struct nfsd4_change_info)); /* set replay cache */ - fh_copy_shallow(&open->op_stateowner->so_replay.rp_openfh, + fh_copy_shallow(&open->op_openowner->oo_owner.so_replay.rp_openfh, ¤t_fh->fh_handle); open->op_truncate = (open->op_iattr.ia_valid & ATTR_SIZE) && (open->op_iattr.ia_size == 0); + /* + * In the delegation case, the client is telling us about an + * open that it *already* performed locally, some time ago. We + * should let it succeed now if possible. + * + * In the case of a CLAIM_FH open, on the other hand, the client + * may be counting on us to enforce permissions (the Linux 4.1 + * client uses this for normal opens, for example). + */ + if (open->op_claim_type == NFS4_OPEN_CLAIM_DELEG_CUR_FH) + accmode = NFSD_MAY_OWNER_OVERRIDE; - status = do_open_permission(rqstp, current_fh, open, - NFSD_MAY_OWNER_OVERRIDE); + status = do_open_permission(rqstp, current_fh, open, accmode); return status; } @@ -285,14 +319,27 @@ nfsd4_open(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, __be32 status; struct nfsd4_compoundres *resp; - dprintk("NFSD: nfsd4_open filename %.*s op_stateowner %p\n", + dprintk("NFSD: nfsd4_open filename %.*s op_openowner %p\n", (int)open->op_fname.len, open->op_fname.data, - open->op_stateowner); + open->op_openowner); /* This check required by spec. */ if (open->op_create && open->op_claim_type != NFS4_OPEN_CLAIM_NULL) return nfserr_inval; + /* We don't yet support WANT bits: */ + open->op_share_access &= NFS4_SHARE_ACCESS_MASK; + + open->op_created = 0; + /* + * RFC5661 18.51.3 + * Before RECLAIM_COMPLETE done, server should deny new lock + */ + if (nfsd4_has_session(cstate) && + !cstate->session->se_client->cl_firststate && + open->op_claim_type != NFS4_OPEN_CLAIM_PREVIOUS) + return nfserr_grace; + if (nfsd4_has_session(cstate)) copy_clientid(&open->op_clientid, cstate->session); @@ -302,7 +349,7 @@ nfsd4_open(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, resp = rqstp->rq_resp; status = nfsd4_process_open1(&resp->cstate, open); if (status == nfserr_replay_me) { - struct nfs4_replay *rp = &open->op_stateowner->so_replay; + struct nfs4_replay *rp = &open->op_openowner->oo_owner.so_replay; fh_put(&cstate->current_fh); fh_copy_shallow(&cstate->current_fh.fh_handle, &rp->rp_openfh); @@ -332,32 +379,23 @@ nfsd4_open(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, switch (open->op_claim_type) { case NFS4_OPEN_CLAIM_DELEGATE_CUR: case NFS4_OPEN_CLAIM_NULL: - /* - * (1) set CURRENT_FH to the file being opened, - * creating it if necessary, (2) set open->op_cinfo, - * (3) set open->op_truncate if the file is to be - * truncated after opening, (4) do permission checking. - */ status = do_open_lookup(rqstp, &cstate->current_fh, open); if (status) goto out; break; case NFS4_OPEN_CLAIM_PREVIOUS: - open->op_stateowner->so_confirmed = 1; - /* - * The CURRENT_FH is already set to the file being - * opened. (1) set open->op_cinfo, (2) set - * open->op_truncate if the file is to be truncated - * after opening, (3) do permission checking. - */ + open->op_openowner->oo_flags |= NFS4_OO_CONFIRMED; + case NFS4_OPEN_CLAIM_FH: + case NFS4_OPEN_CLAIM_DELEG_CUR_FH: status = do_open_fhandle(rqstp, &cstate->current_fh, open); if (status) goto out; break; + case NFS4_OPEN_CLAIM_DELEG_PREV_FH: case NFS4_OPEN_CLAIM_DELEGATE_PREV: - open->op_stateowner->so_confirmed = 1; + open->op_openowner->oo_flags |= NFS4_OO_CONFIRMED; dprintk("NFSD: unsupported OPEN claim type %d\n", open->op_claim_type); status = nfserr_notsupp; @@ -374,12 +412,13 @@ nfsd4_open(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, * set, (2) sets open->op_stateid, (3) sets open->op_delegation. */ status = nfsd4_process_open2(rqstp, &cstate->current_fh, open); + WARN_ON(status && open->op_created); out: - if (open->op_stateowner) { - nfs4_get_stateowner(open->op_stateowner); - cstate->replay_owner = open->op_stateowner; - } - nfs4_unlock_state(); + nfsd4_cleanup_open_state(open, status); + if (open->op_openowner) + cstate->replay_owner = &open->op_openowner->oo_owner; + else + nfs4_unlock_state(); return status; } @@ -460,17 +499,12 @@ static __be32 nfsd4_commit(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_commit *commit) { - __be32 status; - u32 *p = (u32 *)commit->co_verf.data; *p++ = nfssvc_boot.tv_sec; *p++ = nfssvc_boot.tv_usec; - status = nfsd_commit(rqstp, &cstate->current_fh, commit->co_offset, + return nfsd_commit(rqstp, &cstate->current_fh, commit->co_offset, commit->co_count); - if (status == nfserr_symlink) - status = nfserr_inval; - return status; } static __be32 @@ -485,8 +519,6 @@ nfsd4_create(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, status = fh_verify(rqstp, &cstate->current_fh, S_IFDIR, NFSD_MAY_CREATE); - if (status == nfserr_symlink) - status = nfserr_notdir; if (status) return status; @@ -497,15 +529,6 @@ nfsd4_create(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, switch (create->cr_type) { case NF4LNK: - /* ugh! we have to null-terminate the linktext, or - * vfs_symlink() will choke. it is always safe to - * null-terminate by brute force, since at worst we - * will overwrite the first byte of the create namelen - * in the XDR buffer, which has already been extracted - * during XDR decode. - */ - create->cr_linkname[create->cr_linklen] = 0; - status = nfsd_symlink(rqstp, &cstate->current_fh, create->cr_name, create->cr_namelen, create->cr_linkname, create->cr_linklen, @@ -712,8 +735,6 @@ nfsd4_remove(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, return nfserr_grace; status = nfsd_unlink(rqstp, &cstate->current_fh, 0, remove->rm_name, remove->rm_namelen); - if (status == nfserr_symlink) - return nfserr_notdir; if (!status) { fh_unlock(&cstate->current_fh); set_change_info(&remove->rm_cinfo, &cstate->current_fh); @@ -744,8 +765,6 @@ nfsd4_rename(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, (S_ISDIR(cstate->save_fh.fh_dentry->d_inode->i_mode) && S_ISDIR(cstate->current_fh.fh_dentry->d_inode->i_mode))) status = nfserr_exist; - else if (status == nfserr_symlink) - status = nfserr_notdir; if (!status) { set_change_info(&rename->rn_sinfo, &cstate->current_fh); @@ -863,14 +882,14 @@ nfsd4_write(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, nfs4_lock_state(); status = nfs4_preprocess_stateid_op(cstate, stateid, WR_STATE, &filp); - if (filp) - get_file(filp); - nfs4_unlock_state(); - if (status) { + nfs4_unlock_state(); dprintk("NFSD: nfsd4_write: couldn't process stateid!\n"); return status; } + if (filp) + get_file(filp); + nfs4_unlock_state(); cnt = write->wr_buflen; write->wr_how_written = write->wr_stable_how; @@ -886,8 +905,6 @@ nfsd4_write(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, write->wr_bytes_written = cnt; - if (status == nfserr_symlink) - status = nfserr_inval; return status; } @@ -988,6 +1005,8 @@ static inline void nfsd4_increment_op_stats(u32 opnum) typedef __be32(*nfsd4op_func)(struct svc_rqst *, struct nfsd4_compound_state *, void *); +typedef u32(*nfsd4op_rsize)(struct svc_rqst *, struct nfsd4_op *op); + enum nfsd4_op_flags { ALLOWED_WITHOUT_FH = 1 << 0, /* No current filehandle required */ ALLOWED_ON_ABSENT_FS = 1 << 1, /* ops processed on absent fs */ @@ -995,12 +1014,31 @@ enum nfsd4_op_flags { /* For rfc 5661 section 2.6.3.1.1: */ OP_HANDLES_WRONGSEC = 1 << 3, OP_IS_PUTFH_LIKE = 1 << 4, + /* + * These are the ops whose result size we estimate before + * encoding, to avoid performing an op then not being able to + * respond or cache a response. This includes writes and setattrs + * as well as the operations usually called "nonidempotent": + */ + OP_MODIFIES_SOMETHING = 1 << 5, + /* + * Cache compounds containing these ops in the xid-based drc: + * We use the DRC for compounds containing non-idempotent + * operations, *except* those that are 4.1-specific (since + * sessions provide their own EOS), and except for stateful + * operations other than setclientid and setclientid_confirm + * (since sequence numbers provide EOS for open, lock, etc in + * the v4.0 case). + */ + OP_CACHEME = 1 << 6, }; struct nfsd4_operation { nfsd4op_func op_func; u32 op_flags; char *op_name; + /* Try to get response size before operation */ + nfsd4op_rsize op_rsize_bop; }; static struct nfsd4_operation nfsd4_ops[]; @@ -1045,6 +1083,11 @@ static inline struct nfsd4_operation *OPDESC(struct nfsd4_op *op) return &nfsd4_ops[op->opnum]; } +bool nfsd4_cache_this_op(struct nfsd4_op *op) +{ + return OPDESC(op)->op_flags & OP_CACHEME; +} + static bool need_wrongsec_check(struct svc_rqst *rqstp) { struct nfsd4_compoundres *resp = rqstp->rq_resp; @@ -1068,7 +1111,8 @@ static bool need_wrongsec_check(struct svc_rqst *rqstp) */ if (argp->opcnt == resp->opcnt) return false; - + if (next->opnum == OP_ILLEGAL) + return false; nextd = OPDESC(next); /* * Rest of 2.6.3.1.1: certain operations will return WRONGSEC @@ -1090,6 +1134,7 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, struct nfsd4_operation *opdesc; struct nfsd4_compound_state *cstate = &resp->cstate; int slack_bytes; + u32 plen = 0; __be32 status; resp->xbuf = &rqstp->rq_res; @@ -1168,6 +1213,21 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, goto encode_op; } + /* If op is non-idempotent */ + if (opdesc->op_flags & OP_MODIFIES_SOMETHING) { + plen = opdesc->op_rsize_bop(rqstp, op); + /* + * If there's still another operation, make sure + * we'll have space to at least encode an error: + */ + if (resp->opcnt < args->opcnt) + plen += COMPOUND_ERR_SLACK_SPACE; + op->status = nfsd4_check_resp_size(resp, plen); + } + + if (op->status) + goto encode_op; + if (opdesc->op_func) op->status = opdesc->op_func(rqstp, cstate, &op->u); else @@ -1197,7 +1257,7 @@ encode_op: be32_to_cpu(status)); if (cstate->replay_owner) { - nfs4_put_stateowner(cstate->replay_owner); + nfs4_unlock_state(); cstate->replay_owner = NULL; } /* XXX Ugh, we need to get rid of this kind of special case: */ @@ -1212,13 +1272,151 @@ encode_op: fh_put(&resp->cstate.save_fh); BUG_ON(resp->cstate.replay_owner); out: - nfsd4_release_compoundargs(args); /* Reset deferral mechanism for RPC deferrals */ rqstp->rq_usedeferral = 1; dprintk("nfsv4 compound returned %d\n", ntohl(status)); return status; } +#define op_encode_hdr_size (2) +#define op_encode_stateid_maxsz (XDR_QUADLEN(NFS4_STATEID_SIZE)) +#define op_encode_verifier_maxsz (XDR_QUADLEN(NFS4_VERIFIER_SIZE)) +#define op_encode_change_info_maxsz (5) +#define nfs4_fattr_bitmap_maxsz (4) + +#define op_encode_lockowner_maxsz (1 + XDR_QUADLEN(IDMAP_NAMESZ)) +#define op_encode_lock_denied_maxsz (8 + op_encode_lockowner_maxsz) + +#define nfs4_owner_maxsz (1 + XDR_QUADLEN(IDMAP_NAMESZ)) + +#define op_encode_ace_maxsz (3 + nfs4_owner_maxsz) +#define op_encode_delegation_maxsz (1 + op_encode_stateid_maxsz + 1 + \ + op_encode_ace_maxsz) + +#define op_encode_channel_attrs_maxsz (6 + 1 + 1) + +static inline u32 nfsd4_only_status_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size) * sizeof(__be32); +} + +static inline u32 nfsd4_status_stateid_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_stateid_maxsz)* sizeof(__be32); +} + +static inline u32 nfsd4_commit_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_verifier_maxsz) * sizeof(__be32); +} + +static inline u32 nfsd4_create_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_change_info_maxsz + + nfs4_fattr_bitmap_maxsz) * sizeof(__be32); +} + +static inline u32 nfsd4_link_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_change_info_maxsz) + * sizeof(__be32); +} + +static inline u32 nfsd4_lock_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_lock_denied_maxsz) + * sizeof(__be32); +} + +static inline u32 nfsd4_open_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_stateid_maxsz + + op_encode_change_info_maxsz + 1 + + nfs4_fattr_bitmap_maxsz + + op_encode_delegation_maxsz) * sizeof(__be32); +} + +static inline u32 nfsd4_read_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + u32 maxcount = 0, rlen = 0; + + maxcount = svc_max_payload(rqstp); + rlen = op->u.read.rd_length; + + if (rlen > maxcount) + rlen = maxcount; + + return (op_encode_hdr_size + 2) * sizeof(__be32) + rlen; +} + +static inline u32 nfsd4_readdir_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + u32 rlen = op->u.readdir.rd_maxcount; + + if (rlen > PAGE_SIZE) + rlen = PAGE_SIZE; + + return (op_encode_hdr_size + op_encode_verifier_maxsz) + * sizeof(__be32) + rlen; +} + +static inline u32 nfsd4_remove_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_change_info_maxsz) + * sizeof(__be32); +} + +static inline u32 nfsd4_rename_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_change_info_maxsz + + op_encode_change_info_maxsz) * sizeof(__be32); +} + +static inline u32 nfsd4_setattr_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + nfs4_fattr_bitmap_maxsz) * sizeof(__be32); +} + +static inline u32 nfsd4_setclientid_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + 2 + XDR_QUADLEN(NFS4_VERIFIER_SIZE)) * + sizeof(__be32); +} + +static inline u32 nfsd4_write_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + op_encode_verifier_maxsz) * sizeof(__be32); +} + +static inline u32 nfsd4_exchange_id_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + 2 + 1 + /* eir_clientid, eir_sequenceid */\ + 1 + 1 + 0 + /* eir_flags, spr_how, SP4_NONE (for now) */\ + 2 + /*eir_server_owner.so_minor_id */\ + /* eir_server_owner.so_major_id<> */\ + XDR_QUADLEN(NFS4_OPAQUE_LIMIT) + 1 +\ + /* eir_server_scope<> */\ + XDR_QUADLEN(NFS4_OPAQUE_LIMIT) + 1 +\ + 1 + /* eir_server_impl_id array length */\ + 0 /* ignored eir_server_impl_id contents */) * sizeof(__be32); +} + +static inline u32 nfsd4_bind_conn_to_session_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + \ + XDR_QUADLEN(NFS4_MAX_SESSIONID_LEN) + /* bctsr_sessid */\ + 2 /* bctsr_dir, use_conn_in_rdma_mode */) * sizeof(__be32); +} + +static inline u32 nfsd4_create_session_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op) +{ + return (op_encode_hdr_size + \ + XDR_QUADLEN(NFS4_MAX_SESSIONID_LEN) + /* sessionid */\ + 2 + /* csr_sequence, csr_flags */\ + op_encode_channel_attrs_maxsz + \ + op_encode_channel_attrs_maxsz) * sizeof(__be32); +} + static struct nfsd4_operation nfsd4_ops[] = { [OP_ACCESS] = { .op_func = (nfsd4op_func)nfsd4_access, @@ -1226,19 +1424,27 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_CLOSE] = { .op_func = (nfsd4op_func)nfsd4_close, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_CLOSE", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_status_stateid_rsize, }, [OP_COMMIT] = { .op_func = (nfsd4op_func)nfsd4_commit, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_COMMIT", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_commit_rsize, }, [OP_CREATE] = { .op_func = (nfsd4op_func)nfsd4_create, + .op_flags = OP_MODIFIES_SOMETHING | OP_CACHEME, .op_name = "OP_CREATE", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_create_rsize, }, [OP_DELEGRETURN] = { .op_func = (nfsd4op_func)nfsd4_delegreturn, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_DELEGRETURN", + .op_rsize_bop = nfsd4_only_status_rsize, }, [OP_GETATTR] = { .op_func = (nfsd4op_func)nfsd4_getattr, @@ -1251,11 +1457,16 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_LINK] = { .op_func = (nfsd4op_func)nfsd4_link, + .op_flags = ALLOWED_ON_ABSENT_FS | OP_MODIFIES_SOMETHING + | OP_CACHEME, .op_name = "OP_LINK", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_link_rsize, }, [OP_LOCK] = { .op_func = (nfsd4op_func)nfsd4_lock, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_LOCK", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_lock_rsize, }, [OP_LOCKT] = { .op_func = (nfsd4op_func)nfsd4_lockt, @@ -1263,7 +1474,9 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_LOCKU] = { .op_func = (nfsd4op_func)nfsd4_locku, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_LOCKU", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_status_stateid_rsize, }, [OP_LOOKUP] = { .op_func = (nfsd4op_func)nfsd4_lookup, @@ -1281,42 +1494,54 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_OPEN] = { .op_func = (nfsd4op_func)nfsd4_open, - .op_flags = OP_HANDLES_WRONGSEC, + .op_flags = OP_HANDLES_WRONGSEC | OP_MODIFIES_SOMETHING, .op_name = "OP_OPEN", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_open_rsize, }, [OP_OPEN_CONFIRM] = { .op_func = (nfsd4op_func)nfsd4_open_confirm, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_OPEN_CONFIRM", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_status_stateid_rsize, }, [OP_OPEN_DOWNGRADE] = { .op_func = (nfsd4op_func)nfsd4_open_downgrade, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_OPEN_DOWNGRADE", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_status_stateid_rsize, }, [OP_PUTFH] = { .op_func = (nfsd4op_func)nfsd4_putfh, .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS - | OP_IS_PUTFH_LIKE, + | OP_IS_PUTFH_LIKE | OP_MODIFIES_SOMETHING, .op_name = "OP_PUTFH", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_PUTPUBFH] = { .op_func = (nfsd4op_func)nfsd4_putrootfh, .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS - | OP_IS_PUTFH_LIKE, + | OP_IS_PUTFH_LIKE | OP_MODIFIES_SOMETHING, .op_name = "OP_PUTPUBFH", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_PUTROOTFH] = { .op_func = (nfsd4op_func)nfsd4_putrootfh, .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS - | OP_IS_PUTFH_LIKE, + | OP_IS_PUTFH_LIKE | OP_MODIFIES_SOMETHING, .op_name = "OP_PUTROOTFH", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_READ] = { .op_func = (nfsd4op_func)nfsd4_read, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_READ", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_read_rsize, }, [OP_READDIR] = { .op_func = (nfsd4op_func)nfsd4_readdir, + .op_flags = OP_MODIFIES_SOMETHING, .op_name = "OP_READDIR", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_readdir_rsize, }, [OP_READLINK] = { .op_func = (nfsd4op_func)nfsd4_readlink, @@ -1324,27 +1549,36 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_REMOVE] = { .op_func = (nfsd4op_func)nfsd4_remove, + .op_flags = OP_MODIFIES_SOMETHING | OP_CACHEME, .op_name = "OP_REMOVE", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_remove_rsize, }, [OP_RENAME] = { - .op_name = "OP_RENAME", .op_func = (nfsd4op_func)nfsd4_rename, + .op_flags = OP_MODIFIES_SOMETHING | OP_CACHEME, + .op_name = "OP_RENAME", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_rename_rsize, }, [OP_RENEW] = { .op_func = (nfsd4op_func)nfsd4_renew, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_MODIFIES_SOMETHING, .op_name = "OP_RENEW", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, + }, [OP_RESTOREFH] = { .op_func = (nfsd4op_func)nfsd4_restorefh, .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS - | OP_IS_PUTFH_LIKE, + | OP_IS_PUTFH_LIKE | OP_MODIFIES_SOMETHING, .op_name = "OP_RESTOREFH", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_SAVEFH] = { .op_func = (nfsd4op_func)nfsd4_savefh, - .op_flags = OP_HANDLES_WRONGSEC, + .op_flags = OP_HANDLES_WRONGSEC | OP_MODIFIES_SOMETHING, .op_name = "OP_SAVEFH", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_SECINFO] = { .op_func = (nfsd4op_func)nfsd4_secinfo, @@ -1354,16 +1588,22 @@ static struct nfsd4_operation nfsd4_ops[] = { [OP_SETATTR] = { .op_func = (nfsd4op_func)nfsd4_setattr, .op_name = "OP_SETATTR", + .op_flags = OP_MODIFIES_SOMETHING | OP_CACHEME, + .op_rsize_bop = (nfsd4op_rsize)nfsd4_setattr_rsize, }, [OP_SETCLIENTID] = { .op_func = (nfsd4op_func)nfsd4_setclientid, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_MODIFIES_SOMETHING | OP_CACHEME, .op_name = "OP_SETCLIENTID", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_setclientid_rsize, }, [OP_SETCLIENTID_CONFIRM] = { .op_func = (nfsd4op_func)nfsd4_setclientid_confirm, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_MODIFIES_SOMETHING | OP_CACHEME, .op_name = "OP_SETCLIENTID_CONFIRM", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_VERIFY] = { .op_func = (nfsd4op_func)nfsd4_verify, @@ -1371,50 +1611,81 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_WRITE] = { .op_func = (nfsd4op_func)nfsd4_write, + .op_flags = OP_MODIFIES_SOMETHING | OP_CACHEME, .op_name = "OP_WRITE", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_write_rsize, }, [OP_RELEASE_LOCKOWNER] = { .op_func = (nfsd4op_func)nfsd4_release_lockowner, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_MODIFIES_SOMETHING, .op_name = "OP_RELEASE_LOCKOWNER", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, /* NFSv4.1 operations */ [OP_EXCHANGE_ID] = { .op_func = (nfsd4op_func)nfsd4_exchange_id, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP + | OP_MODIFIES_SOMETHING, .op_name = "OP_EXCHANGE_ID", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_exchange_id_rsize, }, [OP_BIND_CONN_TO_SESSION] = { .op_func = (nfsd4op_func)nfsd4_bind_conn_to_session, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP + | OP_MODIFIES_SOMETHING, .op_name = "OP_BIND_CONN_TO_SESSION", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_bind_conn_to_session_rsize, }, [OP_CREATE_SESSION] = { .op_func = (nfsd4op_func)nfsd4_create_session, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP + | OP_MODIFIES_SOMETHING, .op_name = "OP_CREATE_SESSION", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_create_session_rsize, }, [OP_DESTROY_SESSION] = { .op_func = (nfsd4op_func)nfsd4_destroy_session, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP + | OP_MODIFIES_SOMETHING, .op_name = "OP_DESTROY_SESSION", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_SEQUENCE] = { .op_func = (nfsd4op_func)nfsd4_sequence, .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP, .op_name = "OP_SEQUENCE", }, + [OP_DESTROY_CLIENTID] = { + .op_func = (nfsd4op_func)nfsd4_destroy_clientid, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_AS_FIRST_OP + | OP_MODIFIES_SOMETHING, + .op_name = "OP_DESTROY_CLIENTID", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, + }, [OP_RECLAIM_COMPLETE] = { .op_func = (nfsd4op_func)nfsd4_reclaim_complete, - .op_flags = ALLOWED_WITHOUT_FH, + .op_flags = ALLOWED_WITHOUT_FH | OP_MODIFIES_SOMETHING, .op_name = "OP_RECLAIM_COMPLETE", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, }, [OP_SECINFO_NO_NAME] = { .op_func = (nfsd4op_func)nfsd4_secinfo_no_name, .op_flags = OP_HANDLES_WRONGSEC, .op_name = "OP_SECINFO_NO_NAME", }, + [OP_TEST_STATEID] = { + .op_func = (nfsd4op_func)nfsd4_test_stateid, + .op_flags = ALLOWED_WITHOUT_FH, + .op_name = "OP_TEST_STATEID", + }, + [OP_FREE_STATEID] = { + .op_func = (nfsd4op_func)nfsd4_free_stateid, + .op_flags = ALLOWED_WITHOUT_FH | OP_MODIFIES_SOMETHING, + .op_name = "OP_FREE_STATEID", + .op_rsize_bop = (nfsd4op_rsize)nfsd4_only_status_rsize, + }, }; static const char *nfsd4_op_name(unsigned opnum) @@ -1427,16 +1698,6 @@ static const char *nfsd4_op_name(unsigned opnum) #define nfsd4_voidres nfsd4_voidargs struct nfsd4_voidargs { int dummy; }; -/* - * TODO: At the present time, the NFSv4 server does not do XID caching - * of requests. Implementing XID caching would not be a serious problem, - * although it would require a mild change in interfaces since one - * doesn't know whether an NFSv4 request is idempotent until after the - * XDR decode. However, XID caching totally confuses pynfs (Peter - * Astrand's regression testsuite for NFSv4 servers), which reuses - * XID's liberally, so I've left it unimplemented until pynfs generates - * better XID's. - */ static struct svc_procedure nfsd_procedures4[2] = { [NFSPROC4_NULL] = { .pc_func = (svc_procfunc) nfsd4_proc_null, @@ -1452,6 +1713,7 @@ static struct svc_procedure nfsd_procedures4[2] = { .pc_encode = (kxdrproc_t) nfs4svc_encode_compoundres, .pc_argsize = sizeof(struct nfsd4_compoundargs), .pc_ressize = sizeof(struct nfsd4_compoundres), + .pc_release = nfsd4_release_compoundargs, .pc_cachetype = RC_NOCACHE, .pc_xdrressize = NFSD_BUFSIZE/4, }, diff --git a/fs/nfsd/nfs4recover.c b/fs/nfsd/nfs4recover.c index be26814..ed083b9 100644 --- a/fs/nfsd/nfs4recover.c +++ b/fs/nfsd/nfs4recover.c @@ -45,6 +45,7 @@ /* Globals */ static struct file *rec_file; +static char user_recovery_dirname[PATH_MAX] = "/var/lib/nfs/v4recovery"; static int nfs4_save_creds(const struct cred **original_creds) @@ -129,6 +130,7 @@ nfsd4_create_clid_dir(struct nfs4_client *clp) if (!rec_file || clp->cl_firststate) return 0; + clp->cl_firststate = 1; status = nfs4_save_creds(&original_cred); if (status < 0) return status; @@ -143,10 +145,8 @@ nfsd4_create_clid_dir(struct nfs4_client *clp) goto out_unlock; } status = -EEXIST; - if (dentry->d_inode) { - dprintk("NFSD: nfsd4_create_clid_dir: DIRECTORY EXISTS\n"); + if (dentry->d_inode) goto out_put; - } status = mnt_want_write(rec_file->f_path.mnt); if (status) goto out_put; @@ -156,12 +156,14 @@ out_put: dput(dentry); out_unlock: mutex_unlock(&dir->d_inode->i_mutex); - if (status == 0) { - clp->cl_firststate = 1; + if (status == 0) vfs_fsync(rec_file, 0); - } + else + printk(KERN_ERR "NFSD: failed to write recovery record" + " (err %d); please check that %s exists" + " and is writeable", status, + user_recovery_dirname); nfs4_reset_creds(original_cred); - dprintk("NFSD: nfsd4_create_clid_dir returns %d\n", status); return status; } @@ -191,52 +193,42 @@ nfsd4_build_namelist(void *arg, const char *name, int namlen, } static int -nfsd4_list_rec_dir(struct dentry *dir, recdir_func *f) +nfsd4_list_rec_dir(recdir_func *f) { const struct cred *original_cred; - struct file *filp; + struct dentry *dir = rec_file->f_path.dentry; LIST_HEAD(names); - struct name_list *entry; - struct dentry *dentry; int status; - if (!rec_file) - return 0; - status = nfs4_save_creds(&original_cred); if (status < 0) return status; - filp = dentry_open(dget(dir), mntget(rec_file->f_path.mnt), O_RDONLY, - current_cred()); - status = PTR_ERR(filp); - if (IS_ERR(filp)) - goto out; - status = vfs_readdir(filp, nfsd4_build_namelist, &names); - fput(filp); + status = vfs_llseek(rec_file, 0, SEEK_SET); + if (status < 0) { + nfs4_reset_creds(original_cred); + return status; + } + + status = vfs_readdir(rec_file, nfsd4_build_namelist, &names); mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT); while (!list_empty(&names)) { + struct name_list *entry; entry = list_entry(names.next, struct name_list, list); - - dentry = lookup_one_len(entry->name, dir, HEXDIR_LEN-1); - if (IS_ERR(dentry)) { - status = PTR_ERR(dentry); - break; + if (!status) { + struct dentry *dentry; + dentry = lookup_one_len(entry->name, dir, HEXDIR_LEN-1); + if (IS_ERR(dentry)) { + status = PTR_ERR(dentry); + break; + } + status = f(dir, dentry); + dput(dentry); } - status = f(dir, dentry); - dput(dentry); - if (status) - break; list_del(&entry->list); kfree(entry); } mutex_unlock(&dir->d_inode->i_mutex); -out: - while (!list_empty(&names)) { - entry = list_entry(names.next, struct name_list, list); - list_del(&entry->list); - kfree(entry); - } nfs4_reset_creds(original_cred); return status; } @@ -322,7 +314,7 @@ nfsd4_recdir_purge_old(void) { status = mnt_want_write(rec_file->f_path.mnt); if (status) goto out; - status = nfsd4_list_rec_dir(rec_file->f_path.dentry, purge_old); + status = nfsd4_list_rec_dir(purge_old); if (status == 0) vfs_fsync(rec_file, 0); mnt_drop_write(rec_file->f_path.mnt); @@ -352,7 +344,7 @@ nfsd4_recdir_load(void) { if (!rec_file) return 0; - status = nfsd4_list_rec_dir(rec_file->f_path.dentry, load_recdir); + status = nfsd4_list_rec_dir(load_recdir); if (status) printk("nfsd4: failed loading clients from recovery" " directory %s\n", rec_file->f_path.dentry->d_name.name); @@ -364,13 +356,13 @@ nfsd4_recdir_load(void) { */ void -nfsd4_init_recdir(char *rec_dirname) +nfsd4_init_recdir() { const struct cred *original_cred; int status; printk("NFSD: Using %s as the NFSv4 state recovery directory\n", - rec_dirname); + user_recovery_dirname); BUG_ON(rec_file); @@ -382,10 +374,10 @@ nfsd4_init_recdir(char *rec_dirname) return; } - rec_file = filp_open(rec_dirname, O_RDONLY | O_DIRECTORY, 0); + rec_file = filp_open(user_recovery_dirname, O_RDONLY | O_DIRECTORY, 0); if (IS_ERR(rec_file)) { printk("NFSD: unable to find recovery directory %s\n", - rec_dirname); + user_recovery_dirname); rec_file = NULL; } @@ -400,3 +392,30 @@ nfsd4_shutdown_recdir(void) fput(rec_file); rec_file = NULL; } + +/* + * Change the NFSv4 recovery directory to recdir. + */ +int +nfs4_reset_recoverydir(char *recdir) +{ + int status; + struct path path; + + status = kern_path(recdir, LOOKUP_FOLLOW, &path); + if (status) + return status; + status = -ENOTDIR; + if (S_ISDIR(path.dentry->d_inode->i_mode)) { + strcpy(user_recovery_dirname, recdir); + status = 0; + } + path_put(&path); + return status; +} + +char * +nfs4_recoverydir(void) +{ + return user_recovery_dirname; +} diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 4ec38df..d455ea0 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -37,6 +37,7 @@ #include <linux/slab.h> #include <linux/namei.h> #include <linux/swap.h> +#include <linux/pagemap.h> #include <linux/sunrpc/svcauth_gss.h> #include <linux/sunrpc/clnt.h> #include "xdr4.h" @@ -48,9 +49,6 @@ time_t nfsd4_lease = 90; /* default lease time */ time_t nfsd4_grace = 90; static time_t boot_time; -static u32 current_ownerid = 1; -static u32 current_fileid = 1; -static u32 current_delegid = 1; static stateid_t zerostateid; /* bits all 0 */ static stateid_t onestateid; /* bits all 1 */ static u64 current_sessionid = 1; @@ -59,10 +57,7 @@ static u64 current_sessionid = 1; #define ONE_STATEID(stateid) (!memcmp((stateid), &onestateid, sizeof(stateid_t))) /* forward declarations */ -static struct nfs4_stateid * find_stateid(stateid_t *stid, int flags); -static struct nfs4_delegation * find_delegation_stateid(struct inode *ino, stateid_t *stid); -static char user_recovery_dirname[PATH_MAX] = "/var/lib/nfs/v4recovery"; -static void nfs4_set_recdir(char *recdir); +static int check_for_locks(struct nfs4_file *filp, struct nfs4_lockowner *lowner); /* Locking: */ @@ -76,7 +71,8 @@ static DEFINE_MUTEX(client_mutex); */ static DEFINE_SPINLOCK(recall_lock); -static struct kmem_cache *stateowner_slab = NULL; +static struct kmem_cache *openowner_slab = NULL; +static struct kmem_cache *lockowner_slab = NULL; static struct kmem_cache *file_slab = NULL; static struct kmem_cache *stateid_slab = NULL; static struct kmem_cache *deleg_slab = NULL; @@ -108,6 +104,11 @@ opaque_hashval(const void *ptr, int nbytes) static struct list_head del_recall_lru; +static void nfsd4_free_file(struct nfs4_file *f) +{ + kmem_cache_free(file_slab, f); +} + static inline void put_nfs4_file(struct nfs4_file *fi) { @@ -115,7 +116,7 @@ put_nfs4_file(struct nfs4_file *fi) list_del(&fi->fi_hash); spin_unlock(&recall_lock); iput(fi->fi_inode); - kmem_cache_free(file_slab, fi); + nfsd4_free_file(fi); } } @@ -132,35 +133,33 @@ unsigned int max_delegations; * Open owner state (share locks) */ -/* hash tables for nfs4_stateowner */ -#define OWNER_HASH_BITS 8 -#define OWNER_HASH_SIZE (1 << OWNER_HASH_BITS) -#define OWNER_HASH_MASK (OWNER_HASH_SIZE - 1) +/* hash tables for open owners */ +#define OPEN_OWNER_HASH_BITS 8 +#define OPEN_OWNER_HASH_SIZE (1 << OPEN_OWNER_HASH_BITS) +#define OPEN_OWNER_HASH_MASK (OPEN_OWNER_HASH_SIZE - 1) -#define ownerid_hashval(id) \ - ((id) & OWNER_HASH_MASK) -#define ownerstr_hashval(clientid, ownername) \ - (((clientid) + opaque_hashval((ownername.data), (ownername.len))) & OWNER_HASH_MASK) +static unsigned int open_ownerstr_hashval(u32 clientid, struct xdr_netobj *ownername) +{ + unsigned int ret; -static struct list_head ownerid_hashtbl[OWNER_HASH_SIZE]; -static struct list_head ownerstr_hashtbl[OWNER_HASH_SIZE]; + ret = opaque_hashval(ownername->data, ownername->len); + ret += clientid; + return ret & OPEN_OWNER_HASH_MASK; +} + +static struct list_head open_ownerstr_hashtbl[OPEN_OWNER_HASH_SIZE]; /* hash table for nfs4_file */ #define FILE_HASH_BITS 8 #define FILE_HASH_SIZE (1 << FILE_HASH_BITS) -/* hash table for (open)nfs4_stateid */ -#define STATEID_HASH_BITS 10 -#define STATEID_HASH_SIZE (1 << STATEID_HASH_BITS) -#define STATEID_HASH_MASK (STATEID_HASH_SIZE - 1) - -#define file_hashval(x) \ - hash_ptr(x, FILE_HASH_BITS) -#define stateid_hashval(owner_id, file_id) \ - (((owner_id) + (file_id)) & STATEID_HASH_MASK) +static unsigned int file_hashval(struct inode *ino) +{ + /* XXX: why are we hashing on inode pointer, anyway? */ + return hash_ptr(ino, FILE_HASH_BITS); +} static struct list_head file_hashtbl[FILE_HASH_SIZE]; -static struct list_head stateid_hashtbl[STATEID_HASH_SIZE]; static void __nfs4_file_get_access(struct nfs4_file *fp, int oflag) { @@ -203,8 +202,73 @@ static void nfs4_file_put_access(struct nfs4_file *fp, int oflag) __nfs4_file_put_access(fp, oflag); } +static inline int get_new_stid(struct nfs4_stid *stid) +{ + static int min_stateid = 0; + struct idr *stateids = &stid->sc_client->cl_stateids; + int new_stid; + int error; + + error = idr_get_new_above(stateids, stid, min_stateid, &new_stid); + /* + * Note: the necessary preallocation was done in + * nfs4_alloc_stateid(). The idr code caps the number of + * preallocations that can exist at a time, but the state lock + * prevents anyone from using ours before we get here: + */ + BUG_ON(error); + /* + * It shouldn't be a problem to reuse an opaque stateid value. + * I don't think it is for 4.1. But with 4.0 I worry that, for + * example, a stray write retransmission could be accepted by + * the server when it should have been rejected. Therefore, + * adopt a trick from the sctp code to attempt to maximize the + * amount of time until an id is reused, by ensuring they always + * "increase" (mod INT_MAX): + */ + + min_stateid = new_stid+1; + if (min_stateid == INT_MAX) + min_stateid = 0; + return new_stid; +} + +static void init_stid(struct nfs4_stid *stid, struct nfs4_client *cl, unsigned char type) +{ + stateid_t *s = &stid->sc_stateid; + int new_id; + + stid->sc_type = type; + stid->sc_client = cl; + s->si_opaque.so_clid = cl->cl_clientid; + new_id = get_new_stid(stid); + s->si_opaque.so_id = (u32)new_id; + /* Will be incremented before return to client: */ + s->si_generation = 0; +} + +static struct nfs4_stid *nfs4_alloc_stid(struct nfs4_client *cl, struct kmem_cache *slab) +{ + struct idr *stateids = &cl->cl_stateids; + + if (!idr_pre_get(stateids, GFP_KERNEL)) + return NULL; + /* + * Note: if we fail here (or any time between now and the time + * we actually get the new idr), we won't need to undo the idr + * preallocation, since the idr code caps the number of + * preallocated entries. + */ + return kmem_cache_alloc(slab, GFP_KERNEL); +} + +static struct nfs4_ol_stateid * nfs4_alloc_stateid(struct nfs4_client *clp) +{ + return openlockstateid(nfs4_alloc_stid(clp, stateid_slab)); +} + static struct nfs4_delegation * -alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_fh *current_fh, u32 type) +alloc_init_deleg(struct nfs4_client *clp, struct nfs4_ol_stateid *stp, struct svc_fh *current_fh, u32 type) { struct nfs4_delegation *dp; struct nfs4_file *fp = stp->st_file; @@ -221,21 +285,23 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f return NULL; if (num_delegations > max_delegations) return NULL; - dp = kmem_cache_alloc(deleg_slab, GFP_KERNEL); + dp = delegstateid(nfs4_alloc_stid(clp, deleg_slab)); if (dp == NULL) return dp; + init_stid(&dp->dl_stid, clp, NFS4_DELEG_STID); + /* + * delegation seqid's are never incremented. The 4.1 special + * meaning of seqid 0 isn't meaningful, really, but let's avoid + * 0 anyway just for consistency and use 1: + */ + dp->dl_stid.sc_stateid.si_generation = 1; num_delegations++; INIT_LIST_HEAD(&dp->dl_perfile); INIT_LIST_HEAD(&dp->dl_perclnt); INIT_LIST_HEAD(&dp->dl_recall_lru); - dp->dl_client = clp; get_nfs4_file(fp); dp->dl_file = fp; dp->dl_type = type; - dp->dl_stateid.si_boot = boot_time; - dp->dl_stateid.si_stateownerid = current_delegid++; - dp->dl_stateid.si_fileid = 0; - dp->dl_stateid.si_generation = 0; fh_copy_shallow(&dp->dl_fh, ¤t_fh->fh_handle); dp->dl_time = 0; atomic_set(&dp->dl_count, 1); @@ -264,12 +330,29 @@ static void nfs4_put_deleg_lease(struct nfs4_file *fp) } } +static void unhash_stid(struct nfs4_stid *s) +{ + struct idr *stateids = &s->sc_client->cl_stateids; + + idr_remove(stateids, s->sc_stateid.si_opaque.so_id); +} + +static void +hash_delegation_locked(struct nfs4_delegation *dp, struct nfs4_file *fp) +{ + lockdep_assert_held(&recall_lock); + + list_add(&dp->dl_perfile, &fp->fi_delegations); + list_add(&dp->dl_perclnt, &dp->dl_stid.sc_client->cl_delegations); +} + /* Called under the state lock. */ static void unhash_delegation(struct nfs4_delegation *dp) { - list_del_init(&dp->dl_perclnt); + unhash_stid(&dp->dl_stid); spin_lock(&recall_lock); + list_del_init(&dp->dl_perclnt); list_del_init(&dp->dl_perfile); list_del_init(&dp->dl_recall_lru); spin_unlock(&recall_lock); @@ -289,10 +372,16 @@ static DEFINE_SPINLOCK(client_lock); #define CLIENT_HASH_SIZE (1 << CLIENT_HASH_BITS) #define CLIENT_HASH_MASK (CLIENT_HASH_SIZE - 1) -#define clientid_hashval(id) \ - ((id) & CLIENT_HASH_MASK) -#define clientstr_hashval(name) \ - (opaque_hashval((name), 8) & CLIENT_HASH_MASK) +static unsigned int clientid_hashval(u32 id) +{ + return id & CLIENT_HASH_MASK; +} + +static unsigned int clientstr_hashval(const char *name) +{ + return opaque_hashval(name, 8) & CLIENT_HASH_MASK; +} + /* * reclaim_str_hashtbl[] holds known client info from previous reset/reboot * used in reboot/reset lease grace period processing @@ -359,7 +448,7 @@ set_deny(unsigned int *deny, unsigned long bmap) { } static int -test_share(struct nfs4_stateid *stp, struct nfsd4_open *open) { +test_share(struct nfs4_ol_stateid *stp, struct nfsd4_open *open) { unsigned int access, deny; set_access(&access, stp->st_access_bmap); @@ -382,14 +471,13 @@ static int nfs4_access_to_omode(u32 access) BUG(); } -static void unhash_generic_stateid(struct nfs4_stateid *stp) +static void unhash_generic_stateid(struct nfs4_ol_stateid *stp) { - list_del(&stp->st_hash); list_del(&stp->st_perfile); list_del(&stp->st_perstateowner); } -static void free_generic_stateid(struct nfs4_stateid *stp) +static void close_generic_stateid(struct nfs4_ol_stateid *stp) { int i; @@ -398,84 +486,106 @@ static void free_generic_stateid(struct nfs4_stateid *stp) if (test_bit(i, &stp->st_access_bmap)) nfs4_file_put_access(stp->st_file, nfs4_access_to_omode(i)); + __clear_bit(i, &stp->st_access_bmap); } } put_nfs4_file(stp->st_file); + stp->st_file = NULL; +} + +static void free_generic_stateid(struct nfs4_ol_stateid *stp) +{ kmem_cache_free(stateid_slab, stp); } -static void release_lock_stateid(struct nfs4_stateid *stp) +static void release_lock_stateid(struct nfs4_ol_stateid *stp) { struct file *file; unhash_generic_stateid(stp); + unhash_stid(&stp->st_stid); file = find_any_file(stp->st_file); if (file) - locks_remove_posix(file, (fl_owner_t)stp->st_stateowner); + locks_remove_posix(file, (fl_owner_t)lockowner(stp->st_stateowner)); + close_generic_stateid(stp); free_generic_stateid(stp); } -static void unhash_lockowner(struct nfs4_stateowner *sop) +static void unhash_lockowner(struct nfs4_lockowner *lo) { - struct nfs4_stateid *stp; + struct nfs4_ol_stateid *stp; - list_del(&sop->so_idhash); - list_del(&sop->so_strhash); - list_del(&sop->so_perstateid); - while (!list_empty(&sop->so_stateids)) { - stp = list_first_entry(&sop->so_stateids, - struct nfs4_stateid, st_perstateowner); + list_del(&lo->lo_owner.so_strhash); + list_del(&lo->lo_perstateid); + while (!list_empty(&lo->lo_owner.so_stateids)) { + stp = list_first_entry(&lo->lo_owner.so_stateids, + struct nfs4_ol_stateid, st_perstateowner); release_lock_stateid(stp); } } -static void release_lockowner(struct nfs4_stateowner *sop) +static void release_lockowner(struct nfs4_lockowner *lo) { - unhash_lockowner(sop); - nfs4_put_stateowner(sop); + unhash_lockowner(lo); + nfs4_free_lockowner(lo); } static void -release_stateid_lockowners(struct nfs4_stateid *open_stp) +release_stateid_lockowners(struct nfs4_ol_stateid *open_stp) { - struct nfs4_stateowner *lock_sop; + struct nfs4_lockowner *lo; while (!list_empty(&open_stp->st_lockowners)) { - lock_sop = list_entry(open_stp->st_lockowners.next, - struct nfs4_stateowner, so_perstateid); - /* list_del(&open_stp->st_lockowners); */ - BUG_ON(lock_sop->so_is_open_owner); - release_lockowner(lock_sop); + lo = list_entry(open_stp->st_lockowners.next, + struct nfs4_lockowner, lo_perstateid); + release_lockowner(lo); } } -static void release_open_stateid(struct nfs4_stateid *stp) +static void unhash_open_stateid(struct nfs4_ol_stateid *stp) { unhash_generic_stateid(stp); release_stateid_lockowners(stp); + close_generic_stateid(stp); +} + +static void release_open_stateid(struct nfs4_ol_stateid *stp) +{ + unhash_open_stateid(stp); + unhash_stid(&stp->st_stid); free_generic_stateid(stp); } -static void unhash_openowner(struct nfs4_stateowner *sop) +static void unhash_openowner(struct nfs4_openowner *oo) { - struct nfs4_stateid *stp; + struct nfs4_ol_stateid *stp; - list_del(&sop->so_idhash); - list_del(&sop->so_strhash); - list_del(&sop->so_perclient); - list_del(&sop->so_perstateid); /* XXX: necessary? */ - while (!list_empty(&sop->so_stateids)) { - stp = list_first_entry(&sop->so_stateids, - struct nfs4_stateid, st_perstateowner); + list_del(&oo->oo_owner.so_strhash); + list_del(&oo->oo_perclient); + while (!list_empty(&oo->oo_owner.so_stateids)) { + stp = list_first_entry(&oo->oo_owner.so_stateids, + struct nfs4_ol_stateid, st_perstateowner); release_open_stateid(stp); } } -static void release_openowner(struct nfs4_stateowner *sop) +static void release_last_closed_stateid(struct nfs4_openowner *oo) { - unhash_openowner(sop); - list_del(&sop->so_close_lru); - nfs4_put_stateowner(sop); + struct nfs4_ol_stateid *s = oo->oo_last_closed_stid; + + if (s) { + unhash_stid(&s->st_stid); + free_generic_stateid(s); + oo->oo_last_closed_stid = NULL; + } +} + +static void release_openowner(struct nfs4_openowner *oo) +{ + unhash_openowner(oo); + list_del(&oo->oo_close_lru); + release_last_closed_stateid(oo); + nfs4_free_openowner(oo); } #define SESSION_HASH_SIZE 512 @@ -840,9 +950,6 @@ renew_client_locked(struct nfs4_client *clp) return; } - /* - * Move client to the end to the LRU list. - */ dprintk("renewing client (clientid %08x/%08x)\n", clp->cl_clientid.cl_boot, clp->cl_clientid.cl_id); @@ -888,6 +995,18 @@ static struct nfs4_client *alloc_client(struct xdr_netobj name) } memcpy(clp->cl_name.data, name.data, name.len); clp->cl_name.len = name.len; + INIT_LIST_HEAD(&clp->cl_sessions); + idr_init(&clp->cl_stateids); + atomic_set(&clp->cl_refcount, 0); + clp->cl_cb_state = NFSD4_CB_UNKNOWN; + INIT_LIST_HEAD(&clp->cl_idhash); + INIT_LIST_HEAD(&clp->cl_strhash); + INIT_LIST_HEAD(&clp->cl_openowners); + INIT_LIST_HEAD(&clp->cl_delegations); + INIT_LIST_HEAD(&clp->cl_lru); + INIT_LIST_HEAD(&clp->cl_callbacks); + spin_lock_init(&clp->cl_lock); + rpc_init_wait_queue(&clp->cl_cb_waitq, "Backchannel slot table"); return clp; } @@ -901,10 +1020,13 @@ free_client(struct nfs4_client *clp) list_del(&ses->se_perclnt); nfsd4_put_session(ses); } + rpc_destroy_wait_queue(&clp->cl_cb_waitq); if (clp->cl_cred.cr_group_info) put_group_info(clp->cl_cred.cr_group_info); kfree(clp->cl_principal); kfree(clp->cl_name.data); + idr_remove_all(&clp->cl_stateids); + idr_destroy(&clp->cl_stateids); kfree(clp); } @@ -940,7 +1062,7 @@ unhash_client_locked(struct nfs4_client *clp) static void expire_client(struct nfs4_client *clp) { - struct nfs4_stateowner *sop; + struct nfs4_openowner *oo; struct nfs4_delegation *dp; struct list_head reaplist; @@ -958,8 +1080,8 @@ expire_client(struct nfs4_client *clp) unhash_delegation(dp); } while (!list_empty(&clp->cl_openowners)) { - sop = list_entry(clp->cl_openowners.next, struct nfs4_stateowner, so_perclient); - release_openowner(sop); + oo = list_entry(clp->cl_openowners.next, struct nfs4_openowner, oo_perclient); + release_openowner(oo); } nfsd4_shutdown_callback(clp); if (clp->cl_cb_conn.cb_xprt) @@ -1035,6 +1157,23 @@ static void gen_confirm(struct nfs4_client *clp) *p++ = i++; } +static struct nfs4_stid *find_stateid(struct nfs4_client *cl, stateid_t *t) +{ + return idr_find(&cl->cl_stateids, t->si_opaque.so_id); +} + +static struct nfs4_stid *find_stateid_by_type(struct nfs4_client *cl, stateid_t *t, char typemask) +{ + struct nfs4_stid *s; + + s = find_stateid(cl, t); + if (!s) + return NULL; + if (typemask & s->sc_type) + return s; + return NULL; +} + static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir, struct svc_rqst *rqstp, nfs4_verifier *verf) { @@ -1046,7 +1185,6 @@ static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir, if (clp == NULL) return NULL; - INIT_LIST_HEAD(&clp->cl_sessions); princ = svc_gss_principal(rqstp); if (princ) { @@ -1058,19 +1196,9 @@ static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir, } memcpy(clp->cl_recdir, recdir, HEXDIR_LEN); - atomic_set(&clp->cl_refcount, 0); - clp->cl_cb_state = NFSD4_CB_UNKNOWN; - INIT_LIST_HEAD(&clp->cl_idhash); - INIT_LIST_HEAD(&clp->cl_strhash); - INIT_LIST_HEAD(&clp->cl_openowners); - INIT_LIST_HEAD(&clp->cl_delegations); - INIT_LIST_HEAD(&clp->cl_lru); - INIT_LIST_HEAD(&clp->cl_callbacks); - spin_lock_init(&clp->cl_lock); INIT_WORK(&clp->cl_cb_null.cb_work, nfsd4_do_callback_rpc); clp->cl_time = get_seconds(); clear_bit(0, &clp->cl_cb_slot_busy); - rpc_init_wait_queue(&clp->cl_cb_waitq, "Backchannel slot table"); copy_verf(clp, verf); rpc_copy_addr((struct sockaddr *) &clp->cl_addr, sa); clp->cl_flavor = rqstp->rq_flavor; @@ -1080,17 +1208,6 @@ static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir, return clp; } -static int check_name(struct xdr_netobj name) -{ - if (name.len == 0) - return 0; - if (name.len > NFS4_OPAQUE_LIMIT) { - dprintk("NFSD: check_name: name too long(%d)!\n", name.len); - return 0; - } - return 1; -} - static void add_to_unconfirmed(struct nfs4_client *clp, unsigned int strhashval) { @@ -1122,8 +1239,10 @@ find_confirmed_client(clientid_t *clid) unsigned int idhashval = clientid_hashval(clid->cl_id); list_for_each_entry(clp, &conf_id_hashtbl[idhashval], cl_idhash) { - if (same_clid(&clp->cl_clientid, clid)) + if (same_clid(&clp->cl_clientid, clid)) { + renew_client(clp); return clp; + } } return NULL; } @@ -1170,20 +1289,6 @@ find_unconfirmed_client_by_str(const char *dname, unsigned int hashval) return NULL; } -static void rpc_svcaddr2sockaddr(struct sockaddr *sa, unsigned short family, union svc_addr_u *svcaddr) -{ - switch (family) { - case AF_INET: - ((struct sockaddr_in *)sa)->sin_family = AF_INET; - ((struct sockaddr_in *)sa)->sin_addr = svcaddr->addr; - return; - case AF_INET6: - ((struct sockaddr_in6 *)sa)->sin6_family = AF_INET6; - ((struct sockaddr_in6 *)sa)->sin6_addr = svcaddr->addr6; - return; - } -} - static void gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se, struct svc_rqst *rqstp) { @@ -1215,7 +1320,7 @@ gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se, struct svc_r conn->cb_prog = se->se_callback_prog; conn->cb_ident = se->se_callback_ident; - rpc_svcaddr2sockaddr((struct sockaddr *)&conn->cb_saddr, expected_family, &rqstp->rq_daddr); + memcpy(&conn->cb_saddr, &rqstp->rq_daddr, rqstp->rq_daddrlen); return; out_err: conn->cb_addr.ss_family = AF_UNSPEC; @@ -1347,7 +1452,7 @@ nfsd4_exchange_id(struct svc_rqst *rqstp, __func__, rqstp, exid, exid->clname.len, exid->clname.data, addr_str, exid->flags, exid->spa_how); - if (!check_name(exid->clname) || (exid->flags & ~EXCHGID4_FLAG_MASK_A)) + if (exid->flags & ~EXCHGID4_FLAG_MASK_A) return nfserr_inval; /* Currently only support SP4_NONE */ @@ -1503,6 +1608,29 @@ nfsd4_replay_create_session(struct nfsd4_create_session *cr_ses, return slot->sl_status; } +#define NFSD_MIN_REQ_HDR_SEQ_SZ ((\ + 2 * 2 + /* credential,verifier: AUTH_NULL, length 0 */ \ + 1 + /* MIN tag is length with zero, only length */ \ + 3 + /* version, opcount, opcode */ \ + XDR_QUADLEN(NFS4_MAX_SESSIONID_LEN) + \ + /* seqid, slotID, slotID, cache */ \ + 4 ) * sizeof(__be32)) + +#define NFSD_MIN_RESP_HDR_SEQ_SZ ((\ + 2 + /* verifier: AUTH_NULL, length 0 */\ + 1 + /* status */ \ + 1 + /* MIN tag is length with zero, only length */ \ + 3 + /* opcount, opcode, opstatus*/ \ + XDR_QUADLEN(NFS4_MAX_SESSIONID_LEN) + \ + /* seqid, slotID, slotID, slotID, status */ \ + 5 ) * sizeof(__be32)) + +static __be32 check_forechannel_attrs(struct nfsd4_channel_attrs fchannel) +{ + return fchannel.maxreq_sz < NFSD_MIN_REQ_HDR_SEQ_SZ + || fchannel.maxresp_sz < NFSD_MIN_RESP_HDR_SEQ_SZ; +} + __be32 nfsd4_create_session(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, @@ -1571,6 +1699,10 @@ nfsd4_create_session(struct svc_rqst *rqstp, cr_ses->flags &= ~SESSION4_PERSIST; cr_ses->flags &= ~SESSION4_RDMA; + status = nfserr_toosmall; + if (check_forechannel_attrs(cr_ses->fore_channel)) + goto out; + status = nfserr_jukebox; new = alloc_init_session(rqstp, conf, cr_ses); if (!new) @@ -1732,6 +1864,14 @@ static bool nfsd4_session_too_many_ops(struct svc_rqst *rqstp, struct nfsd4_sess return args->opcnt > session->se_fchannel.maxops; } +static bool nfsd4_request_too_big(struct svc_rqst *rqstp, + struct nfsd4_session *session) +{ + struct xdr_buf *xb = &rqstp->rq_arg; + + return xb->len > session->se_fchannel.maxreq_sz; +} + __be32 nfsd4_sequence(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, @@ -1764,6 +1904,10 @@ nfsd4_sequence(struct svc_rqst *rqstp, if (nfsd4_session_too_many_ops(rqstp, session)) goto out; + status = nfserr_req_too_big; + if (nfsd4_request_too_big(rqstp, session)) + goto out; + status = nfserr_badslot; if (seq->slotid >= session->se_fchannel.maxreqs) goto out; @@ -1807,8 +1951,16 @@ out: nfsd4_get_session(cstate->session); atomic_inc(&clp->cl_refcount); - if (clp->cl_cb_state == NFSD4_CB_DOWN) - seq->status_flags |= SEQ4_STATUS_CB_PATH_DOWN; + switch (clp->cl_cb_state) { + case NFSD4_CB_DOWN: + seq->status_flags = SEQ4_STATUS_CB_PATH_DOWN; + break; + case NFSD4_CB_FAULT: + seq->status_flags = SEQ4_STATUS_BACKCHANNEL_FAULT; + break; + default: + seq->status_flags = 0; + } } kfree(conn); spin_unlock(&client_lock); @@ -1816,6 +1968,50 @@ out: return status; } +static inline bool has_resources(struct nfs4_client *clp) +{ + return !list_empty(&clp->cl_openowners) + || !list_empty(&clp->cl_delegations) + || !list_empty(&clp->cl_sessions); +} + +__be32 +nfsd4_destroy_clientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_destroy_clientid *dc) +{ + struct nfs4_client *conf, *unconf, *clp; + int status = 0; + + nfs4_lock_state(); + unconf = find_unconfirmed_client(&dc->clientid); + conf = find_confirmed_client(&dc->clientid); + + if (conf) { + clp = conf; + + if (!is_client_expired(conf) && has_resources(conf)) { + status = nfserr_clientid_busy; + goto out; + } + + /* rfc5661 18.50.3 */ + if (cstate->session && conf == cstate->session->se_client) { + status = nfserr_clientid_busy; + goto out; + } + } else if (unconf) + clp = unconf; + else { + status = nfserr_stale_clientid; + goto out; + } + + expire_client(clp); +out: + nfs4_unlock_state(); + dprintk("%s return %d\n", __func__, ntohl(status)); + return status; +} + __be32 nfsd4_reclaim_complete(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_reclaim_complete *rc) { @@ -1858,19 +2054,13 @@ __be32 nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_setclientid *setclid) { - struct xdr_netobj clname = { - .len = setclid->se_namelen, - .data = setclid->se_name, - }; + struct xdr_netobj clname = setclid->se_name; nfs4_verifier clverifier = setclid->se_verf; unsigned int strhashval; struct nfs4_client *conf, *unconf, *new; __be32 status; char dname[HEXDIR_LEN]; - if (!check_name(clname)) - return nfserr_inval; - status = nfs4_make_rec_clidname(dname, &clname); if (status) return status; @@ -2074,31 +2264,28 @@ out: return status; } +static struct nfs4_file *nfsd4_alloc_file(void) +{ + return kmem_cache_alloc(file_slab, GFP_KERNEL); +} + /* OPEN Share state helper functions */ -static inline struct nfs4_file * -alloc_init_file(struct inode *ino) +static void nfsd4_init_file(struct nfs4_file *fp, struct inode *ino) { - struct nfs4_file *fp; unsigned int hashval = file_hashval(ino); - fp = kmem_cache_alloc(file_slab, GFP_KERNEL); - if (fp) { - atomic_set(&fp->fi_ref, 1); - INIT_LIST_HEAD(&fp->fi_hash); - INIT_LIST_HEAD(&fp->fi_stateids); - INIT_LIST_HEAD(&fp->fi_delegations); - fp->fi_inode = igrab(ino); - fp->fi_id = current_fileid++; - fp->fi_had_conflict = false; - fp->fi_lease = NULL; - memset(fp->fi_fds, 0, sizeof(fp->fi_fds)); - memset(fp->fi_access, 0, sizeof(fp->fi_access)); - spin_lock(&recall_lock); - list_add(&fp->fi_hash, &file_hashtbl[hashval]); - spin_unlock(&recall_lock); - return fp; - } - return NULL; + atomic_set(&fp->fi_ref, 1); + INIT_LIST_HEAD(&fp->fi_hash); + INIT_LIST_HEAD(&fp->fi_stateids); + INIT_LIST_HEAD(&fp->fi_delegations); + fp->fi_inode = igrab(ino); + fp->fi_had_conflict = false; + fp->fi_lease = NULL; + memset(fp->fi_fds, 0, sizeof(fp->fi_fds)); + memset(fp->fi_access, 0, sizeof(fp->fi_access)); + spin_lock(&recall_lock); + list_add(&fp->fi_hash, &file_hashtbl[hashval]); + spin_unlock(&recall_lock); } static void @@ -2113,7 +2300,8 @@ nfsd4_free_slab(struct kmem_cache **slab) void nfsd4_free_slabs(void) { - nfsd4_free_slab(&stateowner_slab); + nfsd4_free_slab(&openowner_slab); + nfsd4_free_slab(&lockowner_slab); nfsd4_free_slab(&file_slab); nfsd4_free_slab(&stateid_slab); nfsd4_free_slab(&deleg_slab); @@ -2122,16 +2310,20 @@ nfsd4_free_slabs(void) static int nfsd4_init_slabs(void) { - stateowner_slab = kmem_cache_create("nfsd4_stateowners", - sizeof(struct nfs4_stateowner), 0, 0, NULL); - if (stateowner_slab == NULL) + openowner_slab = kmem_cache_create("nfsd4_openowners", + sizeof(struct nfs4_openowner), 0, 0, NULL); + if (openowner_slab == NULL) + goto out_nomem; + lockowner_slab = kmem_cache_create("nfsd4_lockowners", + sizeof(struct nfs4_lockowner), 0, 0, NULL); + if (lockowner_slab == NULL) goto out_nomem; file_slab = kmem_cache_create("nfsd4_files", sizeof(struct nfs4_file), 0, 0, NULL); if (file_slab == NULL) goto out_nomem; stateid_slab = kmem_cache_create("nfsd4_stateids", - sizeof(struct nfs4_stateid), 0, 0, NULL); + sizeof(struct nfs4_ol_stateid), 0, 0, NULL); if (stateid_slab == NULL) goto out_nomem; deleg_slab = kmem_cache_create("nfsd4_delegations", @@ -2145,97 +2337,94 @@ out_nomem: return -ENOMEM; } -void -nfs4_free_stateowner(struct kref *kref) +void nfs4_free_openowner(struct nfs4_openowner *oo) { - struct nfs4_stateowner *sop = - container_of(kref, struct nfs4_stateowner, so_ref); - kfree(sop->so_owner.data); - kmem_cache_free(stateowner_slab, sop); + kfree(oo->oo_owner.so_owner.data); + kmem_cache_free(openowner_slab, oo); } -static inline struct nfs4_stateowner * -alloc_stateowner(struct xdr_netobj *owner) +void nfs4_free_lockowner(struct nfs4_lockowner *lo) { - struct nfs4_stateowner *sop; + kfree(lo->lo_owner.so_owner.data); + kmem_cache_free(lockowner_slab, lo); +} - if ((sop = kmem_cache_alloc(stateowner_slab, GFP_KERNEL))) { - if ((sop->so_owner.data = kmalloc(owner->len, GFP_KERNEL))) { - memcpy(sop->so_owner.data, owner->data, owner->len); - sop->so_owner.len = owner->len; - kref_init(&sop->so_ref); - return sop; - } - kmem_cache_free(stateowner_slab, sop); - } - return NULL; +static void init_nfs4_replay(struct nfs4_replay *rp) +{ + rp->rp_status = nfserr_serverfault; + rp->rp_buflen = 0; + rp->rp_buf = rp->rp_ibuf; } -static struct nfs4_stateowner * -alloc_init_open_stateowner(unsigned int strhashval, struct nfs4_client *clp, struct nfsd4_open *open) { +static inline void *alloc_stateowner(struct kmem_cache *slab, struct xdr_netobj *owner, struct nfs4_client *clp) +{ struct nfs4_stateowner *sop; - struct nfs4_replay *rp; - unsigned int idhashval; - if (!(sop = alloc_stateowner(&open->op_owner))) + sop = kmem_cache_alloc(slab, GFP_KERNEL); + if (!sop) + return NULL; + + sop->so_owner.data = kmemdup(owner->data, owner->len, GFP_KERNEL); + if (!sop->so_owner.data) { + kmem_cache_free(slab, sop); return NULL; - idhashval = ownerid_hashval(current_ownerid); - INIT_LIST_HEAD(&sop->so_idhash); - INIT_LIST_HEAD(&sop->so_strhash); - INIT_LIST_HEAD(&sop->so_perclient); + } + sop->so_owner.len = owner->len; + INIT_LIST_HEAD(&sop->so_stateids); - INIT_LIST_HEAD(&sop->so_perstateid); /* not used */ - INIT_LIST_HEAD(&sop->so_close_lru); - sop->so_time = 0; - list_add(&sop->so_idhash, &ownerid_hashtbl[idhashval]); - list_add(&sop->so_strhash, &ownerstr_hashtbl[strhashval]); - list_add(&sop->so_perclient, &clp->cl_openowners); - sop->so_is_open_owner = 1; - sop->so_id = current_ownerid++; sop->so_client = clp; - sop->so_seqid = open->op_seqid; - sop->so_confirmed = 0; - rp = &sop->so_replay; - rp->rp_status = nfserr_serverfault; - rp->rp_buflen = 0; - rp->rp_buf = rp->rp_ibuf; + init_nfs4_replay(&sop->so_replay); return sop; } -static inline void -init_stateid(struct nfs4_stateid *stp, struct nfs4_file *fp, struct nfsd4_open *open) { - struct nfs4_stateowner *sop = open->op_stateowner; - unsigned int hashval = stateid_hashval(sop->so_id, fp->fi_id); +static void hash_openowner(struct nfs4_openowner *oo, struct nfs4_client *clp, unsigned int strhashval) +{ + list_add(&oo->oo_owner.so_strhash, &open_ownerstr_hashtbl[strhashval]); + list_add(&oo->oo_perclient, &clp->cl_openowners); +} - INIT_LIST_HEAD(&stp->st_hash); - INIT_LIST_HEAD(&stp->st_perstateowner); +static struct nfs4_openowner * +alloc_init_open_stateowner(unsigned int strhashval, struct nfs4_client *clp, struct nfsd4_open *open) { + struct nfs4_openowner *oo; + + oo = alloc_stateowner(openowner_slab, &open->op_owner, clp); + if (!oo) + return NULL; + oo->oo_owner.so_is_open_owner = 1; + oo->oo_owner.so_seqid = open->op_seqid; + oo->oo_flags = NFS4_OO_NEW; + oo->oo_time = 0; + oo->oo_last_closed_stid = NULL; + INIT_LIST_HEAD(&oo->oo_close_lru); + hash_openowner(oo, clp, strhashval); + return oo; +} + +static void init_open_stateid(struct nfs4_ol_stateid *stp, struct nfs4_file *fp, struct nfsd4_open *open) { + struct nfs4_openowner *oo = open->op_openowner; + struct nfs4_client *clp = oo->oo_owner.so_client; + + init_stid(&stp->st_stid, clp, NFS4_OPEN_STID); INIT_LIST_HEAD(&stp->st_lockowners); - INIT_LIST_HEAD(&stp->st_perfile); - list_add(&stp->st_hash, &stateid_hashtbl[hashval]); - list_add(&stp->st_perstateowner, &sop->so_stateids); + list_add(&stp->st_perstateowner, &oo->oo_owner.so_stateids); list_add(&stp->st_perfile, &fp->fi_stateids); - stp->st_stateowner = sop; + stp->st_stateowner = &oo->oo_owner; get_nfs4_file(fp); stp->st_file = fp; - stp->st_stateid.si_boot = boot_time; - stp->st_stateid.si_stateownerid = sop->so_id; - stp->st_stateid.si_fileid = fp->fi_id; - stp->st_stateid.si_generation = 0; stp->st_access_bmap = 0; stp->st_deny_bmap = 0; - __set_bit(open->op_share_access & ~NFS4_SHARE_WANT_MASK, - &stp->st_access_bmap); + __set_bit(open->op_share_access, &stp->st_access_bmap); __set_bit(open->op_share_deny, &stp->st_deny_bmap); stp->st_openstp = NULL; } static void -move_to_close_lru(struct nfs4_stateowner *sop) +move_to_close_lru(struct nfs4_openowner *oo) { - dprintk("NFSD: move_to_close_lru nfs4_stateowner %p\n", sop); + dprintk("NFSD: move_to_close_lru nfs4_openowner %p\n", oo); - list_move_tail(&sop->so_close_lru, &close_lru); - sop->so_time = get_seconds(); + list_move_tail(&oo->oo_close_lru, &close_lru); + oo->oo_time = get_seconds(); } static int @@ -2247,14 +2436,18 @@ same_owner_str(struct nfs4_stateowner *sop, struct xdr_netobj *owner, (sop->so_client->cl_clientid.cl_id == clid->cl_id); } -static struct nfs4_stateowner * +static struct nfs4_openowner * find_openstateowner_str(unsigned int hashval, struct nfsd4_open *open) { - struct nfs4_stateowner *so = NULL; + struct nfs4_stateowner *so; + struct nfs4_openowner *oo; - list_for_each_entry(so, &ownerstr_hashtbl[hashval], so_strhash) { - if (same_owner_str(so, &open->op_owner, &open->op_clientid)) - return so; + list_for_each_entry(so, &open_ownerstr_hashtbl[hashval], so_strhash) { + if (same_owner_str(so, &open->op_owner, &open->op_clientid)) { + oo = openowner(so); + renew_client(oo->oo_owner.so_client); + return oo; + } } return NULL; } @@ -2278,31 +2471,6 @@ find_file(struct inode *ino) return NULL; } -static inline int access_valid(u32 x, u32 minorversion) -{ - if ((x & NFS4_SHARE_ACCESS_MASK) < NFS4_SHARE_ACCESS_READ) - return 0; - if ((x & NFS4_SHARE_ACCESS_MASK) > NFS4_SHARE_ACCESS_BOTH) - return 0; - x &= ~NFS4_SHARE_ACCESS_MASK; - if (minorversion && x) { - if ((x & NFS4_SHARE_WANT_MASK) > NFS4_SHARE_WANT_CANCEL) - return 0; - if ((x & NFS4_SHARE_WHEN_MASK) > NFS4_SHARE_PUSH_DELEG_WHEN_UNCONTENDED) - return 0; - x &= ~(NFS4_SHARE_WANT_MASK | NFS4_SHARE_WHEN_MASK); - } - if (x) - return 0; - return 1; -} - -static inline int deny_valid(u32 x) -{ - /* Note: unlike access bits, deny bits may be zero. */ - return x <= NFS4_SHARE_DENY_BOTH; -} - /* * Called to check deny when READ with all zero stateid or * WRITE with all zero or all one stateid @@ -2312,7 +2480,7 @@ nfs4_share_conflict(struct svc_fh *current_fh, unsigned int deny_type) { struct inode *ino = current_fh->fh_dentry->d_inode; struct nfs4_file *fp; - struct nfs4_stateid *stp; + struct nfs4_ol_stateid *stp; __be32 ret; dprintk("NFSD: nfs4_share_conflict\n"); @@ -2383,10 +2551,20 @@ int nfsd_change_deleg_cb(struct file_lock **onlist, int arg) } static const struct lock_manager_operations nfsd_lease_mng_ops = { - .fl_break = nfsd_break_deleg_cb, - .fl_change = nfsd_change_deleg_cb, + .lm_break = nfsd_break_deleg_cb, + .lm_change = nfsd_change_deleg_cb, }; +static __be32 nfsd4_check_seqid(struct nfsd4_compound_state *cstate, struct nfs4_stateowner *so, u32 seqid) +{ + if (nfsd4_has_session(cstate)) + return nfs_ok; + if (seqid == so->so_seqid - 1) + return nfserr_replay_me; + if (seqid == so->so_seqid) + return nfs_ok; + return nfserr_bad_seqid; +} __be32 nfsd4_process_open1(struct nfsd4_compound_state *cstate, @@ -2395,57 +2573,49 @@ nfsd4_process_open1(struct nfsd4_compound_state *cstate, clientid_t *clientid = &open->op_clientid; struct nfs4_client *clp = NULL; unsigned int strhashval; - struct nfs4_stateowner *sop = NULL; - - if (!check_name(open->op_owner)) - return nfserr_inval; + struct nfs4_openowner *oo = NULL; + __be32 status; if (STALE_CLIENTID(&open->op_clientid)) return nfserr_stale_clientid; + /* + * In case we need it later, after we've already created the + * file and don't want to risk a further failure: + */ + open->op_file = nfsd4_alloc_file(); + if (open->op_file == NULL) + return nfserr_jukebox; - strhashval = ownerstr_hashval(clientid->cl_id, open->op_owner); - sop = find_openstateowner_str(strhashval, open); - open->op_stateowner = sop; - if (!sop) { - /* Make sure the client's lease hasn't expired. */ + strhashval = open_ownerstr_hashval(clientid->cl_id, &open->op_owner); + oo = find_openstateowner_str(strhashval, open); + open->op_openowner = oo; + if (!oo) { clp = find_confirmed_client(clientid); if (clp == NULL) return nfserr_expired; - goto renew; + goto new_owner; } - /* When sessions are used, skip open sequenceid processing */ - if (nfsd4_has_session(cstate)) - goto renew; - if (!sop->so_confirmed) { + if (!(oo->oo_flags & NFS4_OO_CONFIRMED)) { /* Replace unconfirmed owners without checking for replay. */ - clp = sop->so_client; - release_openowner(sop); - open->op_stateowner = NULL; - goto renew; - } - if (open->op_seqid == sop->so_seqid - 1) { - if (sop->so_replay.rp_buflen) - return nfserr_replay_me; - /* The original OPEN failed so spectacularly - * that we don't even have replay data saved! - * Therefore, we have no choice but to continue - * processing this OPEN; presumably, we'll - * fail again for the same reason. - */ - dprintk("nfsd4_process_open1: replay with no replay cache\n"); - goto renew; - } - if (open->op_seqid != sop->so_seqid) - return nfserr_bad_seqid; -renew: - if (open->op_stateowner == NULL) { - sop = alloc_init_open_stateowner(strhashval, clp, open); - if (sop == NULL) - return nfserr_jukebox; - open->op_stateowner = sop; + clp = oo->oo_owner.so_client; + release_openowner(oo); + open->op_openowner = NULL; + goto new_owner; } - list_del_init(&sop->so_close_lru); - renew_client(sop->so_client); + status = nfsd4_check_seqid(cstate, &oo->oo_owner, open->op_seqid); + if (status) + return status; + clp = oo->oo_owner.so_client; + goto alloc_stateid; +new_owner: + oo = alloc_init_open_stateowner(strhashval, clp, open); + if (oo == NULL) + return nfserr_jukebox; + open->op_openowner = oo; +alloc_stateid: + open->op_stp = nfs4_alloc_stateid(clp); + if (!open->op_stp) + return nfserr_jukebox; return nfs_ok; } @@ -2458,36 +2628,37 @@ nfs4_check_delegmode(struct nfs4_delegation *dp, int flags) return nfs_ok; } -static struct nfs4_delegation * -find_delegation_file(struct nfs4_file *fp, stateid_t *stid) +static int share_access_to_flags(u32 share_access) { - struct nfs4_delegation *dp; + share_access &= ~NFS4_SHARE_WANT_MASK; - spin_lock(&recall_lock); - list_for_each_entry(dp, &fp->fi_delegations, dl_perfile) - if (dp->dl_stateid.si_stateownerid == stid->si_stateownerid) { - spin_unlock(&recall_lock); - return dp; - } - spin_unlock(&recall_lock); - return NULL; + return share_access == NFS4_SHARE_ACCESS_READ ? RD_STATE : WR_STATE; } -static int share_access_to_flags(u32 share_access) +static struct nfs4_delegation *find_deleg_stateid(struct nfs4_client *cl, stateid_t *s) { - share_access &= ~NFS4_SHARE_WANT_MASK; + struct nfs4_stid *ret; - return share_access == NFS4_SHARE_ACCESS_READ ? RD_STATE : WR_STATE; + ret = find_stateid_by_type(cl, s, NFS4_DELEG_STID); + if (!ret) + return NULL; + return delegstateid(ret); +} + +static bool nfsd4_is_deleg_cur(struct nfsd4_open *open) +{ + return open->op_claim_type == NFS4_OPEN_CLAIM_DELEGATE_CUR || + open->op_claim_type == NFS4_OPEN_CLAIM_DELEG_CUR_FH; } static __be32 -nfs4_check_deleg(struct nfs4_file *fp, struct nfsd4_open *open, +nfs4_check_deleg(struct nfs4_client *cl, struct nfs4_file *fp, struct nfsd4_open *open, struct nfs4_delegation **dp) { int flags; __be32 status = nfserr_bad_stateid; - *dp = find_delegation_file(fp, &open->op_delegate_stateid); + *dp = find_deleg_stateid(cl, &open->op_delegate_stateid); if (*dp == NULL) goto out; flags = share_access_to_flags(open->op_share_access); @@ -2495,41 +2666,37 @@ nfs4_check_deleg(struct nfs4_file *fp, struct nfsd4_open *open, if (status) *dp = NULL; out: - if (open->op_claim_type != NFS4_OPEN_CLAIM_DELEGATE_CUR) + if (!nfsd4_is_deleg_cur(open)) return nfs_ok; if (status) return status; - open->op_stateowner->so_confirmed = 1; + open->op_openowner->oo_flags |= NFS4_OO_CONFIRMED; return nfs_ok; } static __be32 -nfs4_check_open(struct nfs4_file *fp, struct nfsd4_open *open, struct nfs4_stateid **stpp) +nfs4_check_open(struct nfs4_file *fp, struct nfsd4_open *open, struct nfs4_ol_stateid **stpp) { - struct nfs4_stateid *local; - __be32 status = nfserr_share_denied; - struct nfs4_stateowner *sop = open->op_stateowner; + struct nfs4_ol_stateid *local; + struct nfs4_openowner *oo = open->op_openowner; list_for_each_entry(local, &fp->fi_stateids, st_perfile) { /* ignore lock owners */ if (local->st_stateowner->so_is_open_owner == 0) continue; /* remember if we have seen this open owner */ - if (local->st_stateowner == sop) + if (local->st_stateowner == &oo->oo_owner) *stpp = local; /* check for conflicting share reservations */ if (!test_share(local, open)) - goto out; + return nfserr_share_denied; } - status = 0; -out: - return status; + return nfs_ok; } -static inline struct nfs4_stateid * -nfs4_alloc_stateid(void) +static void nfs4_free_stateid(struct nfs4_ol_stateid *s) { - return kmem_cache_alloc(stateid_slab, GFP_KERNEL); + kmem_cache_free(stateid_slab, s); } static inline int nfs4_access_to_access(u32 nfs4_access) @@ -2550,12 +2717,6 @@ static __be32 nfs4_get_vfs_file(struct svc_rqst *rqstp, struct nfs4_file *fp, int oflag = nfs4_access_to_omode(open->op_share_access); int access = nfs4_access_to_access(open->op_share_access); - /* CLAIM_DELEGATE_CUR is used in response to a broken lease; - * allowing it to break the lease and return EAGAIN leaves the - * client unable to make progress in returning the delegation */ - if (open->op_claim_type == NFS4_OPEN_CLAIM_DELEGATE_CUR) - access |= NFSD_MAY_NOT_BREAK_LEASE; - if (!fp->fi_fds[oflag]) { status = nfsd_open(rqstp, cur_fh, S_IFREG, access, &fp->fi_fds[oflag]); @@ -2567,27 +2728,6 @@ static __be32 nfs4_get_vfs_file(struct svc_rqst *rqstp, struct nfs4_file *fp, return nfs_ok; } -static __be32 -nfs4_new_open(struct svc_rqst *rqstp, struct nfs4_stateid **stpp, - struct nfs4_file *fp, struct svc_fh *cur_fh, - struct nfsd4_open *open) -{ - struct nfs4_stateid *stp; - __be32 status; - - stp = nfs4_alloc_stateid(); - if (stp == NULL) - return nfserr_jukebox; - - status = nfs4_get_vfs_file(rqstp, fp, cur_fh, open); - if (status) { - kmem_cache_free(stateid_slab, stp); - return status; - } - *stpp = stp; - return 0; -} - static inline __be32 nfsd4_truncate(struct svc_rqst *rqstp, struct svc_fh *fh, struct nfsd4_open *open) @@ -2604,9 +2744,9 @@ nfsd4_truncate(struct svc_rqst *rqstp, struct svc_fh *fh, } static __be32 -nfs4_upgrade_open(struct svc_rqst *rqstp, struct nfs4_file *fp, struct svc_fh *cur_fh, struct nfs4_stateid *stp, struct nfsd4_open *open) +nfs4_upgrade_open(struct svc_rqst *rqstp, struct nfs4_file *fp, struct svc_fh *cur_fh, struct nfs4_ol_stateid *stp, struct nfsd4_open *open) { - u32 op_share_access = open->op_share_access & ~NFS4_SHARE_WANT_MASK; + u32 op_share_access = open->op_share_access; bool new_access; __be32 status; @@ -2635,8 +2775,8 @@ nfs4_upgrade_open(struct svc_rqst *rqstp, struct nfs4_file *fp, struct svc_fh *c static void nfs4_set_claim_prev(struct nfsd4_open *open) { - open->op_stateowner->so_confirmed = 1; - open->op_stateowner->so_client->cl_firststate = 1; + open->op_openowner->oo_flags |= NFS4_OO_CONFIRMED; + open->op_openowner->oo_owner.so_client->cl_firststate = 1; } /* Should we give out recallable state?: */ @@ -2679,10 +2819,8 @@ static int nfs4_setlease(struct nfs4_delegation *dp, int flag) if (!fl) return -ENOMEM; fl->fl_file = find_readable_file(fp); - list_add(&dp->dl_perclnt, &dp->dl_client->cl_delegations); status = vfs_setlease(fl->fl_file, fl->fl_type, &fl); if (status) { - list_del_init(&dp->dl_perclnt); locks_free_lock(fl); return -ENOMEM; } @@ -2690,7 +2828,9 @@ static int nfs4_setlease(struct nfs4_delegation *dp, int flag) fp->fi_deleg_file = fl->fl_file; get_file(fp->fi_deleg_file); atomic_set(&fp->fi_delegees, 1); - list_add(&dp->dl_perfile, &fp->fi_delegations); + spin_lock(&recall_lock); + hash_delegation_locked(dp, fp); + spin_unlock(&recall_lock); return 0; } @@ -2706,9 +2846,8 @@ static int nfs4_set_delegation(struct nfs4_delegation *dp, int flag) return -EAGAIN; } atomic_inc(&fp->fi_delegees); - list_add(&dp->dl_perfile, &fp->fi_delegations); + hash_delegation_locked(dp, fp); spin_unlock(&recall_lock); - list_add(&dp->dl_perclnt, &dp->dl_client->cl_delegations); return 0; } @@ -2716,14 +2855,14 @@ static int nfs4_set_delegation(struct nfs4_delegation *dp, int flag) * Attempt to hand out a delegation. */ static void -nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_stateid *stp) +nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_ol_stateid *stp) { struct nfs4_delegation *dp; - struct nfs4_stateowner *sop = stp->st_stateowner; + struct nfs4_openowner *oo = container_of(stp->st_stateowner, struct nfs4_openowner, oo_owner); int cb_up; int status, flag = 0; - cb_up = nfsd4_cb_channel_good(sop->so_client); + cb_up = nfsd4_cb_channel_good(oo->oo_owner.so_client); flag = NFS4_OPEN_DELEGATE_NONE; open->op_recall = 0; switch (open->op_claim_type) { @@ -2739,7 +2878,7 @@ nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_sta * had the chance to reclaim theirs.... */ if (locks_in_grace()) goto out; - if (!cb_up || !sop->so_confirmed) + if (!cb_up || !(oo->oo_flags & NFS4_OO_CONFIRMED)) goto out; if (open->op_share_access & NFS4_SHARE_ACCESS_WRITE) flag = NFS4_OPEN_DELEGATE_WRITE; @@ -2750,17 +2889,17 @@ nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_sta goto out; } - dp = alloc_init_deleg(sop->so_client, stp, fh, flag); + dp = alloc_init_deleg(oo->oo_owner.so_client, stp, fh, flag); if (dp == NULL) goto out_no_deleg; status = nfs4_set_delegation(dp, flag); if (status) goto out_free; - memcpy(&open->op_delegate_stateid, &dp->dl_stateid, sizeof(dp->dl_stateid)); + memcpy(&open->op_delegate_stateid, &dp->dl_stid.sc_stateid, sizeof(dp->dl_stid.sc_stateid)); dprintk("NFSD: delegation stateid=" STATEID_FMT "\n", - STATEID_VAL(&dp->dl_stateid)); + STATEID_VAL(&dp->dl_stid.sc_stateid)); out: if (open->op_claim_type == NFS4_OPEN_CLAIM_PREVIOUS && flag == NFS4_OPEN_DELEGATE_NONE @@ -2782,16 +2921,13 @@ __be32 nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_open *open) { struct nfsd4_compoundres *resp = rqstp->rq_resp; + struct nfs4_client *cl = open->op_openowner->oo_owner.so_client; struct nfs4_file *fp = NULL; struct inode *ino = current_fh->fh_dentry->d_inode; - struct nfs4_stateid *stp = NULL; + struct nfs4_ol_stateid *stp = NULL; struct nfs4_delegation *dp = NULL; __be32 status; - status = nfserr_inval; - if (!access_valid(open->op_share_access, resp->cstate.minorversion) - || !deny_valid(open->op_share_deny)) - goto out; /* * Lookup file; if found, lookup stateid and check open request, * and check for delegations in the process of being recalled. @@ -2801,17 +2937,17 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf if (fp) { if ((status = nfs4_check_open(fp, open, &stp))) goto out; - status = nfs4_check_deleg(fp, open, &dp); + status = nfs4_check_deleg(cl, fp, open, &dp); if (status) goto out; } else { status = nfserr_bad_stateid; - if (open->op_claim_type == NFS4_OPEN_CLAIM_DELEGATE_CUR) + if (nfsd4_is_deleg_cur(open)) goto out; status = nfserr_jukebox; - fp = alloc_init_file(ino); - if (fp == NULL) - goto out; + fp = open->op_file; + open->op_file = NULL; + nfsd4_init_file(fp, ino); } /* @@ -2823,24 +2959,24 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf status = nfs4_upgrade_open(rqstp, fp, current_fh, stp, open); if (status) goto out; - update_stateid(&stp->st_stateid); } else { - status = nfs4_new_open(rqstp, &stp, fp, current_fh, open); + status = nfs4_get_vfs_file(rqstp, fp, current_fh, open); if (status) goto out; - init_stateid(stp, fp, open); + stp = open->op_stp; + open->op_stp = NULL; + init_open_stateid(stp, fp, open); status = nfsd4_truncate(rqstp, current_fh, open); if (status) { release_open_stateid(stp); goto out; } - if (nfsd4_has_session(&resp->cstate)) - update_stateid(&stp->st_stateid); } - memcpy(&open->op_stateid, &stp->st_stateid, sizeof(stateid_t)); + update_stateid(&stp->st_stid.sc_stateid); + memcpy(&open->op_stateid, &stp->st_stid.sc_stateid, sizeof(stateid_t)); if (nfsd4_has_session(&resp->cstate)) - open->op_stateowner->so_confirmed = 1; + open->op_openowner->oo_flags |= NFS4_OO_CONFIRMED; /* * Attempt to hand out a delegation. No error return, because the @@ -2851,7 +2987,7 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf status = nfs_ok; dprintk("%s: stateid=" STATEID_FMT "\n", __func__, - STATEID_VAL(&stp->st_stateid)); + STATEID_VAL(&stp->st_stid.sc_stateid)); out: if (fp) put_nfs4_file(fp); @@ -2861,13 +2997,34 @@ out: * To finish the open response, we just need to set the rflags. */ open->op_rflags = NFS4_OPEN_RESULT_LOCKTYPE_POSIX; - if (!open->op_stateowner->so_confirmed && + if (!(open->op_openowner->oo_flags & NFS4_OO_CONFIRMED) && !nfsd4_has_session(&resp->cstate)) open->op_rflags |= NFS4_OPEN_RESULT_CONFIRM; return status; } +void nfsd4_cleanup_open_state(struct nfsd4_open *open, __be32 status) +{ + if (open->op_openowner) { + struct nfs4_openowner *oo = open->op_openowner; + + if (!list_empty(&oo->oo_owner.so_stateids)) + list_del_init(&oo->oo_close_lru); + if (oo->oo_flags & NFS4_OO_NEW) { + if (status) { + release_openowner(oo); + open->op_openowner = NULL; + } else + oo->oo_flags &= ~NFS4_OO_NEW; + } + } + if (open->op_file) + nfsd4_free_file(open->op_file); + if (open->op_stp) + nfs4_free_stateid(open->op_stp); +} + __be32 nfsd4_renew(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, clientid_t *clid) @@ -2888,7 +3045,6 @@ nfsd4_renew(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, dprintk("nfsd4_renew: clientid not found!\n"); goto out; } - renew_client(clp); status = nfserr_cb_path_down; if (!list_empty(&clp->cl_delegations) && clp->cl_cb_state != NFSD4_CB_UP) @@ -2920,7 +3076,7 @@ static time_t nfs4_laundromat(void) { struct nfs4_client *clp; - struct nfs4_stateowner *sop; + struct nfs4_openowner *oo; struct nfs4_delegation *dp; struct list_head *pos, *next, reaplist; time_t cutoff = get_seconds() - nfsd4_lease; @@ -2977,16 +3133,14 @@ nfs4_laundromat(void) } test_val = nfsd4_lease; list_for_each_safe(pos, next, &close_lru) { - sop = list_entry(pos, struct nfs4_stateowner, so_close_lru); - if (time_after((unsigned long)sop->so_time, (unsigned long)cutoff)) { - u = sop->so_time - cutoff; + oo = container_of(pos, struct nfs4_openowner, oo_close_lru); + if (time_after((unsigned long)oo->oo_time, (unsigned long)cutoff)) { + u = oo->oo_time - cutoff; if (test_val > u) test_val = u; break; } - dprintk("NFSD: purging unused open stateowner (so_id %d)\n", - sop->so_id); - release_openowner(sop); + release_openowner(oo); } if (clientid_val < NFSD_LAUNDROMAT_MINTIMEOUT) clientid_val = NFSD_LAUNDROMAT_MINTIMEOUT; @@ -3008,30 +3162,17 @@ laundromat_main(struct work_struct *not_used) queue_delayed_work(laundry_wq, &laundromat_work, t*HZ); } -static struct nfs4_stateowner * -search_close_lru(u32 st_id, int flags) -{ - struct nfs4_stateowner *local = NULL; - - if (flags & CLOSE_STATE) { - list_for_each_entry(local, &close_lru, so_close_lru) { - if (local->so_id == st_id) - return local; - } - } - return NULL; -} - -static inline int -nfs4_check_fh(struct svc_fh *fhp, struct nfs4_stateid *stp) +static inline __be32 nfs4_check_fh(struct svc_fh *fhp, struct nfs4_ol_stateid *stp) { - return fhp->fh_dentry->d_inode != stp->st_file->fi_inode; + if (fhp->fh_dentry->d_inode != stp->st_file->fi_inode) + return nfserr_bad_stateid; + return nfs_ok; } static int STALE_STATEID(stateid_t *stateid) { - if (stateid->si_boot == boot_time) + if (stateid->si_opaque.so_clid.cl_boot == boot_time) return 0; dprintk("NFSD: stale stateid " STATEID_FMT "!\n", STATEID_VAL(stateid)); @@ -3054,7 +3195,7 @@ access_permit_write(unsigned long access_bmap) } static -__be32 nfs4_check_openmode(struct nfs4_stateid *stp, int flags) +__be32 nfs4_check_openmode(struct nfs4_ol_stateid *stp, int flags) { __be32 status = nfserr_openmode; @@ -3097,37 +3238,83 @@ grace_disallows_io(struct inode *inode) return locks_in_grace() && mandatory_lock(inode); } -static int check_stateid_generation(stateid_t *in, stateid_t *ref, int flags) +/* Returns true iff a is later than b: */ +static bool stateid_generation_after(stateid_t *a, stateid_t *b) +{ + return (s32)a->si_generation - (s32)b->si_generation > 0; +} + +static int check_stateid_generation(stateid_t *in, stateid_t *ref, bool has_session) { /* * When sessions are used the stateid generation number is ignored * when it is zero. */ - if ((flags & HAS_SESSION) && in->si_generation == 0) - goto out; + if (has_session && in->si_generation == 0) + return nfs_ok; + + if (in->si_generation == ref->si_generation) + return nfs_ok; /* If the client sends us a stateid from the future, it's buggy: */ - if (in->si_generation > ref->si_generation) + if (stateid_generation_after(in, ref)) return nfserr_bad_stateid; /* - * The following, however, can happen. For example, if the - * client sends an open and some IO at the same time, the open - * may bump si_generation while the IO is still in flight. - * Thanks to hard links and renames, the client never knows what - * file an open will affect. So it could avoid that situation - * only by serializing all opens and IO from the same open - * owner. To recover from the old_stateid error, the client - * will just have to retry the IO: + * However, we could see a stateid from the past, even from a + * non-buggy client. For example, if the client sends a lock + * while some IO is outstanding, the lock may bump si_generation + * while the IO is still in flight. The client could avoid that + * situation by waiting for responses on all the IO requests, + * but better performance may result in retrying IO that + * receives an old_stateid error if requests are rarely + * reordered in flight: */ - if (in->si_generation < ref->si_generation) - return nfserr_old_stateid; -out: + return nfserr_old_stateid; +} + +static __be32 nfsd4_check_openowner_confirmed(struct nfs4_ol_stateid *ols) +{ + if (ols->st_stateowner->so_is_open_owner && + !(openowner(ols->st_stateowner)->oo_flags & NFS4_OO_CONFIRMED)) + return nfserr_bad_stateid; return nfs_ok; } -static int is_delegation_stateid(stateid_t *stateid) +__be32 nfs4_validate_stateid(struct nfs4_client *cl, stateid_t *stateid) { - return stateid->si_fileid == 0; + struct nfs4_stid *s; + __be32 status; + + if (STALE_STATEID(stateid)) + return nfserr_stale_stateid; + + s = find_stateid(cl, stateid); + if (!s) + return nfserr_stale_stateid; + status = check_stateid_generation(stateid, &s->sc_stateid, 1); + if (status) + return status; + if (!(s->sc_type & (NFS4_OPEN_STID | NFS4_LOCK_STID))) + return nfs_ok; + return nfsd4_check_openowner_confirmed(openlockstateid(s)); +} + +static __be32 nfsd4_lookup_stateid(stateid_t *stateid, unsigned char typemask, struct nfs4_stid **s) +{ + struct nfs4_client *cl; + + if (ZERO_STATEID(stateid) || ONE_STATEID(stateid)) + return nfserr_bad_stateid; + if (STALE_STATEID(stateid)) + return nfserr_stale_stateid; + cl = find_confirmed_client(&stateid->si_opaque.so_clid); + if (!cl) + return nfserr_expired; + *s = find_stateid_by_type(cl, stateid, typemask); + if (!*s) + return nfserr_bad_stateid; + return nfs_ok; + } /* @@ -3137,7 +3324,8 @@ __be32 nfs4_preprocess_stateid_op(struct nfsd4_compound_state *cstate, stateid_t *stateid, int flags, struct file **filpp) { - struct nfs4_stateid *stp = NULL; + struct nfs4_stid *s; + struct nfs4_ol_stateid *stp = NULL; struct nfs4_delegation *dp = NULL; struct svc_fh *current_fh = &cstate->current_fh; struct inode *ino = current_fh->fh_dentry->d_inode; @@ -3149,66 +3337,115 @@ nfs4_preprocess_stateid_op(struct nfsd4_compound_state *cstate, if (grace_disallows_io(ino)) return nfserr_grace; - if (nfsd4_has_session(cstate)) - flags |= HAS_SESSION; - if (ZERO_STATEID(stateid) || ONE_STATEID(stateid)) return check_special_stateids(current_fh, stateid, flags); - status = nfserr_stale_stateid; - if (STALE_STATEID(stateid)) + status = nfsd4_lookup_stateid(stateid, NFS4_DELEG_STID|NFS4_OPEN_STID|NFS4_LOCK_STID, &s); + if (status) + return status; + status = check_stateid_generation(stateid, &s->sc_stateid, nfsd4_has_session(cstate)); + if (status) goto out; - - /* - * We assume that any stateid that has the current boot time, - * but that we can't find, is expired: - */ - status = nfserr_expired; - if (is_delegation_stateid(stateid)) { - dp = find_delegation_stateid(ino, stateid); - if (!dp) - goto out; - status = check_stateid_generation(stateid, &dp->dl_stateid, - flags); - if (status) - goto out; + switch (s->sc_type) { + case NFS4_DELEG_STID: + dp = delegstateid(s); status = nfs4_check_delegmode(dp, flags); if (status) goto out; - renew_client(dp->dl_client); if (filpp) { *filpp = dp->dl_file->fi_deleg_file; BUG_ON(!*filpp); } - } else { /* open or lock stateid */ - stp = find_stateid(stateid, flags); - if (!stp) - goto out; - status = nfserr_bad_stateid; - if (nfs4_check_fh(current_fh, stp)) - goto out; - if (!stp->st_stateowner->so_confirmed) + break; + case NFS4_OPEN_STID: + case NFS4_LOCK_STID: + stp = openlockstateid(s); + status = nfs4_check_fh(current_fh, stp); + if (status) goto out; - status = check_stateid_generation(stateid, &stp->st_stateid, - flags); + status = nfsd4_check_openowner_confirmed(stp); if (status) goto out; status = nfs4_check_openmode(stp, flags); if (status) goto out; - renew_client(stp->st_stateowner->so_client); if (filpp) { if (flags & RD_STATE) *filpp = find_readable_file(stp->st_file); else *filpp = find_writeable_file(stp->st_file); } + break; + default: + return nfserr_bad_stateid; } status = nfs_ok; out: return status; } +static __be32 +nfsd4_free_lock_stateid(struct nfs4_ol_stateid *stp) +{ + struct nfs4_lockowner *lo = lockowner(stp->st_stateowner); + + if (check_for_locks(stp->st_file, lo)) + return nfserr_locks_held; + /* + * Currently there's a 1-1 lock stateid<->lockowner + * correspondance, and we have to delete the lockowner when we + * delete the lock stateid: + */ + release_lockowner(lo); + return nfs_ok; +} + +/* + * Test if the stateid is valid + */ +__be32 +nfsd4_test_stateid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, + struct nfsd4_test_stateid *test_stateid) +{ + /* real work is done during encoding */ + return nfs_ok; +} + +__be32 +nfsd4_free_stateid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, + struct nfsd4_free_stateid *free_stateid) +{ + stateid_t *stateid = &free_stateid->fr_stateid; + struct nfs4_stid *s; + struct nfs4_client *cl = cstate->session->se_client; + __be32 ret = nfserr_bad_stateid; + + nfs4_lock_state(); + s = find_stateid(cl, stateid); + if (!s) + goto out; + switch (s->sc_type) { + case NFS4_DELEG_STID: + ret = nfserr_locks_held; + goto out; + case NFS4_OPEN_STID: + case NFS4_LOCK_STID: + ret = check_stateid_generation(stateid, &s->sc_stateid, 1); + if (ret) + goto out; + if (s->sc_type == NFS4_LOCK_STID) + ret = nfsd4_free_lock_stateid(openlockstateid(s)); + else + ret = nfserr_locks_held; + break; + default: + ret = nfserr_bad_stateid; + } +out: + nfs4_unlock_state(); + return ret; +} + static inline int setlkflg (int type) { @@ -3216,124 +3453,64 @@ setlkflg (int type) RD_STATE : WR_STATE; } +static __be32 nfs4_seqid_op_checks(struct nfsd4_compound_state *cstate, stateid_t *stateid, u32 seqid, struct nfs4_ol_stateid *stp) +{ + struct svc_fh *current_fh = &cstate->current_fh; + struct nfs4_stateowner *sop = stp->st_stateowner; + __be32 status; + + status = nfsd4_check_seqid(cstate, sop, seqid); + if (status) + return status; + if (stp->st_stid.sc_type == NFS4_CLOSED_STID) + /* + * "Closed" stateid's exist *only* to return + * nfserr_replay_me from the previous step. + */ + return nfserr_bad_stateid; + status = check_stateid_generation(stateid, &stp->st_stid.sc_stateid, nfsd4_has_session(cstate)); + if (status) + return status; + return nfs4_check_fh(current_fh, stp); +} + /* * Checks for sequence id mutating operations. */ static __be32 nfs4_preprocess_seqid_op(struct nfsd4_compound_state *cstate, u32 seqid, - stateid_t *stateid, int flags, - struct nfs4_stateowner **sopp, - struct nfs4_stateid **stpp, struct nfsd4_lock *lock) + stateid_t *stateid, char typemask, + struct nfs4_ol_stateid **stpp) { - struct nfs4_stateid *stp; - struct nfs4_stateowner *sop; - struct svc_fh *current_fh = &cstate->current_fh; __be32 status; + struct nfs4_stid *s; dprintk("NFSD: %s: seqid=%d stateid = " STATEID_FMT "\n", __func__, seqid, STATEID_VAL(stateid)); *stpp = NULL; - *sopp = NULL; - - if (ZERO_STATEID(stateid) || ONE_STATEID(stateid)) { - dprintk("NFSD: preprocess_seqid_op: magic stateid!\n"); - return nfserr_bad_stateid; - } - - if (STALE_STATEID(stateid)) - return nfserr_stale_stateid; - - if (nfsd4_has_session(cstate)) - flags |= HAS_SESSION; - - /* - * We return BAD_STATEID if filehandle doesn't match stateid, - * the confirmed flag is incorrecly set, or the generation - * number is incorrect. - */ - stp = find_stateid(stateid, flags); - if (stp == NULL) { - /* - * Also, we should make sure this isn't just the result of - * a replayed close: - */ - sop = search_close_lru(stateid->si_stateownerid, flags); - /* It's not stale; let's assume it's expired: */ - if (sop == NULL) - return nfserr_expired; - *sopp = sop; - goto check_replay; - } - - *stpp = stp; - *sopp = sop = stp->st_stateowner; - - if (lock) { - clientid_t *lockclid = &lock->v.new.clientid; - struct nfs4_client *clp = sop->so_client; - int lkflg = 0; - __be32 status; - - lkflg = setlkflg(lock->lk_type); - - if (lock->lk_is_new) { - if (!sop->so_is_open_owner) - return nfserr_bad_stateid; - if (!(flags & HAS_SESSION) && - !same_clid(&clp->cl_clientid, lockclid)) - return nfserr_bad_stateid; - /* stp is the open stateid */ - status = nfs4_check_openmode(stp, lkflg); - if (status) - return status; - } else { - /* stp is the lock stateid */ - status = nfs4_check_openmode(stp->st_openstp, lkflg); - if (status) - return status; - } - } + status = nfsd4_lookup_stateid(stateid, typemask, &s); + if (status) + return status; + *stpp = openlockstateid(s); + cstate->replay_owner = (*stpp)->st_stateowner; - if (nfs4_check_fh(current_fh, stp)) { - dprintk("NFSD: preprocess_seqid_op: fh-stateid mismatch!\n"); - return nfserr_bad_stateid; - } + return nfs4_seqid_op_checks(cstate, stateid, seqid, *stpp); +} - /* - * We now validate the seqid and stateid generation numbers. - * For the moment, we ignore the possibility of - * generation number wraparound. - */ - if (!(flags & HAS_SESSION) && seqid != sop->so_seqid) - goto check_replay; +static __be32 nfs4_preprocess_confirmed_seqid_op(struct nfsd4_compound_state *cstate, u32 seqid, stateid_t *stateid, struct nfs4_ol_stateid **stpp) +{ + __be32 status; + struct nfs4_openowner *oo; - if (sop->so_confirmed && flags & CONFIRM) { - dprintk("NFSD: preprocess_seqid_op: expected" - " unconfirmed stateowner!\n"); - return nfserr_bad_stateid; - } - if (!sop->so_confirmed && !(flags & CONFIRM)) { - dprintk("NFSD: preprocess_seqid_op: stateowner not" - " confirmed yet!\n"); - return nfserr_bad_stateid; - } - status = check_stateid_generation(stateid, &stp->st_stateid, flags); + status = nfs4_preprocess_seqid_op(cstate, seqid, stateid, + NFS4_OPEN_STID, stpp); if (status) return status; - renew_client(sop->so_client); + oo = openowner((*stpp)->st_stateowner); + if (!(oo->oo_flags & NFS4_OO_CONFIRMED)) + return nfserr_bad_stateid; return nfs_ok; - -check_replay: - if (seqid == sop->so_seqid - 1) { - dprintk("NFSD: preprocess_seqid_op: retransmission?\n"); - /* indicate replay to calling function */ - return nfserr_replay_me; - } - dprintk("NFSD: preprocess_seqid_op: bad seqid (expected %d, got %d)\n", - sop->so_seqid, seqid); - *sopp = NULL; - return nfserr_bad_seqid; } __be32 @@ -3341,8 +3518,8 @@ nfsd4_open_confirm(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_open_confirm *oc) { __be32 status; - struct nfs4_stateowner *sop; - struct nfs4_stateid *stp; + struct nfs4_openowner *oo; + struct nfs4_ol_stateid *stp; dprintk("NFSD: nfsd4_open_confirm on file %.*s\n", (int)cstate->current_fh.fh_dentry->d_name.len, @@ -3354,39 +3531,52 @@ nfsd4_open_confirm(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, nfs4_lock_state(); - if ((status = nfs4_preprocess_seqid_op(cstate, + status = nfs4_preprocess_seqid_op(cstate, oc->oc_seqid, &oc->oc_req_stateid, - CONFIRM | OPEN_STATE, - &oc->oc_stateowner, &stp, NULL))) - goto out; - - sop = oc->oc_stateowner; - sop->so_confirmed = 1; - update_stateid(&stp->st_stateid); - memcpy(&oc->oc_resp_stateid, &stp->st_stateid, sizeof(stateid_t)); + NFS4_OPEN_STID, &stp); + if (status) + goto out; + oo = openowner(stp->st_stateowner); + status = nfserr_bad_stateid; + if (oo->oo_flags & NFS4_OO_CONFIRMED) + goto out; + oo->oo_flags |= NFS4_OO_CONFIRMED; + update_stateid(&stp->st_stid.sc_stateid); + memcpy(&oc->oc_resp_stateid, &stp->st_stid.sc_stateid, sizeof(stateid_t)); dprintk("NFSD: %s: success, seqid=%d stateid=" STATEID_FMT "\n", - __func__, oc->oc_seqid, STATEID_VAL(&stp->st_stateid)); + __func__, oc->oc_seqid, STATEID_VAL(&stp->st_stid.sc_stateid)); - nfsd4_create_clid_dir(sop->so_client); + nfsd4_create_clid_dir(oo->oo_owner.so_client); + status = nfs_ok; out: - if (oc->oc_stateowner) { - nfs4_get_stateowner(oc->oc_stateowner); - cstate->replay_owner = oc->oc_stateowner; - } - nfs4_unlock_state(); + if (!cstate->replay_owner) + nfs4_unlock_state(); return status; } -static inline void nfs4_file_downgrade(struct nfs4_stateid *stp, unsigned int to_access) +static inline void nfs4_stateid_downgrade_bit(struct nfs4_ol_stateid *stp, u32 access) { - int i; + if (!test_bit(access, &stp->st_access_bmap)) + return; + nfs4_file_put_access(stp->st_file, nfs4_access_to_omode(access)); + __clear_bit(access, &stp->st_access_bmap); +} - for (i = 1; i < 4; i++) { - if (test_bit(i, &stp->st_access_bmap) - && ((i & to_access) != i)) { - nfs4_file_put_access(stp->st_file, nfs4_access_to_omode(i)); - __clear_bit(i, &stp->st_access_bmap); - } +static inline void nfs4_stateid_downgrade(struct nfs4_ol_stateid *stp, u32 to_access) +{ + switch (to_access) { + case NFS4_SHARE_ACCESS_READ: + nfs4_stateid_downgrade_bit(stp, NFS4_SHARE_ACCESS_WRITE); + nfs4_stateid_downgrade_bit(stp, NFS4_SHARE_ACCESS_BOTH); + break; + case NFS4_SHARE_ACCESS_WRITE: + nfs4_stateid_downgrade_bit(stp, NFS4_SHARE_ACCESS_READ); + nfs4_stateid_downgrade_bit(stp, NFS4_SHARE_ACCESS_BOTH); + break; + case NFS4_SHARE_ACCESS_BOTH: + break; + default: + BUG(); } } @@ -3406,26 +3596,20 @@ nfsd4_open_downgrade(struct svc_rqst *rqstp, struct nfsd4_open_downgrade *od) { __be32 status; - struct nfs4_stateid *stp; + struct nfs4_ol_stateid *stp; dprintk("NFSD: nfsd4_open_downgrade on file %.*s\n", (int)cstate->current_fh.fh_dentry->d_name.len, cstate->current_fh.fh_dentry->d_name.name); - if (!access_valid(od->od_share_access, cstate->minorversion) - || !deny_valid(od->od_share_deny)) - return nfserr_inval; /* We don't yet support WANT bits: */ od->od_share_access &= NFS4_SHARE_ACCESS_MASK; nfs4_lock_state(); - if ((status = nfs4_preprocess_seqid_op(cstate, - od->od_seqid, - &od->od_stateid, - OPEN_STATE, - &od->od_stateowner, &stp, NULL))) + status = nfs4_preprocess_confirmed_seqid_op(cstate, od->od_seqid, + &od->od_stateid, &stp); + if (status) goto out; - status = nfserr_inval; if (!test_bit(od->od_share_access, &stp->st_access_bmap)) { dprintk("NFSD:access not a subset current bitmap: 0x%lx, input access=%08x\n", @@ -3437,22 +3621,45 @@ nfsd4_open_downgrade(struct svc_rqst *rqstp, stp->st_deny_bmap, od->od_share_deny); goto out; } - nfs4_file_downgrade(stp, od->od_share_access); + nfs4_stateid_downgrade(stp, od->od_share_access); reset_union_bmap_deny(od->od_share_deny, &stp->st_deny_bmap); - update_stateid(&stp->st_stateid); - memcpy(&od->od_stateid, &stp->st_stateid, sizeof(stateid_t)); + update_stateid(&stp->st_stid.sc_stateid); + memcpy(&od->od_stateid, &stp->st_stid.sc_stateid, sizeof(stateid_t)); status = nfs_ok; out: - if (od->od_stateowner) { - nfs4_get_stateowner(od->od_stateowner); - cstate->replay_owner = od->od_stateowner; - } - nfs4_unlock_state(); + if (!cstate->replay_owner) + nfs4_unlock_state(); return status; } +void nfsd4_purge_closed_stateid(struct nfs4_stateowner *so) +{ + struct nfs4_openowner *oo; + struct nfs4_ol_stateid *s; + + if (!so->so_is_open_owner) + return; + oo = openowner(so); + s = oo->oo_last_closed_stid; + if (!s) + return; + if (!(oo->oo_flags & NFS4_OO_PURGE_CLOSE)) { + /* Release the last_closed_stid on the next seqid bump: */ + oo->oo_flags |= NFS4_OO_PURGE_CLOSE; + return; + } + oo->oo_flags &= ~NFS4_OO_PURGE_CLOSE; + release_last_closed_stateid(oo); +} + +static void nfsd4_close_open_stateid(struct nfs4_ol_stateid *s) +{ + unhash_open_stateid(s); + s->st_stid.sc_type = NFS4_CLOSED_STID; +} + /* * nfs4_unlock_state() called after encode */ @@ -3461,39 +3668,38 @@ nfsd4_close(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_close *close) { __be32 status; - struct nfs4_stateid *stp; + struct nfs4_openowner *oo; + struct nfs4_ol_stateid *stp; dprintk("NFSD: nfsd4_close on file %.*s\n", (int)cstate->current_fh.fh_dentry->d_name.len, cstate->current_fh.fh_dentry->d_name.name); nfs4_lock_state(); - /* check close_lru for replay */ - if ((status = nfs4_preprocess_seqid_op(cstate, - close->cl_seqid, - &close->cl_stateid, - OPEN_STATE | CLOSE_STATE, - &close->cl_stateowner, &stp, NULL))) + status = nfs4_preprocess_seqid_op(cstate, close->cl_seqid, + &close->cl_stateid, + NFS4_OPEN_STID|NFS4_CLOSED_STID, + &stp); + if (status) goto out; + oo = openowner(stp->st_stateowner); status = nfs_ok; - update_stateid(&stp->st_stateid); - memcpy(&close->cl_stateid, &stp->st_stateid, sizeof(stateid_t)); + update_stateid(&stp->st_stid.sc_stateid); + memcpy(&close->cl_stateid, &stp->st_stid.sc_stateid, sizeof(stateid_t)); - /* release_stateid() calls nfsd_close() if needed */ - release_open_stateid(stp); + nfsd4_close_open_stateid(stp); + release_last_closed_stateid(oo); + oo->oo_last_closed_stid = stp; /* place unused nfs4_stateowners on so_close_lru list to be * released by the laundromat service after the lease period * to enable us to handle CLOSE replay */ - if (list_empty(&close->cl_stateowner->so_stateids)) - move_to_close_lru(close->cl_stateowner); + if (list_empty(&oo->oo_owner.so_stateids)) + move_to_close_lru(oo); out: - if (close->cl_stateowner) { - nfs4_get_stateowner(close->cl_stateowner); - cstate->replay_owner = close->cl_stateowner; - } - nfs4_unlock_state(); + if (!cstate->replay_owner) + nfs4_unlock_state(); return status; } @@ -3503,34 +3709,22 @@ nfsd4_delegreturn(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, { struct nfs4_delegation *dp; stateid_t *stateid = &dr->dr_stateid; + struct nfs4_stid *s; struct inode *inode; __be32 status; - int flags = 0; if ((status = fh_verify(rqstp, &cstate->current_fh, S_IFREG, 0))) return status; inode = cstate->current_fh.fh_dentry->d_inode; - if (nfsd4_has_session(cstate)) - flags |= HAS_SESSION; nfs4_lock_state(); - status = nfserr_bad_stateid; - if (ZERO_STATEID(stateid) || ONE_STATEID(stateid)) - goto out; - status = nfserr_stale_stateid; - if (STALE_STATEID(stateid)) - goto out; - status = nfserr_bad_stateid; - if (!is_delegation_stateid(stateid)) - goto out; - status = nfserr_expired; - dp = find_delegation_stateid(inode, stateid); - if (!dp) + status = nfsd4_lookup_stateid(stateid, NFS4_DELEG_STID, &s); + if (status) goto out; - status = check_stateid_generation(stateid, &dp->dl_stateid, flags); + dp = delegstateid(s); + status = check_stateid_generation(stateid, &dp->dl_stid.sc_stateid, nfsd4_has_session(cstate)); if (status) goto out; - renew_client(dp->dl_client); unhash_delegation(dp); out: @@ -3568,9 +3762,6 @@ last_byte_offset(u64 start, u64 len) return end > start ? end - 1: NFS4_MAX_UINT64; } -#define lockownerid_hashval(id) \ - ((id) & LOCK_HASH_MASK) - static inline unsigned int lock_ownerstr_hashval(struct inode *inode, u32 cl_id, struct xdr_netobj *ownername) @@ -3580,55 +3771,7 @@ lock_ownerstr_hashval(struct inode *inode, u32 cl_id, & LOCK_HASH_MASK; } -static struct list_head lock_ownerid_hashtbl[LOCK_HASH_SIZE]; static struct list_head lock_ownerstr_hashtbl[LOCK_HASH_SIZE]; -static struct list_head lockstateid_hashtbl[STATEID_HASH_SIZE]; - -static struct nfs4_stateid * -find_stateid(stateid_t *stid, int flags) -{ - struct nfs4_stateid *local; - u32 st_id = stid->si_stateownerid; - u32 f_id = stid->si_fileid; - unsigned int hashval; - - dprintk("NFSD: find_stateid flags 0x%x\n",flags); - if (flags & (LOCK_STATE | RD_STATE | WR_STATE)) { - hashval = stateid_hashval(st_id, f_id); - list_for_each_entry(local, &lockstateid_hashtbl[hashval], st_hash) { - if ((local->st_stateid.si_stateownerid == st_id) && - (local->st_stateid.si_fileid == f_id)) - return local; - } - } - - if (flags & (OPEN_STATE | RD_STATE | WR_STATE)) { - hashval = stateid_hashval(st_id, f_id); - list_for_each_entry(local, &stateid_hashtbl[hashval], st_hash) { - if ((local->st_stateid.si_stateownerid == st_id) && - (local->st_stateid.si_fileid == f_id)) - return local; - } - } - return NULL; -} - -static struct nfs4_delegation * -find_delegation_stateid(struct inode *ino, stateid_t *stid) -{ - struct nfs4_file *fp; - struct nfs4_delegation *dl; - - dprintk("NFSD: %s: stateid=" STATEID_FMT "\n", __func__, - STATEID_VAL(stid)); - - fp = find_file(ino); - if (!fp) - return NULL; - dl = find_delegation_file(fp, stid); - put_nfs4_file(fp); - return dl; -} /* * TODO: Linux file offsets are _signed_ 64-bit quantities, which means that @@ -3655,15 +3798,21 @@ static const struct lock_manager_operations nfsd_posix_mng_ops = { static inline void nfs4_set_lock_denied(struct file_lock *fl, struct nfsd4_lock_denied *deny) { - struct nfs4_stateowner *sop; + struct nfs4_lockowner *lo; if (fl->fl_lmops == &nfsd_posix_mng_ops) { - sop = (struct nfs4_stateowner *) fl->fl_owner; - kref_get(&sop->so_ref); - deny->ld_sop = sop; - deny->ld_clientid = sop->so_client->cl_clientid; + lo = (struct nfs4_lockowner *) fl->fl_owner; + deny->ld_owner.data = kmemdup(lo->lo_owner.so_owner.data, + lo->lo_owner.so_owner.len, GFP_KERNEL); + if (!deny->ld_owner.data) + /* We just don't care that much */ + goto nevermind; + deny->ld_owner.len = lo->lo_owner.so_owner.len; + deny->ld_clientid = lo->lo_owner.so_client->cl_clientid; } else { - deny->ld_sop = NULL; +nevermind: + deny->ld_owner.len = 0; + deny->ld_owner.data = NULL; deny->ld_clientid.cl_boot = 0; deny->ld_clientid.cl_id = 0; } @@ -3676,20 +3825,43 @@ nfs4_set_lock_denied(struct file_lock *fl, struct nfsd4_lock_denied *deny) deny->ld_type = NFS4_WRITE_LT; } -static struct nfs4_stateowner * -find_lockstateowner_str(struct inode *inode, clientid_t *clid, +static bool same_lockowner_ino(struct nfs4_lockowner *lo, struct inode *inode, clientid_t *clid, struct xdr_netobj *owner) +{ + struct nfs4_ol_stateid *lst; + + if (!same_owner_str(&lo->lo_owner, owner, clid)) + return false; + if (list_empty(&lo->lo_owner.so_stateids)) { + WARN_ON_ONCE(1); + return false; + } + lst = list_first_entry(&lo->lo_owner.so_stateids, + struct nfs4_ol_stateid, st_perstateowner); + return lst->st_file->fi_inode == inode; +} + +static struct nfs4_lockowner * +find_lockowner_str(struct inode *inode, clientid_t *clid, struct xdr_netobj *owner) { unsigned int hashval = lock_ownerstr_hashval(inode, clid->cl_id, owner); + struct nfs4_lockowner *lo; struct nfs4_stateowner *op; list_for_each_entry(op, &lock_ownerstr_hashtbl[hashval], so_strhash) { - if (same_owner_str(op, owner, clid)) - return op; + lo = lockowner(op); + if (same_lockowner_ino(lo, inode, clid, owner)) + return lo; } return NULL; } +static void hash_lockowner(struct nfs4_lockowner *lo, unsigned int strhashval, struct nfs4_client *clp, struct nfs4_ol_stateid *open_stp) +{ + list_add(&lo->lo_owner.so_strhash, &lock_ownerstr_hashtbl[strhashval]); + list_add(&lo->lo_perstateid, &open_stp->st_lockowners); +} + /* * Alloc a lock owner structure. * Called in nfsd4_lock - therefore, OPEN and OPEN_CONFIRM (if needed) has @@ -3698,67 +3870,40 @@ find_lockstateowner_str(struct inode *inode, clientid_t *clid, * strhashval = lock_ownerstr_hashval */ -static struct nfs4_stateowner * -alloc_init_lock_stateowner(unsigned int strhashval, struct nfs4_client *clp, struct nfs4_stateid *open_stp, struct nfsd4_lock *lock) { - struct nfs4_stateowner *sop; - struct nfs4_replay *rp; - unsigned int idhashval; +static struct nfs4_lockowner * +alloc_init_lock_stateowner(unsigned int strhashval, struct nfs4_client *clp, struct nfs4_ol_stateid *open_stp, struct nfsd4_lock *lock) { + struct nfs4_lockowner *lo; - if (!(sop = alloc_stateowner(&lock->lk_new_owner))) + lo = alloc_stateowner(lockowner_slab, &lock->lk_new_owner, clp); + if (!lo) return NULL; - idhashval = lockownerid_hashval(current_ownerid); - INIT_LIST_HEAD(&sop->so_idhash); - INIT_LIST_HEAD(&sop->so_strhash); - INIT_LIST_HEAD(&sop->so_perclient); - INIT_LIST_HEAD(&sop->so_stateids); - INIT_LIST_HEAD(&sop->so_perstateid); - INIT_LIST_HEAD(&sop->so_close_lru); /* not used */ - sop->so_time = 0; - list_add(&sop->so_idhash, &lock_ownerid_hashtbl[idhashval]); - list_add(&sop->so_strhash, &lock_ownerstr_hashtbl[strhashval]); - list_add(&sop->so_perstateid, &open_stp->st_lockowners); - sop->so_is_open_owner = 0; - sop->so_id = current_ownerid++; - sop->so_client = clp; + INIT_LIST_HEAD(&lo->lo_owner.so_stateids); + lo->lo_owner.so_is_open_owner = 0; /* It is the openowner seqid that will be incremented in encode in the * case of new lockowners; so increment the lock seqid manually: */ - sop->so_seqid = lock->lk_new_lock_seqid + 1; - sop->so_confirmed = 1; - rp = &sop->so_replay; - rp->rp_status = nfserr_serverfault; - rp->rp_buflen = 0; - rp->rp_buf = rp->rp_ibuf; - return sop; + lo->lo_owner.so_seqid = lock->lk_new_lock_seqid + 1; + hash_lockowner(lo, strhashval, clp, open_stp); + return lo; } -static struct nfs4_stateid * -alloc_init_lock_stateid(struct nfs4_stateowner *sop, struct nfs4_file *fp, struct nfs4_stateid *open_stp) +static struct nfs4_ol_stateid * +alloc_init_lock_stateid(struct nfs4_lockowner *lo, struct nfs4_file *fp, struct nfs4_ol_stateid *open_stp) { - struct nfs4_stateid *stp; - unsigned int hashval = stateid_hashval(sop->so_id, fp->fi_id); + struct nfs4_ol_stateid *stp; + struct nfs4_client *clp = lo->lo_owner.so_client; - stp = nfs4_alloc_stateid(); + stp = nfs4_alloc_stateid(clp); if (stp == NULL) - goto out; - INIT_LIST_HEAD(&stp->st_hash); - INIT_LIST_HEAD(&stp->st_perfile); - INIT_LIST_HEAD(&stp->st_perstateowner); - INIT_LIST_HEAD(&stp->st_lockowners); /* not used */ - list_add(&stp->st_hash, &lockstateid_hashtbl[hashval]); + return NULL; + init_stid(&stp->st_stid, clp, NFS4_LOCK_STID); list_add(&stp->st_perfile, &fp->fi_stateids); - list_add(&stp->st_perstateowner, &sop->so_stateids); - stp->st_stateowner = sop; + list_add(&stp->st_perstateowner, &lo->lo_owner.so_stateids); + stp->st_stateowner = &lo->lo_owner; get_nfs4_file(fp); stp->st_file = fp; - stp->st_stateid.si_boot = boot_time; - stp->st_stateid.si_stateownerid = sop->so_id; - stp->st_stateid.si_fileid = fp->fi_id; - stp->st_stateid.si_generation = 0; stp->st_access_bmap = 0; stp->st_deny_bmap = open_stp->st_deny_bmap; stp->st_openstp = open_stp; - -out: return stp; } @@ -3769,7 +3914,7 @@ check_lock_length(u64 offset, u64 length) LOFF_OVERFLOW(offset, length))); } -static void get_lock_access(struct nfs4_stateid *lock_stp, u32 access) +static void get_lock_access(struct nfs4_ol_stateid *lock_stp, u32 access) { struct nfs4_file *fp = lock_stp->st_file; int oflag = nfs4_access_to_omode(access); @@ -3787,15 +3932,16 @@ __be32 nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_lock *lock) { - struct nfs4_stateowner *open_sop = NULL; - struct nfs4_stateowner *lock_sop = NULL; - struct nfs4_stateid *lock_stp; + struct nfs4_openowner *open_sop = NULL; + struct nfs4_lockowner *lock_sop = NULL; + struct nfs4_ol_stateid *lock_stp; struct nfs4_file *fp; struct file *filp = NULL; struct file_lock file_lock; struct file_lock conflock; __be32 status = 0; unsigned int strhashval; + int lkflg; int err; dprintk("NFSD: nfsd4_lock: start=%Ld length=%Ld\n", @@ -3819,7 +3965,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, * Use open owner and open stateid to create lock owner and * lock stateid. */ - struct nfs4_stateid *open_stp = NULL; + struct nfs4_ol_stateid *open_stp = NULL; status = nfserr_stale_clientid; if (!nfsd4_has_session(cstate) && @@ -3827,26 +3973,29 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, goto out; /* validate and update open stateid and open seqid */ - status = nfs4_preprocess_seqid_op(cstate, + status = nfs4_preprocess_confirmed_seqid_op(cstate, lock->lk_new_open_seqid, &lock->lk_new_open_stateid, - OPEN_STATE, - &lock->lk_replay_owner, &open_stp, - lock); + &open_stp); if (status) goto out; - open_sop = lock->lk_replay_owner; + open_sop = openowner(open_stp->st_stateowner); + status = nfserr_bad_stateid; + if (!nfsd4_has_session(cstate) && + !same_clid(&open_sop->oo_owner.so_client->cl_clientid, + &lock->v.new.clientid)) + goto out; /* create lockowner and lock stateid */ fp = open_stp->st_file; - strhashval = lock_ownerstr_hashval(fp->fi_inode, - open_sop->so_client->cl_clientid.cl_id, + strhashval = lock_ownerstr_hashval(fp->fi_inode, + open_sop->oo_owner.so_client->cl_clientid.cl_id, &lock->v.new.owner); /* XXX: Do we need to check for duplicate stateowners on * the same file, or should they just be allowed (and * create new stateids)? */ status = nfserr_jukebox; lock_sop = alloc_init_lock_stateowner(strhashval, - open_sop->so_client, open_stp, lock); + open_sop->oo_owner.so_client, open_stp, lock); if (lock_sop == NULL) goto out; lock_stp = alloc_init_lock_stateid(lock_sop, fp, open_stp); @@ -3855,16 +4004,20 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, } else { /* lock (lock owner + lock stateid) already exists */ status = nfs4_preprocess_seqid_op(cstate, - lock->lk_old_lock_seqid, - &lock->lk_old_lock_stateid, - LOCK_STATE, - &lock->lk_replay_owner, &lock_stp, lock); + lock->lk_old_lock_seqid, + &lock->lk_old_lock_stateid, + NFS4_LOCK_STID, &lock_stp); if (status) goto out; - lock_sop = lock->lk_replay_owner; + lock_sop = lockowner(lock_stp->st_stateowner); fp = lock_stp->st_file; } - /* lock->lk_replay_owner and lock_stp have been created or found */ + /* lock_sop and lock_stp have been created or found */ + + lkflg = setlkflg(lock->lk_type); + status = nfs4_check_openmode(lock_stp, lkflg); + if (status) + goto out; status = nfserr_grace; if (locks_in_grace() && !lock->lk_reclaim) @@ -3915,8 +4068,8 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, err = vfs_lock_file(filp, F_SETLK, &file_lock, &conflock); switch (-err) { case 0: /* success! */ - update_stateid(&lock_stp->st_stateid); - memcpy(&lock->lk_resp_stateid, &lock_stp->st_stateid, + update_stateid(&lock_stp->st_stid.sc_stateid); + memcpy(&lock->lk_resp_stateid, &lock_stp->st_stid.sc_stateid, sizeof(stateid_t)); status = 0; break; @@ -3936,11 +4089,8 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, out: if (status && lock->lk_is_new && lock_sop) release_lockowner(lock_sop); - if (lock->lk_replay_owner) { - nfs4_get_stateowner(lock->lk_replay_owner); - cstate->replay_owner = lock->lk_replay_owner; - } - nfs4_unlock_state(); + if (!cstate->replay_owner) + nfs4_unlock_state(); return status; } @@ -3970,6 +4120,7 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, { struct inode *inode; struct file_lock file_lock; + struct nfs4_lockowner *lo; __be32 status; if (locks_in_grace()) @@ -3978,19 +4129,14 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (check_lock_length(lockt->lt_offset, lockt->lt_length)) return nfserr_inval; - lockt->lt_stateowner = NULL; nfs4_lock_state(); status = nfserr_stale_clientid; if (!nfsd4_has_session(cstate) && STALE_CLIENTID(&lockt->lt_clientid)) goto out; - if ((status = fh_verify(rqstp, &cstate->current_fh, S_IFREG, 0))) { - dprintk("NFSD: nfsd4_lockt: fh_verify() failed!\n"); - if (status == nfserr_symlink) - status = nfserr_inval; + if ((status = fh_verify(rqstp, &cstate->current_fh, S_IFREG, 0))) goto out; - } inode = cstate->current_fh.fh_dentry->d_inode; locks_init_lock(&file_lock); @@ -4009,10 +4155,9 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, goto out; } - lockt->lt_stateowner = find_lockstateowner_str(inode, - &lockt->lt_clientid, &lockt->lt_owner); - if (lockt->lt_stateowner) - file_lock.fl_owner = (fl_owner_t)lockt->lt_stateowner; + lo = find_lockowner_str(inode, &lockt->lt_clientid, &lockt->lt_owner); + if (lo) + file_lock.fl_owner = (fl_owner_t)lo; file_lock.fl_pid = current->tgid; file_lock.fl_flags = FL_POSIX; @@ -4038,7 +4183,7 @@ __be32 nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_locku *locku) { - struct nfs4_stateid *stp; + struct nfs4_ol_stateid *stp; struct file *filp = NULL; struct file_lock file_lock; __be32 status; @@ -4053,13 +4198,10 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, nfs4_lock_state(); - if ((status = nfs4_preprocess_seqid_op(cstate, - locku->lu_seqid, - &locku->lu_stateid, - LOCK_STATE, - &locku->lu_stateowner, &stp, NULL))) + status = nfs4_preprocess_seqid_op(cstate, locku->lu_seqid, + &locku->lu_stateid, NFS4_LOCK_STID, &stp); + if (status) goto out; - filp = find_any_file(stp->st_file); if (!filp) { status = nfserr_lock_range; @@ -4068,7 +4210,7 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, BUG_ON(!filp); locks_init_lock(&file_lock); file_lock.fl_type = F_UNLCK; - file_lock.fl_owner = (fl_owner_t) locku->lu_stateowner; + file_lock.fl_owner = (fl_owner_t)lockowner(stp->st_stateowner); file_lock.fl_pid = current->tgid; file_lock.fl_file = filp; file_lock.fl_flags = FL_POSIX; @@ -4089,15 +4231,12 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, /* * OK, unlock succeeded; the only thing left to do is update the stateid. */ - update_stateid(&stp->st_stateid); - memcpy(&locku->lu_stateid, &stp->st_stateid, sizeof(stateid_t)); + update_stateid(&stp->st_stid.sc_stateid); + memcpy(&locku->lu_stateid, &stp->st_stid.sc_stateid, sizeof(stateid_t)); out: - if (locku->lu_stateowner) { - nfs4_get_stateowner(locku->lu_stateowner); - cstate->replay_owner = locku->lu_stateowner; - } - nfs4_unlock_state(); + if (!cstate->replay_owner) + nfs4_unlock_state(); return status; out_nfserr: @@ -4111,7 +4250,7 @@ out_nfserr: * 0: no locks held by lockowner */ static int -check_for_locks(struct nfs4_file *filp, struct nfs4_stateowner *lowner) +check_for_locks(struct nfs4_file *filp, struct nfs4_lockowner *lowner) { struct file_lock **flpp; struct inode *inode = filp->fi_inode; @@ -4136,7 +4275,8 @@ nfsd4_release_lockowner(struct svc_rqst *rqstp, { clientid_t *clid = &rlockowner->rl_clientid; struct nfs4_stateowner *sop; - struct nfs4_stateid *stp; + struct nfs4_lockowner *lo; + struct nfs4_ol_stateid *stp; struct xdr_netobj *owner = &rlockowner->rl_owner; struct list_head matches; int i; @@ -4160,16 +4300,15 @@ nfsd4_release_lockowner(struct svc_rqst *rqstp, * data structures. */ INIT_LIST_HEAD(&matches); for (i = 0; i < LOCK_HASH_SIZE; i++) { - list_for_each_entry(sop, &lock_ownerid_hashtbl[i], so_idhash) { + list_for_each_entry(sop, &lock_ownerstr_hashtbl[i], so_strhash) { if (!same_owner_str(sop, owner, clid)) continue; list_for_each_entry(stp, &sop->so_stateids, st_perstateowner) { - if (check_for_locks(stp->st_file, sop)) + lo = lockowner(sop); + if (check_for_locks(stp->st_file, lo)) goto out; - /* Note: so_perclient unused for lockowners, - * so it's OK to fool with here. */ - list_add(&sop->so_perclient, &matches); + list_add(&lo->lo_list, &matches); } } } @@ -4178,12 +4317,12 @@ nfsd4_release_lockowner(struct svc_rqst *rqstp, * have been checked. */ status = nfs_ok; while (!list_empty(&matches)) { - sop = list_entry(matches.next, struct nfs4_stateowner, - so_perclient); + lo = list_entry(matches.next, struct nfs4_lockowner, + lo_list); /* unhash_stateowner deletes so_perclient only * for openowners. */ - list_del(&sop->so_perclient); - release_lockowner(sop); + list_del(&lo->lo_list); + release_lockowner(lo); } out: nfs4_unlock_state(); @@ -4305,16 +4444,10 @@ nfs4_state_init(void) for (i = 0; i < FILE_HASH_SIZE; i++) { INIT_LIST_HEAD(&file_hashtbl[i]); } - for (i = 0; i < OWNER_HASH_SIZE; i++) { - INIT_LIST_HEAD(&ownerstr_hashtbl[i]); - INIT_LIST_HEAD(&ownerid_hashtbl[i]); - } - for (i = 0; i < STATEID_HASH_SIZE; i++) { - INIT_LIST_HEAD(&stateid_hashtbl[i]); - INIT_LIST_HEAD(&lockstateid_hashtbl[i]); + for (i = 0; i < OPEN_OWNER_HASH_SIZE; i++) { + INIT_LIST_HEAD(&open_ownerstr_hashtbl[i]); } for (i = 0; i < LOCK_HASH_SIZE; i++) { - INIT_LIST_HEAD(&lock_ownerid_hashtbl[i]); INIT_LIST_HEAD(&lock_ownerstr_hashtbl[i]); } memset(&onestateid, ~0, sizeof(stateid_t)); @@ -4331,7 +4464,7 @@ nfsd4_load_reboot_recovery_data(void) int status; nfs4_lock_state(); - nfsd4_init_recdir(user_recovery_dirname); + nfsd4_init_recdir(); status = nfsd4_recdir_load(); nfs4_unlock_state(); if (status) @@ -4440,40 +4573,3 @@ nfs4_state_shutdown(void) nfs4_unlock_state(); nfsd4_destroy_callback_queue(); } - -/* - * user_recovery_dirname is protected by the nfsd_mutex since it's only - * accessed when nfsd is starting. - */ -static void -nfs4_set_recdir(char *recdir) -{ - strcpy(user_recovery_dirname, recdir); -} - -/* - * Change the NFSv4 recovery directory to recdir. - */ -int -nfs4_reset_recoverydir(char *recdir) -{ - int status; - struct path path; - - status = kern_path(recdir, LOOKUP_FOLLOW, &path); - if (status) - return status; - status = -ENOTDIR; - if (S_ISDIR(path.dentry->d_inode->i_mode)) { - nfs4_set_recdir(recdir); - status = 0; - } - path_put(&path); - return status; -} - -char * -nfs4_recoverydir(void) -{ - return user_recovery_dirname; -} diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 45f53ae..9d2c52b 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -44,13 +44,15 @@ #include <linux/namei.h> #include <linux/statfs.h> #include <linux/utsname.h> +#include <linux/pagemap.h> #include <linux/sunrpc/svcauth_gss.h> #include "idmap.h" #include "acl.h" #include "xdr4.h" #include "vfs.h" - +#include "state.h" +#include "cache.h" #define NFSDDBG_FACILITY NFSDDBG_XDR @@ -131,6 +133,22 @@ xdr_error: \ } \ } while (0) +static void save_buf(struct nfsd4_compoundargs *argp, struct nfsd4_saved_compoundargs *savep) +{ + savep->p = argp->p; + savep->end = argp->end; + savep->pagelen = argp->pagelen; + savep->pagelist = argp->pagelist; +} + +static void restore_buf(struct nfsd4_compoundargs *argp, struct nfsd4_saved_compoundargs *savep) +{ + argp->p = savep->p; + argp->end = savep->end; + argp->pagelen = savep->pagelen; + argp->pagelist = savep->pagelist; +} + static __be32 *read_buf(struct nfsd4_compoundargs *argp, u32 nbytes) { /* We want more bytes than seem to be available. @@ -159,8 +177,8 @@ static __be32 *read_buf(struct nfsd4_compoundargs *argp, u32 nbytes) */ memcpy(p, argp->p, avail); /* step to next page */ - argp->pagelist++; argp->p = page_address(argp->pagelist[0]); + argp->pagelist++; if (argp->pagelen < PAGE_SIZE) { argp->end = argp->p + (argp->pagelen>>2); argp->pagelen = 0; @@ -432,7 +450,6 @@ nfsd4_decode_close(struct nfsd4_compoundargs *argp, struct nfsd4_close *close) { DECODE_HEAD; - close->cl_stateowner = NULL; READ_BUF(4); READ32(close->cl_seqid); return nfsd4_decode_stateid(argp, &close->cl_stateid); @@ -465,7 +482,18 @@ nfsd4_decode_create(struct nfsd4_compoundargs *argp, struct nfsd4_create *create READ_BUF(4); READ32(create->cr_linklen); READ_BUF(create->cr_linklen); - SAVEMEM(create->cr_linkname, create->cr_linklen); + /* + * The VFS will want a null-terminated string, and + * null-terminating in place isn't safe since this might + * end on a page boundary: + */ + create->cr_linkname = + kmalloc(create->cr_linklen + 1, GFP_KERNEL); + if (!create->cr_linkname) + return nfserr_jukebox; + memcpy(create->cr_linkname, p, create->cr_linklen); + create->cr_linkname[create->cr_linklen] = '\0'; + defer_free(argp, kfree, create->cr_linkname); break; case NF4BLK: case NF4CHR: @@ -527,7 +555,6 @@ nfsd4_decode_lock(struct nfsd4_compoundargs *argp, struct nfsd4_lock *lock) { DECODE_HEAD; - lock->lk_replay_owner = NULL; /* * type, reclaim(boolean), offset, length, new_lock_owner(boolean) */ @@ -587,7 +614,6 @@ nfsd4_decode_locku(struct nfsd4_compoundargs *argp, struct nfsd4_locku *locku) { DECODE_HEAD; - locku->lu_stateowner = NULL; READ_BUF(8); READ32(locku->lu_type); if ((locku->lu_type < NFS4_READ_LT) || (locku->lu_type > NFS4_WRITEW_LT)) @@ -618,6 +644,83 @@ nfsd4_decode_lookup(struct nfsd4_compoundargs *argp, struct nfsd4_lookup *lookup DECODE_TAIL; } +static __be32 nfsd4_decode_share_access(struct nfsd4_compoundargs *argp, u32 *x) +{ + __be32 *p; + u32 w; + + READ_BUF(4); + READ32(w); + *x = w; + switch (w & NFS4_SHARE_ACCESS_MASK) { + case NFS4_SHARE_ACCESS_READ: + case NFS4_SHARE_ACCESS_WRITE: + case NFS4_SHARE_ACCESS_BOTH: + break; + default: + return nfserr_bad_xdr; + } + w &= ~NFS4_SHARE_ACCESS_MASK; + if (!w) + return nfs_ok; + if (!argp->minorversion) + return nfserr_bad_xdr; + switch (w & NFS4_SHARE_WANT_MASK) { + case NFS4_SHARE_WANT_NO_PREFERENCE: + case NFS4_SHARE_WANT_READ_DELEG: + case NFS4_SHARE_WANT_WRITE_DELEG: + case NFS4_SHARE_WANT_ANY_DELEG: + case NFS4_SHARE_WANT_NO_DELEG: + case NFS4_SHARE_WANT_CANCEL: + break; + default: + return nfserr_bad_xdr; + } + w &= ~NFS4_SHARE_WANT_MASK; + if (!w) + return nfs_ok; + switch (w) { + case NFS4_SHARE_SIGNAL_DELEG_WHEN_RESRC_AVAIL: + case NFS4_SHARE_PUSH_DELEG_WHEN_UNCONTENDED: + case (NFS4_SHARE_SIGNAL_DELEG_WHEN_RESRC_AVAIL | + NFS4_SHARE_PUSH_DELEG_WHEN_UNCONTENDED): + return nfs_ok; + } +xdr_error: + return nfserr_bad_xdr; +} + +static __be32 nfsd4_decode_share_deny(struct nfsd4_compoundargs *argp, u32 *x) +{ + __be32 *p; + + READ_BUF(4); + READ32(*x); + /* Note: unlinke access bits, deny bits may be zero. */ + if (*x & ~NFS4_SHARE_DENY_BOTH) + return nfserr_bad_xdr; + return nfs_ok; +xdr_error: + return nfserr_bad_xdr; +} + +static __be32 nfsd4_decode_opaque(struct nfsd4_compoundargs *argp, struct xdr_netobj *o) +{ + __be32 *p; + + READ_BUF(4); + READ32(o->len); + + if (o->len == 0 || o->len > NFS4_OPAQUE_LIMIT) + return nfserr_bad_xdr; + + READ_BUF(o->len); + SAVEMEM(o->data, o->len); + return nfs_ok; +xdr_error: + return nfserr_bad_xdr; +} + static __be32 nfsd4_decode_open(struct nfsd4_compoundargs *argp, struct nfsd4_open *open) { @@ -625,19 +728,23 @@ nfsd4_decode_open(struct nfsd4_compoundargs *argp, struct nfsd4_open *open) memset(open->op_bmval, 0, sizeof(open->op_bmval)); open->op_iattr.ia_valid = 0; - open->op_stateowner = NULL; + open->op_openowner = NULL; /* seqid, share_access, share_deny, clientid, ownerlen */ - READ_BUF(16 + sizeof(clientid_t)); + READ_BUF(4); READ32(open->op_seqid); - READ32(open->op_share_access); - READ32(open->op_share_deny); + status = nfsd4_decode_share_access(argp, &open->op_share_access); + if (status) + goto xdr_error; + status = nfsd4_decode_share_deny(argp, &open->op_share_deny); + if (status) + goto xdr_error; + READ_BUF(sizeof(clientid_t)); COPYMEM(&open->op_clientid, sizeof(clientid_t)); - READ32(open->op_owner.len); - - /* owner, open_flag */ - READ_BUF(open->op_owner.len + 4); - SAVEMEM(open->op_owner.data, open->op_owner.len); + status = nfsd4_decode_opaque(argp, &open->op_owner); + if (status) + goto xdr_error; + READ_BUF(4); READ32(open->op_create); switch (open->op_create) { case NFS4_OPEN_NOCREATE: @@ -703,6 +810,19 @@ nfsd4_decode_open(struct nfsd4_compoundargs *argp, struct nfsd4_open *open) if ((status = check_filename(open->op_fname.data, open->op_fname.len, nfserr_inval))) return status; break; + case NFS4_OPEN_CLAIM_FH: + case NFS4_OPEN_CLAIM_DELEG_PREV_FH: + if (argp->minorversion < 1) + goto xdr_error; + /* void */ + break; + case NFS4_OPEN_CLAIM_DELEG_CUR_FH: + if (argp->minorversion < 1) + goto xdr_error; + status = nfsd4_decode_stateid(argp, &open->op_delegate_stateid); + if (status) + return status; + break; default: goto xdr_error; } @@ -715,7 +835,6 @@ nfsd4_decode_open_confirm(struct nfsd4_compoundargs *argp, struct nfsd4_open_con { DECODE_HEAD; - open_conf->oc_stateowner = NULL; status = nfsd4_decode_stateid(argp, &open_conf->oc_req_stateid); if (status) return status; @@ -730,15 +849,17 @@ nfsd4_decode_open_downgrade(struct nfsd4_compoundargs *argp, struct nfsd4_open_d { DECODE_HEAD; - open_down->od_stateowner = NULL; status = nfsd4_decode_stateid(argp, &open_down->od_stateid); if (status) return status; - READ_BUF(12); + READ_BUF(4); READ32(open_down->od_seqid); - READ32(open_down->od_share_access); - READ32(open_down->od_share_deny); - + status = nfsd4_decode_share_access(argp, &open_down->od_share_access); + if (status) + return status; + status = nfsd4_decode_share_deny(argp, &open_down->od_share_deny); + if (status) + return status; DECODE_TAIL; } @@ -879,12 +1000,13 @@ nfsd4_decode_setclientid(struct nfsd4_compoundargs *argp, struct nfsd4_setclient { DECODE_HEAD; - READ_BUF(12); + READ_BUF(8); COPYMEM(setclientid->se_verf.data, 8); - READ32(setclientid->se_namelen); - READ_BUF(setclientid->se_namelen + 8); - SAVEMEM(setclientid->se_name, setclientid->se_namelen); + status = nfsd4_decode_opaque(argp, &setclientid->se_name); + if (status) + return nfserr_bad_xdr; + READ_BUF(8); READ32(setclientid->se_callback_prog); READ32(setclientid->se_callback_netid_len); @@ -1027,11 +1149,9 @@ nfsd4_decode_exchange_id(struct nfsd4_compoundargs *argp, READ_BUF(NFS4_VERIFIER_SIZE); COPYMEM(exid->verifier.data, NFS4_VERIFIER_SIZE); - READ_BUF(4); - READ32(exid->clname.len); - - READ_BUF(exid->clname.len); - SAVEMEM(exid->clname.data, exid->clname.len); + status = nfsd4_decode_opaque(argp, &exid->clname); + if (status) + return nfserr_bad_xdr; READ_BUF(4); READ32(exid->flags); @@ -1240,6 +1360,19 @@ nfsd4_decode_destroy_session(struct nfsd4_compoundargs *argp, } static __be32 +nfsd4_decode_free_stateid(struct nfsd4_compoundargs *argp, + struct nfsd4_free_stateid *free_stateid) +{ + DECODE_HEAD; + + READ_BUF(sizeof(stateid_t)); + READ32(free_stateid->fr_stateid.si_generation); + COPYMEM(&free_stateid->fr_stateid.si_opaque, sizeof(stateid_opaque_t)); + + DECODE_TAIL; +} + +static __be32 nfsd4_decode_sequence(struct nfsd4_compoundargs *argp, struct nfsd4_sequence *seq) { @@ -1255,6 +1388,50 @@ nfsd4_decode_sequence(struct nfsd4_compoundargs *argp, DECODE_TAIL; } +static __be32 +nfsd4_decode_test_stateid(struct nfsd4_compoundargs *argp, struct nfsd4_test_stateid *test_stateid) +{ + unsigned int nbytes; + stateid_t si; + int i; + __be32 *p; + __be32 status; + + READ_BUF(4); + test_stateid->ts_num_ids = ntohl(*p++); + + nbytes = test_stateid->ts_num_ids * sizeof(stateid_t); + if (nbytes > (u32)((char *)argp->end - (char *)argp->p)) + goto xdr_error; + + test_stateid->ts_saved_args = argp; + save_buf(argp, &test_stateid->ts_savedp); + + for (i = 0; i < test_stateid->ts_num_ids; i++) { + status = nfsd4_decode_stateid(argp, &si); + if (status) + return status; + } + + status = 0; +out: + return status; +xdr_error: + dprintk("NFSD: xdr error (%s:%d)\n", __FILE__, __LINE__); + status = nfserr_bad_xdr; + goto out; +} + +static __be32 nfsd4_decode_destroy_clientid(struct nfsd4_compoundargs *argp, struct nfsd4_destroy_clientid *dc) +{ + DECODE_HEAD; + + READ_BUF(8); + COPYMEM(&dc->clientid, 8); + + DECODE_TAIL; +} + static __be32 nfsd4_decode_reclaim_complete(struct nfsd4_compoundargs *argp, struct nfsd4_reclaim_complete *rc) { DECODE_HEAD; @@ -1364,7 +1541,7 @@ static nfsd4_dec nfsd41_dec_ops[] = { [OP_EXCHANGE_ID] = (nfsd4_dec)nfsd4_decode_exchange_id, [OP_CREATE_SESSION] = (nfsd4_dec)nfsd4_decode_create_session, [OP_DESTROY_SESSION] = (nfsd4_dec)nfsd4_decode_destroy_session, - [OP_FREE_STATEID] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_FREE_STATEID] = (nfsd4_dec)nfsd4_decode_free_stateid, [OP_GET_DIR_DELEGATION] = (nfsd4_dec)nfsd4_decode_notsupp, [OP_GETDEVICEINFO] = (nfsd4_dec)nfsd4_decode_notsupp, [OP_GETDEVICELIST] = (nfsd4_dec)nfsd4_decode_notsupp, @@ -1374,9 +1551,9 @@ static nfsd4_dec nfsd41_dec_ops[] = { [OP_SECINFO_NO_NAME] = (nfsd4_dec)nfsd4_decode_secinfo_no_name, [OP_SEQUENCE] = (nfsd4_dec)nfsd4_decode_sequence, [OP_SET_SSV] = (nfsd4_dec)nfsd4_decode_notsupp, - [OP_TEST_STATEID] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_TEST_STATEID] = (nfsd4_dec)nfsd4_decode_test_stateid, [OP_WANT_DELEGATION] = (nfsd4_dec)nfsd4_decode_notsupp, - [OP_DESTROY_CLIENTID] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_DESTROY_CLIENTID] = (nfsd4_dec)nfsd4_decode_destroy_clientid, [OP_RECLAIM_COMPLETE] = (nfsd4_dec)nfsd4_decode_reclaim_complete, }; @@ -1396,6 +1573,7 @@ nfsd4_decode_compound(struct nfsd4_compoundargs *argp) DECODE_HEAD; struct nfsd4_op *op; struct nfsd4_minorversion_ops *ops; + bool cachethis = false; int i; /* @@ -1477,7 +1655,16 @@ nfsd4_decode_compound(struct nfsd4_compoundargs *argp) argp->opcnt = i+1; break; } + /* + * We'll try to cache the result in the DRC if any one + * op in the compound wants to be cached: + */ + cachethis |= nfsd4_cache_this_op(op); } + /* Sessions make the DRC unnecessary: */ + if (argp->minorversion) + cachethis = false; + argp->rqstp->rq_cachetype = cachethis ? RC_REPLBUFF : RC_NOCACHE; DECODE_TAIL; } @@ -1542,18 +1729,6 @@ static void write_cinfo(__be32 **p, struct nfsd4_change_info *c) \ save = resp->p; -static bool seqid_mutating_err(__be32 err) -{ - /* rfc 3530 section 8.1.5: */ - return err != nfserr_stale_clientid && - err != nfserr_stale_stateid && - err != nfserr_bad_stateid && - err != nfserr_bad_seqid && - err != nfserr_bad_xdr && - err != nfserr_resource && - err != nfserr_nofilehandle; -} - /* * Routine for encoding the result of a "seqid-mutating" NFSv4 operation. This * is where sequence id's are incremented, and the replay cache is filled. @@ -1561,15 +1736,20 @@ static bool seqid_mutating_err(__be32 err) * we know whether the error to be returned is a sequence id mutating error. */ -#define ENCODE_SEQID_OP_TAIL(stateowner) do { \ - if (seqid_mutating_err(nfserr) && stateowner) { \ - stateowner->so_seqid++; \ - stateowner->so_replay.rp_status = nfserr; \ - stateowner->so_replay.rp_buflen = \ - (((char *)(resp)->p - (char *)save)); \ - memcpy(stateowner->so_replay.rp_buf, save, \ - stateowner->so_replay.rp_buflen); \ - } } while (0); +static void encode_seqid_op_tail(struct nfsd4_compoundres *resp, __be32 *save, __be32 nfserr) +{ + struct nfs4_stateowner *stateowner = resp->cstate.replay_owner; + + if (seqid_mutating_err(ntohl(nfserr)) && stateowner) { + stateowner->so_seqid++; + stateowner->so_replay.rp_status = nfserr; + stateowner->so_replay.rp_buflen = + (char *)resp->p - (char *)save; + memcpy(stateowner->so_replay.rp_buf, save, + stateowner->so_replay.rp_buflen); + nfsd4_purge_closed_stateid(stateowner); + } +} /* Encode as an array of strings the string given with components * separated @sep. @@ -1628,36 +1808,89 @@ static __be32 nfsd4_encode_fs_location4(struct nfsd4_fs_location *location, } /* - * Return the path to an export point in the pseudo filesystem namespace - * Returned string is safe to use as long as the caller holds a reference - * to @exp. + * Encode a path in RFC3530 'pathname4' format */ -static char *nfsd4_path(struct svc_rqst *rqstp, struct svc_export *exp, __be32 *stat) +static __be32 nfsd4_encode_path(const struct path *root, + const struct path *path, __be32 **pp, int *buflen) { - struct svc_fh tmp_fh; - char *path = NULL, *rootpath; - size_t rootlen; + struct path cur = { + .mnt = path->mnt, + .dentry = path->dentry, + }; + __be32 *p = *pp; + struct dentry **components = NULL; + unsigned int ncomponents = 0; + __be32 err = nfserr_jukebox; - fh_init(&tmp_fh, NFS4_FHSIZE); - *stat = exp_pseudoroot(rqstp, &tmp_fh); - if (*stat) - return NULL; - rootpath = tmp_fh.fh_export->ex_pathname; + dprintk("nfsd4_encode_components("); + + path_get(&cur); + /* First walk the path up to the nfsd root, and store the + * dentries/path components in an array. + */ + for (;;) { + if (cur.dentry == root->dentry && cur.mnt == root->mnt) + break; + if (cur.dentry == cur.mnt->mnt_root) { + if (follow_up(&cur)) + continue; + goto out_free; + } + if ((ncomponents & 15) == 0) { + struct dentry **new; + new = krealloc(components, + sizeof(*new) * (ncomponents + 16), + GFP_KERNEL); + if (!new) + goto out_free; + components = new; + } + components[ncomponents++] = cur.dentry; + cur.dentry = dget_parent(cur.dentry); + } - path = exp->ex_pathname; + *buflen -= 4; + if (*buflen < 0) + goto out_free; + WRITE32(ncomponents); - rootlen = strlen(rootpath); - if (strncmp(path, rootpath, rootlen)) { - dprintk("nfsd: fs_locations failed;" - "%s is not contained in %s\n", path, rootpath); - *stat = nfserr_notsupp; - path = NULL; - goto out; + while (ncomponents) { + struct dentry *dentry = components[ncomponents - 1]; + unsigned int len = dentry->d_name.len; + + *buflen -= 4 + (XDR_QUADLEN(len) << 2); + if (*buflen < 0) + goto out_free; + WRITE32(len); + WRITEMEM(dentry->d_name.name, len); + dprintk("/%s", dentry->d_name.name); + dput(dentry); + ncomponents--; } - path += rootlen; -out: - fh_put(&tmp_fh); - return path; + + *pp = p; + err = 0; +out_free: + dprintk(")\n"); + while (ncomponents) + dput(components[--ncomponents]); + kfree(components); + path_put(&cur); + return err; +} + +static __be32 nfsd4_encode_fsloc_fsroot(struct svc_rqst *rqstp, + const struct path *path, __be32 **pp, int *buflen) +{ + struct svc_export *exp_ps; + __be32 res; + + exp_ps = rqst_find_fsidzero_export(rqstp); + if (IS_ERR(exp_ps)) + return nfserrno(PTR_ERR(exp_ps)); + res = nfsd4_encode_path(&exp_ps->ex_path, path, pp, buflen); + exp_put(exp_ps); + return res; } /* @@ -1671,11 +1904,8 @@ static __be32 nfsd4_encode_fs_locations(struct svc_rqst *rqstp, int i; __be32 *p = *pp; struct nfsd4_fs_locations *fslocs = &exp->ex_fslocs; - char *root = nfsd4_path(rqstp, exp, &status); - if (status) - return status; - status = nfsd4_encode_components('/', root, &p, buflen); + status = nfsd4_encode_fsloc_fsroot(rqstp, &exp->ex_path, &p, buflen); if (status) return status; if ((*buflen -= 4) < 0) @@ -1691,12 +1921,19 @@ static __be32 nfsd4_encode_fs_locations(struct svc_rqst *rqstp, return 0; } -static u32 nfs4_ftypes[16] = { - NF4BAD, NF4FIFO, NF4CHR, NF4BAD, - NF4DIR, NF4BAD, NF4BLK, NF4BAD, - NF4REG, NF4BAD, NF4LNK, NF4BAD, - NF4SOCK, NF4BAD, NF4LNK, NF4BAD, -}; +static u32 nfs4_file_type(umode_t mode) +{ + switch (mode & S_IFMT) { + case S_IFIFO: return NF4FIFO; + case S_IFCHR: return NF4CHR; + case S_IFDIR: return NF4DIR; + case S_IFBLK: return NF4BLK; + case S_IFLNK: return NF4LNK; + case S_IFREG: return NF4REG; + case S_IFSOCK: return NF4SOCK; + default: return NF4BAD; + }; +} static __be32 nfsd4_encode_name(struct svc_rqst *rqstp, int whotype, uid_t id, int group, @@ -1809,8 +2046,8 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp, err = vfs_getattr(exp->ex_path.mnt, dentry, &stat); if (err) goto out_nfserr; - if ((bmval0 & (FATTR4_WORD0_FILES_FREE | FATTR4_WORD0_FILES_TOTAL | - FATTR4_WORD0_MAXNAME)) || + if ((bmval0 & (FATTR4_WORD0_FILES_AVAIL | FATTR4_WORD0_FILES_FREE | + FATTR4_WORD0_FILES_TOTAL | FATTR4_WORD0_MAXNAME)) || (bmval1 & (FATTR4_WORD1_SPACE_AVAIL | FATTR4_WORD1_SPACE_FREE | FATTR4_WORD1_SPACE_TOTAL))) { err = vfs_statfs(&path, &statfs); @@ -1885,7 +2122,7 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp, if (bmval0 & FATTR4_WORD0_TYPE) { if ((buflen -= 4) < 0) goto out_resource; - dummy = nfs4_ftypes[(stat.mode & S_IFMT) >> 12]; + dummy = nfs4_file_type(stat.mode); if (dummy == NF4BAD) goto out_serverfault; WRITE32(dummy); @@ -2187,6 +2424,8 @@ out_acl: WRITE64(stat.ino); } if (bmval2 & FATTR4_WORD2_SUPPATTR_EXCLCREAT) { + if ((buflen -= 16) < 0) + goto out_resource; WRITE32(3); WRITE32(NFSD_SUPPATTR_EXCLCREAT_WORD0); WRITE32(NFSD_SUPPATTR_EXCLCREAT_WORD1); @@ -2416,7 +2655,7 @@ nfsd4_encode_close(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_c if (!nfserr) nfsd4_encode_stateid(resp, &close->cl_stateid); - ENCODE_SEQID_OP_TAIL(close->cl_stateowner); + encode_seqid_op_tail(resp, save, nfserr); return nfserr; } @@ -2492,17 +2731,18 @@ nfsd4_encode_getfh(struct nfsd4_compoundres *resp, __be32 nfserr, struct svc_fh static void nfsd4_encode_lock_denied(struct nfsd4_compoundres *resp, struct nfsd4_lock_denied *ld) { + struct xdr_netobj *conf = &ld->ld_owner; __be32 *p; - RESERVE_SPACE(32 + XDR_LEN(ld->ld_sop ? ld->ld_sop->so_owner.len : 0)); + RESERVE_SPACE(32 + XDR_LEN(conf->len)); WRITE64(ld->ld_start); WRITE64(ld->ld_length); WRITE32(ld->ld_type); - if (ld->ld_sop) { + if (conf->len) { WRITEMEM(&ld->ld_clientid, 8); - WRITE32(ld->ld_sop->so_owner.len); - WRITEMEM(ld->ld_sop->so_owner.data, ld->ld_sop->so_owner.len); - kref_put(&ld->ld_sop->so_ref, nfs4_free_stateowner); + WRITE32(conf->len); + WRITEMEM(conf->data, conf->len); + kfree(conf->data); } else { /* non - nfsv4 lock in conflict, no clientid nor owner */ WRITE64((u64)0); /* clientid */ WRITE32(0); /* length of owner name */ @@ -2520,7 +2760,7 @@ nfsd4_encode_lock(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_lo else if (nfserr == nfserr_denied) nfsd4_encode_lock_denied(resp, &lock->lk_denied); - ENCODE_SEQID_OP_TAIL(lock->lk_replay_owner); + encode_seqid_op_tail(resp, save, nfserr); return nfserr; } @@ -2540,7 +2780,7 @@ nfsd4_encode_locku(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_l if (!nfserr) nfsd4_encode_stateid(resp, &locku->lu_stateid); - ENCODE_SEQID_OP_TAIL(locku->lu_stateowner); + encode_seqid_op_tail(resp, save, nfserr); return nfserr; } @@ -2621,7 +2861,7 @@ nfsd4_encode_open(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_op } /* XXX save filehandle here */ out: - ENCODE_SEQID_OP_TAIL(open->op_stateowner); + encode_seqid_op_tail(resp, save, nfserr); return nfserr; } @@ -2633,7 +2873,7 @@ nfsd4_encode_open_confirm(struct nfsd4_compoundres *resp, __be32 nfserr, struct if (!nfserr) nfsd4_encode_stateid(resp, &oc->oc_resp_stateid); - ENCODE_SEQID_OP_TAIL(oc->oc_stateowner); + encode_seqid_op_tail(resp, save, nfserr); return nfserr; } @@ -2645,7 +2885,7 @@ nfsd4_encode_open_downgrade(struct nfsd4_compoundres *resp, __be32 nfserr, struc if (!nfserr) nfsd4_encode_stateid(resp, &od->od_stateid); - ENCODE_SEQID_OP_TAIL(od->od_stateowner); + encode_seqid_op_tail(resp, save, nfserr); return nfserr; } @@ -2692,8 +2932,6 @@ nfsd4_encode_read(struct nfsd4_compoundres *resp, __be32 nfserr, read->rd_offset, resp->rqstp->rq_vec, read->rd_vlen, &maxcount); - if (nfserr == nfserr_symlink) - nfserr = nfserr_inval; if (nfserr) return nfserr; eof = (read->rd_offset + maxcount >= @@ -2823,8 +3061,6 @@ nfsd4_encode_readdir(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4 readdir->common.err == nfserr_toosmall && readdir->buffer == page) nfserr = nfserr_toosmall; - if (nfserr == nfserr_symlink) - nfserr = nfserr_notdir; if (nfserr) goto err_no_verf; @@ -3128,6 +3364,21 @@ nfsd4_encode_destroy_session(struct nfsd4_compoundres *resp, int nfserr, } static __be32 +nfsd4_encode_free_stateid(struct nfsd4_compoundres *resp, int nfserr, + struct nfsd4_free_stateid *free_stateid) +{ + __be32 *p; + + if (nfserr) + return nfserr; + + RESERVE_SPACE(4); + WRITE32(nfserr); + ADJUST_ARGS(); + return nfserr; +} + +static __be32 nfsd4_encode_sequence(struct nfsd4_compoundres *resp, int nfserr, struct nfsd4_sequence *seq) { @@ -3140,9 +3391,9 @@ nfsd4_encode_sequence(struct nfsd4_compoundres *resp, int nfserr, WRITEMEM(seq->sessionid.data, NFS4_MAX_SESSIONID_LEN); WRITE32(seq->seqid); WRITE32(seq->slotid); - WRITE32(seq->maxslots); - /* For now: target_maxslots = maxslots */ - WRITE32(seq->maxslots); + /* Note slotid's are numbered from zero: */ + WRITE32(seq->maxslots - 1); /* sr_highest_slotid */ + WRITE32(seq->maxslots - 1); /* sr_target_highest_slotid */ WRITE32(seq->status_flags); ADJUST_ARGS(); @@ -3150,6 +3401,37 @@ nfsd4_encode_sequence(struct nfsd4_compoundres *resp, int nfserr, return 0; } +__be32 +nfsd4_encode_test_stateid(struct nfsd4_compoundres *resp, int nfserr, + struct nfsd4_test_stateid *test_stateid) +{ + struct nfsd4_compoundargs *argp; + struct nfs4_client *cl = resp->cstate.session->se_client; + stateid_t si; + __be32 *p; + int i; + int valid; + + restore_buf(test_stateid->ts_saved_args, &test_stateid->ts_savedp); + argp = test_stateid->ts_saved_args; + + RESERVE_SPACE(4); + *p++ = htonl(test_stateid->ts_num_ids); + resp->p = p; + + nfs4_lock_state(); + for (i = 0; i < test_stateid->ts_num_ids; i++) { + nfsd4_decode_stateid(argp, &si); + valid = nfs4_validate_stateid(cl, &si); + RESERVE_SPACE(4); + *p++ = valid; + resp->p = p; + } + nfs4_unlock_state(); + + return nfserr; +} + static __be32 nfsd4_encode_noop(struct nfsd4_compoundres *resp, __be32 nfserr, void *p) { @@ -3208,7 +3490,7 @@ static nfsd4_enc nfsd4_enc_ops[] = { [OP_EXCHANGE_ID] = (nfsd4_enc)nfsd4_encode_exchange_id, [OP_CREATE_SESSION] = (nfsd4_enc)nfsd4_encode_create_session, [OP_DESTROY_SESSION] = (nfsd4_enc)nfsd4_encode_destroy_session, - [OP_FREE_STATEID] = (nfsd4_enc)nfsd4_encode_noop, + [OP_FREE_STATEID] = (nfsd4_enc)nfsd4_encode_free_stateid, [OP_GET_DIR_DELEGATION] = (nfsd4_enc)nfsd4_encode_noop, [OP_GETDEVICEINFO] = (nfsd4_enc)nfsd4_encode_noop, [OP_GETDEVICELIST] = (nfsd4_enc)nfsd4_encode_noop, @@ -3218,7 +3500,7 @@ static nfsd4_enc nfsd4_enc_ops[] = { [OP_SECINFO_NO_NAME] = (nfsd4_enc)nfsd4_encode_secinfo_no_name, [OP_SEQUENCE] = (nfsd4_enc)nfsd4_encode_sequence, [OP_SET_SSV] = (nfsd4_enc)nfsd4_encode_noop, - [OP_TEST_STATEID] = (nfsd4_enc)nfsd4_encode_noop, + [OP_TEST_STATEID] = (nfsd4_enc)nfsd4_encode_test_stateid, [OP_WANT_DELEGATION] = (nfsd4_enc)nfsd4_encode_noop, [OP_DESTROY_CLIENTID] = (nfsd4_enc)nfsd4_encode_noop, [OP_RECLAIM_COMPLETE] = (nfsd4_enc)nfsd4_encode_noop, @@ -3226,34 +3508,29 @@ static nfsd4_enc nfsd4_enc_ops[] = { /* * Calculate the total amount of memory that the compound response has taken - * after encoding the current operation. + * after encoding the current operation with pad. * - * pad: add on 8 bytes for the next operation's op_code and status so that - * there is room to cache a failure on the next operation. + * pad: if operation is non-idempotent, pad was calculate by op_rsize_bop() + * which was specified at nfsd4_operation, else pad is zero. * - * Compare this length to the session se_fmaxresp_cached. + * Compare this length to the session se_fmaxresp_sz and se_fmaxresp_cached. * * Our se_fmaxresp_cached will always be a multiple of PAGE_SIZE, and so * will be at least a page and will therefore hold the xdr_buf head. */ -static int nfsd4_check_drc_limit(struct nfsd4_compoundres *resp) +int nfsd4_check_resp_size(struct nfsd4_compoundres *resp, u32 pad) { - int status = 0; struct xdr_buf *xb = &resp->rqstp->rq_res; - struct nfsd4_compoundargs *args = resp->rqstp->rq_argp; struct nfsd4_session *session = NULL; struct nfsd4_slot *slot = resp->cstate.slot; - u32 length, tlen = 0, pad = 8; + u32 length, tlen = 0; if (!nfsd4_has_session(&resp->cstate)) - return status; + return 0; session = resp->cstate.session; - if (session == NULL || slot->sl_cachethis == 0) - return status; - - if (resp->opcnt >= args->opcnt) - pad = 0; /* this is the last operation */ + if (session == NULL) + return 0; if (xb->page_len == 0) { length = (char *)resp->p - (char *)xb->head[0].iov_base + pad; @@ -3266,10 +3543,14 @@ static int nfsd4_check_drc_limit(struct nfsd4_compoundres *resp) dprintk("%s length %u, xb->page_len %u tlen %u pad %u\n", __func__, length, xb->page_len, tlen, pad); - if (length <= session->se_fchannel.maxresp_cached) - return status; - else + if (length > session->se_fchannel.maxresp_sz) + return nfserr_rep_too_big; + + if (slot->sl_cachethis == 1 && + length > session->se_fchannel.maxresp_cached) return nfserr_rep_too_big_to_cache; + + return 0; } void @@ -3289,8 +3570,8 @@ nfsd4_encode_operation(struct nfsd4_compoundres *resp, struct nfsd4_op *op) !nfsd4_enc_ops[op->opnum]); op->status = nfsd4_enc_ops[op->opnum](resp, op->status, &op->u); /* nfsd4_check_drc_limit guarantees enough room for error status */ - if (!op->status && nfsd4_check_drc_limit(resp)) - op->status = nfserr_rep_too_big_to_cache; + if (!op->status) + op->status = nfsd4_check_resp_size(resp, 0); status: /* * Note: We write the status directly, instead of using WRITE32(), @@ -3331,8 +3612,11 @@ nfs4svc_encode_voidres(struct svc_rqst *rqstp, __be32 *p, void *dummy) return xdr_ressize_check(rqstp, p); } -void nfsd4_release_compoundargs(struct nfsd4_compoundargs *args) +int nfsd4_release_compoundargs(void *rq, __be32 *p, void *resp) { + struct svc_rqst *rqstp = rq; + struct nfsd4_compoundargs *args = rqstp->rq_argp; + if (args->ops != args->iops) { kfree(args->ops); args->ops = args->iops; @@ -3345,13 +3629,12 @@ void nfsd4_release_compoundargs(struct nfsd4_compoundargs *args) tb->release(tb->buf); kfree(tb); } + return 1; } int nfs4svc_decode_compoundargs(struct svc_rqst *rqstp, __be32 *p, struct nfsd4_compoundargs *args) { - __be32 status; - args->p = p; args->end = rqstp->rq_arg.head[0].iov_base + rqstp->rq_arg.head[0].iov_len; args->pagelist = rqstp->rq_arg.pages; @@ -3361,11 +3644,7 @@ nfs4svc_decode_compoundargs(struct svc_rqst *rqstp, __be32 *p, struct nfsd4_comp args->ops = args->iops; args->rqstp = rqstp; - status = nfsd4_decode_compound(args); - if (status) { - nfsd4_release_compoundargs(args); - } - return !status; + return !nfsd4_decode_compound(args); } int diff --git a/fs/nfsd/nfscache.c b/fs/nfsd/nfscache.c index 4666a20..2cbac34 100644 --- a/fs/nfsd/nfscache.c +++ b/fs/nfsd/nfscache.c @@ -118,7 +118,7 @@ hash_refile(struct svc_cacherep *rp) * Note that no operation within the loop may sleep. */ int -nfsd_cache_lookup(struct svc_rqst *rqstp, int type) +nfsd_cache_lookup(struct svc_rqst *rqstp) { struct hlist_node *hn; struct hlist_head *rh; @@ -128,6 +128,7 @@ nfsd_cache_lookup(struct svc_rqst *rqstp, int type) vers = rqstp->rq_vers, proc = rqstp->rq_proc; unsigned long age; + int type = rqstp->rq_cachetype; int rtn; rqstp->rq_cacherep = NULL; diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 2b1449d..c45a2ea 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -9,11 +9,11 @@ #include <linux/ctype.h> #include <linux/sunrpc/svcsock.h> -#include <linux/nfsd/syscall.h> #include <linux/lockd/lockd.h> #include <linux/sunrpc/clnt.h> #include <linux/sunrpc/gss_api.h> #include <linux/sunrpc/gss_krb5_enctypes.h> +#include <linux/module.h> #include "idmap.h" #include "nfsd.h" @@ -24,15 +24,6 @@ */ enum { NFSD_Root = 1, -#ifdef CONFIG_NFSD_DEPRECATED - NFSD_Svc, - NFSD_Add, - NFSD_Del, - NFSD_Export, - NFSD_Unexport, - NFSD_Getfd, - NFSD_Getfs, -#endif NFSD_List, NFSD_Export_features, NFSD_Fh, @@ -59,15 +50,6 @@ enum { /* * write() for these nodes. */ -#ifdef CONFIG_NFSD_DEPRECATED -static ssize_t write_svc(struct file *file, char *buf, size_t size); -static ssize_t write_add(struct file *file, char *buf, size_t size); -static ssize_t write_del(struct file *file, char *buf, size_t size); -static ssize_t write_export(struct file *file, char *buf, size_t size); -static ssize_t write_unexport(struct file *file, char *buf, size_t size); -static ssize_t write_getfd(struct file *file, char *buf, size_t size); -static ssize_t write_getfs(struct file *file, char *buf, size_t size); -#endif static ssize_t write_filehandle(struct file *file, char *buf, size_t size); static ssize_t write_unlock_ip(struct file *file, char *buf, size_t size); static ssize_t write_unlock_fs(struct file *file, char *buf, size_t size); @@ -83,15 +65,6 @@ static ssize_t write_recoverydir(struct file *file, char *buf, size_t size); #endif static ssize_t (*write_op[])(struct file *, char *, size_t) = { -#ifdef CONFIG_NFSD_DEPRECATED - [NFSD_Svc] = write_svc, - [NFSD_Add] = write_add, - [NFSD_Del] = write_del, - [NFSD_Export] = write_export, - [NFSD_Unexport] = write_unexport, - [NFSD_Getfd] = write_getfd, - [NFSD_Getfs] = write_getfs, -#endif [NFSD_Fh] = write_filehandle, [NFSD_FO_UnlockIP] = write_unlock_ip, [NFSD_FO_UnlockFS] = write_unlock_fs, @@ -130,16 +103,6 @@ static ssize_t nfsctl_transaction_write(struct file *file, const char __user *bu static ssize_t nfsctl_transaction_read(struct file *file, char __user *buf, size_t size, loff_t *pos) { -#ifdef CONFIG_NFSD_DEPRECATED - static int warned; - if (file->f_dentry->d_name.name[0] == '.' && !warned) { - printk(KERN_INFO - "Warning: \"%s\" uses deprecated NFSD interface: %s." - " This will be removed in 2.6.40\n", - current->comm, file->f_dentry->d_name.name); - warned = 1; - } -#endif if (! file->private_data) { /* An attempt to read a transaction file without writing * causes a 0-byte write so that the file can return @@ -226,303 +189,6 @@ static const struct file_operations pool_stats_operations = { * payload - write methods */ -#ifdef CONFIG_NFSD_DEPRECATED -/** - * write_svc - Start kernel's NFSD server - * - * Deprecated. /proc/fs/nfsd/threads is preferred. - * Function remains to support old versions of nfs-utils. - * - * Input: - * buf: struct nfsctl_svc - * svc_port: port number of this - * server's listener - * svc_nthreads: number of threads to start - * size: size in bytes of passed in nfsctl_svc - * Output: - * On success: returns zero - * On error: return code is negative errno value - */ -static ssize_t write_svc(struct file *file, char *buf, size_t size) -{ - struct nfsctl_svc *data; - int err; - if (size < sizeof(*data)) - return -EINVAL; - data = (struct nfsctl_svc*) buf; - err = nfsd_svc(data->svc_port, data->svc_nthreads); - if (err < 0) - return err; - return 0; -} - -/** - * write_add - Add or modify client entry in auth unix cache - * - * Deprecated. /proc/net/rpc/auth.unix.ip is preferred. - * Function remains to support old versions of nfs-utils. - * - * Input: - * buf: struct nfsctl_client - * cl_ident: '\0'-terminated C string - * containing domain name - * of client - * cl_naddr: no. of items in cl_addrlist - * cl_addrlist: array of client addresses - * cl_fhkeytype: ignored - * cl_fhkeylen: ignored - * cl_fhkey: ignored - * size: size in bytes of passed in nfsctl_client - * Output: - * On success: returns zero - * On error: return code is negative errno value - * - * Note: Only AF_INET client addresses are passed in, since - * nfsctl_client.cl_addrlist contains only in_addr fields for addresses. - */ -static ssize_t write_add(struct file *file, char *buf, size_t size) -{ - struct nfsctl_client *data; - if (size < sizeof(*data)) - return -EINVAL; - data = (struct nfsctl_client *)buf; - return exp_addclient(data); -} - -/** - * write_del - Remove client from auth unix cache - * - * Deprecated. /proc/net/rpc/auth.unix.ip is preferred. - * Function remains to support old versions of nfs-utils. - * - * Input: - * buf: struct nfsctl_client - * cl_ident: '\0'-terminated C string - * containing domain name - * of client - * cl_naddr: ignored - * cl_addrlist: ignored - * cl_fhkeytype: ignored - * cl_fhkeylen: ignored - * cl_fhkey: ignored - * size: size in bytes of passed in nfsctl_client - * Output: - * On success: returns zero - * On error: return code is negative errno value - * - * Note: Only AF_INET client addresses are passed in, since - * nfsctl_client.cl_addrlist contains only in_addr fields for addresses. - */ -static ssize_t write_del(struct file *file, char *buf, size_t size) -{ - struct nfsctl_client *data; - if (size < sizeof(*data)) - return -EINVAL; - data = (struct nfsctl_client *)buf; - return exp_delclient(data); -} - -/** - * write_export - Export part or all of a local file system - * - * Deprecated. /proc/net/rpc/{nfsd.export,nfsd.fh} are preferred. - * Function remains to support old versions of nfs-utils. - * - * Input: - * buf: struct nfsctl_export - * ex_client: '\0'-terminated C string - * containing domain name - * of client allowed to access - * this export - * ex_path: '\0'-terminated C string - * containing pathname of - * directory in local file system - * ex_dev: fsid to use for this export - * ex_ino: ignored - * ex_flags: export flags for this export - * ex_anon_uid: UID to use for anonymous - * requests - * ex_anon_gid: GID to use for anonymous - * requests - * size: size in bytes of passed in nfsctl_export - * Output: - * On success: returns zero - * On error: return code is negative errno value - */ -static ssize_t write_export(struct file *file, char *buf, size_t size) -{ - struct nfsctl_export *data; - if (size < sizeof(*data)) - return -EINVAL; - data = (struct nfsctl_export*)buf; - return exp_export(data); -} - -/** - * write_unexport - Unexport a previously exported file system - * - * Deprecated. /proc/net/rpc/{nfsd.export,nfsd.fh} are preferred. - * Function remains to support old versions of nfs-utils. - * - * Input: - * buf: struct nfsctl_export - * ex_client: '\0'-terminated C string - * containing domain name - * of client no longer allowed - * to access this export - * ex_path: '\0'-terminated C string - * containing pathname of - * directory in local file system - * ex_dev: ignored - * ex_ino: ignored - * ex_flags: ignored - * ex_anon_uid: ignored - * ex_anon_gid: ignored - * size: size in bytes of passed in nfsctl_export - * Output: - * On success: returns zero - * On error: return code is negative errno value - */ -static ssize_t write_unexport(struct file *file, char *buf, size_t size) -{ - struct nfsctl_export *data; - - if (size < sizeof(*data)) - return -EINVAL; - data = (struct nfsctl_export*)buf; - return exp_unexport(data); -} - -/** - * write_getfs - Get a variable-length NFS file handle by path - * - * Deprecated. /proc/fs/nfsd/filehandle is preferred. - * Function remains to support old versions of nfs-utils. - * - * Input: - * buf: struct nfsctl_fsparm - * gd_addr: socket address of client - * gd_path: '\0'-terminated C string - * containing pathname of - * directory in local file system - * gd_maxlen: maximum size of returned file - * handle - * size: size in bytes of passed in nfsctl_fsparm - * Output: - * On success: passed-in buffer filled with a knfsd_fh structure - * (a variable-length raw NFS file handle); - * return code is the size in bytes of the file handle - * On error: return code is negative errno value - * - * Note: Only AF_INET client addresses are passed in, since gd_addr - * is the same size as a struct sockaddr_in. - */ -static ssize_t write_getfs(struct file *file, char *buf, size_t size) -{ - struct nfsctl_fsparm *data; - struct sockaddr_in *sin; - struct auth_domain *clp; - int err = 0; - struct knfsd_fh *res; - struct in6_addr in6; - - if (size < sizeof(*data)) - return -EINVAL; - data = (struct nfsctl_fsparm*)buf; - err = -EPROTONOSUPPORT; - if (data->gd_addr.sa_family != AF_INET) - goto out; - sin = (struct sockaddr_in *)&data->gd_addr; - if (data->gd_maxlen > NFS3_FHSIZE) - data->gd_maxlen = NFS3_FHSIZE; - - res = (struct knfsd_fh*)buf; - - exp_readlock(); - - ipv6_addr_set_v4mapped(sin->sin_addr.s_addr, &in6); - - clp = auth_unix_lookup(&init_net, &in6); - if (!clp) - err = -EPERM; - else { - err = exp_rootfh(clp, data->gd_path, res, data->gd_maxlen); - auth_domain_put(clp); - } - exp_readunlock(); - if (err == 0) - err = res->fh_size + offsetof(struct knfsd_fh, fh_base); - out: - return err; -} - -/** - * write_getfd - Get a fixed-length NFS file handle by path (used by mountd) - * - * Deprecated. /proc/fs/nfsd/filehandle is preferred. - * Function remains to support old versions of nfs-utils. - * - * Input: - * buf: struct nfsctl_fdparm - * gd_addr: socket address of client - * gd_path: '\0'-terminated C string - * containing pathname of - * directory in local file system - * gd_version: fdparm structure version - * size: size in bytes of passed in nfsctl_fdparm - * Output: - * On success: passed-in buffer filled with nfsctl_res - * (a fixed-length raw NFS file handle); - * return code is the size in bytes of the file handle - * On error: return code is negative errno value - * - * Note: Only AF_INET client addresses are passed in, since gd_addr - * is the same size as a struct sockaddr_in. - */ -static ssize_t write_getfd(struct file *file, char *buf, size_t size) -{ - struct nfsctl_fdparm *data; - struct sockaddr_in *sin; - struct auth_domain *clp; - int err = 0; - struct knfsd_fh fh; - char *res; - struct in6_addr in6; - - if (size < sizeof(*data)) - return -EINVAL; - data = (struct nfsctl_fdparm*)buf; - err = -EPROTONOSUPPORT; - if (data->gd_addr.sa_family != AF_INET) - goto out; - err = -EINVAL; - if (data->gd_version < 2 || data->gd_version > NFSSVC_MAXVERS) - goto out; - - res = buf; - sin = (struct sockaddr_in *)&data->gd_addr; - exp_readlock(); - - ipv6_addr_set_v4mapped(sin->sin_addr.s_addr, &in6); - - clp = auth_unix_lookup(&init_net, &in6); - if (!clp) - err = -EPERM; - else { - err = exp_rootfh(clp, data->gd_path, &fh, NFS_FHSIZE); - auth_domain_put(clp); - } - exp_readunlock(); - - if (err == 0) { - memset(res,0, NFS_FHSIZE); - memcpy(res, &fh.fh_base, fh.fh_size); - err = NFS_FHSIZE; - } - out: - return err; -} -#endif /* CONFIG_NFSD_DEPRECATED */ /** * write_unlock_ip - Release all locks used by a client @@ -1397,15 +1063,6 @@ static ssize_t write_recoverydir(struct file *file, char *buf, size_t size) static int nfsd_fill_super(struct super_block * sb, void * data, int silent) { static struct tree_descr nfsd_files[] = { -#ifdef CONFIG_NFSD_DEPRECATED - [NFSD_Svc] = {".svc", &transaction_ops, S_IWUSR}, - [NFSD_Add] = {".add", &transaction_ops, S_IWUSR}, - [NFSD_Del] = {".del", &transaction_ops, S_IWUSR}, - [NFSD_Export] = {".export", &transaction_ops, S_IWUSR}, - [NFSD_Unexport] = {".unexport", &transaction_ops, S_IWUSR}, - [NFSD_Getfd] = {".getfd", &transaction_ops, S_IWUSR|S_IRUSR}, - [NFSD_Getfs] = {".getfs", &transaction_ops, S_IWUSR|S_IRUSR}, -#endif [NFSD_List] = {"exports", &exports_operations, S_IRUGO}, [NFSD_Export_features] = {"export_features", &export_features_operations, S_IRUGO}, diff --git a/fs/nfsd/nfsd.h b/fs/nfsd/nfsd.h index 7ecfa24..58134a2 100644 --- a/fs/nfsd/nfsd.h +++ b/fs/nfsd/nfsd.h @@ -11,13 +11,39 @@ #include <linux/types.h> #include <linux/mount.h> +#include <linux/nfs.h> +#include <linux/nfs2.h> +#include <linux/nfs3.h> +#include <linux/nfs4.h> +#include <linux/sunrpc/msg_prot.h> + #include <linux/nfsd/debug.h> #include <linux/nfsd/export.h> #include <linux/nfsd/stats.h> + /* * nfsd version */ #define NFSD_SUPPORTED_MINOR_VERSION 1 +/* + * Maximum blocksizes supported by daemon under various circumstances. + */ +#define NFSSVC_MAXBLKSIZE RPCSVC_MAXPAYLOAD +/* NFSv2 is limited by the protocol specification, see RFC 1094 */ +#define NFSSVC_MAXBLKSIZE_V2 (8*1024) + + +/* + * Largest number of bytes we need to allocate for an NFS + * call or reply. Used to control buffer sizes. We use + * the length of v3 WRITE, READDIR and READDIR replies + * which are an RPC header, up to 26 XDR units of reply + * data, and some page data. + * + * Note that accuracy here doesn't matter too much as the + * size is rounded up to a page size when allocating space. + */ +#define NFSD_BUFSIZE ((RPC_MAX_HEADER_WITH_AUTH+26)*XDR_UNIT + NFSSVC_MAXBLKSIZE) struct readdir_cd { __be32 err; /* 0, nfserr, or nfserr_eof */ @@ -335,6 +361,13 @@ static inline u32 nfsd_suppattrs2(u32 minorversion) #define NFSD_SUPPATTR_EXCLCREAT_WORD2 \ NFSD_WRITEABLE_ATTRS_WORD2 +extern int nfsd4_is_junction(struct dentry *dentry); +#else +static inline int nfsd4_is_junction(struct dentry *dentry) +{ + return 0; +} + #endif /* CONFIG_NFSD_V4 */ #endif /* LINUX_NFSD_NFSD_H */ diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index 90c6aa6..c763de5 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -59,28 +59,25 @@ static int nfsd_acceptable(void *expv, struct dentry *dentry) * the write call). */ static inline __be32 -nfsd_mode_check(struct svc_rqst *rqstp, umode_t mode, int type) +nfsd_mode_check(struct svc_rqst *rqstp, umode_t mode, int requested) { - /* Type can be negative when creating hardlinks - not to a dir */ - if (type > 0 && (mode & S_IFMT) != type) { - if (rqstp->rq_vers == 4 && (mode & S_IFMT) == S_IFLNK) - return nfserr_symlink; - else if (type == S_IFDIR) - return nfserr_notdir; - else if ((mode & S_IFMT) == S_IFDIR) - return nfserr_isdir; - else - return nfserr_inval; - } - if (type < 0 && (mode & S_IFMT) == -type) { - if (rqstp->rq_vers == 4 && (mode & S_IFMT) == S_IFLNK) - return nfserr_symlink; - else if (type == -S_IFDIR) - return nfserr_isdir; - else - return nfserr_notdir; - } - return 0; + mode &= S_IFMT; + + if (requested == 0) /* the caller doesn't care */ + return nfs_ok; + if (mode == requested) + return nfs_ok; + /* + * v4 has an error more specific than err_notdir which we should + * return in preference to err_notdir: + */ + if (rqstp->rq_vers == 4 && mode == S_IFLNK) + return nfserr_symlink; + if (requested == S_IFDIR) + return nfserr_notdir; + if (mode == S_IFDIR) + return nfserr_isdir; + return nfserr_inval; } static __be32 nfsd_setuser_and_check_port(struct svc_rqst *rqstp, diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c index 18743c4..7595582 100644 --- a/fs/nfsd/nfssvc.c +++ b/fs/nfsd/nfssvc.c @@ -8,6 +8,7 @@ #include <linux/sched.h> #include <linux/freezer.h> +#include <linux/module.h> #include <linux/fs_struct.h> #include <linux/swap.h> @@ -256,6 +257,8 @@ static void nfsd_last_thread(struct svc_serv *serv) nfsd_serv = NULL; nfsd_shutdown(); + svc_rpcb_cleanup(serv); + printk(KERN_WARNING "nfsd: last server has exited, flushing export " "cache\n"); nfsd_export_flush(); @@ -528,16 +531,9 @@ nfsd(void *vrqstp) continue; } - - /* Lock the export hash tables for reading. */ - exp_readlock(); - validate_process_creds(); svc_process(rqstp); validate_process_creds(); - - /* Unlock export hash tables */ - exp_readunlock(); } /* Clear signals before calling svc_exit_thread() */ @@ -577,8 +573,22 @@ nfsd_dispatch(struct svc_rqst *rqstp, __be32 *statp) rqstp->rq_vers, rqstp->rq_proc); proc = rqstp->rq_procinfo; + /* + * Give the xdr decoder a chance to change this if it wants + * (necessary in the NFSv4.0 compound case) + */ + rqstp->rq_cachetype = proc->pc_cachetype; + /* Decode arguments */ + xdr = proc->pc_decode; + if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base, + rqstp->rq_argp)) { + dprintk("nfsd: failed to decode arguments!\n"); + *statp = rpc_garbage_args; + return 1; + } + /* Check whether we have this call in the cache. */ - switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) { + switch (nfsd_cache_lookup(rqstp)) { case RC_INTR: case RC_DROPIT: return 0; @@ -588,16 +598,6 @@ nfsd_dispatch(struct svc_rqst *rqstp, __be32 *statp) /* do it */ } - /* Decode arguments */ - xdr = proc->pc_decode; - if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base, - rqstp->rq_argp)) { - dprintk("nfsd: failed to decode arguments!\n"); - nfsd_cache_update(rqstp, RC_NOCACHE, NULL); - *statp = rpc_garbage_args; - return 1; - } - /* need to grab the location to store the status, as * nfsv4 does some encoding while processing */ @@ -633,7 +633,7 @@ nfsd_dispatch(struct svc_rqst *rqstp, __be32 *statp) } /* Store reply in cache. */ - nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1); + nfsd_cache_update(rqstp, rqstp->rq_cachetype, statp + 1); return 1; } diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h index 858c7ba..a3cf384 100644 --- a/fs/nfsd/state.h +++ b/fs/nfsd/state.h @@ -35,6 +35,7 @@ #ifndef _NFSD4_STATE_H #define _NFSD4_STATE_H +#include <linux/idr.h> #include <linux/sunrpc/svc_xprt.h> #include <linux/nfsd/nfsfh.h> #include "nfsfh.h" @@ -45,24 +46,20 @@ typedef struct { } clientid_t; typedef struct { - u32 so_boot; - u32 so_stateownerid; - u32 so_fileid; + clientid_t so_clid; + u32 so_id; } stateid_opaque_t; typedef struct { u32 si_generation; stateid_opaque_t si_opaque; } stateid_t; -#define si_boot si_opaque.so_boot -#define si_stateownerid si_opaque.so_stateownerid -#define si_fileid si_opaque.so_fileid #define STATEID_FMT "(%08x/%08x/%08x/%08x)" #define STATEID_VAL(s) \ - (s)->si_boot, \ - (s)->si_stateownerid, \ - (s)->si_fileid, \ + (s)->si_opaque.so_clid.cl_boot, \ + (s)->si_opaque.so_clid.cl_id, \ + (s)->si_opaque.so_id, \ (s)->si_generation struct nfsd4_callback { @@ -76,17 +73,27 @@ struct nfsd4_callback { bool cb_done; }; +struct nfs4_stid { +#define NFS4_OPEN_STID 1 +#define NFS4_LOCK_STID 2 +#define NFS4_DELEG_STID 4 +/* For an open stateid kept around *only* to process close replays: */ +#define NFS4_CLOSED_STID 8 + unsigned char sc_type; + stateid_t sc_stateid; + struct nfs4_client *sc_client; +}; + struct nfs4_delegation { + struct nfs4_stid dl_stid; /* must be first field */ struct list_head dl_perfile; struct list_head dl_perclnt; struct list_head dl_recall_lru; /* delegation recalled */ atomic_t dl_count; /* ref count */ - struct nfs4_client *dl_client; struct nfs4_file *dl_file; u32 dl_type; time_t dl_time; /* For recall: */ - stateid_t dl_stateid; struct knfsd_fh dl_fh; int dl_retries; struct nfsd4_callback dl_recall; @@ -104,6 +111,11 @@ struct nfs4_cb_conn { struct svc_xprt *cb_xprt; /* minorversion 1 only */ }; +static inline struct nfs4_delegation *delegstateid(struct nfs4_stid *s) +{ + return container_of(s, struct nfs4_delegation, dl_stid); +} + /* Maximum number of slots per session. 160 is useful for long haul TCP */ #define NFSD_MAX_SLOTS_PER_SESSION 160 /* Maximum number of operations per session compound */ @@ -220,6 +232,7 @@ struct nfs4_client { struct list_head cl_idhash; /* hash by cl_clientid.id */ struct list_head cl_strhash; /* hash by cl_name */ struct list_head cl_openowners; + struct idr cl_stateids; /* stateid lookup */ struct list_head cl_delegations; struct list_head cl_lru; /* tail queue */ struct xdr_netobj cl_name; /* id generated by client */ @@ -245,6 +258,7 @@ struct nfs4_client { #define NFSD4_CB_UP 0 #define NFSD4_CB_UNKNOWN 1 #define NFSD4_CB_DOWN 2 +#define NFSD4_CB_FAULT 3 int cl_cb_state; struct nfsd4_callback cl_cb_null; struct nfsd4_session *cl_cb_session; @@ -293,6 +307,9 @@ static inline void update_stateid(stateid_t *stateid) { stateid->si_generation++; + /* Wraparound recommendation from 3530bis-13 9.1.3.2: */ + if (stateid->si_generation == 0) + stateid->si_generation = 1; } /* A reasonable value for REPLAY_ISIZE was estimated as follows: @@ -312,49 +329,57 @@ struct nfs4_replay { __be32 rp_status; unsigned int rp_buflen; char *rp_buf; - unsigned intrp_allocated; struct knfsd_fh rp_openfh; char rp_ibuf[NFSD4_REPLAY_ISIZE]; }; -/* -* nfs4_stateowner can either be an open_owner, or a lock_owner -* -* so_idhash: stateid_hashtbl[] for open owner, lockstateid_hashtbl[] -* for lock_owner -* so_strhash: ownerstr_hashtbl[] for open_owner, lock_ownerstr_hashtbl[] -* for lock_owner -* so_perclient: nfs4_client->cl_perclient entry - used when nfs4_client -* struct is reaped. -* so_perfilestate: heads the list of nfs4_stateid (either open or lock) -* and is used to ensure no dangling nfs4_stateid references when we -* release a stateowner. -* so_perlockowner: (open) nfs4_stateid->st_perlockowner entry - used when -* close is called to reap associated byte-range locks -* so_close_lru: (open) stateowner is placed on this list instead of being -* reaped (when so_perfilestate is empty) to hold the last close replay. -* reaped by laundramat thread after lease period. -*/ struct nfs4_stateowner { - struct kref so_ref; - struct list_head so_idhash; /* hash by so_id */ struct list_head so_strhash; /* hash by op_name */ - struct list_head so_perclient; struct list_head so_stateids; - struct list_head so_perstateid; /* for lockowners only */ - struct list_head so_close_lru; /* tail queue */ - time_t so_time; /* time of placement on so_close_lru */ - int so_is_open_owner; /* 1=openowner,0=lockowner */ - u32 so_id; struct nfs4_client * so_client; /* after increment in ENCODE_SEQID_OP_TAIL, represents the next * sequence id expected from the client: */ u32 so_seqid; struct xdr_netobj so_owner; /* open owner name */ - int so_confirmed; /* successful OPEN_CONFIRM? */ struct nfs4_replay so_replay; + bool so_is_open_owner; }; +struct nfs4_openowner { + struct nfs4_stateowner oo_owner; /* must be first field */ + struct list_head oo_perclient; + /* + * We keep around openowners a little while after last close, + * which saves clients from having to confirm, and allows us to + * handle close replays if they come soon enough. The close_lru + * is a list of such openowners, to be reaped by the laundromat + * thread eventually if they remain unused: + */ + struct list_head oo_close_lru; + struct nfs4_ol_stateid *oo_last_closed_stid; + time_t oo_time; /* time of placement on so_close_lru */ +#define NFS4_OO_CONFIRMED 1 +#define NFS4_OO_PURGE_CLOSE 2 +#define NFS4_OO_NEW 4 + unsigned char oo_flags; +}; + +struct nfs4_lockowner { + struct nfs4_stateowner lo_owner; /* must be first element */ + struct list_head lo_perstateid; /* for lockowners only */ + struct list_head lo_list; /* for temporary uses */ +}; + +static inline struct nfs4_openowner * openowner(struct nfs4_stateowner *so) +{ + return container_of(so, struct nfs4_openowner, oo_owner); +} + +static inline struct nfs4_lockowner * lockowner(struct nfs4_stateowner *so) +{ + return container_of(so, struct nfs4_lockowner, lo_owner); +} + /* * nfs4_file: a file opened by some number of (open) nfs4_stateowners. * o fi_perfile list is used to search for conflicting @@ -368,17 +393,17 @@ struct nfs4_file { /* One each for O_RDONLY, O_WRONLY, O_RDWR: */ struct file * fi_fds[3]; /* - * Each open or lock stateid contributes 1 to either - * fi_access[O_RDONLY], fi_access[O_WRONLY], or both, depending - * on open or lock mode: + * Each open or lock stateid contributes 0-4 to the counts + * below depending on which bits are set in st_access_bitmap: + * 1 to fi_access[O_RDONLY] if NFS4_SHARE_ACCES_READ is set + * + 1 to fi_access[O_WRONLY] if NFS4_SHARE_ACCESS_WRITE is set + * + 1 to both of the above if NFS4_SHARE_ACCESS_BOTH is set. */ atomic_t fi_access[2]; struct file *fi_deleg_file; struct file_lock *fi_lease; atomic_t fi_delegees; struct inode *fi_inode; - u32 fi_id; /* used with stateowner->so_id - * for stateid_hashtbl hash */ bool fi_had_conflict; }; @@ -408,44 +433,27 @@ static inline struct file *find_any_file(struct nfs4_file *f) return f->fi_fds[O_RDONLY]; } -/* -* nfs4_stateid can either be an open stateid or (eventually) a lock stateid -* -* (open)nfs4_stateid: one per (open)nfs4_stateowner, nfs4_file -* -* st_hash: stateid_hashtbl[] entry or lockstateid_hashtbl entry -* st_perfile: file_hashtbl[] entry. -* st_perfile_state: nfs4_stateowner->so_perfilestate -* st_perlockowner: (open stateid) list of lock nfs4_stateowners -* st_access_bmap: used only for open stateid -* st_deny_bmap: used only for open stateid -* st_openstp: open stateid lock stateid was derived from -* -* XXX: open stateids and lock stateids have diverged sufficiently that -* we should consider defining separate structs for the two cases. -*/ - -struct nfs4_stateid { - struct list_head st_hash; +/* "ol" stands for "Open or Lock". Better suggestions welcome. */ +struct nfs4_ol_stateid { + struct nfs4_stid st_stid; /* must be first field */ struct list_head st_perfile; struct list_head st_perstateowner; struct list_head st_lockowners; struct nfs4_stateowner * st_stateowner; struct nfs4_file * st_file; - stateid_t st_stateid; unsigned long st_access_bmap; unsigned long st_deny_bmap; - struct nfs4_stateid * st_openstp; + struct nfs4_ol_stateid * st_openstp; }; +static inline struct nfs4_ol_stateid *openlockstateid(struct nfs4_stid *s) +{ + return container_of(s, struct nfs4_ol_stateid, st_stid); +} + /* flags for preprocess_seqid_op() */ -#define HAS_SESSION 0x00000001 -#define CONFIRM 0x00000002 -#define OPEN_STATE 0x00000004 -#define LOCK_STATE 0x00000008 #define RD_STATE 0x00000010 #define WR_STATE 0x00000020 -#define CLOSE_STATE 0x00000040 struct nfsd4_compound_state; @@ -455,7 +463,8 @@ extern void nfs4_lock_state(void); extern void nfs4_unlock_state(void); extern int nfs4_in_grace(void); extern __be32 nfs4_check_open_reclaim(clientid_t *clid); -extern void nfs4_free_stateowner(struct kref *kref); +extern void nfs4_free_openowner(struct nfs4_openowner *); +extern void nfs4_free_lockowner(struct nfs4_lockowner *); extern int set_callback_cred(void); extern void nfsd4_probe_callback(struct nfs4_client *clp); extern void nfsd4_probe_callback_sync(struct nfs4_client *clp); @@ -467,7 +476,7 @@ extern void nfsd4_destroy_callback_queue(void); extern void nfsd4_shutdown_callback(struct nfs4_client *); extern void nfs4_put_delegation(struct nfs4_delegation *dp); extern __be32 nfs4_make_rec_clidname(char *clidname, struct xdr_netobj *clname); -extern void nfsd4_init_recdir(char *recdir_name); +extern void nfsd4_init_recdir(void); extern int nfsd4_recdir_load(void); extern void nfsd4_shutdown_recdir(void); extern int nfs4_client_to_reclaim(const char *name); @@ -476,17 +485,7 @@ extern void nfsd4_recdir_purge_old(void); extern int nfsd4_create_clid_dir(struct nfs4_client *clp); extern void nfsd4_remove_clid_dir(struct nfs4_client *clp); extern void release_session_client(struct nfsd4_session *); - -static inline void -nfs4_put_stateowner(struct nfs4_stateowner *so) -{ - kref_put(&so->so_ref, nfs4_free_stateowner); -} - -static inline void -nfs4_get_stateowner(struct nfs4_stateowner *so) -{ - kref_get(&so->so_ref); -} +extern __be32 nfs4_validate_stateid(struct nfs4_client *, stateid_t *); +extern void nfsd4_purge_closed_stateid(struct nfs4_stateowner *); #endif /* NFSD4_STATE_H */ diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index acf88ae..e2e7914 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -168,6 +168,8 @@ int nfsd_mountpoint(struct dentry *dentry, struct svc_export *exp) { if (d_mountpoint(dentry)) return 1; + if (nfsd4_is_junction(dentry)) + return 1; if (!(exp->ex_flags & NFSEXP_V4ROOT)) return 0; return dentry->d_inode != NULL; @@ -295,41 +297,12 @@ commit_metadata(struct svc_fh *fhp) } /* - * Set various file attributes. - * N.B. After this call fhp needs an fh_put + * Go over the attributes and take care of the small differences between + * NFS semantics and what Linux expects. */ -__be32 -nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, - int check_guard, time_t guardtime) +static void +nfsd_sanitize_attrs(struct inode *inode, struct iattr *iap) { - struct dentry *dentry; - struct inode *inode; - int accmode = NFSD_MAY_SATTR; - int ftype = 0; - __be32 err; - int host_err; - int size_change = 0; - - if (iap->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_SIZE)) - accmode |= NFSD_MAY_WRITE|NFSD_MAY_OWNER_OVERRIDE; - if (iap->ia_valid & ATTR_SIZE) - ftype = S_IFREG; - - /* Get inode */ - err = fh_verify(rqstp, fhp, ftype, accmode); - if (err) - goto out; - - dentry = fhp->fh_dentry; - inode = dentry->d_inode; - - /* Ignore any mode updates on symlinks */ - if (S_ISLNK(inode->i_mode)) - iap->ia_valid &= ~ATTR_MODE; - - if (!iap->ia_valid) - goto out; - /* * NFSv2 does not differentiate between "set-[ac]time-to-now" * which only requires access, and "set-[ac]time-to-X" which @@ -339,8 +312,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, * convert to "set to now" instead of "set to explicit time" * * We only call inode_change_ok as the last test as technically - * it is not an interface that we should be using. It is only - * valid if the filesystem does not define it's own i_op->setattr. + * it is not an interface that we should be using. */ #define BOTH_TIME_SET (ATTR_ATIME_SET | ATTR_MTIME_SET) #define MAX_TOUCH_TIME_ERROR (30*60) @@ -366,30 +338,6 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, iap->ia_valid &= ~BOTH_TIME_SET; } } - - /* - * The size case is special. - * It changes the file as well as the attributes. - */ - if (iap->ia_valid & ATTR_SIZE) { - if (iap->ia_size < inode->i_size) { - err = nfsd_permission(rqstp, fhp->fh_export, dentry, - NFSD_MAY_TRUNC|NFSD_MAY_OWNER_OVERRIDE); - if (err) - goto out; - } - - host_err = get_write_access(inode); - if (host_err) - goto out_nfserr; - - size_change = 1; - host_err = locks_verify_truncate(inode, NULL, iap->ia_size); - if (host_err) { - put_write_access(inode); - goto out_nfserr; - } - } /* sanitize the mode change */ if (iap->ia_valid & ATTR_MODE) { @@ -412,32 +360,120 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, iap->ia_valid |= (ATTR_KILL_SUID | ATTR_KILL_SGID); } } +} - /* Change the attributes. */ +static __be32 +nfsd_get_write_access(struct svc_rqst *rqstp, struct svc_fh *fhp, + struct iattr *iap) +{ + struct inode *inode = fhp->fh_dentry->d_inode; + int host_err; - iap->ia_valid |= ATTR_CTIME; + if (iap->ia_size < inode->i_size) { + __be32 err; + + err = nfsd_permission(rqstp, fhp->fh_export, fhp->fh_dentry, + NFSD_MAY_TRUNC | NFSD_MAY_OWNER_OVERRIDE); + if (err) + return err; + } + + host_err = get_write_access(inode); + if (host_err) + goto out_nfserrno; + + host_err = locks_verify_truncate(inode, NULL, iap->ia_size); + if (host_err) + goto out_put_write_access; + return 0; + +out_put_write_access: + put_write_access(inode); +out_nfserrno: + return nfserrno(host_err); +} + +/* + * Set various file attributes. After this call fhp needs an fh_put. + */ +__be32 +nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, + int check_guard, time_t guardtime) +{ + struct dentry *dentry; + struct inode *inode; + int accmode = NFSD_MAY_SATTR; + int ftype = 0; + __be32 err; + int host_err; + bool get_write_count; + int size_change = 0; - err = nfserr_notsync; - if (!check_guard || guardtime == inode->i_ctime.tv_sec) { - host_err = nfsd_break_lease(inode); + if (iap->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_SIZE)) + accmode |= NFSD_MAY_WRITE|NFSD_MAY_OWNER_OVERRIDE; + if (iap->ia_valid & ATTR_SIZE) + ftype = S_IFREG; + + /* Callers that do fh_verify should do the fh_want_write: */ + get_write_count = !fhp->fh_dentry; + + /* Get inode */ + err = fh_verify(rqstp, fhp, ftype, accmode); + if (err) + goto out; + if (get_write_count) { + host_err = fh_want_write(fhp); if (host_err) - goto out_nfserr; - fh_lock(fhp); + return nfserrno(host_err); + } - host_err = notify_change(dentry, iap); - err = nfserrno(host_err); - fh_unlock(fhp); + dentry = fhp->fh_dentry; + inode = dentry->d_inode; + + /* Ignore any mode updates on symlinks */ + if (S_ISLNK(inode->i_mode)) + iap->ia_valid &= ~ATTR_MODE; + + if (!iap->ia_valid) + goto out; + + nfsd_sanitize_attrs(inode, iap); + + /* + * The size case is special, it changes the file in addition to the + * attributes. + */ + if (iap->ia_valid & ATTR_SIZE) { + err = nfsd_get_write_access(rqstp, fhp, iap); + if (err) + goto out; + size_change = 1; } + + iap->ia_valid |= ATTR_CTIME; + + if (check_guard && guardtime != inode->i_ctime.tv_sec) { + err = nfserr_notsync; + goto out_put_write_access; + } + + host_err = nfsd_break_lease(inode); + if (host_err) + goto out_put_write_access_nfserror; + + fh_lock(fhp); + host_err = notify_change(dentry, iap); + fh_unlock(fhp); + +out_put_write_access_nfserror: + err = nfserrno(host_err); +out_put_write_access: if (size_change) put_write_access(inode); if (!err) commit_metadata(fhp); out: return err; - -out_nfserr: - err = nfserrno(host_err); - goto out; } #if defined(CONFIG_NFSD_V2_ACL) || \ @@ -472,6 +508,9 @@ set_nfsv4_acl_one(struct dentry *dentry, struct posix_acl *pacl, char *key) char *buf = NULL; int error = 0; + if (!pacl) + return vfs_setxattr(dentry, key, NULL, 0, 0); + buflen = posix_acl_xattr_size(pacl->a_count); buf = kmalloc(buflen, GFP_KERNEL); error = -ENOMEM; @@ -502,7 +541,7 @@ nfsd4_set_nfs4_acl(struct svc_rqst *rqstp, struct svc_fh *fhp, unsigned int flags = 0; /* Get inode */ - error = fh_verify(rqstp, fhp, 0 /* S_IFREG */, NFSD_MAY_SATTR); + error = fh_verify(rqstp, fhp, 0, NFSD_MAY_SATTR); if (error) return error; @@ -592,6 +631,22 @@ nfsd4_get_nfs4_acl(struct svc_rqst *rqstp, struct dentry *dentry, struct nfs4_ac return error; } +#define NFSD_XATTR_JUNCTION_PREFIX XATTR_TRUSTED_PREFIX "junction." +#define NFSD_XATTR_JUNCTION_TYPE NFSD_XATTR_JUNCTION_PREFIX "type" +int nfsd4_is_junction(struct dentry *dentry) +{ + struct inode *inode = dentry->d_inode; + + if (inode == NULL) + return 0; + if (inode->i_mode & S_IXUGO) + return 0; + if (!(inode->i_mode & S_ISVTX)) + return 0; + if (vfs_getxattr(dentry, NFSD_XATTR_JUNCTION_TYPE, NULL, 0) <= 0) + return 0; + return 1; +} #endif /* defined(CONFIG_NFSD_V4) */ #ifdef CONFIG_NFSD_V3 @@ -708,12 +763,13 @@ static int nfsd_open_break_lease(struct inode *inode, int access) /* * Open an existing file or directory. - * The access argument indicates the type of open (read/write/lock) + * The may_flags argument indicates the type of open (read/write/lock) + * and additional flags. * N.B. After this call fhp needs an fh_put */ __be32 nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, - int access, struct file **filp) + int may_flags, struct file **filp) { struct dentry *dentry; struct inode *inode; @@ -728,7 +784,7 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, * and (hopefully) checked permission - so allow OWNER_OVERRIDE * in case a chmod has now revoked permission. */ - err = fh_verify(rqstp, fhp, type, access | NFSD_MAY_OWNER_OVERRIDE); + err = fh_verify(rqstp, fhp, type, may_flags | NFSD_MAY_OWNER_OVERRIDE); if (err) goto out; @@ -739,7 +795,7 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, * or any access when mandatory locking enabled */ err = nfserr_perm; - if (IS_APPEND(inode) && (access & NFSD_MAY_WRITE)) + if (IS_APPEND(inode) && (may_flags & NFSD_MAY_WRITE)) goto out; /* * We must ignore files (but only files) which might have mandatory @@ -752,22 +808,30 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, if (!inode->i_fop) goto out; - host_err = nfsd_open_break_lease(inode, access); + host_err = nfsd_open_break_lease(inode, may_flags); if (host_err) /* NOMEM or WOULDBLOCK */ goto out_nfserr; - if (access & NFSD_MAY_WRITE) { - if (access & NFSD_MAY_READ) + if (may_flags & NFSD_MAY_WRITE) { + if (may_flags & NFSD_MAY_READ) flags = O_RDWR|O_LARGEFILE; else flags = O_WRONLY|O_LARGEFILE; } *filp = dentry_open(dget(dentry), mntget(fhp->fh_export->ex_path.mnt), flags, current_cred()); - if (IS_ERR(*filp)) + if (IS_ERR(*filp)) { host_err = PTR_ERR(*filp); - else - host_err = ima_file_check(*filp, access); + *filp = NULL; + } else { + host_err = ima_file_check(*filp, may_flags); + + if (may_flags & NFSD_MAY_64BIT_COOKIE) + (*filp)->f_mode |= FMODE_64BITHASH; + else + (*filp)->f_mode |= FMODE_32BITHASH; + } + out_nfserr: err = nfserrno(host_err); out: @@ -1352,7 +1416,7 @@ __be32 do_nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, char *fname, int flen, struct iattr *iap, struct svc_fh *resfhp, int createmode, u32 *verifier, - int *truncp, int *created) + bool *truncp, bool *created) { struct dentry *dentry, *dchild = NULL; struct inode *dirp; @@ -1421,7 +1485,7 @@ do_nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, switch (createmode) { case NFS3_CREATE_UNCHECKED: if (! S_ISREG(dchild->d_inode->i_mode)) - err = nfserr_exist; + goto out; else if (truncp) { /* in nfsv4, we need to treat this case a little * differently. we don't want to truncate the @@ -1440,13 +1504,19 @@ do_nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, case NFS3_CREATE_EXCLUSIVE: if ( dchild->d_inode->i_mtime.tv_sec == v_mtime && dchild->d_inode->i_atime.tv_sec == v_atime - && dchild->d_inode->i_size == 0 ) + && dchild->d_inode->i_size == 0 ) { + if (created) + *created = 1; break; + } case NFS4_CREATE_EXCLUSIVE4_1: if ( dchild->d_inode->i_mtime.tv_sec == v_mtime && dchild->d_inode->i_atime.tv_sec == v_atime - && dchild->d_inode->i_size == 0 ) + && dchild->d_inode->i_size == 0 ) { + if (created) + *created = 1; goto set_attr; + } /* fallthru */ case NFS3_CREATE_GUARDED: err = nfserr_exist; @@ -1632,10 +1702,12 @@ nfsd_link(struct svc_rqst *rqstp, struct svc_fh *ffhp, err = fh_verify(rqstp, ffhp, S_IFDIR, NFSD_MAY_CREATE); if (err) goto out; - err = fh_verify(rqstp, tfhp, -S_IFDIR, NFSD_MAY_NOP); + err = fh_verify(rqstp, tfhp, 0, NFSD_MAY_NOP); if (err) goto out; - + err = nfserr_isdir; + if (S_ISDIR(tfhp->fh_dentry->d_inode->i_mode)) + goto out; err = nfserr_perm; if (!len) goto out; @@ -1989,8 +2061,13 @@ nfsd_readdir(struct svc_rqst *rqstp, struct svc_fh *fhp, loff_t *offsetp, __be32 err; struct file *file; loff_t offset = *offsetp; + int may_flags = NFSD_MAY_READ; + + /* NFSv2 only supports 32 bit cookies */ + if (rqstp->rq_vers > 2) + may_flags |= NFSD_MAY_64BIT_COOKIE; - err = nfsd_open(rqstp, fhp, S_IFDIR, NFSD_MAY_READ, &file); + err = nfsd_open(rqstp, fhp, S_IFDIR, may_flags, &file); if (err) goto out; diff --git a/fs/nfsd/vfs.h b/fs/nfsd/vfs.h index a22e40e..c7fe962 100644 --- a/fs/nfsd/vfs.h +++ b/fs/nfsd/vfs.h @@ -10,22 +10,24 @@ /* * Flags for nfsd_permission */ -#define NFSD_MAY_NOP 0 -#define NFSD_MAY_EXEC 1 /* == MAY_EXEC */ -#define NFSD_MAY_WRITE 2 /* == MAY_WRITE */ -#define NFSD_MAY_READ 4 /* == MAY_READ */ -#define NFSD_MAY_SATTR 8 -#define NFSD_MAY_TRUNC 16 -#define NFSD_MAY_LOCK 32 -#define NFSD_MAY_MASK 63 +#define NFSD_MAY_NOP 0 +#define NFSD_MAY_EXEC 0x001 /* == MAY_EXEC */ +#define NFSD_MAY_WRITE 0x002 /* == MAY_WRITE */ +#define NFSD_MAY_READ 0x004 /* == MAY_READ */ +#define NFSD_MAY_SATTR 0x008 +#define NFSD_MAY_TRUNC 0x010 +#define NFSD_MAY_LOCK 0x020 +#define NFSD_MAY_MASK 0x03f /* extra hints to permission and open routines: */ -#define NFSD_MAY_OWNER_OVERRIDE 64 -#define NFSD_MAY_LOCAL_ACCESS 128 /* IRIX doing local access check on device special file*/ -#define NFSD_MAY_BYPASS_GSS_ON_ROOT 256 -#define NFSD_MAY_NOT_BREAK_LEASE 512 -#define NFSD_MAY_BYPASS_GSS 1024 -#define NFSD_MAY_READ_IF_EXEC 2048 +#define NFSD_MAY_OWNER_OVERRIDE 0x040 +#define NFSD_MAY_LOCAL_ACCESS 0x080 /* for device special files */ +#define NFSD_MAY_BYPASS_GSS_ON_ROOT 0x100 +#define NFSD_MAY_NOT_BREAK_LEASE 0x200 +#define NFSD_MAY_BYPASS_GSS 0x400 +#define NFSD_MAY_READ_IF_EXEC 0x800 + +#define NFSD_MAY_64BIT_COOKIE 0x1000 /* 64 bit readdir cookies for >= NFSv3 */ #define NFSD_MAY_CREATE (NFSD_MAY_EXEC|NFSD_MAY_WRITE) #define NFSD_MAY_REMOVE (NFSD_MAY_EXEC|NFSD_MAY_WRITE|NFSD_MAY_TRUNC) @@ -62,7 +64,7 @@ __be32 nfsd_access(struct svc_rqst *, struct svc_fh *, u32 *, u32 *); __be32 do_nfsd_create(struct svc_rqst *, struct svc_fh *, char *name, int len, struct iattr *attrs, struct svc_fh *res, int createmode, - u32 *verifier, int *truncp, int *created); + u32 *verifier, bool *truncp, bool *created); __be32 nfsd_commit(struct svc_rqst *, struct svc_fh *, loff_t, unsigned long); #endif /* CONFIG_NFSD_V3 */ @@ -106,4 +108,14 @@ struct posix_acl *nfsd_get_posix_acl(struct svc_fh *, int); int nfsd_set_posix_acl(struct svc_fh *, int, struct posix_acl *); #endif +static inline int fh_want_write(struct svc_fh *fh) +{ + return mnt_want_write(fh->fh_export->ex_path.mnt); +} + +static inline void fh_drop_write(struct svc_fh *fh) +{ + mnt_drop_write(fh->fh_export->ex_path.mnt); +} + #endif /* LINUX_NFSD_VFS_H */ diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h index 366401e..2364747 100644 --- a/fs/nfsd/xdr4.h +++ b/fs/nfsd/xdr4.h @@ -81,7 +81,6 @@ struct nfsd4_access { struct nfsd4_close { u32 cl_seqid; /* request */ stateid_t cl_stateid; /* request+response */ - struct nfs4_stateowner * cl_stateowner; /* response */ }; struct nfsd4_commit { @@ -131,7 +130,7 @@ struct nfsd4_link { struct nfsd4_lock_denied { clientid_t ld_clientid; - struct nfs4_stateowner *ld_sop; + struct xdr_netobj ld_owner; u64 ld_start; u64 ld_length; u32 ld_type; @@ -165,9 +164,6 @@ struct nfsd4_lock { } ok; struct nfsd4_lock_denied denied; } u; - /* The lk_replay_owner is the open owner in the open_to_lock_owner - * case and the lock owner otherwise: */ - struct nfs4_stateowner *lk_replay_owner; }; #define lk_new_open_seqid v.new.open_seqid #define lk_new_open_stateid v.new.open_stateid @@ -188,7 +184,6 @@ struct nfsd4_lockt { struct xdr_netobj lt_owner; u64 lt_offset; u64 lt_length; - struct nfs4_stateowner * lt_stateowner; struct nfsd4_lock_denied lt_denied; }; @@ -199,7 +194,6 @@ struct nfsd4_locku { stateid_t lu_stateid; u64 lu_offset; u64 lu_length; - struct nfs4_stateowner *lu_stateowner; }; @@ -232,8 +226,11 @@ struct nfsd4_open { u32 op_recall; /* recall */ struct nfsd4_change_info op_cinfo; /* response */ u32 op_rflags; /* response */ - int op_truncate; /* used during processing */ - struct nfs4_stateowner *op_stateowner; /* used during processing */ + bool op_truncate; /* used during processing */ + bool op_created; /* used during processing */ + struct nfs4_openowner *op_openowner; /* used during processing */ + struct nfs4_file *op_file; /* used during processing */ + struct nfs4_ol_stateid *op_stp; /* used during processing */ struct nfs4_acl *op_acl; }; #define op_iattr iattr @@ -243,7 +240,6 @@ struct nfsd4_open_confirm { stateid_t oc_req_stateid /* request */; u32 oc_seqid /* request */; stateid_t oc_resp_stateid /* response */; - struct nfs4_stateowner * oc_stateowner; /* response */ }; struct nfsd4_open_downgrade { @@ -251,7 +247,6 @@ struct nfsd4_open_downgrade { u32 od_seqid; u32 od_share_access; u32 od_share_deny; - struct nfs4_stateowner *od_stateowner; }; @@ -325,8 +320,7 @@ struct nfsd4_setattr { struct nfsd4_setclientid { nfs4_verifier se_verf; /* request */ - u32 se_namelen; /* request */ - char * se_name; /* request */ + struct xdr_netobj se_name; u32 se_callback_prog; /* request */ u32 se_callback_netid_len; /* request */ char * se_callback_netid_val; /* request */ @@ -342,6 +336,24 @@ struct nfsd4_setclientid_confirm { nfs4_verifier sc_confirm; }; +struct nfsd4_saved_compoundargs { + __be32 *p; + __be32 *end; + int pagelen; + struct page **pagelist; +}; + +struct nfsd4_test_stateid { + __be32 ts_num_ids; + struct nfsd4_compoundargs *ts_saved_args; + struct nfsd4_saved_compoundargs ts_savedp; +}; + +struct nfsd4_free_stateid { + stateid_t fr_stateid; /* request */ + __be32 fr_status; /* response */ +}; + /* also used for NVERIFY */ struct nfsd4_verify { u32 ve_bmval[3]; /* request */ @@ -386,6 +398,10 @@ struct nfsd4_destroy_session { struct nfs4_sessionid sessionid; }; +struct nfsd4_destroy_clientid { + clientid_t clientid; +}; + struct nfsd4_reclaim_complete { u32 rca_one_fs; }; @@ -432,10 +448,14 @@ struct nfsd4_op { struct nfsd4_destroy_session destroy_session; struct nfsd4_sequence sequence; struct nfsd4_reclaim_complete reclaim_complete; + struct nfsd4_test_stateid test_stateid; + struct nfsd4_free_stateid free_stateid; } u; struct nfs4_replay * replay; }; +bool nfsd4_cache_this_op(struct nfsd4_op *); + struct nfsd4_compoundargs { /* scratch variables for XDR decode */ __be32 * p; @@ -458,6 +478,7 @@ struct nfsd4_compoundargs { u32 opcnt; struct nfsd4_op *ops; struct nfsd4_op iops[8]; + int cachetype; }; struct nfsd4_compoundres { @@ -508,6 +529,7 @@ int nfs4svc_decode_compoundargs(struct svc_rqst *, __be32 *, struct nfsd4_compoundargs *); int nfs4svc_encode_compoundres(struct svc_rqst *, __be32 *, struct nfsd4_compoundres *); +int nfsd4_check_resp_size(struct nfsd4_compoundres *, u32); void nfsd4_encode_operation(struct nfsd4_compoundres *, struct nfsd4_op *); void nfsd4_encode_replay(struct nfsd4_compoundres *resp, struct nfsd4_op *op); __be32 nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp, @@ -534,11 +556,13 @@ extern __be32 nfsd4_sequence(struct svc_rqst *, extern __be32 nfsd4_destroy_session(struct svc_rqst *, struct nfsd4_compound_state *, struct nfsd4_destroy_session *); +extern __be32 nfsd4_destroy_clientid(struct svc_rqst *, struct nfsd4_compound_state *, struct nfsd4_destroy_clientid *); __be32 nfsd4_reclaim_complete(struct svc_rqst *, struct nfsd4_compound_state *, struct nfsd4_reclaim_complete *); extern __be32 nfsd4_process_open1(struct nfsd4_compound_state *, struct nfsd4_open *open); extern __be32 nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_open *open); +extern void nfsd4_cleanup_open_state(struct nfsd4_open *open, __be32 status); extern __be32 nfsd4_open_confirm(struct svc_rqst *rqstp, struct nfsd4_compound_state *, struct nfsd4_open_confirm *oc); extern __be32 nfsd4_close(struct svc_rqst *rqstp, @@ -559,11 +583,15 @@ extern __be32 nfsd4_release_lockowner(struct svc_rqst *rqstp, struct nfsd4_compound_state *, struct nfsd4_release_lockowner *rlockowner); -extern void nfsd4_release_compoundargs(struct nfsd4_compoundargs *); +extern int nfsd4_release_compoundargs(void *rq, __be32 *p, void *resp); extern __be32 nfsd4_delegreturn(struct svc_rqst *rqstp, struct nfsd4_compound_state *, struct nfsd4_delegreturn *dr); extern __be32 nfsd4_renew(struct svc_rqst *rqstp, struct nfsd4_compound_state *, clientid_t *clid); +extern __be32 nfsd4_test_stateid(struct svc_rqst *rqstp, + struct nfsd4_compound_state *, struct nfsd4_test_stateid *test_stateid); +extern __be32 nfsd4_free_stateid(struct svc_rqst *rqstp, + struct nfsd4_compound_state *, struct nfsd4_free_stateid *free_stateid); #endif /* diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index b2e3ff3..090d8ce 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -31,6 +31,8 @@ #include "alloc.h" #include "dat.h" +static void __nilfs_btree_init(struct nilfs_bmap *bmap); + static struct nilfs_btree_path *nilfs_btree_alloc_path(void) { struct nilfs_btree_path *path; @@ -368,6 +370,34 @@ static int nilfs_btree_node_broken(const struct nilfs_btree_node *node, return ret; } +/** + * nilfs_btree_root_broken - verify consistency of btree root node + * @node: btree root node to be examined + * @ino: inode number + * + * Return Value: If node is broken, 1 is returned. Otherwise, 0 is returned. + */ +static int nilfs_btree_root_broken(const struct nilfs_btree_node *node, + unsigned long ino) +{ + int level, flags, nchildren; + int ret = 0; + + level = nilfs_btree_node_get_level(node); + flags = nilfs_btree_node_get_flags(node); + nchildren = nilfs_btree_node_get_nchildren(node); + + if (unlikely(level < NILFS_BTREE_LEVEL_NODE_MIN || + level >= NILFS_BTREE_LEVEL_MAX || + nchildren < 0 || + nchildren > NILFS_BTREE_ROOT_NCHILDREN_MAX)) { + pr_crit("NILFS: bad btree root (inode number=%lu): level = %d, flags = 0x%x, nchildren = %d\n", + ino, level, flags, nchildren); + ret = 1; + } + return ret; +} + int nilfs_btree_broken_node_block(struct buffer_head *bh) { int ret; @@ -1713,7 +1743,7 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *btree, /* convert and insert */ dat = NILFS_BMAP_USE_VBN(btree) ? nilfs_bmap_get_dat(btree) : NULL; - nilfs_btree_init(btree); + __nilfs_btree_init(btree); if (nreq != NULL) { nilfs_bmap_commit_alloc_ptr(btree, dreq, dat); nilfs_bmap_commit_alloc_ptr(btree, nreq, dat); @@ -2294,12 +2324,23 @@ static const struct nilfs_bmap_operations nilfs_btree_ops_gc = { .bop_gather_data = NULL, }; -int nilfs_btree_init(struct nilfs_bmap *bmap) +static void __nilfs_btree_init(struct nilfs_bmap *bmap) { bmap->b_ops = &nilfs_btree_ops; bmap->b_nchildren_per_block = NILFS_BTREE_NODE_NCHILDREN_MAX(nilfs_btree_node_size(bmap)); - return 0; +} + +int nilfs_btree_init(struct nilfs_bmap *bmap) +{ + int ret = 0; + + __nilfs_btree_init(bmap); + + if (nilfs_btree_root_broken(nilfs_btree_get_root(bmap), + bmap->b_inode->i_ino)) + ret = -EIO; + return ret; } void nilfs_btree_init_gc(struct nilfs_bmap *bmap) diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c index d7eeca6..2660152 100644 --- a/fs/nilfs2/file.c +++ b/fs/nilfs2/file.c @@ -27,7 +27,7 @@ #include "nilfs.h" #include "segment.h" -int nilfs_sync_file(struct file *file, int datasync) +int nilfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) { /* * Called from fsync() system call @@ -40,8 +40,15 @@ int nilfs_sync_file(struct file *file, int datasync) struct inode *inode = file->f_mapping->host; int err; - if (!nilfs_inode_dirty(inode)) + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + mutex_lock(&inode->i_mutex); + + if (!nilfs_inode_dirty(inode)) { + mutex_unlock(&inode->i_mutex); return 0; + } if (datasync) err = nilfs_construct_dsync_segment(inode->i_sb, inode, 0, @@ -49,6 +56,7 @@ int nilfs_sync_file(struct file *file, int datasync) else err = nilfs_construct_segment(inode->i_sb); + mutex_unlock(&inode->i_mutex); return err; } diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 373cd7b..b2d8a96 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -24,6 +24,7 @@ #include <linux/buffer_head.h> #include <linux/gfp.h> #include <linux/mpage.h> +#include <linux/pagemap.h> #include <linux/writeback.h> #include <linux/uio.h> #include "nilfs.h" @@ -195,10 +196,10 @@ static int nilfs_writepage(struct page *page, struct writeback_control *wbc) static int nilfs_set_page_dirty(struct page *page) { + struct inode *inode = page->mapping->host; int ret = __set_page_dirty_nobuffers(page); if (page_has_buffers(page)) { - struct inode *inode = page->mapping->host; unsigned nr_dirty = 0; struct buffer_head *bh, *head; @@ -221,6 +222,10 @@ static int nilfs_set_page_dirty(struct page *page) if (nr_dirty) nilfs_set_file_dirty(inode, nr_dirty); + } else if (ret) { + unsigned nr_dirty = 1 << (PAGE_CACHE_SHIFT - inode->i_blkbits); + + nilfs_set_file_dirty(inode, nr_dirty); } return ret; } @@ -278,8 +283,8 @@ nilfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, return 0; /* Needs synchronization with the cleaner */ - size = blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov, - offset, nr_segs, nilfs_get_block, NULL); + size = blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs, + nilfs_get_block); /* * In case of error extending write may have instantiated a few @@ -373,7 +378,7 @@ struct inode *nilfs_new_inode(struct inode *dir, int mode) failed_acl: failed_bmap: - inode->i_nlink = 0; + clear_nlink(inode); iput(inode); /* raw_inode will be deleted through generic_delete_inode() */ goto failed; @@ -415,7 +420,7 @@ int nilfs_read_inode_common(struct inode *inode, inode->i_mode = le16_to_cpu(raw_inode->i_mode); inode->i_uid = (uid_t)le32_to_cpu(raw_inode->i_uid); inode->i_gid = (gid_t)le32_to_cpu(raw_inode->i_gid); - inode->i_nlink = le16_to_cpu(raw_inode->i_links_count); + set_nlink(inode, le16_to_cpu(raw_inode->i_links_count)); inode->i_size = le64_to_cpu(raw_inode->i_size); inode->i_atime.tv_sec = le64_to_cpu(raw_inode->i_mtime); inode->i_ctime.tv_sec = le64_to_cpu(raw_inode->i_ctime); @@ -797,6 +802,8 @@ int nilfs_setattr(struct dentry *dentry, struct iattr *iattr) if ((iattr->ia_valid & ATTR_SIZE) && iattr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + err = vmtruncate(inode, iattr->ia_size); if (unlikely(err)) goto out_err; @@ -818,14 +825,14 @@ out_err: return err; } -int nilfs_permission(struct inode *inode, int mask, unsigned int flags) +int nilfs_permission(struct inode *inode, int mask) { struct nilfs_root *root = NILFS_I(inode)->i_root; if ((mask & MAY_WRITE) && root && root->cno != NILFS_CPTREE_CURRENT_CNO) return -EROFS; /* snapshot is not writable */ - return generic_permission(inode, mask, flags, NULL); + return generic_permission(inode, mask); } int nilfs_load_inode_block(struct inode *inode, struct buffer_head **pbh) diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index cee648e..2b5e695 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -625,6 +625,9 @@ static int nilfs_ioctl_clean_segments(struct inode *inode, struct file *filp, if (argv[n].v_nmembs > nsegs * nilfs->ns_blocks_per_segment) goto out_free; + if (argv[n].v_nmembs >= UINT_MAX / argv[n].v_size) + goto out_free; + len = argv[n].v_size * argv[n].v_nmembs; base = (void __user *)(unsigned long)argv[n].v_base; if (len == 0) { diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 546849b..768982d 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -72,12 +72,7 @@ nilfs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd) return ERR_PTR(-ENAMETOOLONG); ino = nilfs_inode_by_name(dir, &dentry->d_name); - inode = NULL; - if (ino) { - inode = nilfs_iget(dir->i_sb, NILFS_I(dir)->i_root, ino); - if (IS_ERR(inode)) - return ERR_CAST(inode); - } + inode = ino ? nilfs_iget(dir->i_sb, NILFS_I(dir)->i_root, ino) : NULL; return d_splice_alias(inode, dentry); } @@ -294,7 +289,7 @@ static int nilfs_do_unlink(struct inode *dir, struct dentry *dentry) nilfs_warning(inode->i_sb, __func__, "deleting nonexistent file (%lu), %d\n", inode->i_ino, inode->i_nlink); - inode->i_nlink = 1; + set_nlink(inode, 1); } err = nilfs_delete_entry(de, page); if (err) diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index f02b9ad..7a2e7b8 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -128,7 +128,6 @@ enum { * @ti_save: Backup of journal_info field of task_struct * @ti_flags: Flags * @ti_count: Nest level - * @ti_garbage: List of inode to be put when releasing semaphore */ struct nilfs_transaction_info { u32 ti_magic; @@ -137,7 +136,6 @@ struct nilfs_transaction_info { one of other filesystems has a bug. */ unsigned short ti_flags; unsigned short ti_count; - struct list_head ti_garbage; }; /* ti_magic */ @@ -235,7 +233,7 @@ extern void nilfs_set_link(struct inode *, struct nilfs_dir_entry *, struct page *, struct inode *); /* file.c */ -extern int nilfs_sync_file(struct file *, int); +extern int nilfs_sync_file(struct file *, loff_t, loff_t, int); /* ioctl.c */ long nilfs_ioctl(struct file *, unsigned int, unsigned long); @@ -264,7 +262,7 @@ extern void nilfs_update_inode(struct inode *, struct buffer_head *); extern void nilfs_truncate(struct inode *); extern void nilfs_evict_inode(struct inode *); extern int nilfs_setattr(struct dentry *, struct iattr *); -int nilfs_permission(struct inode *inode, int mask, unsigned int flags); +int nilfs_permission(struct inode *inode, int mask); int nilfs_load_inode_block(struct inode *inode, struct buffer_head **pbh); extern int nilfs_inode_dirty(struct inode *); int nilfs_set_file_dirty(struct inode *inode, unsigned nr_dirty); @@ -276,10 +274,10 @@ int nilfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, /* super.c */ extern struct inode *nilfs_alloc_inode(struct super_block *); extern void nilfs_destroy_inode(struct inode *); -extern void nilfs_error(struct super_block *, const char *, const char *, ...) - __attribute__ ((format (printf, 3, 4))); -extern void nilfs_warning(struct super_block *, const char *, const char *, ...) - __attribute__ ((format (printf, 3, 4))); +extern __printf(3, 4) +void nilfs_error(struct super_block *, const char *, const char *, ...); +extern __printf(3, 4) +void nilfs_warning(struct super_block *, const char *, const char *, ...); extern struct nilfs_super_block * nilfs_read_super_block(struct super_block *, u64, int, struct buffer_head **); extern int nilfs_store_magic_and_option(struct super_block *, diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index 65221a0..16eacec 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -94,6 +94,7 @@ void nilfs_forget_buffer(struct buffer_head *bh) clear_buffer_nilfs_volatile(bh); clear_buffer_nilfs_checked(bh); clear_buffer_nilfs_redirected(bh); + clear_buffer_async_write(bh); clear_buffer_dirty(bh); if (nilfs_page_buffers_clean(page)) __nilfs_clear_page_dirty(page); @@ -390,6 +391,7 @@ void nilfs_clear_dirty_pages(struct address_space *mapping) bh = head = page_buffers(page); do { lock_buffer(bh); + clear_buffer_async_write(bh); clear_buffer_dirty(bh); clear_buffer_nilfs_volatile(bh); clear_buffer_nilfs_checked(bh); diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 6f24e67..6bba106 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -302,7 +302,6 @@ static void nilfs_transaction_lock(struct super_block *sb, ti->ti_count = 0; ti->ti_save = cur_ti; ti->ti_magic = NILFS_TI_MAGIC; - INIT_LIST_HEAD(&ti->ti_garbage); current->journal_info = ti; for (;;) { @@ -329,8 +328,6 @@ static void nilfs_transaction_unlock(struct super_block *sb) up_write(&nilfs->ns_segctor_sem); current->journal_info = ti->ti_save; - if (!list_empty(&ti->ti_garbage)) - nilfs_dispose_list(nilfs, &ti->ti_garbage, 0); } static void *nilfs_segctor_map_segsum_entry(struct nilfs_sc_info *sci, @@ -662,7 +659,7 @@ static size_t nilfs_lookup_dirty_data_buffers(struct inode *inode, bh = head = page_buffers(page); do { - if (!buffer_dirty(bh)) + if (!buffer_dirty(bh) || buffer_async_write(bh)) continue; get_bh(bh); list_add_tail(&bh->b_assoc_buffers, listp); @@ -696,7 +693,8 @@ static void nilfs_lookup_dirty_node_buffers(struct inode *inode, for (i = 0; i < pagevec_count(&pvec); i++) { bh = head = page_buffers(pvec.pages[i]); do { - if (buffer_dirty(bh)) { + if (buffer_dirty(bh) && + !buffer_async_write(bh)) { get_bh(bh); list_add_tail(&bh->b_assoc_buffers, listp); @@ -742,6 +740,15 @@ static void nilfs_dispose_list(struct the_nilfs *nilfs, } } +static void nilfs_iput_work_func(struct work_struct *work) +{ + struct nilfs_sc_info *sci = container_of(work, struct nilfs_sc_info, + sc_iput_work); + struct the_nilfs *nilfs = sci->sc_super->s_fs_info; + + nilfs_dispose_list(nilfs, &sci->sc_iput_queue, 0); +} + static int nilfs_test_metadata_dirty(struct the_nilfs *nilfs, struct nilfs_root *root) { @@ -1436,17 +1443,19 @@ static int nilfs_segctor_collect(struct nilfs_sc_info *sci, nilfs_clear_logs(&sci->sc_segbufs); - err = nilfs_segctor_extend_segments(sci, nilfs, nadd); - if (unlikely(err)) - return err; - if (sci->sc_stage.flags & NILFS_CF_SUFREED) { err = nilfs_sufile_cancel_freev(nilfs->ns_sufile, sci->sc_freesegs, sci->sc_nfreesegs, NULL); WARN_ON(err); /* do not happen */ + sci->sc_stage.flags &= ~NILFS_CF_SUFREED; } + + err = nilfs_segctor_extend_segments(sci, nilfs, nadd); + if (unlikely(err)) + return err; + nadd = min_t(int, nadd << 1, SC_MAX_SEGDELTA); sci->sc_stage = prev_stage; } @@ -1576,6 +1585,7 @@ static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci) list_for_each_entry(bh, &segbuf->sb_segsum_buffers, b_assoc_buffers) { + set_buffer_async_write(bh); if (bh->b_page != bd_page) { if (bd_page) { lock_page(bd_page); @@ -1589,6 +1599,7 @@ static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci) list_for_each_entry(bh, &segbuf->sb_payload_buffers, b_assoc_buffers) { + set_buffer_async_write(bh); if (bh == segbuf->sb_super_root) { if (bh->b_page != bd_page) { lock_page(bd_page); @@ -1674,6 +1685,7 @@ static void nilfs_abort_logs(struct list_head *logs, int err) list_for_each_entry(segbuf, logs, sb_list) { list_for_each_entry(bh, &segbuf->sb_segsum_buffers, b_assoc_buffers) { + clear_buffer_async_write(bh); if (bh->b_page != bd_page) { if (bd_page) end_page_writeback(bd_page); @@ -1683,6 +1695,7 @@ static void nilfs_abort_logs(struct list_head *logs, int err) list_for_each_entry(bh, &segbuf->sb_payload_buffers, b_assoc_buffers) { + clear_buffer_async_write(bh); if (bh == segbuf->sb_super_root) { if (bh->b_page != bd_page) { end_page_writeback(bd_page); @@ -1752,6 +1765,7 @@ static void nilfs_segctor_complete_write(struct nilfs_sc_info *sci) b_assoc_buffers) { set_buffer_uptodate(bh); clear_buffer_dirty(bh); + clear_buffer_async_write(bh); if (bh->b_page != bd_page) { if (bd_page) end_page_writeback(bd_page); @@ -1773,6 +1787,7 @@ static void nilfs_segctor_complete_write(struct nilfs_sc_info *sci) b_assoc_buffers) { set_buffer_uptodate(bh); clear_buffer_dirty(bh); + clear_buffer_async_write(bh); clear_buffer_delay(bh); clear_buffer_nilfs_volatile(bh); clear_buffer_nilfs_redirected(bh); @@ -1887,8 +1902,9 @@ static int nilfs_segctor_collect_dirty_files(struct nilfs_sc_info *sci, static void nilfs_segctor_drop_written_files(struct nilfs_sc_info *sci, struct the_nilfs *nilfs) { - struct nilfs_transaction_info *ti = current->journal_info; struct nilfs_inode_info *ii, *n; + int during_mount = !(sci->sc_super->s_flags & MS_ACTIVE); + int defer_iput = false; spin_lock(&nilfs->ns_inode_lock); list_for_each_entry_safe(ii, n, &sci->sc_dirty_files, i_dirty) { @@ -1899,9 +1915,24 @@ static void nilfs_segctor_drop_written_files(struct nilfs_sc_info *sci, clear_bit(NILFS_I_BUSY, &ii->i_state); brelse(ii->i_bh); ii->i_bh = NULL; - list_move_tail(&ii->i_dirty, &ti->ti_garbage); + list_del_init(&ii->i_dirty); + if (!ii->vfs_inode.i_nlink || during_mount) { + /* + * Defer calling iput() to avoid deadlocks if + * i_nlink == 0 or mount is not yet finished. + */ + list_add_tail(&ii->i_dirty, &sci->sc_iput_queue); + defer_iput = true; + } else { + spin_unlock(&nilfs->ns_inode_lock); + iput(&ii->vfs_inode); + spin_lock(&nilfs->ns_inode_lock); + } } spin_unlock(&nilfs->ns_inode_lock); + + if (defer_iput) + schedule_work(&sci->sc_iput_work); } /* @@ -2568,6 +2599,8 @@ static struct nilfs_sc_info *nilfs_segctor_new(struct super_block *sb, INIT_LIST_HEAD(&sci->sc_segbufs); INIT_LIST_HEAD(&sci->sc_write_logs); INIT_LIST_HEAD(&sci->sc_gc_inodes); + INIT_LIST_HEAD(&sci->sc_iput_queue); + INIT_WORK(&sci->sc_iput_work, nilfs_iput_work_func); init_timer(&sci->sc_timer); sci->sc_interval = HZ * NILFS_SC_DEFAULT_TIMEOUT; @@ -2594,6 +2627,8 @@ static void nilfs_segctor_write_out(struct nilfs_sc_info *sci) ret = nilfs_segctor_construct(sci, SC_LSEG_SR); nilfs_transaction_unlock(sci->sc_super); + flush_work(&sci->sc_iput_work); + } while (ret && retrycount-- > 0); } @@ -2618,6 +2653,9 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci) || sci->sc_seq_request != sci->sc_seq_done); spin_unlock(&sci->sc_state_lock); + if (flush_work(&sci->sc_iput_work)) + flag = true; + if (flag || !nilfs_segctor_confirm(sci)) nilfs_segctor_write_out(sci); @@ -2627,6 +2665,12 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci) nilfs_dispose_list(nilfs, &sci->sc_dirty_files, 1); } + if (!list_empty(&sci->sc_iput_queue)) { + nilfs_warning(sci->sc_super, __func__, + "iput queue is not empty\n"); + nilfs_dispose_list(nilfs, &sci->sc_iput_queue, 1); + } + WARN_ON(!list_empty(&sci->sc_segbufs)); WARN_ON(!list_empty(&sci->sc_write_logs)); diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h index 38a1d00..a48d6de 100644 --- a/fs/nilfs2/segment.h +++ b/fs/nilfs2/segment.h @@ -26,6 +26,7 @@ #include <linux/types.h> #include <linux/fs.h> #include <linux/buffer_head.h> +#include <linux/workqueue.h> #include <linux/nilfs2_fs.h> #include "nilfs.h" @@ -92,6 +93,8 @@ struct nilfs_segsum_pointer { * @sc_nblk_inc: Block count of current generation * @sc_dirty_files: List of files to be written * @sc_gc_inodes: List of GC inodes having blocks to be written + * @sc_iput_queue: list of inodes for which iput should be done + * @sc_iput_work: work struct to defer iput call * @sc_freesegs: array of segment numbers to be freed * @sc_nfreesegs: number of segments on @sc_freesegs * @sc_dsync_inode: inode whose data pages are written for a sync operation @@ -135,6 +138,8 @@ struct nilfs_sc_info { struct list_head sc_dirty_files; struct list_head sc_gc_inodes; + struct list_head sc_iput_queue; + struct work_struct sc_iput_work; __u64 *sc_freesegs; size_t sc_nfreesegs; diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c index e913ad1..a721907 100644 --- a/fs/ocfs2/acl.c +++ b/fs/ocfs2/acl.c @@ -247,7 +247,7 @@ static int ocfs2_set_acl(handle_t *handle, case ACL_TYPE_ACCESS: name_index = OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS; if (acl) { - mode_t mode = inode->i_mode; + umode_t mode = inode->i_mode; ret = posix_acl_equiv_mode(acl, &mode); if (ret < 0) return ret; @@ -290,47 +290,32 @@ static int ocfs2_set_acl(handle_t *handle, return ret; } -int ocfs2_check_acl(struct inode *inode, int mask, unsigned int flags) +struct posix_acl *ocfs2_iop_get_acl(struct inode *inode, int type) { struct ocfs2_super *osb; struct buffer_head *di_bh = NULL; struct posix_acl *acl; int ret = -EAGAIN; - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - osb = OCFS2_SB(inode->i_sb); if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return ret; + return NULL; ret = ocfs2_read_inode_block(inode, &di_bh); - if (ret < 0) { - mlog_errno(ret); - return ret; - } + if (ret < 0) + return ERR_PTR(ret); - acl = ocfs2_get_acl_nolock(inode, ACL_TYPE_ACCESS, di_bh); + acl = ocfs2_get_acl_nolock(inode, type, di_bh); brelse(di_bh); - if (IS_ERR(acl)) { - mlog_errno(PTR_ERR(acl)); - return PTR_ERR(acl); - } - if (acl) { - ret = posix_acl_permission(inode, acl, mask); - posix_acl_release(acl); - return ret; - } - - return -EAGAIN; + return acl; } int ocfs2_acl_chmod(struct inode *inode) { struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - struct posix_acl *acl, *clone; + struct posix_acl *acl; int ret; if (S_ISLNK(inode->i_mode)) @@ -342,15 +327,12 @@ int ocfs2_acl_chmod(struct inode *inode) acl = ocfs2_get_acl(inode, ACL_TYPE_ACCESS); if (IS_ERR(acl) || !acl) return PTR_ERR(acl); - clone = posix_acl_clone(acl, GFP_KERNEL); + ret = posix_acl_chmod(&acl, GFP_KERNEL, inode->i_mode); + if (ret) + return ret; + ret = ocfs2_set_acl(NULL, inode, NULL, ACL_TYPE_ACCESS, + acl, NULL, NULL); posix_acl_release(acl); - if (!clone) - return -ENOMEM; - ret = posix_acl_chmod_masq(clone, inode->i_mode); - if (!ret) - ret = ocfs2_set_acl(NULL, inode, NULL, ACL_TYPE_ACCESS, - clone, NULL, NULL); - posix_acl_release(clone); return ret; } @@ -369,7 +351,7 @@ int ocfs2_init_acl(handle_t *handle, struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); struct posix_acl *acl = NULL; int ret = 0, ret2; - mode_t mode; + umode_t mode; if (!S_ISLNK(inode->i_mode)) { if (osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) { @@ -388,8 +370,6 @@ int ocfs2_init_acl(handle_t *handle, } } if ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) && acl) { - struct posix_acl *clone; - if (S_ISDIR(inode->i_mode)) { ret = ocfs2_set_acl(handle, inode, di_bh, ACL_TYPE_DEFAULT, acl, @@ -397,27 +377,22 @@ int ocfs2_init_acl(handle_t *handle, if (ret) goto cleanup; } - clone = posix_acl_clone(acl, GFP_NOFS); - ret = -ENOMEM; - if (!clone) - goto cleanup; - mode = inode->i_mode; - ret = posix_acl_create_masq(clone, &mode); - if (ret >= 0) { - ret2 = ocfs2_acl_set_mode(inode, di_bh, handle, mode); - if (ret2) { - mlog_errno(ret2); - ret = ret2; - goto cleanup; - } - if (ret > 0) { - ret = ocfs2_set_acl(handle, inode, - di_bh, ACL_TYPE_ACCESS, - clone, meta_ac, data_ac); - } + ret = posix_acl_create(&acl, GFP_NOFS, &mode); + if (ret < 0) + return ret; + + ret2 = ocfs2_acl_set_mode(inode, di_bh, handle, mode); + if (ret2) { + mlog_errno(ret2); + ret = ret2; + goto cleanup; + } + if (ret > 0) { + ret = ocfs2_set_acl(handle, inode, + di_bh, ACL_TYPE_ACCESS, + acl, meta_ac, data_ac); } - posix_acl_release(clone); } cleanup: posix_acl_release(acl); diff --git a/fs/ocfs2/acl.h b/fs/ocfs2/acl.h index 4fe7c9c..071fbd38 100644 --- a/fs/ocfs2/acl.h +++ b/fs/ocfs2/acl.h @@ -26,7 +26,7 @@ struct ocfs2_acl_entry { __le32 e_id; }; -extern int ocfs2_check_acl(struct inode *, int, unsigned int); +struct posix_acl *ocfs2_iop_get_acl(struct inode *inode, int type); extern int ocfs2_acl_chmod(struct inode *); extern int ocfs2_init_acl(handle_t *, struct inode *, struct inode *, struct buffer_head *, struct buffer_head *, diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 76c8165..31b9463 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -5699,7 +5699,7 @@ int ocfs2_remove_btree_range(struct inode *inode, OCFS2_JOURNAL_ACCESS_WRITE); if (ret) { mlog_errno(ret); - goto out; + goto out_commit; } dquot_free_space_nodirty(inode, diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index ac97bca..16653b2 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -290,7 +290,15 @@ static int ocfs2_readpage(struct file *file, struct page *page) } if (down_read_trylock(&oi->ip_alloc_sem) == 0) { + /* + * Unlock the page and cycle ip_alloc_sem so that we don't + * busyloop waiting for ip_alloc_sem to unlock + */ ret = AOP_TRUNCATED_PAGE; + unlock_page(page); + unlock = 0; + down_read(&oi->ip_alloc_sem); + up_read(&oi->ip_alloc_sem); goto out_inode_unlock; } @@ -551,9 +559,8 @@ bail: /* * ocfs2_dio_end_io is called by the dio core when a dio is finished. We're - * particularly interested in the aio/dio case. Like the core uses - * i_alloc_sem, we use the rw_lock DLM lock to protect io on one node from - * truncation on another. + * particularly interested in the aio/dio case. We use the rw_lock DLM lock + * to protect io on one node from truncation on another. */ static void ocfs2_dio_end_io(struct kiocb *iocb, loff_t offset, @@ -564,13 +571,21 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, { struct inode *inode = iocb->ki_filp->f_path.dentry->d_inode; int level; + wait_queue_head_t *wq = ocfs2_ioend_wq(inode); /* this io's submitter should not have unlocked this before we could */ BUG_ON(!ocfs2_iocb_is_rw_locked(iocb)); - if (ocfs2_iocb_is_sem_locked(iocb)) { - up_read(&inode->i_alloc_sem); + if (ocfs2_iocb_is_sem_locked(iocb)) ocfs2_iocb_clear_sem_locked(iocb); + + if (ocfs2_iocb_is_unaligned_aio(iocb)) { + ocfs2_iocb_clear_unaligned_aio(iocb); + + if (atomic_dec_and_test(&OCFS2_I(inode)->ip_unaligned_aio) && + waitqueue_active(wq)) { + wake_up_all(wq); + } } ocfs2_iocb_clear_rw_locked(iocb); @@ -578,6 +593,7 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, level = ocfs2_iocb_rw_locked_level(iocb); ocfs2_rw_unlock(inode, level); + inode_dio_done(inode); if (is_async) aio_complete(iocb, ret, 0); } @@ -865,6 +881,12 @@ struct ocfs2_write_ctxt { struct page *w_target_page; /* + * w_target_locked is used for page_mkwrite path indicating no unlocking + * against w_target_page in ocfs2_write_end_nolock. + */ + unsigned int w_target_locked:1; + + /* * ocfs2_write_end() uses this to know what the real range to * write in the target should be. */ @@ -895,10 +917,32 @@ void ocfs2_unlock_and_free_pages(struct page **pages, int num_pages) } } -static void ocfs2_free_write_ctxt(struct ocfs2_write_ctxt *wc) +static void ocfs2_unlock_pages(struct ocfs2_write_ctxt *wc) { + int i; + + /* + * w_target_locked is only set to true in the page_mkwrite() case. + * The intent is to allow us to lock the target page from write_begin() + * to write_end(). The caller must hold a ref on w_target_page. + */ + if (wc->w_target_locked) { + BUG_ON(!wc->w_target_page); + for (i = 0; i < wc->w_num_pages; i++) { + if (wc->w_target_page == wc->w_pages[i]) { + wc->w_pages[i] = NULL; + break; + } + } + mark_page_accessed(wc->w_target_page); + page_cache_release(wc->w_target_page); + } ocfs2_unlock_and_free_pages(wc->w_pages, wc->w_num_pages); +} +static void ocfs2_free_write_ctxt(struct ocfs2_write_ctxt *wc) +{ + ocfs2_unlock_pages(wc); brelse(wc->w_di_bh); kfree(wc); } @@ -1134,20 +1178,17 @@ static int ocfs2_grab_pages_for_write(struct address_space *mapping, */ lock_page(mmap_page); + /* Exit and let the caller retry */ if (mmap_page->mapping != mapping) { + WARN_ON(mmap_page->mapping); unlock_page(mmap_page); - /* - * Sanity check - the locking in - * ocfs2_pagemkwrite() should ensure - * that this code doesn't trigger. - */ - ret = -EINVAL; - mlog_errno(ret); + ret = -EAGAIN; goto out; } page_cache_get(mmap_page); wc->w_pages[i] = mmap_page; + wc->w_target_locked = true; } else { wc->w_pages[i] = find_or_create_page(mapping, index, GFP_NOFS); @@ -1162,6 +1203,8 @@ static int ocfs2_grab_pages_for_write(struct address_space *mapping, wc->w_target_page = wc->w_pages[i]; } out: + if (ret) + wc->w_target_locked = false; return ret; } @@ -1819,11 +1862,23 @@ try_again: */ ret = ocfs2_grab_pages_for_write(mapping, wc, wc->w_cpos, pos, len, cluster_of_pages, mmap_page); - if (ret) { + if (ret && ret != -EAGAIN) { mlog_errno(ret); goto out_quota; } + /* + * ocfs2_grab_pages_for_write() returns -EAGAIN if it could not lock + * the target page. In this case, we exit with no error and no target + * page. This will trigger the caller, page_mkwrite(), to re-try + * the operation. + */ + if (ret == -EAGAIN) { + BUG_ON(wc->w_target_page); + ret = 0; + goto out_quota; + } + ret = ocfs2_write_cluster_by_desc(mapping, data_ac, meta_ac, wc, pos, len); if (ret) { @@ -2008,11 +2063,19 @@ out_write_size: di->i_mtime_nsec = di->i_ctime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec); ocfs2_journal_dirty(handle, wc->w_di_bh); + /* unlock pages before dealloc since it needs acquiring j_trans_barrier + * lock, or it will cause a deadlock since journal commit threads holds + * this lock and will ask for the page lock when flushing the data. + * put it here to preserve the unlock order. + */ + ocfs2_unlock_pages(wc); + ocfs2_commit_trans(osb, handle); ocfs2_run_deallocs(osb, &wc->w_dealloc); - ocfs2_free_write_ctxt(wc); + brelse(wc->w_di_bh); + kfree(wc); return copied; } diff --git a/fs/ocfs2/aops.h b/fs/ocfs2/aops.h index 75cf3ad..ffb2da3 100644 --- a/fs/ocfs2/aops.h +++ b/fs/ocfs2/aops.h @@ -78,6 +78,7 @@ enum ocfs2_iocb_lock_bits { OCFS2_IOCB_RW_LOCK = 0, OCFS2_IOCB_RW_LOCK_LEVEL, OCFS2_IOCB_SEM, + OCFS2_IOCB_UNALIGNED_IO, OCFS2_IOCB_NUM_LOCKS }; @@ -91,4 +92,17 @@ enum ocfs2_iocb_lock_bits { clear_bit(OCFS2_IOCB_SEM, (unsigned long *)&iocb->private) #define ocfs2_iocb_is_sem_locked(iocb) \ test_bit(OCFS2_IOCB_SEM, (unsigned long *)&iocb->private) + +#define ocfs2_iocb_set_unaligned_aio(iocb) \ + set_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) +#define ocfs2_iocb_clear_unaligned_aio(iocb) \ + clear_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) +#define ocfs2_iocb_is_unaligned_aio(iocb) \ + test_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) + +#define OCFS2_IOEND_WQ_HASH_SZ 37 +#define ocfs2_ioend_wq(v) (&ocfs2__ioend_wq[((unsigned long)(v)) %\ + OCFS2_IOEND_WQ_HASH_SZ]) +extern wait_queue_head_t ocfs2__ioend_wq[OCFS2_IOEND_WQ_HASH_SZ]; + #endif /* OCFS2_FILE_H */ diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c index 5d18ad1..4f66e00 100644 --- a/fs/ocfs2/buffer_head_io.c +++ b/fs/ocfs2/buffer_head_io.c @@ -90,7 +90,6 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh, * information for this bh as it's not marked locally * uptodate. */ ret = -EIO; - put_bh(bh); mlog_errno(ret); } @@ -420,7 +419,6 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb, if (!buffer_uptodate(bh)) { ret = -EIO; - put_bh(bh); mlog_errno(ret); } diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 9a3e6bb..a4e855e 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -216,6 +216,7 @@ struct o2hb_region { struct list_head hr_all_item; unsigned hr_unclean_stop:1, + hr_aborted_start:1, hr_item_pinned:1, hr_item_dropped:1; @@ -254,6 +255,10 @@ struct o2hb_region { * a more complete api that doesn't lead to this sort of fragility. */ atomic_t hr_steady_iterations; + /* terminate o2hb thread if it does not reach steady state + * (hr_steady_iterations == 0) within hr_unsteady_iterations */ + atomic_t hr_unsteady_iterations; + char hr_dev_name[BDEVNAME_SIZE]; unsigned int hr_timeout_ms; @@ -324,6 +329,10 @@ static void o2hb_write_timeout(struct work_struct *work) static void o2hb_arm_write_timeout(struct o2hb_region *reg) { + /* Arm writeout only after thread reaches steady state */ + if (atomic_read(®->hr_steady_iterations) != 0) + return; + mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n", O2HB_MAX_WRITE_TIMEOUT_MS); @@ -537,9 +546,14 @@ static int o2hb_verify_crc(struct o2hb_region *reg, return read == computed; } -/* We want to make sure that nobody is heartbeating on top of us -- - * this will help detect an invalid configuration. */ -static void o2hb_check_last_timestamp(struct o2hb_region *reg) +/* + * Compare the slot data with what we wrote in the last iteration. + * If the match fails, print an appropriate error message. This is to + * detect errors like... another node hearting on the same slot, + * flaky device that is losing writes, etc. + * Returns 1 if check succeeds, 0 otherwise. + */ +static int o2hb_check_own_slot(struct o2hb_region *reg) { struct o2hb_disk_slot *slot; struct o2hb_disk_heartbeat_block *hb_block; @@ -548,13 +562,13 @@ static void o2hb_check_last_timestamp(struct o2hb_region *reg) slot = ®->hr_slots[o2nm_this_node()]; /* Don't check on our 1st timestamp */ if (!slot->ds_last_time) - return; + return 0; hb_block = slot->ds_raw_block; if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time && le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation && hb_block->hb_node == slot->ds_node_num) - return; + return 1; #define ERRSTR1 "Another node is heartbeating on device" #define ERRSTR2 "Heartbeat generation mismatch on device" @@ -574,6 +588,8 @@ static void o2hb_check_last_timestamp(struct o2hb_region *reg) (unsigned long long)slot->ds_last_time, hb_block->hb_node, (unsigned long long)le64_to_cpu(hb_block->hb_generation), (unsigned long long)le64_to_cpu(hb_block->hb_seq)); + + return 0; } static inline void o2hb_prepare_block(struct o2hb_region *reg, @@ -719,17 +735,24 @@ static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot) o2nm_node_put(node); } -static void o2hb_set_quorum_device(struct o2hb_region *reg, - struct o2hb_disk_slot *slot) +static void o2hb_set_quorum_device(struct o2hb_region *reg) { - assert_spin_locked(&o2hb_live_lock); - if (!o2hb_global_heartbeat_active()) return; - if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + /* Prevent race with o2hb_heartbeat_group_drop_item() */ + if (kthread_should_stop()) + return; + + /* Tag region as quorum only after thread reaches steady state */ + if (atomic_read(®->hr_steady_iterations) != 0) return; + spin_lock(&o2hb_live_lock); + + if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + goto unlock; + /* * A region can be added to the quorum only when it sees all * live nodes heartbeat on it. In other words, the region has been @@ -737,13 +760,10 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg, */ if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap, sizeof(o2hb_live_node_bitmap))) - return; - - if (slot->ds_changed_samples < O2HB_LIVE_THRESHOLD) - return; + goto unlock; - printk(KERN_NOTICE "o2hb: Region %s is now a quorum device\n", - config_item_name(®->hr_item)); + printk(KERN_NOTICE "o2hb: Region %s (%s) is now a quorum device\n", + config_item_name(®->hr_item), reg->hr_dev_name); set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); @@ -754,6 +774,8 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg, if (o2hb_pop_count(&o2hb_quorum_region_bitmap, O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF) o2hb_region_unpin(NULL); +unlock: + spin_unlock(&o2hb_live_lock); } static int o2hb_check_slot(struct o2hb_region *reg, @@ -925,8 +947,6 @@ fire_callbacks: slot->ds_equal_samples = 0; } out: - o2hb_set_quorum_device(reg, slot); - spin_unlock(&o2hb_live_lock); o2hb_run_event_list(&event); @@ -957,7 +977,8 @@ static int o2hb_highest_node(unsigned long *nodes, static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) { - int i, ret, highest_node, change = 0; + int i, ret, highest_node; + int membership_change = 0, own_slot_ok = 0; unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)]; unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; struct o2hb_bio_wait_ctxt write_wc; @@ -966,7 +987,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) sizeof(configured_nodes)); if (ret) { mlog_errno(ret); - return ret; + goto bail; } /* @@ -982,8 +1003,9 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES); if (highest_node >= O2NM_MAX_NODES) { - mlog(ML_NOTICE, "ocfs2_heartbeat: no configured nodes found!\n"); - return -EINVAL; + mlog(ML_NOTICE, "o2hb: No configured nodes found!\n"); + ret = -EINVAL; + goto bail; } /* No sense in reading the slots of nodes that don't exist @@ -993,29 +1015,27 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) ret = o2hb_read_slots(reg, highest_node + 1); if (ret < 0) { mlog_errno(ret); - return ret; + goto bail; } /* With an up to date view of the slots, we can check that no * other node has been improperly configured to heartbeat in * our slot. */ - o2hb_check_last_timestamp(reg); + own_slot_ok = o2hb_check_own_slot(reg); /* fill in the proper info for our next heartbeat */ o2hb_prepare_block(reg, reg->hr_generation); - /* And fire off the write. Note that we don't wait on this I/O - * until later. */ ret = o2hb_issue_node_write(reg, &write_wc); if (ret < 0) { mlog_errno(ret); - return ret; + goto bail; } i = -1; while((i = find_next_bit(configured_nodes, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) { - change |= o2hb_check_slot(reg, ®->hr_slots[i]); + membership_change |= o2hb_check_slot(reg, ®->hr_slots[i]); } /* @@ -1030,18 +1050,39 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) * disk */ mlog(ML_ERROR, "Write error %d on device \"%s\"\n", write_wc.wc_error, reg->hr_dev_name); - return write_wc.wc_error; + ret = write_wc.wc_error; + goto bail; } - o2hb_arm_write_timeout(reg); + /* Skip disarming the timeout if own slot has stale/bad data */ + if (own_slot_ok) { + o2hb_set_quorum_device(reg); + o2hb_arm_write_timeout(reg); + } +bail: /* let the person who launched us know when things are steady */ - if (!change && (atomic_read(®->hr_steady_iterations) != 0)) { - if (atomic_dec_and_test(®->hr_steady_iterations)) + if (atomic_read(®->hr_steady_iterations) != 0) { + if (!ret && own_slot_ok && !membership_change) { + if (atomic_dec_and_test(®->hr_steady_iterations)) + wake_up(&o2hb_steady_queue); + } + } + + if (atomic_read(®->hr_steady_iterations) != 0) { + if (atomic_dec_and_test(®->hr_unsteady_iterations)) { + printk(KERN_NOTICE "o2hb: Unable to stabilize " + "heartbeart on region %s (%s)\n", + config_item_name(®->hr_item), + reg->hr_dev_name); + atomic_set(®->hr_steady_iterations, 0); + reg->hr_aborted_start = 1; wake_up(&o2hb_steady_queue); + ret = -EIO; + } } - return 0; + return ret; } /* Subtract b from a, storing the result in a. a *must* have a larger @@ -1095,7 +1136,8 @@ static int o2hb_thread(void *data) /* Pin node */ o2nm_depend_this_node(); - while (!kthread_should_stop() && !reg->hr_unclean_stop) { + while (!kthread_should_stop() && + !reg->hr_unclean_stop && !reg->hr_aborted_start) { /* We track the time spent inside * o2hb_do_disk_heartbeat so that we avoid more than * hr_timeout_ms between disk writes. On busy systems @@ -1103,10 +1145,7 @@ static int o2hb_thread(void *data) * likely to time itself out. */ do_gettimeofday(&before_hb); - i = 0; - do { - ret = o2hb_do_disk_heartbeat(reg); - } while (ret && ++i < 2); + ret = o2hb_do_disk_heartbeat(reg); do_gettimeofday(&after_hb); elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb); @@ -1117,7 +1156,8 @@ static int o2hb_thread(void *data) after_hb.tv_sec, (unsigned long) after_hb.tv_usec, elapsed_msec); - if (elapsed_msec < reg->hr_timeout_ms) { + if (!kthread_should_stop() && + elapsed_msec < reg->hr_timeout_ms) { /* the kthread api has blocked signals for us so no * need to record the return value. */ msleep_interruptible(reg->hr_timeout_ms - elapsed_msec); @@ -1134,20 +1174,20 @@ static int o2hb_thread(void *data) * to timeout on this region when we could just as easily * write a clear generation - thus indicating to them that * this node has left this region. - * - * XXX: Should we skip this on unclean_stop? */ - o2hb_prepare_block(reg, 0); - ret = o2hb_issue_node_write(reg, &write_wc); - if (ret == 0) { - o2hb_wait_on_io(reg, &write_wc); - } else { - mlog_errno(ret); + */ + if (!reg->hr_unclean_stop && !reg->hr_aborted_start) { + o2hb_prepare_block(reg, 0); + ret = o2hb_issue_node_write(reg, &write_wc); + if (ret == 0) + o2hb_wait_on_io(reg, &write_wc); + else + mlog_errno(ret); } /* Unpin node */ o2nm_undepend_this_node(); - mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread exiting\n"); + mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n"); return 0; } @@ -1158,6 +1198,7 @@ static int o2hb_debug_open(struct inode *inode, struct file *file) struct o2hb_debug_buf *db = inode->i_private; struct o2hb_region *reg; unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long lts; char *buf = NULL; int i = -1; int out = 0; @@ -1194,9 +1235,11 @@ static int o2hb_debug_open(struct inode *inode, struct file *file) case O2HB_DB_TYPE_REGION_ELAPSED_TIME: reg = (struct o2hb_region *)db->db_data; - out += snprintf(buf + out, PAGE_SIZE - out, "%u\n", - jiffies_to_msecs(jiffies - - reg->hr_last_timeout_start)); + lts = reg->hr_last_timeout_start; + /* If 0, it has never been set before */ + if (lts) + lts = jiffies_to_msecs(jiffies - lts); + out += snprintf(buf + out, PAGE_SIZE - out, "%lu\n", lts); goto done; case O2HB_DB_TYPE_REGION_PINNED: @@ -1426,6 +1469,8 @@ static void o2hb_region_release(struct config_item *item) struct page *page; struct o2hb_region *reg = to_o2hb_region(item); + mlog(ML_HEARTBEAT, "hb region release (%s)\n", reg->hr_dev_name); + if (reg->hr_tmp_block) kfree(reg->hr_tmp_block); @@ -1792,7 +1837,10 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, live_threshold <<= 1; spin_unlock(&o2hb_live_lock); } - atomic_set(®->hr_steady_iterations, live_threshold + 1); + ++live_threshold; + atomic_set(®->hr_steady_iterations, live_threshold); + /* unsteady_iterations is double the steady_iterations */ + atomic_set(®->hr_unsteady_iterations, (live_threshold << 1)); hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s", reg->hr_item.ci_name); @@ -1809,14 +1857,12 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, ret = wait_event_interruptible(o2hb_steady_queue, atomic_read(®->hr_steady_iterations) == 0); if (ret) { - /* We got interrupted (hello ptrace!). Clean up */ - spin_lock(&o2hb_live_lock); - hb_task = reg->hr_task; - reg->hr_task = NULL; - spin_unlock(&o2hb_live_lock); + atomic_set(®->hr_steady_iterations, 0); + reg->hr_aborted_start = 1; + } - if (hb_task) - kthread_stop(hb_task); + if (reg->hr_aborted_start) { + ret = -EIO; goto out; } @@ -1833,8 +1879,8 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, ret = -EIO; if (hb_task && o2hb_global_heartbeat_active()) - printk(KERN_NOTICE "o2hb: Heartbeat started on region %s\n", - config_item_name(®->hr_item)); + printk(KERN_NOTICE "o2hb: Heartbeat started on region %s (%s)\n", + config_item_name(®->hr_item), reg->hr_dev_name); out: if (filp) @@ -2092,13 +2138,6 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, /* stop the thread when the user removes the region dir */ spin_lock(&o2hb_live_lock); - if (o2hb_global_heartbeat_active()) { - clear_bit(reg->hr_region_num, o2hb_region_bitmap); - clear_bit(reg->hr_region_num, o2hb_live_region_bitmap); - if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) - quorum_region = 1; - clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); - } hb_task = reg->hr_task; reg->hr_task = NULL; reg->hr_item_dropped = 1; @@ -2107,19 +2146,30 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, if (hb_task) kthread_stop(hb_task); + if (o2hb_global_heartbeat_active()) { + spin_lock(&o2hb_live_lock); + clear_bit(reg->hr_region_num, o2hb_region_bitmap); + clear_bit(reg->hr_region_num, o2hb_live_region_bitmap); + if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + quorum_region = 1; + clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); + spin_unlock(&o2hb_live_lock); + printk(KERN_NOTICE "o2hb: Heartbeat %s on region %s (%s)\n", + ((atomic_read(®->hr_steady_iterations) == 0) ? + "stopped" : "start aborted"), config_item_name(item), + reg->hr_dev_name); + } + /* * If we're racing a dev_write(), we need to wake them. They will * check reg->hr_task */ if (atomic_read(®->hr_steady_iterations) != 0) { + reg->hr_aborted_start = 1; atomic_set(®->hr_steady_iterations, 0); wake_up(&o2hb_steady_queue); } - if (o2hb_global_heartbeat_active()) - printk(KERN_NOTICE "o2hb: Heartbeat stopped on region %s\n", - config_item_name(®->hr_item)); - config_item_put(item); if (!o2hb_global_heartbeat_active() || !quorum_region) diff --git a/fs/ocfs2/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c index 3a58359..dc45deb 100644 --- a/fs/ocfs2/cluster/netdebug.c +++ b/fs/ocfs2/cluster/netdebug.c @@ -47,6 +47,7 @@ #define SC_DEBUG_NAME "sock_containers" #define NST_DEBUG_NAME "send_tracking" #define STATS_DEBUG_NAME "stats" +#define NODES_DEBUG_NAME "connected_nodes" #define SHOW_SOCK_CONTAINERS 0 #define SHOW_SOCK_STATS 1 @@ -55,6 +56,7 @@ static struct dentry *o2net_dentry; static struct dentry *sc_dentry; static struct dentry *nst_dentry; static struct dentry *stats_dentry; +static struct dentry *nodes_dentry; static DEFINE_SPINLOCK(o2net_debug_lock); @@ -491,53 +493,87 @@ static const struct file_operations sc_seq_fops = { .release = sc_fop_release, }; -int o2net_debugfs_init(void) +static int o2net_fill_bitmap(char *buf, int len) { - o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL); - if (!o2net_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + int i = -1, out = 0; - nst_dentry = debugfs_create_file(NST_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &nst_seq_fops); - if (!nst_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + o2net_fill_node_map(map, sizeof(map)); - sc_dentry = debugfs_create_file(SC_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &sc_seq_fops); - if (!sc_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + while ((i = find_next_bit(map, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) + out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i); + out += snprintf(buf + out, PAGE_SIZE - out, "\n"); - stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &stats_seq_fops); - if (!stats_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + return out; +} + +static int nodes_fop_open(struct inode *inode, struct file *file) +{ + char *buf; + + buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + i_size_write(inode, o2net_fill_bitmap(buf, PAGE_SIZE)); + + file->private_data = buf; return 0; -bail: - debugfs_remove(stats_dentry); - debugfs_remove(sc_dentry); - debugfs_remove(nst_dentry); - debugfs_remove(o2net_dentry); - return -ENOMEM; } +static int o2net_debug_release(struct inode *inode, struct file *file) +{ + kfree(file->private_data); + return 0; +} + +static ssize_t o2net_debug_read(struct file *file, char __user *buf, + size_t nbytes, loff_t *ppos) +{ + return simple_read_from_buffer(buf, nbytes, ppos, file->private_data, + i_size_read(file->f_mapping->host)); +} + +static const struct file_operations nodes_fops = { + .open = nodes_fop_open, + .release = o2net_debug_release, + .read = o2net_debug_read, + .llseek = generic_file_llseek, +}; + void o2net_debugfs_exit(void) { + debugfs_remove(nodes_dentry); debugfs_remove(stats_dentry); debugfs_remove(sc_dentry); debugfs_remove(nst_dentry); debugfs_remove(o2net_dentry); } +int o2net_debugfs_init(void) +{ + mode_t mode = S_IFREG|S_IRUSR; + + o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL); + if (o2net_dentry) + nst_dentry = debugfs_create_file(NST_DEBUG_NAME, mode, + o2net_dentry, NULL, &nst_seq_fops); + if (nst_dentry) + sc_dentry = debugfs_create_file(SC_DEBUG_NAME, mode, + o2net_dentry, NULL, &sc_seq_fops); + if (sc_dentry) + stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, mode, + o2net_dentry, NULL, &stats_seq_fops); + if (stats_dentry) + nodes_dentry = debugfs_create_file(NODES_DEBUG_NAME, mode, + o2net_dentry, NULL, &nodes_fops); + if (nodes_dentry) + return 0; + + o2net_debugfs_exit(); + mlog_errno(-ENOMEM); + return -ENOMEM; +} + #endif /* CONFIG_DEBUG_FS */ diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index db5ee4b..044e7b5 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -59,6 +59,7 @@ #include <linux/idr.h> #include <linux/kref.h> #include <linux/net.h> +#include <linux/export.h> #include <net/tcp.h> #include <asm/uaccess.h> @@ -545,7 +546,7 @@ static void o2net_set_nn_state(struct o2net_node *nn, } if (was_valid && !valid) { - printk(KERN_NOTICE "o2net: no longer connected to " + printk(KERN_NOTICE "o2net: No longer connected to " SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc)); o2net_complete_nodes_nsw(nn); } @@ -555,7 +556,7 @@ static void o2net_set_nn_state(struct o2net_node *nn, cancel_delayed_work(&nn->nn_connect_expired); printk(KERN_NOTICE "o2net: %s " SC_NODEF_FMT "\n", o2nm_this_node() > sc->sc_node->nd_num ? - "connected to" : "accepted connection from", + "Connected to" : "Accepted connection from", SC_NODEF_ARGS(sc)); } @@ -643,7 +644,7 @@ static void o2net_state_change(struct sock *sk) o2net_sc_queue_work(sc, &sc->sc_connect_work); break; default: - printk(KERN_INFO "o2net: connection to " SC_NODEF_FMT + printk(KERN_INFO "o2net: Connection to " SC_NODEF_FMT " shutdown, state %d\n", SC_NODEF_ARGS(sc), sk->sk_state); o2net_sc_queue_work(sc, &sc->sc_shutdown_work); @@ -1034,6 +1035,25 @@ static int o2net_tx_can_proceed(struct o2net_node *nn, return ret; } +/* Get a map of all nodes to which this node is currently connected to */ +void o2net_fill_node_map(unsigned long *map, unsigned bytes) +{ + struct o2net_sock_container *sc; + int node, ret; + + BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long))); + + memset(map, 0, bytes); + for (node = 0; node < O2NM_MAX_NODES; ++node) { + o2net_tx_can_proceed(o2net_nn_from_num(node), &sc, &ret); + if (!ret) { + set_bit(node, map); + sc_put(sc); + } + } +} +EXPORT_SYMBOL_GPL(o2net_fill_node_map); + int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, size_t caller_veclen, u8 target_node, int *status) { @@ -1284,11 +1304,11 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num); if (hand->protocol_version != cpu_to_be64(O2NET_PROTOCOL_VERSION)) { - mlog(ML_NOTICE, SC_NODEF_FMT " advertised net protocol " - "version %llu but %llu is required, disconnecting\n", - SC_NODEF_ARGS(sc), - (unsigned long long)be64_to_cpu(hand->protocol_version), - O2NET_PROTOCOL_VERSION); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " Advertised net " + "protocol version %llu but %llu is required. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + (unsigned long long)be64_to_cpu(hand->protocol_version), + O2NET_PROTOCOL_VERSION); /* don't bother reconnecting if its the wrong version. */ o2net_ensure_shutdown(nn, sc, -ENOTCONN); @@ -1302,33 +1322,33 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) */ if (be32_to_cpu(hand->o2net_idle_timeout_ms) != o2net_idle_timeout()) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a network idle timeout of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2net_idle_timeout_ms), - o2net_idle_timeout()); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a network " + "idle timeout of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2net_idle_timeout_ms), + o2net_idle_timeout()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } if (be32_to_cpu(hand->o2net_keepalive_delay_ms) != o2net_keepalive_delay()) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a keepalive delay of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2net_keepalive_delay_ms), - o2net_keepalive_delay()); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a keepalive " + "delay of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2net_keepalive_delay_ms), + o2net_keepalive_delay()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } if (be32_to_cpu(hand->o2hb_heartbeat_timeout_ms) != O2HB_MAX_WRITE_TIMEOUT_MS) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a heartbeat timeout of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2hb_heartbeat_timeout_ms), - O2HB_MAX_WRITE_TIMEOUT_MS); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a heartbeat " + "timeout of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2hb_heartbeat_timeout_ms), + O2HB_MAX_WRITE_TIMEOUT_MS); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } @@ -1539,28 +1559,16 @@ static void o2net_idle_timer(unsigned long data) { struct o2net_sock_container *sc = (struct o2net_sock_container *)data; struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num); - #ifdef CONFIG_DEBUG_FS - ktime_t now = ktime_get(); + unsigned long msecs = ktime_to_ms(ktime_get()) - + ktime_to_ms(sc->sc_tv_timer); +#else + unsigned long msecs = o2net_idle_timeout(); #endif - printk(KERN_NOTICE "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u " - "seconds, shutting it down.\n", SC_NODEF_ARGS(sc), - o2net_idle_timeout() / 1000, - o2net_idle_timeout() % 1000); - -#ifdef CONFIG_DEBUG_FS - mlog(ML_NOTICE, "Here are some times that might help debug the " - "situation: (Timer: %lld, Now %lld, DataReady %lld, Advance %lld-%lld, " - "Key 0x%08x, Func %u, FuncTime %lld-%lld)\n", - (long long)ktime_to_us(sc->sc_tv_timer), (long long)ktime_to_us(now), - (long long)ktime_to_us(sc->sc_tv_data_ready), - (long long)ktime_to_us(sc->sc_tv_advance_start), - (long long)ktime_to_us(sc->sc_tv_advance_stop), - sc->sc_msg_key, sc->sc_msg_type, - (long long)ktime_to_us(sc->sc_tv_func_start), - (long long)ktime_to_us(sc->sc_tv_func_stop)); -#endif + printk(KERN_NOTICE "o2net: Connection to " SC_NODEF_FMT " has been " + "idle for %lu.%lu secs, shutting it down.\n", SC_NODEF_ARGS(sc), + msecs / 1000, msecs % 1000); /* * Initialize the nn_timeout so that the next connection attempt @@ -1693,8 +1701,8 @@ static void o2net_start_connect(struct work_struct *work) out: if (ret) { - mlog(ML_NOTICE, "connect attempt to " SC_NODEF_FMT " failed " - "with errno %d\n", SC_NODEF_ARGS(sc), ret); + printk(KERN_NOTICE "o2net: Connect attempt to " SC_NODEF_FMT + " failed with errno %d\n", SC_NODEF_ARGS(sc), ret); /* 0 err so that another will be queued and attempted * from set_nn_state */ if (sc) @@ -1717,8 +1725,8 @@ static void o2net_connect_expired(struct work_struct *work) spin_lock(&nn->nn_lock); if (!nn->nn_sc_valid) { - mlog(ML_ERROR, "no connection established with node %u after " - "%u.%u seconds, giving up and returning errors.\n", + printk(KERN_NOTICE "o2net: No connection established with " + "node %u after %u.%u seconds, giving up.\n", o2net_num_from_nn(nn), o2net_idle_timeout() / 1000, o2net_idle_timeout() % 1000); @@ -1861,21 +1869,21 @@ static int o2net_accept_one(struct socket *sock) node = o2nm_get_node_by_ip(sin.sin_addr.s_addr); if (node == NULL) { - mlog(ML_NOTICE, "attempt to connect from unknown node at %pI4:%d\n", - &sin.sin_addr.s_addr, ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Attempt to connect from unknown " + "node at %pI4:%d\n", &sin.sin_addr.s_addr, + ntohs(sin.sin_port)); ret = -EINVAL; goto out; } if (o2nm_this_node() >= node->nd_num) { local_node = o2nm_get_node_by_num(o2nm_this_node()); - mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' (" - "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n", - local_node->nd_name, local_node->nd_num, - &(local_node->nd_ipv4_address), - ntohs(local_node->nd_ipv4_port), - node->nd_name, node->nd_num, &sin.sin_addr.s_addr, - ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Unexpected connect attempt seen " + "at node '%s' (%u, %pI4:%d) from node '%s' (%u, " + "%pI4:%d)\n", local_node->nd_name, local_node->nd_num, + &(local_node->nd_ipv4_address), + ntohs(local_node->nd_ipv4_port), node->nd_name, + node->nd_num, &sin.sin_addr.s_addr, ntohs(sin.sin_port)); ret = -EINVAL; goto out; } @@ -1900,10 +1908,10 @@ static int o2net_accept_one(struct socket *sock) ret = 0; spin_unlock(&nn->nn_lock); if (ret) { - mlog(ML_NOTICE, "attempt to connect from node '%s' at " - "%pI4:%d but it already has an open connection\n", - node->nd_name, &sin.sin_addr.s_addr, - ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Attempt to connect from node '%s' " + "at %pI4:%d but it already has an open connection\n", + node->nd_name, &sin.sin_addr.s_addr, + ntohs(sin.sin_port)); goto out; } @@ -1983,7 +1991,7 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port) ret = sock_create(PF_INET, SOCK_STREAM, IPPROTO_TCP, &sock); if (ret < 0) { - mlog(ML_ERROR, "unable to create socket, ret=%d\n", ret); + printk(KERN_ERR "o2net: Error %d while creating socket\n", ret); goto out; } @@ -2000,16 +2008,15 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port) sock->sk->sk_reuse = 1; ret = sock->ops->bind(sock, (struct sockaddr *)&sin, sizeof(sin)); if (ret < 0) { - mlog(ML_ERROR, "unable to bind socket at %pI4:%u, " - "ret=%d\n", &addr, ntohs(port), ret); + printk(KERN_ERR "o2net: Error %d while binding socket at " + "%pI4:%u\n", ret, &addr, ntohs(port)); goto out; } ret = sock->ops->listen(sock, 64); - if (ret < 0) { - mlog(ML_ERROR, "unable to listen on %pI4:%u, ret=%d\n", - &addr, ntohs(port), ret); - } + if (ret < 0) + printk(KERN_ERR "o2net: Error %d while listening on %pI4:%u\n", + ret, &addr, ntohs(port)); out: if (ret) { diff --git a/fs/ocfs2/cluster/tcp.h b/fs/ocfs2/cluster/tcp.h index fd6179e..5bada2a 100644 --- a/fs/ocfs2/cluster/tcp.h +++ b/fs/ocfs2/cluster/tcp.h @@ -106,6 +106,8 @@ int o2net_register_handler(u32 msg_type, u32 key, u32 max_len, struct list_head *unreg_list); void o2net_unregister_handler_list(struct list_head *list); +void o2net_fill_node_map(unsigned long *map, unsigned bytes); + struct o2nm_node; int o2net_register_hb_callbacks(void); void o2net_unregister_hb_callbacks(void); diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c index e5ba348..26977cc 100644 --- a/fs/ocfs2/dcache.c +++ b/fs/ocfs2/dcache.c @@ -175,7 +175,7 @@ struct dentry *ocfs2_find_local_alias(struct inode *inode, spin_lock(&inode->i_lock); list_for_each(p, &inode->i_dentry) { - dentry = list_entry(p, struct dentry, d_alias); + dentry = list_entry(p, struct dentry, d_u.d_alias); spin_lock(&dentry->d_lock); if (ocfs2_match_dentry(dentry, parent_blkno, skip_unhashed)) { diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c index 8582e3f..8fe4e28 100644 --- a/fs/ocfs2/dir.c +++ b/fs/ocfs2/dir.c @@ -1184,8 +1184,7 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir, if (pde) le16_add_cpu(&pde->rec_len, le16_to_cpu(de->rec_len)); - else - de->inode = 0; + de->inode = 0; dir->i_version++; ocfs2_journal_dirty(handle, bh); goto bail; @@ -2292,7 +2291,7 @@ static int ocfs2_fill_new_dir_id(struct ocfs2_super *osb, ocfs2_journal_dirty(handle, di_bh); i_size_write(inode, size); - inode->i_nlink = 2; + set_nlink(inode, 2); inode->i_blocks = ocfs2_inode_sector_count(inode); ret = ocfs2_mark_inode_dirty(handle, inode, di_bh); @@ -2354,7 +2353,7 @@ static int ocfs2_fill_new_dir_el(struct ocfs2_super *osb, ocfs2_journal_dirty(handle, new_bh); i_size_write(inode, inode->i_sb->s_blocksize); - inode->i_nlink = 2; + set_nlink(inode, 2); inode->i_blocks = ocfs2_inode_sector_count(inode); status = ocfs2_mark_inode_dirty(handle, inode, fe_bh); if (status < 0) { diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index d602abb..a5952ce 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -859,8 +859,8 @@ void dlm_complete_recovery_thread(struct dlm_ctxt *dlm); void dlm_wait_for_recovery(struct dlm_ctxt *dlm); void dlm_kick_recovery_thread(struct dlm_ctxt *dlm); int dlm_is_node_dead(struct dlm_ctxt *dlm, u8 node); -int dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout); -int dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout); +void dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout); +void dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout); void dlm_put(struct dlm_ctxt *dlm); struct dlm_ctxt *dlm_grab(struct dlm_ctxt *dlm); @@ -877,9 +877,8 @@ static inline void dlm_lockres_get(struct dlm_lock_resource *res) kref_get(&res->refs); } void dlm_lockres_put(struct dlm_lock_resource *res); -void __dlm_unhash_lockres(struct dlm_lock_resource *res); -void __dlm_insert_lockres(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res); +void __dlm_unhash_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); +void __dlm_insert_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, const char *name, unsigned int len, @@ -902,46 +901,15 @@ struct dlm_lock_resource *dlm_new_lockres(struct dlm_ctxt *dlm, const char *name, unsigned int namelen); -#define dlm_lockres_set_refmap_bit(bit,res) \ - __dlm_lockres_set_refmap_bit(bit,res,__FILE__,__LINE__) -#define dlm_lockres_clear_refmap_bit(bit,res) \ - __dlm_lockres_clear_refmap_bit(bit,res,__FILE__,__LINE__) +void dlm_lockres_set_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit); +void dlm_lockres_clear_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit); -static inline void __dlm_lockres_set_refmap_bit(int bit, - struct dlm_lock_resource *res, - const char *file, - int line) -{ - //printk("%s:%d:%.*s: setting bit %d\n", file, line, - // res->lockname.len, res->lockname.name, bit); - set_bit(bit, res->refmap); -} - -static inline void __dlm_lockres_clear_refmap_bit(int bit, - struct dlm_lock_resource *res, - const char *file, - int line) -{ - //printk("%s:%d:%.*s: clearing bit %d\n", file, line, - // res->lockname.len, res->lockname.name, bit); - clear_bit(bit, res->refmap); -} - -void __dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - const char *file, - int line); -void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - int new_lockres, - const char *file, - int line); -#define dlm_lockres_drop_inflight_ref(d,r) \ - __dlm_lockres_drop_inflight_ref(d,r,__FILE__,__LINE__) -#define dlm_lockres_grab_inflight_ref(d,r) \ - __dlm_lockres_grab_inflight_ref(d,r,0,__FILE__,__LINE__) -#define dlm_lockres_grab_inflight_ref_new(d,r) \ - __dlm_lockres_grab_inflight_ref(d,r,1,__FILE__,__LINE__) +void dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); +void dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); void dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock); void dlm_queue_bast(struct dlm_ctxt *dlm, struct dlm_lock *lock); diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 56f82cb..0e28e24 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -30,6 +30,7 @@ #include <linux/sysctl.h> #include <linux/spinlock.h> #include <linux/debugfs.h> +#include <linux/export.h> #include "cluster/heartbeat.h" #include "cluster/nodemanager.h" diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 6ed6b95..92f2ead 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -157,16 +157,18 @@ static int dlm_protocol_compare(struct dlm_protocol_version *existing, static void dlm_unregister_domain_handlers(struct dlm_ctxt *dlm); -void __dlm_unhash_lockres(struct dlm_lock_resource *lockres) +void __dlm_unhash_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { - if (!hlist_unhashed(&lockres->hash_node)) { - hlist_del_init(&lockres->hash_node); - dlm_lockres_put(lockres); - } + if (hlist_unhashed(&res->hash_node)) + return; + + mlog(0, "%s: Unhash res %.*s\n", dlm->name, res->lockname.len, + res->lockname.name); + hlist_del_init(&res->hash_node); + dlm_lockres_put(res); } -void __dlm_insert_lockres(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res) +void __dlm_insert_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { struct hlist_head *bucket; struct qstr *q; @@ -180,6 +182,9 @@ void __dlm_insert_lockres(struct dlm_ctxt *dlm, dlm_lockres_get(res); hlist_add_head(&res->hash_node, bucket); + + mlog(0, "%s: Hash res %.*s\n", dlm->name, res->lockname.len, + res->lockname.name); } struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, @@ -539,17 +544,17 @@ again: static void __dlm_print_nodes(struct dlm_ctxt *dlm) { - int node = -1; + int node = -1, num = 0; assert_spin_locked(&dlm->spinlock); - printk(KERN_NOTICE "o2dlm: Nodes in domain %s: ", dlm->name); - + printk("( "); while ((node = find_next_bit(dlm->domain_map, O2NM_MAX_NODES, node + 1)) < O2NM_MAX_NODES) { printk("%d ", node); + ++num; } - printk("\n"); + printk(") %u nodes\n", num); } static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, @@ -566,11 +571,10 @@ static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, node = exit_msg->node_idx; - printk(KERN_NOTICE "o2dlm: Node %u leaves domain %s\n", node, dlm->name); - spin_lock(&dlm->spinlock); clear_bit(node, dlm->domain_map); clear_bit(node, dlm->exit_domain_map); + printk(KERN_NOTICE "o2dlm: Node %u leaves domain %s ", node, dlm->name); __dlm_print_nodes(dlm); /* notify anything attached to the heartbeat events */ @@ -755,6 +759,7 @@ void dlm_unregister_domain(struct dlm_ctxt *dlm) dlm_mark_domain_leaving(dlm); dlm_leave_domain(dlm); + printk(KERN_NOTICE "o2dlm: Leaving domain %s\n", dlm->name); dlm_force_free_mles(dlm); dlm_complete_dlm_shutdown(dlm); } @@ -970,7 +975,7 @@ static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data, clear_bit(assert->node_idx, dlm->exit_domain_map); __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); - printk(KERN_NOTICE "o2dlm: Node %u joins domain %s\n", + printk(KERN_NOTICE "o2dlm: Node %u joins domain %s ", assert->node_idx, dlm->name); __dlm_print_nodes(dlm); @@ -1701,8 +1706,10 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm) bail: spin_lock(&dlm->spinlock); __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); - if (!status) + if (!status) { + printk(KERN_NOTICE "o2dlm: Joining domain %s ", dlm->name); __dlm_print_nodes(dlm); + } spin_unlock(&dlm->spinlock); if (ctxt) { @@ -2131,13 +2138,6 @@ struct dlm_ctxt * dlm_register_domain(const char *domain, goto leave; } - if (!o2hb_check_local_node_heartbeating()) { - mlog(ML_ERROR, "the local node has not been configured, or is " - "not heartbeating\n"); - ret = -EPROTO; - goto leave; - } - mlog(0, "register called for domain \"%s\"\n", domain); retry: diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c index 8d39e0f..975810b 100644 --- a/fs/ocfs2/dlm/dlmlock.c +++ b/fs/ocfs2/dlm/dlmlock.c @@ -183,10 +183,6 @@ static enum dlm_status dlmlock_master(struct dlm_ctxt *dlm, kick_thread = 1; } } - /* reduce the inflight count, this may result in the lockres - * being purged below during calc_usage */ - if (lock->ml.node == dlm->node_num) - dlm_lockres_drop_inflight_ref(dlm, res); spin_unlock(&res->spinlock); wake_up(&res->wq); @@ -231,10 +227,16 @@ static enum dlm_status dlmlock_remote(struct dlm_ctxt *dlm, lock->ml.type, res->lockname.len, res->lockname.name, flags); + /* + * Wait if resource is getting recovered, remastered, etc. + * If the resource was remastered and new owner is self, then exit. + */ spin_lock(&res->spinlock); - - /* will exit this call with spinlock held */ __dlm_wait_on_lockres(res); + if (res->owner == dlm->node_num) { + spin_unlock(&res->spinlock); + return DLM_RECOVERING; + } res->state |= DLM_LOCK_RES_IN_PROGRESS; /* add lock to local (secondary) queue */ @@ -319,27 +321,23 @@ static enum dlm_status dlm_send_remote_lock_request(struct dlm_ctxt *dlm, tmpret = o2net_send_message(DLM_CREATE_LOCK_MSG, dlm->key, &create, sizeof(create), res->owner, &status); if (tmpret >= 0) { - // successfully sent and received - ret = status; // this is already a dlm_status + ret = status; if (ret == DLM_REJECTED) { - mlog(ML_ERROR, "%s:%.*s: BUG. this is a stale lockres " - "no longer owned by %u. that node is coming back " - "up currently.\n", dlm->name, create.namelen, + mlog(ML_ERROR, "%s: res %.*s, Stale lockres no longer " + "owned by node %u. That node is coming back up " + "currently.\n", dlm->name, create.namelen, create.name, res->owner); dlm_print_one_lock_resource(res); BUG(); } } else { - mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " - "node %u\n", tmpret, DLM_CREATE_LOCK_MSG, dlm->key, - res->owner); - if (dlm_is_host_down(tmpret)) { + mlog(ML_ERROR, "%s: res %.*s, Error %d send CREATE LOCK to " + "node %u\n", dlm->name, create.namelen, create.name, + tmpret, res->owner); + if (dlm_is_host_down(tmpret)) ret = DLM_RECOVERING; - mlog(0, "node %u died so returning DLM_RECOVERING " - "from lock message!\n", res->owner); - } else { + else ret = dlm_err_to_dlm_status(tmpret); - } } return ret; @@ -440,7 +438,7 @@ struct dlm_lock * dlm_new_lock(int type, u8 node, u64 cookie, /* zero memory only if kernel-allocated */ lksb = kzalloc(sizeof(*lksb), GFP_NOFS); if (!lksb) { - kfree(lock); + kmem_cache_free(dlm_lock_cache, lock); return NULL; } kernel_allocated = 1; @@ -718,18 +716,10 @@ retry_lock: if (status == DLM_RECOVERING || status == DLM_MIGRATING || status == DLM_FORWARD) { - mlog(0, "retrying lock with migration/" - "recovery/in progress\n"); msleep(100); - /* no waiting for dlm_reco_thread */ if (recovery) { if (status != DLM_RECOVERING) goto retry_lock; - - mlog(0, "%s: got RECOVERING " - "for $RECOVERY lock, master " - "was %u\n", dlm->name, - res->owner); /* wait to see the node go down, then * drop down and allow the lockres to * get cleaned up. need to remaster. */ @@ -741,6 +731,14 @@ retry_lock: } } + /* Inflight taken in dlm_get_lock_resource() is dropped here */ + spin_lock(&res->spinlock); + dlm_lockres_drop_inflight_ref(dlm, res); + spin_unlock(&res->spinlock); + + dlm_lockres_calc_usage(dlm, res); + dlm_kick_thread(dlm, res); + if (status != DLM_NORMAL) { lock->lksb->flags &= ~DLM_LKSB_GET_LVB; if (status != DLM_NOTQUEUED) diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 11eefb8..8e48ba5 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -631,39 +631,58 @@ error: return NULL; } -void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - int new_lockres, - const char *file, - int line) +void dlm_lockres_set_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit) { - if (!new_lockres) - assert_spin_locked(&res->spinlock); + assert_spin_locked(&res->spinlock); - if (!test_bit(dlm->node_num, res->refmap)) { - BUG_ON(res->inflight_locks != 0); - dlm_lockres_set_refmap_bit(dlm->node_num, res); - } + mlog(0, "res %.*s, set node %u, %ps()\n", res->lockname.len, + res->lockname.name, bit, __builtin_return_address(0)); + + set_bit(bit, res->refmap); +} + +void dlm_lockres_clear_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit) +{ + assert_spin_locked(&res->spinlock); + + mlog(0, "res %.*s, clr node %u, %ps()\n", res->lockname.len, + res->lockname.name, bit, __builtin_return_address(0)); + + clear_bit(bit, res->refmap); +} + +static void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ res->inflight_locks++; - mlog(0, "%s:%.*s: inflight++: now %u\n", - dlm->name, res->lockname.len, res->lockname.name, - res->inflight_locks); + + mlog(0, "%s: res %.*s, inflight++: now %u, %ps()\n", dlm->name, + res->lockname.len, res->lockname.name, res->inflight_locks, + __builtin_return_address(0)); +} + +void dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + assert_spin_locked(&res->spinlock); + __dlm_lockres_grab_inflight_ref(dlm, res); } -void __dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - const char *file, - int line) +void dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) { assert_spin_locked(&res->spinlock); BUG_ON(res->inflight_locks == 0); + res->inflight_locks--; - mlog(0, "%s:%.*s: inflight--: now %u\n", - dlm->name, res->lockname.len, res->lockname.name, - res->inflight_locks); - if (res->inflight_locks == 0) - dlm_lockres_clear_refmap_bit(dlm->node_num, res); + + mlog(0, "%s: res %.*s, inflight--: now %u, %ps()\n", dlm->name, + res->lockname.len, res->lockname.name, res->inflight_locks, + __builtin_return_address(0)); + wake_up(&res->wq); } @@ -697,7 +716,6 @@ struct dlm_lock_resource * dlm_get_lock_resource(struct dlm_ctxt *dlm, unsigned int hash; int tries = 0; int bit, wait_on_recovery = 0; - int drop_inflight_if_nonlocal = 0; BUG_ON(!lockid); @@ -709,36 +727,46 @@ lookup: spin_lock(&dlm->spinlock); tmpres = __dlm_lookup_lockres_full(dlm, lockid, namelen, hash); if (tmpres) { - int dropping_ref = 0; - spin_unlock(&dlm->spinlock); - spin_lock(&tmpres->spinlock); - /* We wait for the other thread that is mastering the resource */ + + /* + * Right after dlm spinlock was released, dlm_thread could have + * purged the lockres. Check if lockres got unhashed. If so + * start over. + */ + if (hlist_unhashed(&tmpres->hash_node)) { + spin_unlock(&tmpres->spinlock); + dlm_lockres_put(tmpres); + tmpres = NULL; + goto lookup; + } + + /* Wait on the thread that is mastering the resource */ if (tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN) { __dlm_wait_on_lockres(tmpres); BUG_ON(tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN); + spin_unlock(&tmpres->spinlock); + dlm_lockres_put(tmpres); + tmpres = NULL; + goto lookup; } - if (tmpres->owner == dlm->node_num) { - BUG_ON(tmpres->state & DLM_LOCK_RES_DROPPING_REF); - dlm_lockres_grab_inflight_ref(dlm, tmpres); - } else if (tmpres->state & DLM_LOCK_RES_DROPPING_REF) - dropping_ref = 1; - spin_unlock(&tmpres->spinlock); - - /* wait until done messaging the master, drop our ref to allow - * the lockres to be purged, start over. */ - if (dropping_ref) { - spin_lock(&tmpres->spinlock); - __dlm_wait_on_lockres_flags(tmpres, DLM_LOCK_RES_DROPPING_REF); + /* Wait on the resource purge to complete before continuing */ + if (tmpres->state & DLM_LOCK_RES_DROPPING_REF) { + BUG_ON(tmpres->owner == dlm->node_num); + __dlm_wait_on_lockres_flags(tmpres, + DLM_LOCK_RES_DROPPING_REF); spin_unlock(&tmpres->spinlock); dlm_lockres_put(tmpres); tmpres = NULL; goto lookup; } - mlog(0, "found in hash!\n"); + /* Grab inflight ref to pin the resource */ + dlm_lockres_grab_inflight_ref(dlm, tmpres); + + spin_unlock(&tmpres->spinlock); if (res) dlm_lockres_put(res); res = tmpres; @@ -829,8 +857,8 @@ lookup: * but they might own this lockres. wait on them. */ bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) { - mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to " - "recover before lock mastery can begin\n", + mlog(0, "%s: res %.*s, At least one node (%d) " + "to recover before lock mastery can begin\n", dlm->name, namelen, (char *)lockid, bit); wait_on_recovery = 1; } @@ -843,12 +871,9 @@ lookup: /* finally add the lockres to its hash bucket */ __dlm_insert_lockres(dlm, res); - /* since this lockres is new it doesn't not require the spinlock */ - dlm_lockres_grab_inflight_ref_new(dlm, res); - /* if this node does not become the master make sure to drop - * this inflight reference below */ - drop_inflight_if_nonlocal = 1; + /* since this lockres is new it doesn't not require the spinlock */ + __dlm_lockres_grab_inflight_ref(dlm, res); /* get an extra ref on the mle in case this is a BLOCK * if so, the creator of the BLOCK may try to put the last @@ -864,8 +889,8 @@ redo_request: * dlm spinlock would be detectable be a change on the mle, * so we only need to clear out the recovery map once. */ if (dlm_is_recovery_lock(lockid, namelen)) { - mlog(ML_NOTICE, "%s: recovery map is not empty, but " - "must master $RECOVERY lock now\n", dlm->name); + mlog(0, "%s: Recovery map is not empty, but must " + "master $RECOVERY lock now\n", dlm->name); if (!dlm_pre_master_reco_lockres(dlm, res)) wait_on_recovery = 0; else { @@ -883,8 +908,8 @@ redo_request: spin_lock(&dlm->spinlock); bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) { - mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to " - "recover before lock mastery can begin\n", + mlog(0, "%s: res %.*s, At least one node (%d) " + "to recover before lock mastery can begin\n", dlm->name, namelen, (char *)lockid, bit); wait_on_recovery = 1; } else @@ -913,8 +938,8 @@ redo_request: * yet, keep going until it does. this is how the * master will know that asserts are needed back to * the lower nodes. */ - mlog(0, "%s:%.*s: requests only up to %u but master " - "is %u, keep going\n", dlm->name, namelen, + mlog(0, "%s: res %.*s, Requests only up to %u but " + "master is %u, keep going\n", dlm->name, namelen, lockid, nodenum, mle->master); } } @@ -924,13 +949,12 @@ wait: ret = dlm_wait_for_lock_mastery(dlm, res, mle, &blocked); if (ret < 0) { wait_on_recovery = 1; - mlog(0, "%s:%.*s: node map changed, redo the " - "master request now, blocked=%d\n", - dlm->name, res->lockname.len, + mlog(0, "%s: res %.*s, Node map changed, redo the master " + "request now, blocked=%d\n", dlm->name, res->lockname.len, res->lockname.name, blocked); if (++tries > 20) { - mlog(ML_ERROR, "%s:%.*s: spinning on " - "dlm_wait_for_lock_mastery, blocked=%d\n", + mlog(ML_ERROR, "%s: res %.*s, Spinning on " + "dlm_wait_for_lock_mastery, blocked = %d\n", dlm->name, res->lockname.len, res->lockname.name, blocked); dlm_print_one_lock_resource(res); @@ -940,7 +964,8 @@ wait: goto redo_request; } - mlog(0, "lockres mastered by %u\n", res->owner); + mlog(0, "%s: res %.*s, Mastered by %u\n", dlm->name, res->lockname.len, + res->lockname.name, res->owner); /* make sure we never continue without this */ BUG_ON(res->owner == O2NM_MAX_NODES); @@ -952,8 +977,6 @@ wait: wake_waiters: spin_lock(&res->spinlock); - if (res->owner != dlm->node_num && drop_inflight_if_nonlocal) - dlm_lockres_drop_inflight_ref(dlm, res); res->state &= ~DLM_LOCK_RES_IN_PROGRESS; spin_unlock(&res->spinlock); wake_up(&res->wq); @@ -1388,6 +1411,7 @@ int dlm_master_request_handler(struct o2net_msg *msg, u32 len, void *data, int found, ret; int set_maybe; int dispatch_assert = 0; + int dispatched = 0; if (!dlm_grab(dlm)) return DLM_MASTER_RESP_NO; @@ -1426,9 +1450,7 @@ way_up_top: } if (res->owner == dlm->node_num) { - mlog(0, "%s:%.*s: setting bit %u in refmap\n", - dlm->name, namelen, name, request->node_idx); - dlm_lockres_set_refmap_bit(request->node_idx, res); + dlm_lockres_set_refmap_bit(dlm, res, request->node_idx); spin_unlock(&res->spinlock); response = DLM_MASTER_RESP_YES; if (mle) @@ -1493,10 +1515,8 @@ way_up_top: * go back and clean the mles on any * other nodes */ dispatch_assert = 1; - dlm_lockres_set_refmap_bit(request->node_idx, res); - mlog(0, "%s:%.*s: setting bit %u in refmap\n", - dlm->name, namelen, name, - request->node_idx); + dlm_lockres_set_refmap_bit(dlm, res, + request->node_idx); } else response = DLM_MASTER_RESP_NO; } else { @@ -1598,13 +1618,16 @@ send_response: mlog(ML_ERROR, "failed to dispatch assert master work\n"); response = DLM_MASTER_RESP_ERROR; dlm_lockres_put(res); + } else { + dispatched = 1; } } else { if (res) dlm_lockres_put(res); } - dlm_put(dlm); + if (!dispatched) + dlm_put(dlm); return response; } @@ -1702,7 +1725,7 @@ again: "lockres, set the bit in the refmap\n", namelen, lockname, to); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(to, res); + dlm_lockres_set_refmap_bit(dlm, res, to); spin_unlock(&res->spinlock); } } @@ -2022,7 +2045,6 @@ int dlm_dispatch_assert_master(struct dlm_ctxt *dlm, /* queue up work for dlm_assert_master_worker */ - dlm_grab(dlm); /* get an extra ref for the work item */ dlm_init_work_item(dlm, item, dlm_assert_master_worker, NULL); item->u.am.lockres = res; /* already have a ref */ /* can optionally ignore node numbers higher than this node */ @@ -2187,8 +2209,6 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) namelen = res->lockname.len; BUG_ON(namelen > O2NM_MAX_NAME_LEN); - mlog(0, "%s:%.*s: sending deref to %d\n", - dlm->name, namelen, lockname, res->owner); memset(&deref, 0, sizeof(deref)); deref.node_idx = dlm->node_num; deref.namelen = namelen; @@ -2197,14 +2217,12 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) ret = o2net_send_message(DLM_DEREF_LOCKRES_MSG, dlm->key, &deref, sizeof(deref), res->owner, &r); if (ret < 0) - mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " - "node %u\n", ret, DLM_DEREF_LOCKRES_MSG, dlm->key, - res->owner); + mlog(ML_ERROR, "%s: res %.*s, error %d send DEREF to node %u\n", + dlm->name, namelen, lockname, ret, res->owner); else if (r < 0) { /* BAD. other node says I did not have a ref. */ - mlog(ML_ERROR,"while dropping ref on %s:%.*s " - "(master=%u) got %d.\n", dlm->name, namelen, - lockname, res->owner, r); + mlog(ML_ERROR, "%s: res %.*s, DEREF to node %u got %d\n", + dlm->name, namelen, lockname, res->owner, r); dlm_print_one_lock_resource(res); BUG(); } @@ -2260,7 +2278,7 @@ int dlm_deref_lockres_handler(struct o2net_msg *msg, u32 len, void *data, else { BUG_ON(res->state & DLM_LOCK_RES_DROPPING_REF); if (test_bit(node, res->refmap)) { - dlm_lockres_clear_refmap_bit(node, res); + dlm_lockres_clear_refmap_bit(dlm, res, node); cleared = 1; } } @@ -2320,7 +2338,7 @@ static void dlm_deref_lockres_worker(struct dlm_work_item *item, void *data) BUG_ON(res->state & DLM_LOCK_RES_DROPPING_REF); if (test_bit(node, res->refmap)) { __dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG); - dlm_lockres_clear_refmap_bit(node, res); + dlm_lockres_clear_refmap_bit(dlm, res, node); cleared = 1; } spin_unlock(&res->spinlock); @@ -2802,7 +2820,8 @@ static void dlm_remove_nonlocal_locks(struct dlm_ctxt *dlm, BUG_ON(!list_empty(&lock->bast_list)); BUG_ON(lock->ast_pending); BUG_ON(lock->bast_pending); - dlm_lockres_clear_refmap_bit(lock->ml.node, res); + dlm_lockres_clear_refmap_bit(dlm, res, + lock->ml.node); list_del_init(&lock->list); dlm_lock_put(lock); /* In a normal unlock, we would have added a @@ -2823,7 +2842,7 @@ static void dlm_remove_nonlocal_locks(struct dlm_ctxt *dlm, mlog(0, "%s:%.*s: node %u had a ref to this " "migrating lockres, clearing\n", dlm->name, res->lockname.len, res->lockname.name, bit); - dlm_lockres_clear_refmap_bit(bit, res); + dlm_lockres_clear_refmap_bit(dlm, res, bit); } bit++; } @@ -2916,9 +2935,9 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm, &migrate, sizeof(migrate), nodenum, &status); if (ret < 0) { - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_MIGRATE_REQUEST_MSG, - dlm->key, nodenum); + mlog(ML_ERROR, "%s: res %.*s, Error %d send " + "MIGRATE_REQUEST to node %u\n", dlm->name, + migrate.namelen, migrate.name, ret, nodenum); if (!dlm_is_host_down(ret)) { mlog(ML_ERROR, "unhandled error=%d!\n", ret); BUG(); @@ -2937,7 +2956,7 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm, dlm->name, res->lockname.len, res->lockname.name, nodenum); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(nodenum, res); + dlm_lockres_set_refmap_bit(dlm, res, nodenum); spin_unlock(&res->spinlock); } } @@ -3271,7 +3290,7 @@ int dlm_finish_migration(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, * mastery reference here since old_master will briefly have * a reference after the migration completes */ spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(old_master, res); + dlm_lockres_set_refmap_bit(dlm, res, old_master); spin_unlock(&res->spinlock); mlog(0, "now time to do a migrate request to other nodes\n"); diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index 7efab6d..0e5013e 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -362,40 +362,38 @@ static int dlm_is_node_recovered(struct dlm_ctxt *dlm, u8 node) } -int dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout) +void dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout) { - if (timeout) { - mlog(ML_NOTICE, "%s: waiting %dms for notification of " - "death of node %u\n", dlm->name, timeout, node); + if (dlm_is_node_dead(dlm, node)) + return; + + printk(KERN_NOTICE "o2dlm: Waiting on the death of node %u in " + "domain %s\n", node, dlm->name); + + if (timeout) wait_event_timeout(dlm->dlm_reco_thread_wq, - dlm_is_node_dead(dlm, node), - msecs_to_jiffies(timeout)); - } else { - mlog(ML_NOTICE, "%s: waiting indefinitely for notification " - "of death of node %u\n", dlm->name, node); + dlm_is_node_dead(dlm, node), + msecs_to_jiffies(timeout)); + else wait_event(dlm->dlm_reco_thread_wq, dlm_is_node_dead(dlm, node)); - } - /* for now, return 0 */ - return 0; } -int dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout) +void dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout) { - if (timeout) { - mlog(0, "%s: waiting %dms for notification of " - "recovery of node %u\n", dlm->name, timeout, node); + if (dlm_is_node_recovered(dlm, node)) + return; + + printk(KERN_NOTICE "o2dlm: Waiting on the recovery of node %u in " + "domain %s\n", node, dlm->name); + + if (timeout) wait_event_timeout(dlm->dlm_reco_thread_wq, - dlm_is_node_recovered(dlm, node), - msecs_to_jiffies(timeout)); - } else { - mlog(0, "%s: waiting indefinitely for notification " - "of recovery of node %u\n", dlm->name, node); + dlm_is_node_recovered(dlm, node), + msecs_to_jiffies(timeout)); + else wait_event(dlm->dlm_reco_thread_wq, dlm_is_node_recovered(dlm, node)); - } - /* for now, return 0 */ - return 0; } /* callers of the top-level api calls (dlmlock/dlmunlock) should @@ -430,6 +428,8 @@ static void dlm_begin_recovery(struct dlm_ctxt *dlm) { spin_lock(&dlm->spinlock); BUG_ON(dlm->reco.state & DLM_RECO_STATE_ACTIVE); + printk(KERN_NOTICE "o2dlm: Begin recovery on domain %s for node %u\n", + dlm->name, dlm->reco.dead_node); dlm->reco.state |= DLM_RECO_STATE_ACTIVE; spin_unlock(&dlm->spinlock); } @@ -440,9 +440,18 @@ static void dlm_end_recovery(struct dlm_ctxt *dlm) BUG_ON(!(dlm->reco.state & DLM_RECO_STATE_ACTIVE)); dlm->reco.state &= ~DLM_RECO_STATE_ACTIVE; spin_unlock(&dlm->spinlock); + printk(KERN_NOTICE "o2dlm: End recovery on domain %s\n", dlm->name); wake_up(&dlm->reco.event); } +static void dlm_print_recovery_master(struct dlm_ctxt *dlm) +{ + printk(KERN_NOTICE "o2dlm: Node %u (%s) is the Recovery Master for the " + "dead node %u in domain %s\n", dlm->reco.new_master, + (dlm->node_num == dlm->reco.new_master ? "me" : "he"), + dlm->reco.dead_node, dlm->name); +} + static int dlm_do_recovery(struct dlm_ctxt *dlm) { int status = 0; @@ -505,9 +514,8 @@ static int dlm_do_recovery(struct dlm_ctxt *dlm) } mlog(0, "another node will master this recovery session.\n"); } - mlog(0, "dlm=%s (%d), new_master=%u, this node=%u, dead_node=%u\n", - dlm->name, task_pid_nr(dlm->dlm_reco_thread_task), dlm->reco.new_master, - dlm->node_num, dlm->reco.dead_node); + + dlm_print_recovery_master(dlm); /* it is safe to start everything back up here * because all of the dead node's lock resources @@ -518,15 +526,13 @@ static int dlm_do_recovery(struct dlm_ctxt *dlm) return 0; master_here: - mlog(ML_NOTICE, "(%d) Node %u is the Recovery Master for the Dead Node " - "%u for Domain %s\n", task_pid_nr(dlm->dlm_reco_thread_task), - dlm->node_num, dlm->reco.dead_node, dlm->name); + dlm_print_recovery_master(dlm); status = dlm_remaster_locks(dlm, dlm->reco.dead_node); if (status < 0) { /* we should never hit this anymore */ - mlog(ML_ERROR, "error %d remastering locks for node %u, " - "retrying.\n", status, dlm->reco.dead_node); + mlog(ML_ERROR, "%s: Error %d remastering locks for node %u, " + "retrying.\n", dlm->name, status, dlm->reco.dead_node); /* yield a bit to allow any final network messages * to get handled on remaining nodes */ msleep(100); @@ -534,7 +540,10 @@ master_here: /* success! see if any other nodes need recovery */ mlog(0, "DONE mastering recovery of %s:%u here(this=%u)!\n", dlm->name, dlm->reco.dead_node, dlm->node_num); - dlm_reset_recovery(dlm); + spin_lock(&dlm->spinlock); + __dlm_reset_recovery(dlm); + dlm->reco.state &= ~DLM_RECO_STATE_FINALIZE; + spin_unlock(&dlm->spinlock); } dlm_end_recovery(dlm); @@ -567,7 +576,7 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) BUG_ON(ndata->state != DLM_RECO_NODE_DATA_INIT); ndata->state = DLM_RECO_NODE_DATA_REQUESTING; - mlog(0, "requesting lock info from node %u\n", + mlog(0, "%s: Requesting lock info from node %u\n", dlm->name, ndata->node_num); if (ndata->node_num == dlm->node_num) { @@ -640,7 +649,7 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) spin_unlock(&dlm_reco_state_lock); } - mlog(0, "done requesting all lock info\n"); + mlog(0, "%s: Done requesting all lock info\n", dlm->name); /* nodes should be sending reco data now * just need to wait */ @@ -692,6 +701,14 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) if (all_nodes_done) { int ret; + /* Set this flag on recovery master to avoid + * a new recovery for another dead node start + * before the recovery is not done. That may + * cause recovery hung.*/ + spin_lock(&dlm->spinlock); + dlm->reco.state |= DLM_RECO_STATE_FINALIZE; + spin_unlock(&dlm->spinlock); + /* all nodes are now in DLM_RECO_NODE_DATA_DONE state * just send a finalize message to everyone and * clean up */ @@ -802,10 +819,9 @@ static int dlm_request_all_locks(struct dlm_ctxt *dlm, u8 request_from, /* negative status is handled by caller */ if (ret < 0) - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_LOCK_REQUEST_MSG, - dlm->key, request_from); - + mlog(ML_ERROR, "%s: Error %d send LOCK_REQUEST to node %u " + "to recover dead node %u\n", dlm->name, ret, + request_from, dead_node); // return from here, then // sleep until all received or error return ret; @@ -956,9 +972,9 @@ static int dlm_send_all_done_msg(struct dlm_ctxt *dlm, u8 dead_node, u8 send_to) ret = o2net_send_message(DLM_RECO_DATA_DONE_MSG, dlm->key, &done_msg, sizeof(done_msg), send_to, &tmpret); if (ret < 0) { - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_RECO_DATA_DONE_MSG, - dlm->key, send_to); + mlog(ML_ERROR, "%s: Error %d send RECO_DATA_DONE to node %u " + "to recover dead node %u\n", dlm->name, ret, send_to, + dead_node); if (!dlm_is_host_down(ret)) { BUG(); } @@ -1127,9 +1143,11 @@ static int dlm_send_mig_lockres_msg(struct dlm_ctxt *dlm, if (ret < 0) { /* XXX: negative status is not handled. * this will end up killing this node. */ - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_MIG_LOCKRES_MSG, - dlm->key, send_to); + mlog(ML_ERROR, "%s: res %.*s, Error %d send MIG_LOCKRES to " + "node %u (%s)\n", dlm->name, mres->lockname_len, + mres->lockname, ret, send_to, + (orig_flags & DLM_MRES_MIGRATION ? + "migration" : "recovery")); } else { /* might get an -ENOMEM back here */ ret = status; @@ -1671,6 +1689,7 @@ int dlm_master_requery_handler(struct o2net_msg *msg, u32 len, void *data, unsigned int hash; int master = DLM_LOCK_RES_OWNER_UNKNOWN; u32 flags = DLM_ASSERT_MASTER_REQUERY; + int dispatched = 0; if (!dlm_grab(dlm)) { /* since the domain has gone away on this @@ -1692,6 +1711,8 @@ int dlm_master_requery_handler(struct o2net_msg *msg, u32 len, void *data, mlog_errno(-ENOMEM); /* retry!? */ BUG(); + } else { + dispatched = 1; } } else /* put.. incase we are not the master */ dlm_lockres_put(res); @@ -1699,7 +1720,8 @@ int dlm_master_requery_handler(struct o2net_msg *msg, u32 len, void *data, } spin_unlock(&dlm->spinlock); - dlm_put(dlm); + if (!dispatched) + dlm_put(dlm); return master; } @@ -1745,13 +1767,13 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, struct dlm_migratable_lockres *mres) { struct dlm_migratable_lock *ml; - struct list_head *queue; + struct list_head *queue, *iter; struct list_head *tmpq = NULL; struct dlm_lock *newlock = NULL; struct dlm_lockstatus *lksb = NULL; int ret = 0; int i, j, bad; - struct dlm_lock *lock = NULL; + struct dlm_lock *lock; u8 from = O2NM_MAX_NODES; unsigned int added = 0; __be64 c; @@ -1767,7 +1789,7 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, dlm->name, mres->lockname_len, mres->lockname, from); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(from, res); + dlm_lockres_set_refmap_bit(dlm, res, from); spin_unlock(&res->spinlock); added++; break; @@ -1786,14 +1808,16 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, /* MIGRATION ONLY! */ BUG_ON(!(mres->flags & DLM_MRES_MIGRATION)); + lock = NULL; spin_lock(&res->spinlock); for (j = DLM_GRANTED_LIST; j <= DLM_BLOCKED_LIST; j++) { tmpq = dlm_list_idx_to_ptr(res, j); - list_for_each_entry(lock, tmpq, list) { - if (lock->ml.cookie != ml->cookie) - lock = NULL; - else + list_for_each(iter, tmpq) { + lock = list_entry(iter, + struct dlm_lock, list); + if (lock->ml.cookie == ml->cookie) break; + lock = NULL; } if (lock) break; @@ -1965,7 +1989,7 @@ skip_lvb: mlog(0, "%s:%.*s: added lock for node %u, " "setting refmap bit\n", dlm->name, res->lockname.len, res->lockname.name, ml->node); - dlm_lockres_set_refmap_bit(ml->node, res); + dlm_lockres_set_refmap_bit(dlm, res, ml->node); added++; } spin_unlock(&res->spinlock); @@ -2084,6 +2108,9 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm, list_for_each_entry_safe(res, next, &dlm->reco.resources, recovering) { if (res->owner == dead_node) { + mlog(0, "%s: res %.*s, Changing owner from %u to %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->owner, new_master); list_del_init(&res->recovering); spin_lock(&res->spinlock); /* new_master has our reference from @@ -2105,40 +2132,30 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm, for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_lockres_hash(dlm, i); hlist_for_each_entry(res, hash_iter, bucket, hash_node) { - if (res->state & DLM_LOCK_RES_RECOVERING) { - if (res->owner == dead_node) { - mlog(0, "(this=%u) res %.*s owner=%u " - "was not on recovering list, but " - "clearing state anyway\n", - dlm->node_num, res->lockname.len, - res->lockname.name, new_master); - } else if (res->owner == dlm->node_num) { - mlog(0, "(this=%u) res %.*s owner=%u " - "was not on recovering list, " - "owner is THIS node, clearing\n", - dlm->node_num, res->lockname.len, - res->lockname.name, new_master); - } else - continue; + if (!(res->state & DLM_LOCK_RES_RECOVERING)) + continue; - if (!list_empty(&res->recovering)) { - mlog(0, "%s:%.*s: lockres was " - "marked RECOVERING, owner=%u\n", - dlm->name, res->lockname.len, - res->lockname.name, res->owner); - list_del_init(&res->recovering); - dlm_lockres_put(res); - } - spin_lock(&res->spinlock); - /* new_master has our reference from - * the lock state sent during recovery */ - dlm_change_lockres_owner(dlm, res, new_master); - res->state &= ~DLM_LOCK_RES_RECOVERING; - if (__dlm_lockres_has_locks(res)) - __dlm_dirty_lockres(dlm, res); - spin_unlock(&res->spinlock); - wake_up(&res->wq); + if (res->owner != dead_node && + res->owner != dlm->node_num) + continue; + + if (!list_empty(&res->recovering)) { + list_del_init(&res->recovering); + dlm_lockres_put(res); } + + /* new_master has our reference from + * the lock state sent during recovery */ + mlog(0, "%s: res %.*s, Changing owner from %u to %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->owner, new_master); + spin_lock(&res->spinlock); + dlm_change_lockres_owner(dlm, res, new_master); + res->state &= ~DLM_LOCK_RES_RECOVERING; + if (__dlm_lockres_has_locks(res)) + __dlm_dirty_lockres(dlm, res); + spin_unlock(&res->spinlock); + wake_up(&res->wq); } } } @@ -2252,12 +2269,12 @@ static void dlm_free_dead_locks(struct dlm_ctxt *dlm, res->lockname.len, res->lockname.name, freed, dead_node); __dlm_print_one_lock_resource(res); } - dlm_lockres_clear_refmap_bit(dead_node, res); + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); } else if (test_bit(dead_node, res->refmap)) { mlog(0, "%s:%.*s: dead node %u had a ref, but had " "no locks and had not purged before dying\n", dlm->name, res->lockname.len, res->lockname.name, dead_node); - dlm_lockres_clear_refmap_bit(dead_node, res); + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); } /* do not kick thread yet */ @@ -2324,9 +2341,9 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) dlm_revalidate_lvb(dlm, res, dead_node); if (res->owner == dead_node) { if (res->state & DLM_LOCK_RES_DROPPING_REF) { - mlog(ML_NOTICE, "Ignore %.*s for " + mlog(ML_NOTICE, "%s: res %.*s, Skip " "recovery as it is being freed\n", - res->lockname.len, + dlm->name, res->lockname.len, res->lockname.name); } else dlm_move_lockres_to_recovery_list(dlm, @@ -2870,8 +2887,8 @@ int dlm_finalize_reco_handler(struct o2net_msg *msg, u32 len, void *data, BUG(); } dlm->reco.state &= ~DLM_RECO_STATE_FINALIZE; + __dlm_reset_recovery(dlm); spin_unlock(&dlm->spinlock); - dlm_reset_recovery(dlm); dlm_kick_recovery_thread(dlm); break; default: diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index 1d6d1d2..e73c833 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -94,24 +94,26 @@ int __dlm_lockres_unused(struct dlm_lock_resource *res) { int bit; + assert_spin_locked(&res->spinlock); + if (__dlm_lockres_has_locks(res)) return 0; + /* Locks are in the process of being created */ + if (res->inflight_locks) + return 0; + if (!list_empty(&res->dirty) || res->state & DLM_LOCK_RES_DIRTY) return 0; if (res->state & DLM_LOCK_RES_RECOVERING) return 0; + /* Another node has this resource with this node as the master */ bit = find_next_bit(res->refmap, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) return 0; - /* - * since the bit for dlm->node_num is not set, inflight_locks better - * be zero - */ - BUG_ON(res->inflight_locks != 0); return 1; } @@ -185,8 +187,6 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, /* clear our bit from the master's refmap, ignore errors */ ret = dlm_drop_lockres_ref(dlm, res); if (ret < 0) { - mlog(ML_ERROR, "%s: deref %.*s failed %d\n", dlm->name, - res->lockname.len, res->lockname.name, ret); if (!dlm_is_host_down(ret)) BUG(); } @@ -209,7 +209,7 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, BUG(); } - __dlm_unhash_lockres(res); + __dlm_unhash_lockres(dlm, res); /* lockres is not in the hash now. drop the flag and wake up * any processes waiting in dlm_get_lock_resource. */ diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index ab4046f..b5e457c 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -1692,7 +1692,7 @@ int ocfs2_open_lock(struct inode *inode) mlog(0, "inode %llu take PRMODE open lock\n", (unsigned long long)OCFS2_I(inode)->ip_blkno); - if (ocfs2_mount_local(osb)) + if (ocfs2_is_hard_readonly(osb) || ocfs2_mount_local(osb)) goto out; lockres = &OCFS2_I(inode)->ip_open_lockres; @@ -1718,6 +1718,12 @@ int ocfs2_try_open_lock(struct inode *inode, int write) (unsigned long long)OCFS2_I(inode)->ip_blkno, write ? "EXMODE" : "PRMODE"); + if (ocfs2_is_hard_readonly(osb)) { + if (write) + status = -EROFS; + goto out; + } + if (ocfs2_mount_local(osb)) goto out; @@ -2092,7 +2098,7 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode) inode->i_uid = be32_to_cpu(lvb->lvb_iuid); inode->i_gid = be32_to_cpu(lvb->lvb_igid); inode->i_mode = be16_to_cpu(lvb->lvb_imode); - inode->i_nlink = be16_to_cpu(lvb->lvb_inlink); + set_nlink(inode, be16_to_cpu(lvb->lvb_inlink)); ocfs2_unpack_timespec(&inode->i_atime, be64_to_cpu(lvb->lvb_iatime_packed)); ocfs2_unpack_timespec(&inode->i_mtime, @@ -2298,7 +2304,7 @@ int ocfs2_inode_lock_full_nested(struct inode *inode, if (ocfs2_is_hard_readonly(osb)) { if (ex) status = -EROFS; - goto bail; + goto getbh; } if (ocfs2_mount_local(osb)) @@ -2356,7 +2362,7 @@ local: mlog_errno(status); goto bail; } - +getbh: if (ret_bh) { status = ocfs2_assign_bh(inode, ret_bh, local_bh); if (status < 0) { @@ -2631,8 +2637,11 @@ int ocfs2_dentry_lock(struct dentry *dentry, int ex) BUG_ON(!dl); - if (ocfs2_is_hard_readonly(osb)) - return -EROFS; + if (ocfs2_is_hard_readonly(osb)) { + if (ex) + return -EROFS; + return 0; + } if (ocfs2_mount_local(osb)) return 0; @@ -2650,7 +2659,7 @@ void ocfs2_dentry_unlock(struct dentry *dentry, int ex) struct ocfs2_dentry_lock *dl = dentry->d_fsdata; struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); - if (!ocfs2_mount_local(osb)) + if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb)) ocfs2_cluster_unlock(osb, &dl->dl_lockres, level); } @@ -3959,9 +3968,13 @@ static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb) osb->dc_work_sequence = osb->dc_wake_sequence; processed = osb->blocked_lock_count; - while (processed) { - BUG_ON(list_empty(&osb->blocked_lock_list)); - + /* + * blocked lock processing in this loop might call iput which can + * remove items off osb->blocked_lock_list. Downconvert up to + * 'processed' number of locks, but stop short if we had some + * removed in ocfs2_mark_lockres_freeing when downconverting. + */ + while (processed && !list_empty(&osb->blocked_lock_list)) { lockres = list_entry(osb->blocked_lock_list.next, struct ocfs2_lock_res, l_blocked_list); list_del_init(&lockres->l_blocked_list); diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c index 774a032..cf22847 100644 --- a/fs/ocfs2/extent_map.c +++ b/fs/ocfs2/extent_map.c @@ -831,6 +831,102 @@ out: return ret; } +int ocfs2_seek_data_hole_offset(struct file *file, loff_t *offset, int origin) +{ + struct inode *inode = file->f_mapping->host; + int ret; + unsigned int is_last = 0, is_data = 0; + u16 cs_bits = OCFS2_SB(inode->i_sb)->s_clustersize_bits; + u32 cpos, cend, clen, hole_size; + u64 extoff, extlen; + struct buffer_head *di_bh = NULL; + struct ocfs2_extent_rec rec; + + BUG_ON(origin != SEEK_DATA && origin != SEEK_HOLE); + + ret = ocfs2_inode_lock(inode, &di_bh, 0); + if (ret) { + mlog_errno(ret); + goto out; + } + + down_read(&OCFS2_I(inode)->ip_alloc_sem); + + if (*offset >= inode->i_size) { + ret = -ENXIO; + goto out_unlock; + } + + if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) { + if (origin == SEEK_HOLE) + *offset = inode->i_size; + goto out_unlock; + } + + clen = 0; + cpos = *offset >> cs_bits; + cend = ocfs2_clusters_for_bytes(inode->i_sb, inode->i_size); + + while (cpos < cend && !is_last) { + ret = ocfs2_get_clusters_nocache(inode, di_bh, cpos, &hole_size, + &rec, &is_last); + if (ret) { + mlog_errno(ret); + goto out_unlock; + } + + extoff = cpos; + extoff <<= cs_bits; + + if (rec.e_blkno == 0ULL) { + clen = hole_size; + is_data = 0; + } else { + clen = le16_to_cpu(rec.e_leaf_clusters) - + (cpos - le32_to_cpu(rec.e_cpos)); + is_data = (rec.e_flags & OCFS2_EXT_UNWRITTEN) ? 0 : 1; + } + + if ((!is_data && origin == SEEK_HOLE) || + (is_data && origin == SEEK_DATA)) { + if (extoff > *offset) + *offset = extoff; + goto out_unlock; + } + + if (!is_last) + cpos += clen; + } + + if (origin == SEEK_HOLE) { + extoff = cpos; + extoff <<= cs_bits; + extlen = clen; + extlen <<= cs_bits; + + if ((extoff + extlen) > inode->i_size) + extlen = inode->i_size - extoff; + extoff += extlen; + if (extoff > *offset) + *offset = extoff; + goto out_unlock; + } + + ret = -ENXIO; + +out_unlock: + + brelse(di_bh); + + up_read(&OCFS2_I(inode)->ip_alloc_sem); + + ocfs2_inode_unlock(inode, 0); +out: + if (ret && ret != -ENXIO) + ret = -ENXIO; + return ret; +} + int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr, struct buffer_head *bhs[], int flags, int (*validate)(struct super_block *sb, diff --git a/fs/ocfs2/extent_map.h b/fs/ocfs2/extent_map.h index e79d41c..67ea57d 100644 --- a/fs/ocfs2/extent_map.h +++ b/fs/ocfs2/extent_map.h @@ -53,6 +53,8 @@ int ocfs2_extent_map_get_blocks(struct inode *inode, u64 v_blkno, u64 *p_blkno, int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 map_start, u64 map_len); +int ocfs2_seek_data_hole_offset(struct file *file, loff_t *offset, int origin); + int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster, u32 *p_cluster, u32 *num_clusters, struct ocfs2_extent_list *el, diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index b1e35a3..6a7a3d9 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -171,7 +171,8 @@ static int ocfs2_dir_release(struct inode *inode, struct file *file) return 0; } -static int ocfs2_sync_file(struct file *file, int datasync) +static int ocfs2_sync_file(struct file *file, loff_t start, loff_t end, + int datasync) { int err = 0; journal_t *journal; @@ -184,6 +185,16 @@ static int ocfs2_sync_file(struct file *file, int datasync) file->f_path.dentry->d_name.name, (unsigned long long)datasync); + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + /* + * Probably don't need the i_mutex at all in here, just putting it here + * to be consistent with how fsync used to be called, someone more + * familiar with the fs could possibly remove it. + */ + mutex_lock(&inode->i_mutex); if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) { /* * We still have to flush drive's caches to get data to the @@ -200,6 +211,7 @@ static int ocfs2_sync_file(struct file *file, int datasync) bail: if (err) mlog_errno(err); + mutex_unlock(&inode->i_mutex); return (err < 0) ? -EIO : 0; } @@ -1142,6 +1154,8 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) if (status) goto bail_unlock; + inode_dio_wait(inode); + if (i_size_read(inode) > attr->ia_size) { if (ocfs2_should_order_data(inode)) { status = ocfs2_begin_ordered_truncate(inode, @@ -1279,11 +1293,11 @@ bail: return err; } -int ocfs2_permission(struct inode *inode, int mask, unsigned int flags) +int ocfs2_permission(struct inode *inode, int mask) { int ret; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; ret = ocfs2_inode_lock(inode, NULL, 0); @@ -1293,7 +1307,7 @@ int ocfs2_permission(struct inode *inode, int mask, unsigned int flags) goto out; } - ret = generic_permission(inode, mask, flags, ocfs2_check_acl); + ret = generic_permission(inode, mask); ocfs2_inode_unlock(inode, 0); out: @@ -1936,6 +1950,9 @@ static int __ocfs2_change_file_space(struct file *file, struct inode *inode, if (ret < 0) mlog_errno(ret); + if (file && (file->f_flags & O_SYNC)) + handle->h_sync = 1; + ocfs2_commit_trans(osb, handle); out_inode_unlock: @@ -2038,6 +2055,23 @@ out: return ret; } +static void ocfs2_aiodio_wait(struct inode *inode) +{ + wait_queue_head_t *wq = ocfs2_ioend_wq(inode); + + wait_event(*wq, (atomic_read(&OCFS2_I(inode)->ip_unaligned_aio) == 0)); +} + +static int ocfs2_is_io_unaligned(struct inode *inode, size_t count, loff_t pos) +{ + int blockmask = inode->i_sb->s_blocksize - 1; + loff_t final_size = pos + count; + + if ((pos & blockmask) || (final_size & blockmask)) + return 1; + return 0; +} + static int ocfs2_prepare_inode_for_refcount(struct inode *inode, struct file *file, loff_t pos, size_t count, @@ -2216,6 +2250,7 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); int full_coherency = !(osb->s_mount_opt & OCFS2_MOUNT_COHERENCY_BUFFERED); + int unaligned_dio = 0; trace_ocfs2_file_aio_write(inode, file, file->f_path.dentry, (unsigned long long)OCFS2_I(inode)->ip_blkno, @@ -2236,9 +2271,8 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, ocfs2_iocb_clear_sem_locked(iocb); relock: - /* to match setattr's i_mutex -> i_alloc_sem -> rw_lock ordering */ + /* to match setattr's i_mutex -> rw_lock ordering */ if (direct_io) { - down_read(&inode->i_alloc_sem); have_alloc_sem = 1; /* communicate with ocfs2_dio_end_io */ ocfs2_iocb_set_sem_locked(iocb); @@ -2284,13 +2318,16 @@ relock: goto out; } + if (direct_io && !is_sync_kiocb(iocb)) + unaligned_dio = ocfs2_is_io_unaligned(inode, iocb->ki_left, + *ppos); + /* * We can't complete the direct I/O as requested, fall back to * buffered I/O. */ if (direct_io && !can_do_direct) { ocfs2_rw_unlock(inode, rw_level); - up_read(&inode->i_alloc_sem); have_alloc_sem = 0; rw_level = -1; @@ -2299,6 +2336,18 @@ relock: goto relock; } + if (unaligned_dio) { + /* + * Wait on previous unaligned aio to complete before + * proceeding. + */ + ocfs2_aiodio_wait(inode); + + /* Mark the iocb as needing a decrement in ocfs2_dio_end_io */ + atomic_inc(&OCFS2_I(inode)->ip_unaligned_aio); + ocfs2_iocb_set_unaligned_aio(iocb); + } + /* * To later detect whether a journal commit for sync writes is * necessary, we sample i_size, and cluster count here. @@ -2338,10 +2387,14 @@ out_dio: /* buffered aio wouldn't have proper lock coverage today */ BUG_ON(ret == -EIOCBQUEUED && !(file->f_flags & O_DIRECT)); + if (unlikely(written <= 0)) + goto no_sync; + if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode) || ((file->f_flags & O_DIRECT) && !direct_io)) { - ret = filemap_fdatawrite_range(file->f_mapping, pos, - pos + count - 1); + ret = filemap_fdatawrite_range(file->f_mapping, + iocb->ki_pos - written, + iocb->ki_pos - 1); if (ret < 0) written = ret; @@ -2354,15 +2407,16 @@ out_dio: } if (!ret) - ret = filemap_fdatawait_range(file->f_mapping, pos, - pos + count - 1); + ret = filemap_fdatawait_range(file->f_mapping, + iocb->ki_pos - written, + iocb->ki_pos - 1); } +no_sync: /* * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io * function pointer which is called when o_direct io completes so that - * it can unlock our rw lock. (it's the clustered equivalent of - * i_alloc_sem; protects truncate from racing with pending ios). + * it can unlock our rw lock. * Unfortunately there are error cases which call end_io and others * that don't. so we don't have to unlock the rw_lock if either an * async dio is going to do it in the future or an end_io after an @@ -2371,6 +2425,12 @@ out_dio: if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) { rw_level = -1; have_alloc_sem = 0; + unaligned_dio = 0; + } + + if (unaligned_dio) { + ocfs2_iocb_clear_unaligned_aio(iocb); + atomic_dec(&OCFS2_I(inode)->ip_unaligned_aio); } out: @@ -2378,10 +2438,8 @@ out: ocfs2_rw_unlock(inode, rw_level); out_sems: - if (have_alloc_sem) { - up_read(&inode->i_alloc_sem); + if (have_alloc_sem) ocfs2_iocb_clear_sem_locked(iocb); - } mutex_unlock(&inode->i_mutex); @@ -2416,9 +2474,7 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe, struct address_space *mapping = out->f_mapping; struct inode *inode = mapping->host; struct splice_desc sd = { - .total_len = len, .flags = flags, - .pos = *ppos, .u.file = out, }; @@ -2428,6 +2484,12 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe, out->f_path.dentry->d_name.len, out->f_path.dentry->d_name.name, len); + ret = generic_write_checks(out, ppos, &len, 0); + if (ret) + return ret; + sd.total_len = len; + sd.pos = *ppos; + if (pipe->inode) mutex_lock_nested(&pipe->inode->i_mutex, I_MUTEX_PARENT); @@ -2531,7 +2593,6 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, * need locks to protect pending reads from racing with truncate. */ if (filp->f_flags & O_DIRECT) { - down_read(&inode->i_alloc_sem); have_alloc_sem = 1; ocfs2_iocb_set_sem_locked(iocb); @@ -2574,16 +2635,66 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, } bail: - if (have_alloc_sem) { - up_read(&inode->i_alloc_sem); + if (have_alloc_sem) ocfs2_iocb_clear_sem_locked(iocb); - } + if (rw_level != -1) ocfs2_rw_unlock(inode, rw_level); return ret; } +/* Refer generic_file_llseek_unlocked() */ +static loff_t ocfs2_file_llseek(struct file *file, loff_t offset, int origin) +{ + struct inode *inode = file->f_mapping->host; + int ret = 0; + + mutex_lock(&inode->i_mutex); + + switch (origin) { + case SEEK_SET: + break; + case SEEK_END: + offset += inode->i_size; + break; + case SEEK_CUR: + if (offset == 0) { + offset = file->f_pos; + goto out; + } + offset += file->f_pos; + break; + case SEEK_DATA: + case SEEK_HOLE: + ret = ocfs2_seek_data_hole_offset(file, &offset, origin); + if (ret) + goto out; + break; + default: + ret = -EINVAL; + goto out; + } + + if (offset < 0 && !(file->f_mode & FMODE_UNSIGNED_OFFSET)) + ret = -EINVAL; + if (!ret && offset > inode->i_sb->s_maxbytes) + ret = -EINVAL; + if (ret) + goto out; + + if (offset != file->f_pos) { + file->f_pos = offset; + file->f_version = 0; + } + +out: + mutex_unlock(&inode->i_mutex); + if (ret) + return ret; + return offset; +} + const struct inode_operations ocfs2_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, @@ -2593,12 +2704,14 @@ const struct inode_operations ocfs2_file_iops = { .listxattr = ocfs2_listxattr, .removexattr = generic_removexattr, .fiemap = ocfs2_fiemap, + .get_acl = ocfs2_iop_get_acl, }; const struct inode_operations ocfs2_special_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, .permission = ocfs2_permission, + .get_acl = ocfs2_iop_get_acl, }; /* @@ -2606,7 +2719,7 @@ const struct inode_operations ocfs2_special_file_iops = { * ocfs2_fops_no_plocks and ocfs2_dops_no_plocks! */ const struct file_operations ocfs2_fops = { - .llseek = generic_file_llseek, + .llseek = ocfs2_file_llseek, .read = do_sync_read, .write = do_sync_write, .mmap = ocfs2_mmap, @@ -2654,7 +2767,7 @@ const struct file_operations ocfs2_dops = { * the cluster. */ const struct file_operations ocfs2_fops_no_plocks = { - .llseek = generic_file_llseek, + .llseek = ocfs2_file_llseek, .read = do_sync_read, .write = do_sync_write, .mmap = ocfs2_mmap, diff --git a/fs/ocfs2/file.h b/fs/ocfs2/file.h index f5afbbe..97bf761 100644 --- a/fs/ocfs2/file.h +++ b/fs/ocfs2/file.h @@ -61,7 +61,7 @@ int ocfs2_zero_extend(struct inode *inode, struct buffer_head *di_bh, int ocfs2_setattr(struct dentry *dentry, struct iattr *attr); int ocfs2_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); -int ocfs2_permission(struct inode *inode, int mask, unsigned int flags); +int ocfs2_permission(struct inode *inode, int mask); int ocfs2_should_update_atime(struct inode *inode, struct vfsmount *vfsmnt); diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c index b4c8bb6..17454a9 100644 --- a/fs/ocfs2/inode.c +++ b/fs/ocfs2/inode.c @@ -291,7 +291,7 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe, (unsigned long long)OCFS2_I(inode)->ip_blkno, (unsigned long long)le64_to_cpu(fe->i_blkno)); - inode->i_nlink = ocfs2_read_links_count(fe); + set_nlink(inode, ocfs2_read_links_count(fe)); trace_ocfs2_populate_inode(OCFS2_I(inode)->ip_blkno, le32_to_cpu(fe->i_flags)); @@ -951,7 +951,7 @@ static void ocfs2_cleanup_delete_inode(struct inode *inode, trace_ocfs2_cleanup_delete_inode( (unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data); if (sync_data) - write_inode_now(inode, 1); + filemap_write_and_wait(inode->i_mapping); truncate_inode_pages(&inode->i_data, 0); } @@ -1290,7 +1290,7 @@ void ocfs2_refresh_inode(struct inode *inode, OCFS2_I(inode)->ip_dyn_features = le16_to_cpu(fe->i_dyn_features); ocfs2_set_inode_flags(inode); i_size_write(inode, le64_to_cpu(fe->i_size)); - inode->i_nlink = ocfs2_read_links_count(fe); + set_nlink(inode, ocfs2_read_links_count(fe)); inode->i_uid = le32_to_cpu(fe->i_uid); inode->i_gid = le32_to_cpu(fe->i_gid); inode->i_mode = le16_to_cpu(fe->i_mode); diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h index 1c508b1..88924a3 100644 --- a/fs/ocfs2/inode.h +++ b/fs/ocfs2/inode.h @@ -43,6 +43,9 @@ struct ocfs2_inode_info /* protects extended attribute changes on this inode */ struct rw_semaphore ip_xattr_sem; + /* Number of outstanding AIO's which are not page aligned */ + atomic_t ip_unaligned_aio; + /* These fields are protected by ip_lock */ spinlock_t ip_lock; u32 ip_open_count; diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c index bc91072..726ff26 100644 --- a/fs/ocfs2/ioctl.c +++ b/fs/ocfs2/ioctl.c @@ -122,7 +122,7 @@ static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, if ((oldflags & OCFS2_IMMUTABLE_FL) || ((flags ^ oldflags) & (OCFS2_APPEND_FL | OCFS2_IMMUTABLE_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) - goto bail_unlock; + goto bail_commit; } ocfs2_inode->ip_attr = flags; @@ -132,6 +132,7 @@ static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, if (status < 0) mlog_errno(status); +bail_commit: ocfs2_commit_trans(osb, handle); bail_unlock: ocfs2_inode_unlock(inode, 1); @@ -381,7 +382,7 @@ int ocfs2_info_handle_freeinode(struct inode *inode, if (!oifi) { status = -ENOMEM; mlog_errno(status); - goto bail; + goto out_err; } if (o2info_from_user(*oifi, req)) @@ -431,7 +432,7 @@ bail: o2info_set_request_error(&oifi->ifi_req, req); kfree(oifi); - +out_err: return status; } @@ -666,7 +667,7 @@ int ocfs2_info_handle_freefrag(struct inode *inode, if (!oiff) { status = -ENOMEM; mlog_errno(status); - goto bail; + goto out_err; } if (o2info_from_user(*oiff, req)) @@ -716,7 +717,7 @@ bail: o2info_set_request_error(&oiff->iff_req, req); kfree(oiff); - +out_err: return status; } diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 295d564..0a42ae9 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -1544,9 +1544,9 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb, /* we need to run complete recovery for offline orphan slots */ ocfs2_replay_map_set_state(osb, REPLAY_NEEDED); - mlog(ML_NOTICE, "Recovering node %d from slot %d on device (%u,%u)\n", - node_num, slot_num, - MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev)); + printk(KERN_NOTICE "ocfs2: Begin replay journal (node %d, slot %d) on "\ + "device (%u,%u)\n", node_num, slot_num, MAJOR(osb->sb->s_dev), + MINOR(osb->sb->s_dev)); OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters); @@ -1601,6 +1601,9 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb, jbd2_journal_destroy(journal); + printk(KERN_NOTICE "ocfs2: End replay journal (node %d, slot %d) on "\ + "device (%u,%u)\n", node_num, slot_num, MAJOR(osb->sb->s_dev), + MINOR(osb->sb->s_dev)); done: /* drop the lock on this nodes journal */ if (got_lock) @@ -1808,6 +1811,20 @@ static inline unsigned long ocfs2_orphan_scan_timeout(void) * every slot, queuing a recovery of the slot on the ocfs2_wq thread. This * is done to catch any orphans that are left over in orphan directories. * + * It scans all slots, even ones that are in use. It does so to handle the + * case described below: + * + * Node 1 has an inode it was using. The dentry went away due to memory + * pressure. Node 1 closes the inode, but it's on the free list. The node + * has the open lock. + * Node 2 unlinks the inode. It grabs the dentry lock to notify others, + * but node 1 has no dentry and doesn't get the message. It trylocks the + * open lock, sees that another node has a PR, and does nothing. + * Later node 2 runs its orphan dir. It igets the inode, trylocks the + * open lock, sees the PR still, and does nothing. + * Basically, we have to trigger an orphan iput on node 1. The only way + * for this to happen is if node 1 runs node 2's orphan dir. + * * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT * seconds. It gets an EX lock on os_lockres and checks sequence number * stored in LVB. If the sequence number has changed, it means some other diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h index 68cf2f6..a3385b6 100644 --- a/fs/ocfs2/journal.h +++ b/fs/ocfs2/journal.h @@ -441,10 +441,11 @@ static inline int ocfs2_mknod_credits(struct super_block *sb, int is_dir, #define OCFS2_SIMPLE_DIR_EXTEND_CREDITS (2) /* file update (nlink, etc) + directory mtime/ctime + dir entry block + quota - * update on dir + index leaf + dx root update for free list */ + * update on dir + index leaf + dx root update for free list + + * previous dirblock update in the free list */ static inline int ocfs2_link_credits(struct super_block *sb) { - return 2*OCFS2_INODE_UPDATE_CREDITS + 3 + + return 2*OCFS2_INODE_UPDATE_CREDITS + 4 + ocfs2_quota_trans_credits(sb); } diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c index 3e9393c..9cd4108 100644 --- a/fs/ocfs2/mmap.c +++ b/fs/ocfs2/mmap.c @@ -61,7 +61,7 @@ static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf) static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, struct page *page) { - int ret; + int ret = VM_FAULT_NOPAGE; struct inode *inode = file->f_path.dentry->d_inode; struct address_space *mapping = inode->i_mapping; loff_t pos = page_offset(page); @@ -71,32 +71,25 @@ static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, void *fsdata; loff_t size = i_size_read(inode); - /* - * Another node might have truncated while we were waiting on - * cluster locks. - * We don't check size == 0 before the shift. This is borrowed - * from do_generic_file_read. - */ last_index = (size - 1) >> PAGE_CACHE_SHIFT; - if (unlikely(!size || page->index > last_index)) { - ret = -EINVAL; - goto out; - } /* - * The i_size check above doesn't catch the case where nodes - * truncated and then re-extended the file. We'll re-check the - * page mapping after taking the page lock inside of - * ocfs2_write_begin_nolock(). + * There are cases that lead to the page no longer bebongs to the + * mapping. + * 1) pagecache truncates locally due to memory pressure. + * 2) pagecache truncates when another is taking EX lock against + * inode lock. see ocfs2_data_convert_worker. + * + * The i_size check doesn't catch the case where nodes truncated and + * then re-extended the file. We'll re-check the page mapping after + * taking the page lock inside of ocfs2_write_begin_nolock(). + * + * Let VM retry with these cases. */ - if (!PageUptodate(page) || page->mapping != inode->i_mapping) { - /* - * the page has been umapped in ocfs2_data_downconvert_worker. - * So return 0 here and let VFS retry. - */ - ret = 0; + if ((page->mapping != inode->i_mapping) || + (!PageUptodate(page)) || + (page_offset(page) >= size)) goto out; - } /* * Call ocfs2_write_begin() and ocfs2_write_end() to take @@ -116,17 +109,21 @@ static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, if (ret) { if (ret != -ENOSPC) mlog_errno(ret); + if (ret == -ENOMEM) + ret = VM_FAULT_OOM; + else + ret = VM_FAULT_SIGBUS; goto out; } - ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page, - fsdata); - if (ret < 0) { - mlog_errno(ret); + if (!locked_page) { + ret = VM_FAULT_NOPAGE; goto out; } + ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page, + fsdata); BUG_ON(ret != len); - ret = 0; + ret = VM_FAULT_LOCKED; out: return ret; } @@ -168,8 +165,6 @@ static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) out: ocfs2_unblock_signals(&oldset); - if (ret) - ret = VM_FAULT_SIGBUS; return ret; } diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c index cd94270..184c76b 100644 --- a/fs/ocfs2/move_extents.c +++ b/fs/ocfs2/move_extents.c @@ -36,7 +36,6 @@ #include "dir.h" #include "buffer_head_io.h" #include "sysfile.h" -#include "suballoc.h" #include "refcounttree.h" #include "move_extents.h" @@ -746,7 +745,7 @@ static int ocfs2_move_extent(struct ocfs2_move_extents_context *context, */ ocfs2_probe_alloc_group(inode, gd_bh, &goal_bit, len, move_max_hop, new_phys_cpos); - if (!new_phys_cpos) { + if (!*new_phys_cpos) { ret = -ENOSPC; goto out_commit; } diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index e5d738c..a8b2bfe 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -199,9 +199,7 @@ static struct inode *ocfs2_get_init_inode(struct inode *dir, int mode) * these are used by the support functions here and in * callers. */ if (S_ISDIR(mode)) - inode->i_nlink = 2; - else - inode->i_nlink = 1; + set_nlink(inode, 2); inode_init_owner(inode, dir, mode); dquot_initialize(inode); return inode; @@ -1379,7 +1377,7 @@ static int ocfs2_rename(struct inode *old_dir, } if (new_inode) { - new_inode->i_nlink--; + drop_nlink(new_inode); new_inode->i_ctime = CURRENT_TIME; } old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME; @@ -1387,9 +1385,9 @@ static int ocfs2_rename(struct inode *old_dir, if (update_dot_dot) { status = ocfs2_update_entry(old_inode, handle, &old_inode_dot_dot_res, new_dir); - old_dir->i_nlink--; + drop_nlink(old_dir); if (new_inode) { - new_inode->i_nlink--; + drop_nlink(new_inode); } else { inc_nlink(new_dir); mark_inode_dirty(new_dir); @@ -2018,7 +2016,7 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb, orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data; if (S_ISDIR(inode->i_mode)) ocfs2_add_links_count(orphan_fe, 1); - orphan_dir_inode->i_nlink = ocfs2_read_links_count(orphan_fe); + set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe)); ocfs2_journal_dirty(handle, orphan_dir_bh); status = __ocfs2_add_entry(handle, orphan_dir_inode, name, @@ -2116,7 +2114,7 @@ int ocfs2_orphan_del(struct ocfs2_super *osb, orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data; if (S_ISDIR(inode->i_mode)) ocfs2_add_links_count(orphan_fe, -1); - orphan_dir_inode->i_nlink = ocfs2_read_links_count(orphan_fe); + set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe)); ocfs2_journal_dirty(handle, orphan_dir_bh); leave: @@ -2282,7 +2280,7 @@ int ocfs2_create_inode_in_orphan(struct inode *dir, goto leave; } - inode->i_nlink = 0; + clear_nlink(inode); /* do the real work now. */ status = __ocfs2_mknod_locked(dir, inode, 0, &new_di_bh, parent_di_bh, handle, @@ -2437,7 +2435,7 @@ int ocfs2_mv_orphaned_inode_to_new(struct inode *dir, di = (struct ocfs2_dinode *)di_bh->b_data; le32_add_cpu(&di->i_flags, -OCFS2_ORPHANED_FL); di->i_orphaned_slot = 0; - inode->i_nlink = 1; + set_nlink(inode, 1); ocfs2_set_links_count(di, inode->i_nlink); ocfs2_journal_dirty(handle, di_bh); @@ -2498,4 +2496,5 @@ const struct inode_operations ocfs2_dir_iops = { .listxattr = ocfs2_listxattr, .removexattr = generic_removexattr, .fiemap = ocfs2_fiemap, + .get_acl = ocfs2_iop_get_acl, }; diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 4092858..d355e6e 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -836,18 +836,65 @@ static inline unsigned int ocfs2_clusters_to_megabytes(struct super_block *sb, static inline void _ocfs2_set_bit(unsigned int bit, unsigned long *bitmap) { - __test_and_set_bit_le(bit, bitmap); + __set_bit_le(bit, bitmap); } #define ocfs2_set_bit(bit, addr) _ocfs2_set_bit((bit), (unsigned long *)(addr)) static inline void _ocfs2_clear_bit(unsigned int bit, unsigned long *bitmap) { - __test_and_clear_bit_le(bit, bitmap); + __clear_bit_le(bit, bitmap); } #define ocfs2_clear_bit(bit, addr) _ocfs2_clear_bit((bit), (unsigned long *)(addr)) #define ocfs2_test_bit test_bit_le #define ocfs2_find_next_zero_bit find_next_zero_bit_le #define ocfs2_find_next_bit find_next_bit_le + +static inline void *correct_addr_and_bit_unaligned(int *bit, void *addr) +{ +#if BITS_PER_LONG == 64 + *bit += ((unsigned long) addr & 7UL) << 3; + addr = (void *) ((unsigned long) addr & ~7UL); +#elif BITS_PER_LONG == 32 + *bit += ((unsigned long) addr & 3UL) << 3; + addr = (void *) ((unsigned long) addr & ~3UL); +#else +#error "how many bits you are?!" +#endif + return addr; +} + +static inline void ocfs2_set_bit_unaligned(int bit, void *bitmap) +{ + bitmap = correct_addr_and_bit_unaligned(&bit, bitmap); + ocfs2_set_bit(bit, bitmap); +} + +static inline void ocfs2_clear_bit_unaligned(int bit, void *bitmap) +{ + bitmap = correct_addr_and_bit_unaligned(&bit, bitmap); + ocfs2_clear_bit(bit, bitmap); +} + +static inline int ocfs2_test_bit_unaligned(int bit, void *bitmap) +{ + bitmap = correct_addr_and_bit_unaligned(&bit, bitmap); + return ocfs2_test_bit(bit, bitmap); +} + +static inline int ocfs2_find_next_zero_bit_unaligned(void *bitmap, int max, + int start) +{ + int fix = 0, ret, tmpmax; + bitmap = correct_addr_and_bit_unaligned(&fix, bitmap); + tmpmax = max + fix; + start += fix; + + ret = ocfs2_find_next_zero_bit(bitmap, tmpmax, start) - fix; + if (ret > max) + return max; + return ret; +} + #endif /* OCFS2_H */ diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index 92fcd57..a40e5ce 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -712,6 +712,12 @@ static int ocfs2_release_dquot(struct dquot *dquot) */ if (status < 0) mlog_errno(status); + /* + * Clear dq_off so that we search for the structure in quota file next + * time we acquire it. The structure might be deleted and reallocated + * elsewhere by another node while our dquot structure is on freelist. + */ + dquot->dq_off = 0; clear_bit(DQ_ACTIVE_B, &dquot->dq_flags); out_trans: ocfs2_commit_trans(osb, handle); @@ -750,16 +756,17 @@ static int ocfs2_acquire_dquot(struct dquot *dquot) status = ocfs2_lock_global_qf(info, 1); if (status < 0) goto out; - if (!test_bit(DQ_READ_B, &dquot->dq_flags)) { - status = ocfs2_qinfo_lock(info, 0); - if (status < 0) - goto out_dq; - status = qtree_read_dquot(&info->dqi_gi, dquot); - ocfs2_qinfo_unlock(info, 0); - if (status < 0) - goto out_dq; - } - set_bit(DQ_READ_B, &dquot->dq_flags); + status = ocfs2_qinfo_lock(info, 0); + if (status < 0) + goto out_dq; + /* + * We always want to read dquot structure from disk because we don't + * know what happened with it while it was on freelist. + */ + status = qtree_read_dquot(&info->dqi_gi, dquot); + ocfs2_qinfo_unlock(info, 0); + if (status < 0) + goto out_dq; OCFS2_DQUOT(dquot)->dq_use_count++; OCFS2_DQUOT(dquot)->dq_origspace = dquot->dq_dqb.dqb_curspace; diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index dc8007f..b6cfcf2 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -404,7 +404,9 @@ struct ocfs2_quota_recovery *ocfs2_begin_quota_recovery( int status = 0; struct ocfs2_quota_recovery *rec; - mlog(ML_NOTICE, "Beginning quota recovery in slot %u\n", slot_num); + printk(KERN_NOTICE "ocfs2: Beginning quota recovery on device (%s) for " + "slot %u\n", osb->dev_str, slot_num); + rec = ocfs2_alloc_quota_recovery(); if (!rec) return ERR_PTR(-ENOMEM); @@ -549,8 +551,8 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode, goto out_commit; } lock_buffer(qbh); - WARN_ON(!ocfs2_test_bit(bit, dchunk->dqc_bitmap)); - ocfs2_clear_bit(bit, dchunk->dqc_bitmap); + WARN_ON(!ocfs2_test_bit_unaligned(bit, dchunk->dqc_bitmap)); + ocfs2_clear_bit_unaligned(bit, dchunk->dqc_bitmap); le32_add_cpu(&dchunk->dqc_free, 1); unlock_buffer(qbh); ocfs2_journal_dirty(handle, qbh); @@ -596,7 +598,9 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb, struct inode *lqinode; unsigned int flags; - mlog(ML_NOTICE, "Finishing quota recovery in slot %u\n", slot_num); + printk(KERN_NOTICE "ocfs2: Finishing quota recovery on device (%s) for " + "slot %u\n", osb->dev_str, slot_num); + mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); for (type = 0; type < MAXQUOTAS; type++) { if (list_empty(&(rec->r_list[type]))) @@ -612,8 +616,9 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb, /* Someone else is holding the lock? Then he must be * doing the recovery. Just skip the file... */ if (status == -EAGAIN) { - mlog(ML_NOTICE, "skipping quota recovery for slot %d " - "because quota file is locked.\n", slot_num); + printk(KERN_NOTICE "ocfs2: Skipping quota recovery on " + "device (%s) for slot %d because quota file is " + "locked.\n", osb->dev_str, slot_num); status = 0; goto out_put; } else if (status < 0) { @@ -944,7 +949,7 @@ static struct ocfs2_quota_chunk *ocfs2_find_free_entry(struct super_block *sb, * ol_quota_entries_per_block(sb); } - found = ocfs2_find_next_zero_bit(dchunk->dqc_bitmap, len, 0); + found = ocfs2_find_next_zero_bit_unaligned(dchunk->dqc_bitmap, len, 0); /* We failed? */ if (found == len) { mlog(ML_ERROR, "Did not find empty entry in chunk %d with %u" @@ -1208,7 +1213,7 @@ static void olq_alloc_dquot(struct buffer_head *bh, void *private) struct ocfs2_local_disk_chunk *dchunk; dchunk = (struct ocfs2_local_disk_chunk *)bh->b_data; - ocfs2_set_bit(*offset, dchunk->dqc_bitmap); + ocfs2_set_bit_unaligned(*offset, dchunk->dqc_bitmap); le32_add_cpu(&dchunk->dqc_free, -1); } @@ -1289,16 +1294,12 @@ int ocfs2_local_release_dquot(handle_t *handle, struct dquot *dquot) (od->dq_chunk->qc_headerbh->b_data); /* Mark structure as freed */ lock_buffer(od->dq_chunk->qc_headerbh); - ocfs2_clear_bit(offset, dchunk->dqc_bitmap); + ocfs2_clear_bit_unaligned(offset, dchunk->dqc_bitmap); le32_add_cpu(&dchunk->dqc_free, 1); unlock_buffer(od->dq_chunk->qc_headerbh); ocfs2_journal_dirty(handle, od->dq_chunk->qc_headerbh); out: - /* Clear the read bit so that next time someone uses this - * dquot he reads fresh info from disk and allocates local - * dquot structure */ - clear_bit(DQ_READ_B, &dquot->dq_flags); return status; } diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c index 15d29cc..9f32d7c 100644 --- a/fs/ocfs2/refcounttree.c +++ b/fs/ocfs2/refcounttree.c @@ -4368,25 +4368,6 @@ static inline int ocfs2_may_create(struct inode *dir, struct dentry *child) return inode_permission(dir, MAY_WRITE | MAY_EXEC); } -/* copied from user_path_parent. */ -static int ocfs2_user_path_parent(const char __user *path, - struct nameidata *nd, char **name) -{ - char *s = getname(path); - int error; - - if (IS_ERR(s)) - return PTR_ERR(s); - - error = kern_path_parent(s, nd); - if (error) - putname(s); - else - *name = s; - - return error; -} - /** * ocfs2_vfs_reflink - Create a reference-counted link * @@ -4460,10 +4441,8 @@ int ocfs2_reflink_ioctl(struct inode *inode, bool preserve) { struct dentry *new_dentry; - struct nameidata nd; - struct path old_path; + struct path old_path, new_path; int error; - char *to = NULL; if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) return -EOPNOTSUPP; @@ -4474,39 +4453,33 @@ int ocfs2_reflink_ioctl(struct inode *inode, return error; } - error = ocfs2_user_path_parent(newname, &nd, &to); - if (error) { + new_dentry = user_path_create(AT_FDCWD, newname, &new_path, 0); + error = PTR_ERR(new_dentry); + if (IS_ERR(new_dentry)) { mlog_errno(error); goto out; } error = -EXDEV; - if (old_path.mnt != nd.path.mnt) - goto out_release; - new_dentry = lookup_create(&nd, 0); - error = PTR_ERR(new_dentry); - if (IS_ERR(new_dentry)) { + if (old_path.mnt != new_path.mnt) { mlog_errno(error); - goto out_unlock; + goto out_dput; } - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(new_path.mnt); if (error) { mlog_errno(error); goto out_dput; } error = ocfs2_vfs_reflink(old_path.dentry, - nd.path.dentry->d_inode, + new_path.dentry->d_inode, new_dentry, preserve); - mnt_drop_write(nd.path.mnt); + mnt_drop_write(new_path.mnt); out_dput: dput(new_dentry); -out_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); -out_release: - path_put(&nd.path); - putname(to); + mutex_unlock(&new_path.dentry->d_inode->i_mutex); + path_put(&new_path); out: path_put(&old_path); diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 26fc001..1424c15 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -493,8 +493,8 @@ int ocfs2_find_slot(struct ocfs2_super *osb) goto bail; } } else - mlog(ML_NOTICE, "slot %d is already allocated to this node!\n", - slot); + printk(KERN_INFO "ocfs2: Slot %d on device (%s) was already " + "allocated to this node!\n", slot, osb->dev_str); ocfs2_set_slot(si, slot, osb->node_num); osb->slot_num = slot; diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c index 19965b0..9436801 100644 --- a/fs/ocfs2/stack_o2cb.c +++ b/fs/ocfs2/stack_o2cb.c @@ -28,6 +28,7 @@ #include "cluster/masklog.h" #include "cluster/nodemanager.h" #include "cluster/heartbeat.h" +#include "cluster/tcp.h" #include "stackglue.h" @@ -256,6 +257,61 @@ static void o2cb_dump_lksb(struct ocfs2_dlm_lksb *lksb) } /* + * Check if this node is heartbeating and is connected to all other + * heartbeating nodes. + */ +static int o2cb_cluster_check(void) +{ + u8 node_num; + int i; + unsigned long hbmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long netmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + + node_num = o2nm_this_node(); + if (node_num == O2NM_MAX_NODES) { + printk(KERN_ERR "o2cb: This node has not been configured.\n"); + return -EINVAL; + } + + /* + * o2dlm expects o2net sockets to be created. If not, then + * dlm_join_domain() fails with a stack of errors which are both cryptic + * and incomplete. The idea here is to detect upfront whether we have + * managed to connect to all nodes or not. If not, then list the nodes + * to allow the user to check the configuration (incorrect IP, firewall, + * etc.) Yes, this is racy. But its not the end of the world. + */ +#define O2CB_MAP_STABILIZE_COUNT 60 + for (i = 0; i < O2CB_MAP_STABILIZE_COUNT; ++i) { + o2hb_fill_node_map(hbmap, sizeof(hbmap)); + if (!test_bit(node_num, hbmap)) { + printk(KERN_ERR "o2cb: %s heartbeat has not been " + "started.\n", (o2hb_global_heartbeat_active() ? + "Global" : "Local")); + return -EINVAL; + } + o2net_fill_node_map(netmap, sizeof(netmap)); + /* Force set the current node to allow easy compare */ + set_bit(node_num, netmap); + if (!memcmp(hbmap, netmap, sizeof(hbmap))) + return 0; + if (i < O2CB_MAP_STABILIZE_COUNT) + msleep(1000); + } + + printk(KERN_ERR "o2cb: This node could not connect to nodes:"); + i = -1; + while ((i = find_next_bit(hbmap, O2NM_MAX_NODES, + i + 1)) < O2NM_MAX_NODES) { + if (!test_bit(i, netmap)) + printk(" %u", i); + } + printk(".\n"); + + return -ENOTCONN; +} + +/* * Called from the dlm when it's about to evict a node. This is how the * classic stack signals node death. */ @@ -263,8 +319,8 @@ static void o2dlm_eviction_cb(int node_num, void *data) { struct ocfs2_cluster_connection *conn = data; - mlog(ML_NOTICE, "o2dlm has evicted node %d from group %.*s\n", - node_num, conn->cc_namelen, conn->cc_name); + printk(KERN_NOTICE "o2cb: o2dlm has evicted node %d from domain %.*s\n", + node_num, conn->cc_namelen, conn->cc_name); conn->cc_recovery_handler(node_num, conn->cc_recovery_data); } @@ -280,12 +336,11 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) BUG_ON(conn == NULL); BUG_ON(conn->cc_proto == NULL); - /* for now we only have one cluster/node, make sure we see it - * in the heartbeat universe */ - if (!o2hb_check_local_node_heartbeating()) { - if (o2hb_global_heartbeat_active()) - mlog(ML_ERROR, "Global heartbeat not started\n"); - rc = -EINVAL; + /* Ensure cluster stack is up and all nodes are connected */ + rc = o2cb_cluster_check(); + if (rc) { + printk(KERN_ERR "o2cb: Cluster check failed. Fix errors " + "before retrying.\n"); goto out; } diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 56f6102..5fe6b1e 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -54,6 +54,7 @@ #include "ocfs1_fs_compat.h" #include "alloc.h" +#include "aops.h" #include "blockcheck.h" #include "dlmglue.h" #include "export.h" @@ -1107,9 +1108,9 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) ocfs2_set_ro_flag(osb, 1); - printk(KERN_NOTICE "Readonly device detected. No cluster " - "services will be utilized for this mount. Recovery " - "will be skipped.\n"); + printk(KERN_NOTICE "ocfs2: Readonly device (%s) detected. " + "Cluster services will not be used for this mount. " + "Recovery will be skipped.\n", osb->dev_str); } if (!ocfs2_is_hard_readonly(osb)) { @@ -1582,8 +1583,8 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt) seq_printf(s, ",localflocks,"); if (osb->osb_cluster_stack[0]) - seq_printf(s, ",cluster_stack=%.*s", OCFS2_STACK_LABEL_LEN, - osb->osb_cluster_stack); + seq_show_option_n(s, "cluster_stack", osb->osb_cluster_stack, + OCFS2_STACK_LABEL_LEN); if (opts & OCFS2_MOUNT_USRQUOTA) seq_printf(s, ",usrquota"); if (opts & OCFS2_MOUNT_GRPQUOTA) @@ -1616,12 +1617,17 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt) return 0; } +wait_queue_head_t ocfs2__ioend_wq[OCFS2_IOEND_WQ_HASH_SZ]; + static int __init ocfs2_init(void) { - int status; + int status, i; ocfs2_print_version(); + for (i = 0; i < OCFS2_IOEND_WQ_HASH_SZ; i++) + init_waitqueue_head(&ocfs2__ioend_wq[i]); + status = init_ocfs2_uptodate_cache(); if (status < 0) { mlog_errno(status); @@ -1760,7 +1766,7 @@ static void ocfs2_inode_init_once(void *data) ocfs2_extent_map_init(&oi->vfs_inode); INIT_LIST_HEAD(&oi->ip_io_markers); oi->ip_dir_start_lookup = 0; - + atomic_set(&oi->ip_unaligned_aio, 0); init_rwsem(&oi->ip_alloc_sem); init_rwsem(&oi->ip_xattr_sem); mutex_init(&oi->ip_io_mutex); @@ -1974,7 +1980,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) * If we failed before we got a uuid_str yet, we can't stop * heartbeat. Otherwise, do it. */ - if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str) + if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str && + !ocfs2_is_hard_readonly(osb)) hangup_needed = 1; if (osb->cconn) @@ -2353,7 +2360,7 @@ static int ocfs2_initialize_super(struct super_block *sb, mlog_errno(status); goto bail; } - cleancache_init_shared_fs((char *)&uuid_net_key, sb); + cleancache_init_shared_fs((char *)&di->id2.i_super.s_uuid, sb); bail: return status; @@ -2462,8 +2469,8 @@ static int ocfs2_check_volume(struct ocfs2_super *osb) goto finally; } } else { - mlog(ML_NOTICE, "File system was not unmounted cleanly, " - "recovering volume.\n"); + printk(KERN_NOTICE "ocfs2: File system on device (%s) was not " + "unmounted cleanly, recovering it.\n", osb->dev_str); } local = ocfs2_mount_local(osb); diff --git a/fs/ocfs2/super.h b/fs/ocfs2/super.h index 40c7de0..74ff74c 100644 --- a/fs/ocfs2/super.h +++ b/fs/ocfs2/super.h @@ -31,17 +31,15 @@ extern struct workqueue_struct *ocfs2_wq; int ocfs2_publish_get_mount_state(struct ocfs2_super *osb, int node_num); -void __ocfs2_error(struct super_block *sb, - const char *function, - const char *fmt, ...) - __attribute__ ((format (printf, 3, 4))); +__printf(3, 4) +void __ocfs2_error(struct super_block *sb, const char *function, + const char *fmt, ...); #define ocfs2_error(sb, fmt, args...) __ocfs2_error(sb, __PRETTY_FUNCTION__, fmt, ##args) -void __ocfs2_abort(struct super_block *sb, - const char *function, - const char *fmt, ...) - __attribute__ ((format (printf, 3, 4))); +__printf(3, 4) +void __ocfs2_abort(struct super_block *sb, const char *function, + const char *fmt, ...); #define ocfs2_abort(sb, fmt, args...) __ocfs2_abort(sb, __PRETTY_FUNCTION__, fmt, ##args) diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index 61a84cf..bef187b 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -2376,16 +2376,18 @@ static int ocfs2_remove_value_outside(struct inode*inode, } ret = ocfs2_xattr_value_truncate(inode, vb, 0, &ctxt); - if (ret < 0) { - mlog_errno(ret); - break; - } ocfs2_commit_trans(osb, ctxt.handle); if (ctxt.meta_ac) { ocfs2_free_alloc_context(ctxt.meta_ac); ctxt.meta_ac = NULL; } + + if (ret < 0) { + mlog_errno(ret); + break; + } + } if (ctxt.meta_ac) @@ -7195,20 +7197,9 @@ int ocfs2_init_security_and_acl(struct inode *dir, { int ret = 0; struct buffer_head *dir_bh = NULL; - struct ocfs2_security_xattr_info si = { - .enable = 1, - }; - ret = ocfs2_init_security_get(inode, dir, qstr, &si); - if (!ret) { - ret = ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY, - si.name, si.value, si.value_len, - XATTR_CREATE); - if (ret) { - mlog_errno(ret); - goto leave; - } - } else if (ret != -EOPNOTSUPP) { + ret = ocfs2_init_security_get(inode, dir, qstr, NULL); + if (ret) { mlog_errno(ret); goto leave; } @@ -7265,6 +7256,22 @@ static int ocfs2_xattr_security_set(struct dentry *dentry, const char *name, name, value, size, flags); } +int ocfs2_initxattrs(struct inode *inode, const struct xattr *xattr_array, + void *fs_info) +{ + const struct xattr *xattr; + int err = 0; + + for (xattr = xattr_array; xattr->name != NULL; xattr++) { + err = ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY, + xattr->name, xattr->value, + xattr->value_len, XATTR_CREATE); + if (err) + break; + } + return err; +} + int ocfs2_init_security_get(struct inode *inode, struct inode *dir, const struct qstr *qstr, @@ -7273,8 +7280,13 @@ int ocfs2_init_security_get(struct inode *inode, /* check whether ocfs2 support feature xattr */ if (!ocfs2_supports_xattr(OCFS2_SB(dir->i_sb))) return -EOPNOTSUPP; - return security_inode_init_security(inode, dir, qstr, &si->name, - &si->value, &si->value_len); + if (si) + return security_old_inode_init_security(inode, dir, qstr, + &si->name, &si->value, + &si->value_len); + + return security_inode_init_security(inode, dir, qstr, + &ocfs2_initxattrs, NULL); } int ocfs2_init_security_set(handle_t *handle, diff --git a/fs/reiserfs/bitmap.c b/fs/reiserfs/bitmap.c index 483442e..d1aca1d 100644 --- a/fs/reiserfs/bitmap.c +++ b/fs/reiserfs/bitmap.c @@ -214,7 +214,7 @@ static int scan_bitmap_block(struct reiserfs_transaction_handle *th, } /* otherwise we clear all bit were set ... */ while (--i >= *beg) - reiserfs_test_and_clear_le_bit + reiserfs_clear_le_bit (i, bh->b_data); reiserfs_restore_prepared_buffer(s, bh); *beg = org; @@ -1222,15 +1222,11 @@ void reiserfs_cache_bitmap_metadata(struct super_block *sb, info->free_count = 0; while (--cur >= (unsigned long *)bh->b_data) { - int i; - /* 0 and ~0 are special, we can optimize for them */ if (*cur == 0) info->free_count += BITS_PER_LONG; else if (*cur != ~0L) /* A mix, investigate */ - for (i = BITS_PER_LONG - 1; i >= 0; i--) - if (!reiserfs_test_le_bit(i, cur)) - info->free_count++; + info->free_count += BITS_PER_LONG - hweight_long(*cur); } } diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c index 198dabf..8048eea 100644 --- a/fs/reiserfs/dir.c +++ b/fs/reiserfs/dir.c @@ -14,7 +14,8 @@ extern const struct reiserfs_key MIN_KEY; static int reiserfs_readdir(struct file *, void *, filldir_t); -static int reiserfs_dir_fsync(struct file *filp, int datasync); +static int reiserfs_dir_fsync(struct file *filp, loff_t start, loff_t end, + int datasync); const struct file_operations reiserfs_dir_operations = { .llseek = generic_file_llseek, @@ -27,13 +28,21 @@ const struct file_operations reiserfs_dir_operations = { #endif }; -static int reiserfs_dir_fsync(struct file *filp, int datasync) +static int reiserfs_dir_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *inode = filp->f_mapping->host; int err; + + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + mutex_lock(&inode->i_mutex); reiserfs_write_lock(inode->i_sb); err = reiserfs_commit_for_inode(inode); reiserfs_write_unlock(inode->i_sb); + mutex_unlock(&inode->i_mutex); if (err < 0) return err; return 0; @@ -119,6 +128,7 @@ int reiserfs_readdir_dentry(struct dentry *dentry, void *dirent, char *d_name; off_t d_off; ino_t d_ino; + loff_t cur_pos = deh_offset(deh); if (!de_visible(deh)) /* it is hidden entry */ @@ -191,8 +201,9 @@ int reiserfs_readdir_dentry(struct dentry *dentry, void *dirent, if (local_buf != small_buf) { kfree(local_buf); } - // next entry should be looked for with such offset - next_pos = deh_offset(deh) + 1; + + /* deh_offset(deh) may be invalid now. */ + next_pos = cur_pos + 1; if (item_moved(&tmp_ih, &path_to_entry)) { goto research; diff --git a/fs/reiserfs/file.c b/fs/reiserfs/file.c index 91f080c..f011185 100644 --- a/fs/reiserfs/file.c +++ b/fs/reiserfs/file.c @@ -126,7 +126,7 @@ static int reiserfs_file_open(struct inode *inode, struct file *file) return err; } -static void reiserfs_vfs_truncate_file(struct inode *inode) +void reiserfs_vfs_truncate_file(struct inode *inode) { mutex_lock(&(REISERFS_I(inode)->tailpack)); reiserfs_truncate_file(inode, 1); @@ -140,12 +140,18 @@ static void reiserfs_vfs_truncate_file(struct inode *inode) * be removed... */ -static int reiserfs_sync_file(struct file *filp, int datasync) +static int reiserfs_sync_file(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *inode = filp->f_mapping->host; int err; int barrier_done; + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + mutex_lock(&inode->i_mutex); BUG_ON(!S_ISREG(inode->i_mode)); err = sync_mapping_buffers(inode->i_mapping); reiserfs_write_lock(inode->i_sb); @@ -153,6 +159,7 @@ static int reiserfs_sync_file(struct file *filp, int datasync) reiserfs_write_unlock(inode->i_sb); if (barrier_done != 1 && reiserfs_barrier_flush(inode->i_sb)) blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); + mutex_unlock(&inode->i_mutex); if (barrier_done < 0) return barrier_done; return (err < 0) ? -EIO : 0; @@ -305,11 +312,11 @@ const struct file_operations reiserfs_file_operations = { }; const struct inode_operations reiserfs_file_inode_operations = { - .truncate = reiserfs_vfs_truncate_file, .setattr = reiserfs_setattr, .setxattr = reiserfs_setxattr, .getxattr = reiserfs_getxattr, .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .get_acl = reiserfs_get_acl, }; diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c index ebe8db4..fcb07e5 100644 --- a/fs/reiserfs/inode.c +++ b/fs/reiserfs/inode.c @@ -1154,7 +1154,7 @@ static void init_inode(struct inode *inode, struct treepath *path) set_inode_item_key_version(inode, KEY_FORMAT_3_5); set_inode_sd_version(inode, STAT_DATA_V1); inode->i_mode = sd_v1_mode(sd); - inode->i_nlink = sd_v1_nlink(sd); + set_nlink(inode, sd_v1_nlink(sd)); inode->i_uid = sd_v1_uid(sd); inode->i_gid = sd_v1_gid(sd); inode->i_size = sd_v1_size(sd); @@ -1199,7 +1199,7 @@ static void init_inode(struct inode *inode, struct treepath *path) struct stat_data *sd = (struct stat_data *)B_I_PITEM(bh, ih); inode->i_mode = sd_v2_mode(sd); - inode->i_nlink = sd_v2_nlink(sd); + set_nlink(inode, sd_v2_nlink(sd)); inode->i_uid = sd_v2_uid(sd); inode->i_size = sd_v2_size(sd); inode->i_gid = sd_v2_gid(sd); @@ -1444,7 +1444,7 @@ void reiserfs_read_locked_inode(struct inode *inode, /* a stale NFS handle can trigger this without it being an error */ pathrelse(&path_to_sd); reiserfs_make_bad_inode(inode); - inode->i_nlink = 0; + clear_nlink(inode); return; } @@ -1475,6 +1475,11 @@ void reiserfs_read_locked_inode(struct inode *inode, reiserfs_check_path(&path_to_sd); /* init inode should be relsing */ + /* + * Stat data v1 doesn't support ACLs. + */ + if (get_inode_sd_version(inode) == STAT_DATA_V1) + cache_no_acl(inode); } /** @@ -1832,7 +1837,7 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th, #endif /* fill stat data */ - inode->i_nlink = (S_ISDIR(mode) ? 2 : 1); + set_nlink(inode, (S_ISDIR(mode) ? 2 : 1)); /* uid and gid must already be set by the caller for quota init */ @@ -1989,7 +1994,7 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th, make_bad_inode(inode); out_inserted_sd: - inode->i_nlink = 0; + clear_nlink(inode); th->t_trans_id = 0; /* so the caller can't use this handle later */ unlock_new_inode(inode); /* OK to do even if we hadn't locked it */ iput(inode); @@ -3075,9 +3080,8 @@ static ssize_t reiserfs_direct_IO(int rw, struct kiocb *iocb, struct inode *inode = file->f_mapping->host; ssize_t ret; - ret = blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov, - offset, nr_segs, - reiserfs_get_blocks_direct_io, NULL); + ret = blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs, + reiserfs_get_blocks_direct_io); /* * In case of error extending write may have instantiated a few @@ -3087,8 +3091,10 @@ static ssize_t reiserfs_direct_IO(int rw, struct kiocb *iocb, loff_t isize = i_size_read(inode); loff_t end = offset + iov_length(iov, nr_segs); - if (end > isize) - vmtruncate(inode, isize); + if ((end > isize) && inode_newsize_ok(inode, isize) == 0) { + truncate_setsize(inode, isize); + reiserfs_vfs_truncate_file(inode); + } } return ret; @@ -3120,6 +3126,9 @@ int reiserfs_setattr(struct dentry *dentry, struct iattr *attr) error = -EFBIG; goto out; } + + inode_dio_wait(inode); + /* fill in hole pointers in the expanding truncate case. */ if (attr->ia_size > inode->i_size) { error = generic_cont_expand_simple(inode, attr->ia_size); @@ -3199,8 +3208,19 @@ int reiserfs_setattr(struct dentry *dentry, struct iattr *attr) */ reiserfs_write_unlock_once(inode->i_sb, depth); if ((attr->ia_valid & ATTR_SIZE) && - attr->ia_size != i_size_read(inode)) - error = vmtruncate(inode, attr->ia_size); + attr->ia_size != i_size_read(inode)) { + error = inode_newsize_ok(inode, attr->ia_size); + if (!error) { + /* + * Could race against reiserfs_file_release + * if called from NFS, so take tailpack mutex. + */ + mutex_lock(&REISERFS_I(inode)->tailpack); + truncate_setsize(inode, attr->ia_size); + reiserfs_truncate_file(inode, 1); + mutex_unlock(&REISERFS_I(inode)->tailpack); + } + } if (!error) { setattr_copy(inode, attr); diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c index c5e82ec..938a77f 100644 --- a/fs/reiserfs/journal.c +++ b/fs/reiserfs/journal.c @@ -291,14 +291,13 @@ int reiserfs_allocate_list_bitmaps(struct super_block *sb, for (i = 0; i < JOURNAL_NUM_BITMAPS; i++) { jb = jb_array + i; jb->journal_list = NULL; - jb->bitmaps = vmalloc(mem); + jb->bitmaps = vzalloc(mem); if (!jb->bitmaps) { reiserfs_warning(sb, "clm-2000", "unable to " "allocate bitmaps for journal lists"); failed = 1; break; } - memset(jb->bitmaps, 0, mem); } if (failed) { free_list_bitmaps(sb, jb_array); @@ -353,11 +352,10 @@ static struct reiserfs_journal_cnode *allocate_cnodes(int num_cnodes) if (num_cnodes <= 0) { return NULL; } - head = vmalloc(num_cnodes * sizeof(struct reiserfs_journal_cnode)); + head = vzalloc(num_cnodes * sizeof(struct reiserfs_journal_cnode)); if (!head) { return NULL; } - memset(head, 0, num_cnodes * sizeof(struct reiserfs_journal_cnode)); head[0].prev = NULL; head[0].next = head + 1; for (i = 1; i < num_cnodes; i++) { @@ -678,23 +676,19 @@ struct buffer_chunk { static void write_chunk(struct buffer_chunk *chunk) { int i; - get_fs_excl(); for (i = 0; i < chunk->nr; i++) { submit_logged_buffer(chunk->bh[i]); } chunk->nr = 0; - put_fs_excl(); } static void write_ordered_chunk(struct buffer_chunk *chunk) { int i; - get_fs_excl(); for (i = 0; i < chunk->nr; i++) { submit_ordered_buffer(chunk->bh[i]); } chunk->nr = 0; - put_fs_excl(); } static int add_to_chunk(struct buffer_chunk *chunk, struct buffer_head *bh, @@ -986,8 +980,6 @@ static int flush_commit_list(struct super_block *s, return 0; } - get_fs_excl(); - /* before we can put our commit blocks on disk, we have to make sure everyone older than ** us is on disk too */ @@ -1145,7 +1137,6 @@ static int flush_commit_list(struct super_block *s, if (retval) reiserfs_abort(s, retval, "Journal write error in %s", __func__); - put_fs_excl(); return retval; } @@ -1374,8 +1365,6 @@ static int flush_journal_list(struct super_block *s, return 0; } - get_fs_excl(); - /* if all the work is already done, get out of here */ if (atomic_read(&(jl->j_nonzerolen)) <= 0 && atomic_read(&(jl->j_commit_left)) <= 0) { @@ -1597,7 +1586,6 @@ static int flush_journal_list(struct super_block *s, put_journal_list(s, jl); if (flushall) mutex_unlock(&journal->j_flush_mutex); - put_fs_excl(); return err; } @@ -1928,15 +1916,19 @@ static int do_journal_release(struct reiserfs_transaction_handle *th, } reiserfs_mounted_fs_count--; - /* wait for all commits to finish */ - cancel_delayed_work(&SB_JOURNAL(sb)->j_work); /* * We must release the write lock here because * the workqueue job (flush_async_commit) needs this lock */ reiserfs_write_unlock(sb); - flush_workqueue(commit_wq); + /* + * Cancel flushing of old commits. Note that this work will not + * be requeued because superblock is being shutdown and doesn't + * have MS_ACTIVE set. + */ + /* wait for all commits to finish */ + cancel_delayed_work_sync(&SB_JOURNAL(sb)->j_work); if (!reiserfs_mounted_fs_count) { destroy_workqueue(commit_wq); @@ -2695,14 +2687,13 @@ int journal_init(struct super_block *sb, const char *j_dev_name, * dependency inversion warnings. */ reiserfs_write_unlock(sb); - journal = SB_JOURNAL(sb) = vmalloc(sizeof(struct reiserfs_journal)); + journal = SB_JOURNAL(sb) = vzalloc(sizeof(struct reiserfs_journal)); if (!journal) { reiserfs_warning(sb, "journal-1256", "unable to get memory for journal structure"); reiserfs_write_lock(sb); return 1; } - memset(journal, 0, sizeof(struct reiserfs_journal)); INIT_LIST_HEAD(&journal->j_bitmap_nodes); INIT_LIST_HEAD(&journal->j_prealloc_list); INIT_LIST_HEAD(&journal->j_working_list); @@ -3108,7 +3099,6 @@ static int do_journal_begin_r(struct reiserfs_transaction_handle *th, th->t_trans_id = journal->j_trans_id; unlock_journal(sb); INIT_LIST_HEAD(&th->t_list); - get_fs_excl(); return 0; out_fail: @@ -3964,7 +3954,6 @@ static int do_journal_end(struct reiserfs_transaction_handle *th, flush = flags & FLUSH_ALL; wait_on_commit = flags & WAIT; - put_fs_excl(); current->journal_info = th->t_handle_save; reiserfs_check_lock_depth(sb, "journal end"); if (journal->j_len == 0) { @@ -4226,8 +4215,15 @@ static int do_journal_end(struct reiserfs_transaction_handle *th, if (flush) { flush_commit_list(sb, jl, 1); flush_journal_list(sb, jl, 1); - } else if (!(jl->j_state & LIST_COMMIT_PENDING)) - queue_delayed_work(commit_wq, &journal->j_work, HZ / 10); + } else if (!(jl->j_state & LIST_COMMIT_PENDING)) { + /* + * Avoid queueing work when sb is being shut down. Transaction + * will be flushed on journal shutdown. + */ + if (sb->s_flags & MS_ACTIVE) + queue_delayed_work(commit_wq, + &journal->j_work, HZ / 10); + } /* if the next transaction has any chance of wrapping, flush ** transactions that might get overwritten. If any journal lists are very @@ -4316,4 +4312,3 @@ void reiserfs_abort_journal(struct super_block *sb, int errno) dump_stack(); #endif } - diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c index 1186626..80058e8 100644 --- a/fs/reiserfs/namei.c +++ b/fs/reiserfs/namei.c @@ -19,7 +19,7 @@ #include <linux/reiserfs_xattr.h> #include <linux/quotaops.h> -#define INC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) { inc_nlink(i); if (i->i_nlink >= REISERFS_LINK_MAX) i->i_nlink=1; } +#define INC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) { inc_nlink(i); if (i->i_nlink >= REISERFS_LINK_MAX) set_nlink(i, 1); } #define DEC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) drop_nlink(i); // directory item contains array of entry headers. This performs @@ -622,7 +622,7 @@ static int reiserfs_create(struct inode *dir, struct dentry *dentry, int mode, dentry->d_name.len, inode, 1 /*visible */ ); if (retval) { int err; - inode->i_nlink--; + drop_nlink(inode); reiserfs_update_sd(&th, inode); err = journal_end(&th, dir->i_sb, jbegin_count); if (err) @@ -702,7 +702,7 @@ static int reiserfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dentry->d_name.len, inode, 1 /*visible */ ); if (retval) { int err; - inode->i_nlink--; + drop_nlink(inode); reiserfs_update_sd(&th, inode); err = journal_end(&th, dir->i_sb, jbegin_count); if (err) @@ -787,7 +787,7 @@ static int reiserfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) dentry->d_name.len, inode, 1 /*visible */ ); if (retval) { int err; - inode->i_nlink = 0; + clear_nlink(inode); DEC_DIR_INODE_NLINK(dir); reiserfs_update_sd(&th, inode); err = journal_end(&th, dir->i_sb, jbegin_count); @@ -964,7 +964,7 @@ static int reiserfs_unlink(struct inode *dir, struct dentry *dentry) reiserfs_warning(inode->i_sb, "reiserfs-7042", "deleting nonexistent file (%lu), %d", inode->i_ino, inode->i_nlink); - inode->i_nlink = 1; + set_nlink(inode, 1); } drop_nlink(inode); @@ -1086,7 +1086,7 @@ static int reiserfs_symlink(struct inode *parent_dir, dentry->d_name.len, inode, 1 /*visible */ ); if (retval) { int err; - inode->i_nlink--; + drop_nlink(inode); reiserfs_update_sd(&th, inode); err = journal_end(&th, parent_dir->i_sb, jbegin_count); if (err) @@ -1129,7 +1129,7 @@ static int reiserfs_link(struct dentry *old_dentry, struct inode *dir, retval = journal_begin(&th, dir->i_sb, jbegin_count); if (retval) { - inode->i_nlink--; + drop_nlink(inode); reiserfs_write_unlock(dir->i_sb); return retval; } @@ -1144,7 +1144,7 @@ static int reiserfs_link(struct dentry *old_dentry, struct inode *dir, if (retval) { int err; - inode->i_nlink--; + drop_nlink(inode); err = journal_end(&th, dir->i_sb, jbegin_count); reiserfs_write_unlock(dir->i_sb); return err ? err : retval; @@ -1529,6 +1529,7 @@ const struct inode_operations reiserfs_dir_inode_operations = { .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .get_acl = reiserfs_get_acl, }; /* @@ -1545,6 +1546,7 @@ const struct inode_operations reiserfs_symlink_inode_operations = { .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .get_acl = reiserfs_get_acl, }; @@ -1558,5 +1560,5 @@ const struct inode_operations reiserfs_special_inode_operations = { .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, - + .get_acl = reiserfs_get_acl, }; diff --git a/fs/reiserfs/resize.c b/fs/reiserfs/resize.c index b3a94d2..7483279 100644 --- a/fs/reiserfs/resize.c +++ b/fs/reiserfs/resize.c @@ -111,15 +111,13 @@ int reiserfs_resize(struct super_block *s, unsigned long block_count_new) /* allocate additional bitmap blocks, reallocate array of bitmap * block pointers */ bitmap = - vmalloc(sizeof(struct reiserfs_bitmap_info) * bmap_nr_new); + vzalloc(sizeof(struct reiserfs_bitmap_info) * bmap_nr_new); if (!bitmap) { /* Journal bitmaps are still supersized, but the memory isn't * leaked, so I guess it's ok */ printk("reiserfs_resize: unable to allocate memory.\n"); return -ENOMEM; } - memset(bitmap, 0, - sizeof(struct reiserfs_bitmap_info) * bmap_nr_new); for (i = 0; i < bmap_nr; i++) bitmap[i] = old_bitmap[i]; @@ -136,7 +134,7 @@ int reiserfs_resize(struct super_block *s, unsigned long block_count_new) return -EIO; } memset(bh->b_data, 0, sb_blocksize(sb)); - reiserfs_test_and_set_le_bit(0, bh->b_data); + reiserfs_set_le_bit(0, bh->b_data); reiserfs_cache_bitmap_metadata(s, bh, bitmap + i); set_buffer_uptodate(bh); @@ -172,7 +170,7 @@ int reiserfs_resize(struct super_block *s, unsigned long block_count_new) reiserfs_prepare_for_journal(s, bh, 1); for (i = block_r; i < s->s_blocksize * 8; i++) - reiserfs_test_and_clear_le_bit(i, bh->b_data); + reiserfs_clear_le_bit(i, bh->b_data); info->free_count += s->s_blocksize * 8 - block_r; journal_mark_dirty(&th, s, bh); @@ -190,7 +188,7 @@ int reiserfs_resize(struct super_block *s, unsigned long block_count_new) reiserfs_prepare_for_journal(s, bh, 1); for (i = block_r_new; i < s->s_blocksize * 8; i++) - reiserfs_test_and_set_le_bit(i, bh->b_data); + reiserfs_set_le_bit(i, bh->b_data); journal_mark_dirty(&th, s, bh); brelse(bh); diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 7527623..569498a 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -1663,6 +1663,7 @@ static int reiserfs_fill_super(struct super_block *s, void *data, int silent) /* Set default values for options: non-aggressive tails, RO on errors */ REISERFS_SB(s)->s_mount_opt |= (1 << REISERFS_SMALLTAIL); REISERFS_SB(s)->s_mount_opt |= (1 << REISERFS_ERROR_RO); + REISERFS_SB(s)->s_mount_opt |= (1 << REISERFS_BARRIER_FLUSH); /* no preallocation minimum, be smart in reiserfs_file_write instead */ REISERFS_SB(s)->s_alloc_options.preallocmin = 0; diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 6e3ca4e..04eecc4 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -555,11 +555,10 @@ reiserfs_xattr_set_handle(struct reiserfs_transaction_handle *th, reiserfs_write_unlock(inode->i_sb); mutex_lock_nested(&dentry->d_inode->i_mutex, I_MUTEX_XATTR); - down_write(&dentry->d_inode->i_alloc_sem); + inode_dio_wait(dentry->d_inode); reiserfs_write_lock(inode->i_sb); err = reiserfs_setattr(dentry, &newattrs); - up_write(&dentry->d_inode->i_alloc_sem); mutex_unlock(&dentry->d_inode->i_mutex); } else update_ctime(inode); @@ -868,27 +867,6 @@ out: return err; } -static int reiserfs_check_acl(struct inode *inode, int mask, unsigned int flags) -{ - struct posix_acl *acl; - int error = -EAGAIN; /* do regular unix permission checks by default */ - - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - - acl = reiserfs_get_acl(inode, ACL_TYPE_ACCESS); - - if (acl) { - if (!IS_ERR(acl)) { - error = posix_acl_permission(inode, acl, mask); - posix_acl_release(acl); - } else if (PTR_ERR(acl) != -ENODATA) - error = PTR_ERR(acl); - } - - return error; -} - static int create_privroot(struct dentry *dentry) { int err; @@ -952,7 +930,7 @@ static int xattr_mount_check(struct super_block *s) return 0; } -int reiserfs_permission(struct inode *inode, int mask, unsigned int flags) +int reiserfs_permission(struct inode *inode, int mask) { /* * We don't do permission checks on the internal objects. @@ -961,15 +939,7 @@ int reiserfs_permission(struct inode *inode, int mask, unsigned int flags) if (IS_PRIVATE(inode)) return 0; -#ifdef CONFIG_REISERFS_FS_XATTR - /* - * Stat data v1 doesn't support ACLs. - */ - if (get_inode_sd_version(inode) != STAT_DATA_V1) - return generic_permission(inode, mask, flags, - reiserfs_check_acl); -#endif - return generic_permission(inode, mask, flags, NULL); + return generic_permission(inode, mask); } static int xattr_hide_revalidate(struct dentry *dentry, struct nameidata *nd) diff --git a/fs/reiserfs/xattr_acl.c b/fs/reiserfs/xattr_acl.c index 3dc38f1..6da0396 100644 --- a/fs/reiserfs/xattr_acl.c +++ b/fs/reiserfs/xattr_acl.c @@ -272,12 +272,10 @@ reiserfs_set_acl(struct reiserfs_transaction_handle *th, struct inode *inode, case ACL_TYPE_ACCESS: name = POSIX_ACL_XATTR_ACCESS; if (acl) { - mode_t mode = inode->i_mode; - error = posix_acl_equiv_mode(acl, &mode); + error = posix_acl_equiv_mode(acl, &inode->i_mode); if (error < 0) return error; else { - inode->i_mode = mode; if (error == 0) acl = NULL; } @@ -354,10 +352,6 @@ reiserfs_inherit_default_acl(struct reiserfs_transaction_handle *th, return PTR_ERR(acl); if (acl) { - struct posix_acl *acl_copy; - mode_t mode = inode->i_mode; - int need_acl; - /* Copy the default ACL to the default ACL of a new directory */ if (S_ISDIR(inode->i_mode)) { err = reiserfs_set_acl(th, inode, ACL_TYPE_DEFAULT, @@ -368,29 +362,13 @@ reiserfs_inherit_default_acl(struct reiserfs_transaction_handle *th, /* Now we reconcile the new ACL and the mode, potentially modifying both */ - acl_copy = posix_acl_clone(acl, GFP_NOFS); - if (!acl_copy) { - err = -ENOMEM; - goto cleanup; - } - - need_acl = posix_acl_create_masq(acl_copy, &mode); - if (need_acl >= 0) { - if (mode != inode->i_mode) { - inode->i_mode = mode; - } + err = posix_acl_create(&acl, GFP_NOFS, &inode->i_mode); + if (err < 0) + return err; - /* If we need an ACL.. */ - if (need_acl > 0) { - err = reiserfs_set_acl(th, inode, - ACL_TYPE_ACCESS, - acl_copy); - if (err) - goto cleanup_copy; - } - } - cleanup_copy: - posix_acl_release(acl_copy); + /* If we need an ACL.. */ + if (err > 0) + err = reiserfs_set_acl(th, inode, ACL_TYPE_ACCESS, acl); cleanup: posix_acl_release(acl); } else { @@ -445,7 +423,10 @@ int reiserfs_cache_default_acl(struct inode *inode) int reiserfs_acl_chmod(struct inode *inode) { - struct posix_acl *acl, *clone; + struct reiserfs_transaction_handle th; + struct posix_acl *acl; + size_t size; + int depth; int error; if (S_ISLNK(inode->i_mode)) @@ -463,30 +444,22 @@ int reiserfs_acl_chmod(struct inode *inode) return 0; if (IS_ERR(acl)) return PTR_ERR(acl); - clone = posix_acl_clone(acl, GFP_NOFS); - posix_acl_release(acl); - if (!clone) - return -ENOMEM; - error = posix_acl_chmod_masq(clone, inode->i_mode); + error = posix_acl_chmod(&acl, GFP_NOFS, inode->i_mode); + if (error) + return error; + + size = reiserfs_xattr_nblocks(inode, reiserfs_acl_size(acl->a_count)); + depth = reiserfs_write_lock_once(inode->i_sb); + error = journal_begin(&th, inode->i_sb, size * 2); if (!error) { - struct reiserfs_transaction_handle th; - size_t size = reiserfs_xattr_nblocks(inode, - reiserfs_acl_size(clone->a_count)); - int depth; - - depth = reiserfs_write_lock_once(inode->i_sb); - error = journal_begin(&th, inode->i_sb, size * 2); - if (!error) { - int error2; - error = reiserfs_set_acl(&th, inode, ACL_TYPE_ACCESS, - clone); - error2 = journal_end(&th, inode->i_sb, size * 2); - if (error2) - error = error2; - } - reiserfs_write_unlock_once(inode->i_sb, depth); + int error2; + error = reiserfs_set_acl(&th, inode, ACL_TYPE_ACCESS, acl); + error2 = journal_end(&th, inode->i_sb, size * 2); + if (error2) + error = error2; } - posix_acl_release(clone); + reiserfs_write_unlock_once(inode->i_sb, depth); + posix_acl_release(acl); return error; } diff --git a/fs/reiserfs/xattr_security.c b/fs/reiserfs/xattr_security.c index ef66c18..534668f 100644 --- a/fs/reiserfs/xattr_security.c +++ b/fs/reiserfs/xattr_security.c @@ -66,8 +66,8 @@ int reiserfs_security_init(struct inode *dir, struct inode *inode, if (IS_PRIVATE(dir)) return 0; - error = security_inode_init_security(inode, dir, qstr, &sec->name, - &sec->value, &sec->length); + error = security_old_inode_init_security(inode, dir, qstr, &sec->name, + &sec->value, &sec->length); if (error) { if (error == -EOPNOTSUPP) error = 0; diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c index 315de66..bc4f94b 100644 --- a/fs/ubifs/budget.c +++ b/fs/ubifs/budget.c @@ -63,7 +63,7 @@ static void shrink_liability(struct ubifs_info *c, int nr_to_write) { down_read(&c->vfs_sb->s_umount); - writeback_inodes_sb(c->vfs_sb); + writeback_inodes_sb(c->vfs_sb, WB_REASON_FS_FREE_SPACE); up_read(&c->vfs_sb->s_umount); } diff --git a/fs/ubifs/commit.c b/fs/ubifs/commit.c index 87cd0ea..b2ca12f 100644 --- a/fs/ubifs/commit.c +++ b/fs/ubifs/commit.c @@ -78,7 +78,7 @@ static int nothing_to_commit(struct ubifs_info *c) * If the root TNC node is dirty, we definitely have something to * commit. */ - if (c->zroot.znode && test_bit(DIRTY_ZNODE, &c->zroot.znode->flags)) + if (c->zroot.znode && ubifs_zn_dirty(c->zroot.znode)) return 0; /* @@ -166,15 +166,10 @@ static int do_commit(struct ubifs_info *c) err = ubifs_orphan_end_commit(c); if (err) goto out; - old_ltail_lnum = c->ltail_lnum; - err = ubifs_log_end_commit(c, new_ltail_lnum); - if (err) - goto out; err = dbg_check_old_index(c, &zroot); if (err) goto out; - mutex_lock(&c->mst_mutex); c->mst_node->cmt_no = cpu_to_le64(c->cmt_no); c->mst_node->log_lnum = cpu_to_le32(new_ltail_lnum); c->mst_node->root_lnum = cpu_to_le32(zroot.lnum); @@ -203,8 +198,9 @@ static int do_commit(struct ubifs_info *c) c->mst_node->flags |= cpu_to_le32(UBIFS_MST_NO_ORPHS); else c->mst_node->flags &= ~cpu_to_le32(UBIFS_MST_NO_ORPHS); - err = ubifs_write_master(c); - mutex_unlock(&c->mst_mutex); + + old_ltail_lnum = c->ltail_lnum; + err = ubifs_log_end_commit(c, new_ltail_lnum); if (err) goto out; @@ -418,7 +414,7 @@ int ubifs_run_commit(struct ubifs_info *c) spin_lock(&c->cs_lock); if (c->cmt_state == COMMIT_BROKEN) { - err = -EINVAL; + err = -EROFS; goto out; } @@ -444,7 +440,7 @@ int ubifs_run_commit(struct ubifs_info *c) * re-check it. */ if (c->cmt_state == COMMIT_BROKEN) { - err = -EINVAL; + err = -EROFS; goto out_cmt_unlock; } @@ -576,7 +572,7 @@ int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot) struct idx_node *i; size_t sz; - if (!(ubifs_chk_flags & UBIFS_CHK_OLD_IDX)) + if (!dbg_is_chk_index(c)) return 0; INIT_LIST_HEAD(&list); diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c index 0bb2bce..b09ba2d 100644 --- a/fs/ubifs/debug.c +++ b/fs/ubifs/debug.c @@ -27,13 +27,12 @@ * various local functions of those subsystems. */ -#define UBIFS_DBG_PRESERVE_UBI - -#include "ubifs.h" #include <linux/module.h> -#include <linux/moduleparam.h> #include <linux/debugfs.h> #include <linux/math64.h> +#include <linux/uaccess.h> +#include <linux/random.h> +#include "ubifs.h" #ifdef CONFIG_UBIFS_FS_DEBUG @@ -42,15 +41,6 @@ DEFINE_SPINLOCK(dbg_lock); static char dbg_key_buf0[128]; static char dbg_key_buf1[128]; -unsigned int ubifs_chk_flags; -unsigned int ubifs_tst_flags; - -module_param_named(debug_chks, ubifs_chk_flags, uint, S_IRUGO | S_IWUSR); -module_param_named(debug_tsts, ubifs_tst_flags, uint, S_IRUGO | S_IWUSR); - -MODULE_PARM_DESC(debug_chks, "Debug check flags"); -MODULE_PARM_DESC(debug_tsts, "Debug special test flags"); - static const char *get_key_fmt(int fmt) { switch (fmt) { @@ -91,6 +81,28 @@ static const char *get_key_type(int type) } } +static const char *get_dent_type(int type) +{ + switch (type) { + case UBIFS_ITYPE_REG: + return "file"; + case UBIFS_ITYPE_DIR: + return "dir"; + case UBIFS_ITYPE_LNK: + return "symlink"; + case UBIFS_ITYPE_BLK: + return "blkdev"; + case UBIFS_ITYPE_CHR: + return "char dev"; + case UBIFS_ITYPE_FIFO: + return "fifo"; + case UBIFS_ITYPE_SOCK: + return "socket"; + default: + return "unknown/invalid type"; + } +} + static void sprintf_key(const struct ubifs_info *c, const union ubifs_key *key, char *buffer) { @@ -234,9 +246,13 @@ static void dump_ch(const struct ubifs_ch *ch) printk(KERN_DEBUG "\tlen %u\n", le32_to_cpu(ch->len)); } -void dbg_dump_inode(const struct ubifs_info *c, const struct inode *inode) +void dbg_dump_inode(struct ubifs_info *c, const struct inode *inode) { const struct ubifs_inode *ui = ubifs_inode(inode); + struct qstr nm = { .name = NULL }; + union ubifs_key key; + struct ubifs_dent_node *dent, *pdent = NULL; + int count = 2; printk(KERN_DEBUG "Dump in-memory inode:"); printk(KERN_DEBUG "\tinode %lu\n", inode->i_ino); @@ -270,6 +286,32 @@ void dbg_dump_inode(const struct ubifs_info *c, const struct inode *inode) printk(KERN_DEBUG "\tlast_page_read %lu\n", ui->last_page_read); printk(KERN_DEBUG "\tread_in_a_row %lu\n", ui->read_in_a_row); printk(KERN_DEBUG "\tdata_len %d\n", ui->data_len); + + if (!S_ISDIR(inode->i_mode)) + return; + + printk(KERN_DEBUG "List of directory entries:\n"); + ubifs_assert(!mutex_is_locked(&c->tnc_mutex)); + + lowest_dent_key(c, &key, inode->i_ino); + while (1) { + dent = ubifs_tnc_next_ent(c, &key, &nm); + if (IS_ERR(dent)) { + if (PTR_ERR(dent) != -ENOENT) + printk(KERN_DEBUG "error %ld\n", PTR_ERR(dent)); + break; + } + + printk(KERN_DEBUG "\t%d: %s (%s)\n", + count++, dent->name, get_dent_type(dent->type)); + + nm.name = dent->name; + nm.len = le16_to_cpu(dent->nlen); + kfree(pdent); + pdent = dent; + key_read(c, &dent->key, &key); + } + kfree(pdent); } void dbg_dump_node(const struct ubifs_info *c, const void *node) @@ -278,7 +320,7 @@ void dbg_dump_node(const struct ubifs_info *c, const void *node) union ubifs_key key; const struct ubifs_ch *ch = node; - if (dbg_failure_mode) + if (dbg_is_tst_rcvry(c)) return; /* If the magic is incorrect, just hexdump the first bytes */ @@ -828,13 +870,29 @@ void dbg_dump_lpt_info(struct ubifs_info *c) spin_unlock(&dbg_lock); } +void dbg_dump_sleb(const struct ubifs_info *c, + const struct ubifs_scan_leb *sleb, int offs) +{ + struct ubifs_scan_node *snod; + + printk(KERN_DEBUG "(pid %d) start dumping scanned data from LEB %d:%d\n", + current->pid, sleb->lnum, offs); + + list_for_each_entry(snod, &sleb->nodes, list) { + cond_resched(); + printk(KERN_DEBUG "Dumping node at LEB %d:%d len %d\n", sleb->lnum, + snod->offs, snod->len); + dbg_dump_node(c, snod->node); + } +} + void dbg_dump_leb(const struct ubifs_info *c, int lnum) { struct ubifs_scan_leb *sleb; struct ubifs_scan_node *snod; void *buf; - if (dbg_failure_mode) + if (dbg_is_tst_rcvry(c)) return; printk(KERN_DEBUG "(pid %d) start dumping LEB %d\n", @@ -1080,6 +1138,7 @@ out: /** * dbg_check_synced_i_size - check synchronized inode size. + * @c: UBIFS file-system description object * @inode: inode to check * * If inode is clean, synchronized inode size has to be equivalent to current @@ -1087,12 +1146,12 @@ out: * has to be locked). Returns %0 if synchronized inode size if correct, and * %-EINVAL if not. */ -int dbg_check_synced_i_size(struct inode *inode) +int dbg_check_synced_i_size(const struct ubifs_info *c, struct inode *inode) { int err = 0; struct ubifs_inode *ui = ubifs_inode(inode); - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; if (!S_ISREG(inode->i_mode)) return 0; @@ -1125,7 +1184,7 @@ int dbg_check_synced_i_size(struct inode *inode) * Note, it is good idea to make sure the @dir->i_mutex is locked before * calling this function. */ -int dbg_check_dir_size(struct ubifs_info *c, const struct inode *dir) +int dbg_check_dir(struct ubifs_info *c, const struct inode *dir) { unsigned int nlink = 2; union ubifs_key key; @@ -1133,7 +1192,7 @@ int dbg_check_dir_size(struct ubifs_info *c, const struct inode *dir) struct qstr nm = { .name = NULL }; loff_t size = UBIFS_INO_NODE_SZ; - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; if (!S_ISDIR(dir->i_mode)) @@ -1167,12 +1226,14 @@ int dbg_check_dir_size(struct ubifs_info *c, const struct inode *dir) "but calculated size is %llu", dir->i_ino, (unsigned long long)i_size_read(dir), (unsigned long long)size); + dbg_dump_inode(c, dir); dump_stack(); return -EINVAL; } if (dir->i_nlink != nlink) { ubifs_err("directory inode %lu has nlink %u, but calculated " "nlink is %u", dir->i_ino, dir->i_nlink, nlink); + dbg_dump_inode(c, dir); dump_stack(); return -EINVAL; } @@ -1489,7 +1550,7 @@ int dbg_check_tnc(struct ubifs_info *c, int extra) long clean_cnt = 0, dirty_cnt = 0; int err, last; - if (!(ubifs_chk_flags & UBIFS_CHK_TNC)) + if (!dbg_is_chk_index(c)) return 0; ubifs_assert(mutex_is_locked(&c->tnc_mutex)); @@ -1736,7 +1797,7 @@ int dbg_check_idx_size(struct ubifs_info *c, long long idx_size) int err; long long calc = 0; - if (!(ubifs_chk_flags & UBIFS_CHK_IDX_SZ)) + if (!dbg_is_chk_index(c)) return 0; err = dbg_walk_index(c, NULL, add_size, &calc); @@ -2312,7 +2373,7 @@ int dbg_check_filesystem(struct ubifs_info *c) int err; struct fsck_data fsckd; - if (!(ubifs_chk_flags & UBIFS_CHK_FS)) + if (!dbg_is_chk_fs(c)) return 0; fsckd.inodes = RB_ROOT; @@ -2347,7 +2408,7 @@ int dbg_check_data_nodes_order(struct ubifs_info *c, struct list_head *head) struct list_head *cur; struct ubifs_scan_node *sa, *sb; - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; for (cur = head->next; cur->next != head; cur = cur->next) { @@ -2414,7 +2475,7 @@ int dbg_check_nondata_nodes_order(struct ubifs_info *c, struct list_head *head) struct list_head *cur; struct ubifs_scan_node *sa, *sb; - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; for (cur = head->next; cur->next != head; cur = cur->next) { @@ -2491,214 +2552,141 @@ error_dump: return 0; } -int dbg_force_in_the_gaps(void) +static inline int chance(unsigned int n, unsigned int out_of) { - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) - return 0; + return !!((random32() % out_of) + 1 <= n); - return !(random32() & 7); } -/* Failure mode for recovery testing */ - -#define chance(n, d) (simple_rand() <= (n) * 32768LL / (d)) - -struct failure_mode_info { - struct list_head list; - struct ubifs_info *c; -}; - -static LIST_HEAD(fmi_list); -static DEFINE_SPINLOCK(fmi_lock); - -static unsigned int next; - -static int simple_rand(void) +static int power_cut_emulated(struct ubifs_info *c, int lnum, int write) { - if (next == 0) - next = current->pid; - next = next * 1103515245 + 12345; - return (next >> 16) & 32767; -} - -static void failure_mode_init(struct ubifs_info *c) -{ - struct failure_mode_info *fmi; - - fmi = kmalloc(sizeof(struct failure_mode_info), GFP_NOFS); - if (!fmi) { - ubifs_err("Failed to register failure mode - no memory"); - return; - } - fmi->c = c; - spin_lock(&fmi_lock); - list_add_tail(&fmi->list, &fmi_list); - spin_unlock(&fmi_lock); -} - -static void failure_mode_exit(struct ubifs_info *c) -{ - struct failure_mode_info *fmi, *tmp; - - spin_lock(&fmi_lock); - list_for_each_entry_safe(fmi, tmp, &fmi_list, list) - if (fmi->c == c) { - list_del(&fmi->list); - kfree(fmi); - } - spin_unlock(&fmi_lock); -} - -static struct ubifs_info *dbg_find_info(struct ubi_volume_desc *desc) -{ - struct failure_mode_info *fmi; - - spin_lock(&fmi_lock); - list_for_each_entry(fmi, &fmi_list, list) - if (fmi->c->ubi == desc) { - struct ubifs_info *c = fmi->c; - - spin_unlock(&fmi_lock); - return c; - } - spin_unlock(&fmi_lock); - return NULL; -} - -static int in_failure_mode(struct ubi_volume_desc *desc) -{ - struct ubifs_info *c = dbg_find_info(desc); - - if (c && dbg_failure_mode) - return c->dbg->failure_mode; - return 0; -} + struct ubifs_debug_info *d = c->dbg; -static int do_fail(struct ubi_volume_desc *desc, int lnum, int write) -{ - struct ubifs_info *c = dbg_find_info(desc); - struct ubifs_debug_info *d; + ubifs_assert(dbg_is_tst_rcvry(c)); - if (!c || !dbg_failure_mode) - return 0; - d = c->dbg; - if (d->failure_mode) - return 1; - if (!d->fail_cnt) { - /* First call - decide delay to failure */ + if (!d->pc_cnt) { + /* First call - decide delay to the power cut */ if (chance(1, 2)) { - unsigned int delay = 1 << (simple_rand() >> 11); + unsigned long delay; if (chance(1, 2)) { - d->fail_delay = 1; - d->fail_timeout = jiffies + - msecs_to_jiffies(delay); - dbg_rcvry("failing after %ums", delay); + d->pc_delay = 1; + /* Fail withing 1 minute */ + delay = random32() % 60000; + d->pc_timeout = jiffies; + d->pc_timeout += msecs_to_jiffies(delay); + ubifs_warn("failing after %lums", delay); } else { - d->fail_delay = 2; - d->fail_cnt_max = delay; - dbg_rcvry("failing after %u calls", delay); + d->pc_delay = 2; + delay = random32() % 10000; + /* Fail within 10000 operations */ + d->pc_cnt_max = delay; + ubifs_warn("failing after %lu calls", delay); } } - d->fail_cnt += 1; + + d->pc_cnt += 1; } + /* Determine if failure delay has expired */ - if (d->fail_delay == 1) { - if (time_before(jiffies, d->fail_timeout)) + if (d->pc_delay == 1 && time_before(jiffies, d->pc_timeout)) return 0; - } else if (d->fail_delay == 2) - if (d->fail_cnt++ < d->fail_cnt_max) + if (d->pc_delay == 2 && d->pc_cnt++ < d->pc_cnt_max) return 0; + if (lnum == UBIFS_SB_LNUM) { - if (write) { - if (chance(1, 2)) - return 0; - } else if (chance(19, 20)) + if (write && chance(1, 2)) return 0; - dbg_rcvry("failing in super block LEB %d", lnum); + if (chance(19, 20)) + return 0; + ubifs_warn("failing in super block LEB %d", lnum); } else if (lnum == UBIFS_MST_LNUM || lnum == UBIFS_MST_LNUM + 1) { if (chance(19, 20)) return 0; - dbg_rcvry("failing in master LEB %d", lnum); + ubifs_warn("failing in master LEB %d", lnum); } else if (lnum >= UBIFS_LOG_LNUM && lnum <= c->log_last) { - if (write) { - if (chance(99, 100)) - return 0; - } else if (chance(399, 400)) + if (write && chance(99, 100)) + return 0; + if (chance(399, 400)) return 0; - dbg_rcvry("failing in log LEB %d", lnum); + ubifs_warn("failing in log LEB %d", lnum); } else if (lnum >= c->lpt_first && lnum <= c->lpt_last) { - if (write) { - if (chance(7, 8)) - return 0; - } else if (chance(19, 20)) + if (write && chance(7, 8)) return 0; - dbg_rcvry("failing in LPT LEB %d", lnum); + if (chance(19, 20)) + return 0; + ubifs_warn("failing in LPT LEB %d", lnum); } else if (lnum >= c->orph_first && lnum <= c->orph_last) { - if (write) { - if (chance(1, 2)) - return 0; - } else if (chance(9, 10)) + if (write && chance(1, 2)) return 0; - dbg_rcvry("failing in orphan LEB %d", lnum); + if (chance(9, 10)) + return 0; + ubifs_warn("failing in orphan LEB %d", lnum); } else if (lnum == c->ihead_lnum) { if (chance(99, 100)) return 0; - dbg_rcvry("failing in index head LEB %d", lnum); + ubifs_warn("failing in index head LEB %d", lnum); } else if (c->jheads && lnum == c->jheads[GCHD].wbuf.lnum) { if (chance(9, 10)) return 0; - dbg_rcvry("failing in GC head LEB %d", lnum); + ubifs_warn("failing in GC head LEB %d", lnum); } else if (write && !RB_EMPTY_ROOT(&c->buds) && !ubifs_search_bud(c, lnum)) { if (chance(19, 20)) return 0; - dbg_rcvry("failing in non-bud LEB %d", lnum); + ubifs_warn("failing in non-bud LEB %d", lnum); } else if (c->cmt_state == COMMIT_RUNNING_BACKGROUND || c->cmt_state == COMMIT_RUNNING_REQUIRED) { if (chance(999, 1000)) return 0; - dbg_rcvry("failing in bud LEB %d commit running", lnum); + ubifs_warn("failing in bud LEB %d commit running", lnum); } else { if (chance(9999, 10000)) return 0; - dbg_rcvry("failing in bud LEB %d commit not running", lnum); + ubifs_warn("failing in bud LEB %d commit not running", lnum); } - ubifs_err("*** SETTING FAILURE MODE ON (LEB %d) ***", lnum); - d->failure_mode = 1; + + d->pc_happened = 1; + ubifs_warn("========== Power cut emulated =========="); dump_stack(); return 1; } -static void cut_data(const void *buf, int len) +static void cut_data(const void *buf, unsigned int len) { - int flen, i; + unsigned int from, to, i, ffs = chance(1, 2); unsigned char *p = (void *)buf; - flen = (len * (long long)simple_rand()) >> 15; - for (i = flen; i < len; i++) - p[i] = 0xff; -} + from = random32() % (len + 1); + if (chance(1, 2)) + to = random32() % (len - from + 1); + else + to = len; -int dbg_leb_read(struct ubi_volume_desc *desc, int lnum, char *buf, int offset, - int len, int check) -{ - if (in_failure_mode(desc)) - return -EROFS; - return ubi_leb_read(desc, lnum, buf, offset, len, check); + if (from < to) + ubifs_warn("filled bytes %u-%u with %s", from, to - 1, + ffs ? "0xFFs" : "random data"); + + if (ffs) + for (i = from; i < to; i++) + p[i] = 0xFF; + else + for (i = from; i < to; i++) + p[i] = random32() % 0x100; } -int dbg_leb_write(struct ubi_volume_desc *desc, int lnum, const void *buf, - int offset, int len, int dtype) +int dbg_leb_write(struct ubifs_info *c, int lnum, const void *buf, + int offs, int len, int dtype) { int err, failing; - if (in_failure_mode(desc)) + if (c->dbg->pc_happened) return -EROFS; - failing = do_fail(desc, lnum, 1); + + failing = power_cut_emulated(c, lnum, 1); if (failing) cut_data(buf, len); - err = ubi_leb_write(desc, lnum, buf, offset, len, dtype); + err = ubi_leb_write(c->ubi, lnum, buf, offs, len, dtype); if (err) return err; if (failing) @@ -2706,162 +2694,207 @@ int dbg_leb_write(struct ubi_volume_desc *desc, int lnum, const void *buf, return 0; } -int dbg_leb_change(struct ubi_volume_desc *desc, int lnum, const void *buf, +int dbg_leb_change(struct ubifs_info *c, int lnum, const void *buf, int len, int dtype) { int err; - if (do_fail(desc, lnum, 1)) + if (c->dbg->pc_happened) return -EROFS; - err = ubi_leb_change(desc, lnum, buf, len, dtype); + if (power_cut_emulated(c, lnum, 1)) + return -EROFS; + err = ubi_leb_change(c->ubi, lnum, buf, len, dtype); if (err) return err; - if (do_fail(desc, lnum, 1)) + if (power_cut_emulated(c, lnum, 1)) return -EROFS; return 0; } -int dbg_leb_erase(struct ubi_volume_desc *desc, int lnum) +int dbg_leb_unmap(struct ubifs_info *c, int lnum) { int err; - if (do_fail(desc, lnum, 0)) + if (c->dbg->pc_happened) + return -EROFS; + if (power_cut_emulated(c, lnum, 0)) return -EROFS; - err = ubi_leb_erase(desc, lnum); + err = ubi_leb_unmap(c->ubi, lnum); if (err) return err; - if (do_fail(desc, lnum, 0)) + if (power_cut_emulated(c, lnum, 0)) return -EROFS; return 0; } -int dbg_leb_unmap(struct ubi_volume_desc *desc, int lnum) +int dbg_leb_map(struct ubifs_info *c, int lnum, int dtype) { int err; - if (do_fail(desc, lnum, 0)) + if (c->dbg->pc_happened) return -EROFS; - err = ubi_leb_unmap(desc, lnum); + if (power_cut_emulated(c, lnum, 0)) + return -EROFS; + err = ubi_leb_map(c->ubi, lnum, dtype); if (err) return err; - if (do_fail(desc, lnum, 0)) + if (power_cut_emulated(c, lnum, 0)) return -EROFS; return 0; } -int dbg_is_mapped(struct ubi_volume_desc *desc, int lnum) -{ - if (in_failure_mode(desc)) - return -EROFS; - return ubi_is_mapped(desc, lnum); -} +/* + * Root directory for UBIFS stuff in debugfs. Contains sub-directories which + * contain the stuff specific to particular file-system mounts. + */ +static struct dentry *dfs_rootdir; -int dbg_leb_map(struct ubi_volume_desc *desc, int lnum, int dtype) +static int dfs_file_open(struct inode *inode, struct file *file) { - int err; - - if (do_fail(desc, lnum, 0)) - return -EROFS; - err = ubi_leb_map(desc, lnum, dtype); - if (err) - return err; - if (do_fail(desc, lnum, 0)) - return -EROFS; - return 0; + file->private_data = inode->i_private; + return nonseekable_open(inode, file); } /** - * ubifs_debugging_init - initialize UBIFS debugging. - * @c: UBIFS file-system description object + * provide_user_output - provide output to the user reading a debugfs file. + * @val: boolean value for the answer + * @u: the buffer to store the answer at + * @count: size of the buffer + * @ppos: position in the @u output buffer * - * This function initializes debugging-related data for the file system. - * Returns zero in case of success and a negative error code in case of + * This is a simple helper function which stores @val boolean value in the user + * buffer when the user reads one of UBIFS debugfs files. Returns amount of + * bytes written to @u in case of success and a negative error code in case of * failure. */ -int ubifs_debugging_init(struct ubifs_info *c) +static int provide_user_output(int val, char __user *u, size_t count, + loff_t *ppos) { - c->dbg = kzalloc(sizeof(struct ubifs_debug_info), GFP_KERNEL); - if (!c->dbg) - return -ENOMEM; + char buf[3]; - failure_mode_init(c); - return 0; + if (val) + buf[0] = '1'; + else + buf[0] = '0'; + buf[1] = '\n'; + buf[2] = 0x00; + + return simple_read_from_buffer(u, count, ppos, buf, 2); } -/** - * ubifs_debugging_exit - free debugging data. - * @c: UBIFS file-system description object - */ -void ubifs_debugging_exit(struct ubifs_info *c) +static ssize_t dfs_file_read(struct file *file, char __user *u, size_t count, + loff_t *ppos) { - failure_mode_exit(c); - kfree(c->dbg); -} + struct dentry *dent = file->f_path.dentry; + struct ubifs_info *c = file->private_data; + struct ubifs_debug_info *d = c->dbg; + int val; + + if (dent == d->dfs_chk_gen) + val = d->chk_gen; + else if (dent == d->dfs_chk_index) + val = d->chk_index; + else if (dent == d->dfs_chk_orph) + val = d->chk_orph; + else if (dent == d->dfs_chk_lprops) + val = d->chk_lprops; + else if (dent == d->dfs_chk_fs) + val = d->chk_fs; + else if (dent == d->dfs_tst_rcvry) + val = d->tst_rcvry; + else + return -EINVAL; -/* - * Root directory for UBIFS stuff in debugfs. Contains sub-directories which - * contain the stuff specific to particular file-system mounts. - */ -static struct dentry *dfs_rootdir; + return provide_user_output(val, u, count, ppos); +} /** - * dbg_debugfs_init - initialize debugfs file-system. + * interpret_user_input - interpret user debugfs file input. + * @u: user-provided buffer with the input + * @count: buffer size * - * UBIFS uses debugfs file-system to expose various debugging knobs to - * user-space. This function creates "ubifs" directory in the debugfs - * file-system. Returns zero in case of success and a negative error code in - * case of failure. + * This is a helper function which interpret user input to a boolean UBIFS + * debugfs file. Returns %0 or %1 in case of success and a negative error code + * in case of failure. */ -int dbg_debugfs_init(void) +static int interpret_user_input(const char __user *u, size_t count) { - dfs_rootdir = debugfs_create_dir("ubifs", NULL); - if (IS_ERR(dfs_rootdir)) { - int err = PTR_ERR(dfs_rootdir); - ubifs_err("cannot create \"ubifs\" debugfs directory, " - "error %d\n", err); - return err; - } + size_t buf_size; + char buf[8]; - return 0; -} + buf_size = min_t(size_t, count, (sizeof(buf) - 1)); + if (copy_from_user(buf, u, buf_size)) + return -EFAULT; -/** - * dbg_debugfs_exit - remove the "ubifs" directory from debugfs file-system. - */ -void dbg_debugfs_exit(void) -{ - debugfs_remove(dfs_rootdir); -} + if (buf[0] == '1') + return 1; + else if (buf[0] == '0') + return 0; -static int open_debugfs_file(struct inode *inode, struct file *file) -{ - file->private_data = inode->i_private; - return nonseekable_open(inode, file); + return -EINVAL; } -static ssize_t write_debugfs_file(struct file *file, const char __user *buf, - size_t count, loff_t *ppos) +static ssize_t dfs_file_write(struct file *file, const char __user *u, + size_t count, loff_t *ppos) { struct ubifs_info *c = file->private_data; struct ubifs_debug_info *d = c->dbg; + struct dentry *dent = file->f_path.dentry; + int val; - if (file->f_path.dentry == d->dfs_dump_lprops) + /* + * TODO: this is racy - the file-system might have already been + * unmounted and we'd oops in this case. The plan is to fix it with + * help of 'iterate_supers_type()' which we should have in v3.0: when + * a debugfs opened, we rember FS's UUID in file->private_data. Then + * whenever we access the FS via a debugfs file, we iterate all UBIFS + * superblocks and fine the one with the same UUID, and take the + * locking right. + * + * The other way to go suggested by Al Viro is to create a separate + * 'ubifs-debug' file-system instead. + */ + if (file->f_path.dentry == d->dfs_dump_lprops) { dbg_dump_lprops(c); - else if (file->f_path.dentry == d->dfs_dump_budg) + return count; + } + if (file->f_path.dentry == d->dfs_dump_budg) { dbg_dump_budg(c, &c->bi); - else if (file->f_path.dentry == d->dfs_dump_tnc) { + return count; + } + if (file->f_path.dentry == d->dfs_dump_tnc) { mutex_lock(&c->tnc_mutex); dbg_dump_tnc(c); mutex_unlock(&c->tnc_mutex); - } else + return count; + } + + val = interpret_user_input(u, count); + if (val < 0) + return val; + + if (dent == d->dfs_chk_gen) + d->chk_gen = val; + else if (dent == d->dfs_chk_index) + d->chk_index = val; + else if (dent == d->dfs_chk_orph) + d->chk_orph = val; + else if (dent == d->dfs_chk_lprops) + d->chk_lprops = val; + else if (dent == d->dfs_chk_fs) + d->chk_fs = val; + else if (dent == d->dfs_tst_rcvry) + d->tst_rcvry = val; + else return -EINVAL; return count; } static const struct file_operations dfs_fops = { - .open = open_debugfs_file, - .write = write_debugfs_file, + .open = dfs_file_open, + .read = dfs_file_read, + .write = dfs_file_write, .owner = THIS_MODULE, .llseek = no_llseek, }; @@ -2880,12 +2913,20 @@ static const struct file_operations dfs_fops = { */ int dbg_debugfs_init_fs(struct ubifs_info *c) { - int err; + int err, n; const char *fname; struct dentry *dent; struct ubifs_debug_info *d = c->dbg; - sprintf(d->dfs_dir_name, "ubi%d_%d", c->vi.ubi_num, c->vi.vol_id); + n = snprintf(d->dfs_dir_name, UBIFS_DFS_DIR_LEN + 1, UBIFS_DFS_DIR_NAME, + c->vi.ubi_num, c->vi.vol_id); + if (n == UBIFS_DFS_DIR_LEN) { + /* The array size is too small */ + fname = UBIFS_DFS_DIR_NAME; + dent = ERR_PTR(-EINVAL); + goto out; + } + fname = d->dfs_dir_name; dent = debugfs_create_dir(fname, dfs_rootdir); if (IS_ERR_OR_NULL(dent)) @@ -2910,13 +2951,55 @@ int dbg_debugfs_init_fs(struct ubifs_info *c) goto out_remove; d->dfs_dump_tnc = dent; + fname = "chk_general"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, d->dfs_dir, c, + &dfs_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + d->dfs_chk_gen = dent; + + fname = "chk_index"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, d->dfs_dir, c, + &dfs_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + d->dfs_chk_index = dent; + + fname = "chk_orphans"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, d->dfs_dir, c, + &dfs_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + d->dfs_chk_orph = dent; + + fname = "chk_lprops"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, d->dfs_dir, c, + &dfs_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + d->dfs_chk_lprops = dent; + + fname = "chk_fs"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, d->dfs_dir, c, + &dfs_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + d->dfs_chk_fs = dent; + + fname = "tst_recovery"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, d->dfs_dir, c, + &dfs_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + d->dfs_tst_rcvry = dent; + return 0; out_remove: debugfs_remove_recursive(d->dfs_dir); out: err = dent ? PTR_ERR(dent) : -ENODEV; - ubifs_err("cannot create \"%s\" debugfs directory, error %d\n", + ubifs_err("cannot create \"%s\" debugfs file or directory, error %d\n", fname, err); return err; } @@ -2930,4 +3013,179 @@ void dbg_debugfs_exit_fs(struct ubifs_info *c) debugfs_remove_recursive(c->dbg->dfs_dir); } +struct ubifs_global_debug_info ubifs_dbg; + +static struct dentry *dfs_chk_gen; +static struct dentry *dfs_chk_index; +static struct dentry *dfs_chk_orph; +static struct dentry *dfs_chk_lprops; +static struct dentry *dfs_chk_fs; +static struct dentry *dfs_tst_rcvry; + +static ssize_t dfs_global_file_read(struct file *file, char __user *u, + size_t count, loff_t *ppos) +{ + struct dentry *dent = file->f_path.dentry; + int val; + + if (dent == dfs_chk_gen) + val = ubifs_dbg.chk_gen; + else if (dent == dfs_chk_index) + val = ubifs_dbg.chk_index; + else if (dent == dfs_chk_orph) + val = ubifs_dbg.chk_orph; + else if (dent == dfs_chk_lprops) + val = ubifs_dbg.chk_lprops; + else if (dent == dfs_chk_fs) + val = ubifs_dbg.chk_fs; + else if (dent == dfs_tst_rcvry) + val = ubifs_dbg.tst_rcvry; + else + return -EINVAL; + + return provide_user_output(val, u, count, ppos); +} + +static ssize_t dfs_global_file_write(struct file *file, const char __user *u, + size_t count, loff_t *ppos) +{ + struct dentry *dent = file->f_path.dentry; + int val; + + val = interpret_user_input(u, count); + if (val < 0) + return val; + + if (dent == dfs_chk_gen) + ubifs_dbg.chk_gen = val; + else if (dent == dfs_chk_index) + ubifs_dbg.chk_index = val; + else if (dent == dfs_chk_orph) + ubifs_dbg.chk_orph = val; + else if (dent == dfs_chk_lprops) + ubifs_dbg.chk_lprops = val; + else if (dent == dfs_chk_fs) + ubifs_dbg.chk_fs = val; + else if (dent == dfs_tst_rcvry) + ubifs_dbg.tst_rcvry = val; + else + return -EINVAL; + + return count; +} + +static const struct file_operations dfs_global_fops = { + .read = dfs_global_file_read, + .write = dfs_global_file_write, + .owner = THIS_MODULE, + .llseek = no_llseek, +}; + +/** + * dbg_debugfs_init - initialize debugfs file-system. + * + * UBIFS uses debugfs file-system to expose various debugging knobs to + * user-space. This function creates "ubifs" directory in the debugfs + * file-system. Returns zero in case of success and a negative error code in + * case of failure. + */ +int dbg_debugfs_init(void) +{ + int err; + const char *fname; + struct dentry *dent; + + fname = "ubifs"; + dent = debugfs_create_dir(fname, NULL); + if (IS_ERR_OR_NULL(dent)) + goto out; + dfs_rootdir = dent; + + fname = "chk_general"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir, NULL, + &dfs_global_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + dfs_chk_gen = dent; + + fname = "chk_index"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir, NULL, + &dfs_global_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + dfs_chk_index = dent; + + fname = "chk_orphans"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir, NULL, + &dfs_global_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + dfs_chk_orph = dent; + + fname = "chk_lprops"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir, NULL, + &dfs_global_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + dfs_chk_lprops = dent; + + fname = "chk_fs"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir, NULL, + &dfs_global_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + dfs_chk_fs = dent; + + fname = "tst_recovery"; + dent = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir, NULL, + &dfs_global_fops); + if (IS_ERR_OR_NULL(dent)) + goto out_remove; + dfs_tst_rcvry = dent; + + return 0; + +out_remove: + debugfs_remove_recursive(dfs_rootdir); +out: + err = dent ? PTR_ERR(dent) : -ENODEV; + ubifs_err("cannot create \"%s\" debugfs file or directory, error %d\n", + fname, err); + return err; +} + +/** + * dbg_debugfs_exit - remove the "ubifs" directory from debugfs file-system. + */ +void dbg_debugfs_exit(void) +{ + debugfs_remove_recursive(dfs_rootdir); +} + +/** + * ubifs_debugging_init - initialize UBIFS debugging. + * @c: UBIFS file-system description object + * + * This function initializes debugging-related data for the file system. + * Returns zero in case of success and a negative error code in case of + * failure. + */ +int ubifs_debugging_init(struct ubifs_info *c) +{ + c->dbg = kzalloc(sizeof(struct ubifs_debug_info), GFP_KERNEL); + if (!c->dbg) + return -ENOMEM; + + return 0; +} + +/** + * ubifs_debugging_exit - free debugging data. + * @c: UBIFS file-system description object + */ +void ubifs_debugging_exit(struct ubifs_info *c) +{ + kfree(c->dbg); +} + #endif /* CONFIG_UBIFS_FS_DEBUG */ diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h index fd75b63..c9d2941 100644 --- a/fs/ubifs/debug.h +++ b/fs/ubifs/debug.h @@ -31,18 +31,25 @@ typedef int (*dbg_znode_callback)(struct ubifs_info *c, #ifdef CONFIG_UBIFS_FS_DEBUG -#include <linux/random.h> +/* + * The UBIFS debugfs directory name pattern and maximum name length (3 for "ubi" + * + 1 for "_" and plus 2x2 for 2 UBI numbers and 1 for the trailing zero byte. + */ +#define UBIFS_DFS_DIR_NAME "ubi%d_%d" +#define UBIFS_DFS_DIR_LEN (3 + 1 + 2*2 + 1) /** * ubifs_debug_info - per-FS debugging information. * @old_zroot: old index root - used by 'dbg_check_old_index()' * @old_zroot_level: old index root level - used by 'dbg_check_old_index()' * @old_zroot_sqnum: old index root sqnum - used by 'dbg_check_old_index()' - * @failure_mode: failure mode for recovery testing - * @fail_delay: 0=>don't delay, 1=>delay a time, 2=>delay a number of calls - * @fail_timeout: time in jiffies when delay of failure mode expires - * @fail_cnt: current number of calls to failure mode I/O functions - * @fail_cnt_max: number of calls by which to delay failure mode + * + * @pc_happened: non-zero if an emulated power cut happened + * @pc_delay: 0=>don't delay, 1=>delay a time, 2=>delay a number of calls + * @pc_timeout: time in jiffies when delay of failure mode expires + * @pc_cnt: current number of calls to failure mode I/O functions + * @pc_cnt_max: number of calls by which to delay failure mode + * * @chk_lpt_sz: used by LPT tree size checker * @chk_lpt_sz2: used by LPT tree size checker * @chk_lpt_wastage: used by LPT tree size checker @@ -56,21 +63,36 @@ typedef int (*dbg_znode_callback)(struct ubifs_info *c, * @saved_free: saved amount of free space * @saved_idx_gc_cnt: saved value of @c->idx_gc_cnt * + * @chk_gen: if general extra checks are enabled + * @chk_index: if index xtra checks are enabled + * @chk_orph: if orphans extra checks are enabled + * @chk_lprops: if lprops extra checks are enabled + * @chk_fs: if UBIFS contents extra checks are enabled + * @tst_rcvry: if UBIFS recovery testing mode enabled + * * @dfs_dir_name: name of debugfs directory containing this file-system's files * @dfs_dir: direntry object of the file-system debugfs directory * @dfs_dump_lprops: "dump lprops" debugfs knob * @dfs_dump_budg: "dump budgeting information" debugfs knob * @dfs_dump_tnc: "dump TNC" debugfs knob + * @dfs_chk_gen: debugfs knob to enable UBIFS general extra checks + * @dfs_chk_index: debugfs knob to enable UBIFS index extra checks + * @dfs_chk_orph: debugfs knob to enable UBIFS orphans extra checks + * @dfs_chk_lprops: debugfs knob to enable UBIFS LEP properties extra checks + * @dfs_chk_fs: debugfs knob to enable UBIFS contents extra checks + * @dfs_tst_rcvry: debugfs knob to enable UBIFS recovery testing */ struct ubifs_debug_info { struct ubifs_zbranch old_zroot; int old_zroot_level; unsigned long long old_zroot_sqnum; - int failure_mode; - int fail_delay; - unsigned long fail_timeout; - unsigned int fail_cnt; - unsigned int fail_cnt_max; + + int pc_happened; + int pc_delay; + unsigned long pc_timeout; + unsigned int pc_cnt; + unsigned int pc_cnt_max; + long long chk_lpt_sz; long long chk_lpt_sz2; long long chk_lpt_wastage; @@ -84,11 +106,43 @@ struct ubifs_debug_info { long long saved_free; int saved_idx_gc_cnt; - char dfs_dir_name[100]; + unsigned int chk_gen:1; + unsigned int chk_index:1; + unsigned int chk_orph:1; + unsigned int chk_lprops:1; + unsigned int chk_fs:1; + unsigned int tst_rcvry:1; + + char dfs_dir_name[UBIFS_DFS_DIR_LEN + 1]; struct dentry *dfs_dir; struct dentry *dfs_dump_lprops; struct dentry *dfs_dump_budg; struct dentry *dfs_dump_tnc; + struct dentry *dfs_chk_gen; + struct dentry *dfs_chk_index; + struct dentry *dfs_chk_orph; + struct dentry *dfs_chk_lprops; + struct dentry *dfs_chk_fs; + struct dentry *dfs_tst_rcvry; +}; + +/** + * ubifs_global_debug_info - global (not per-FS) UBIFS debugging information. + * + * @chk_gen: if general extra checks are enabled + * @chk_index: if index xtra checks are enabled + * @chk_orph: if orphans extra checks are enabled + * @chk_lprops: if lprops extra checks are enabled + * @chk_fs: if UBIFS contents extra checks are enabled + * @tst_rcvry: if UBIFS recovery testing mode enabled + */ +struct ubifs_global_debug_info { + unsigned int chk_gen:1; + unsigned int chk_index:1; + unsigned int chk_orph:1; + unsigned int chk_lprops:1; + unsigned int chk_fs:1; + unsigned int tst_rcvry:1; }; #define ubifs_assert(expr) do { \ @@ -128,6 +182,8 @@ const char *dbg_key_str1(const struct ubifs_info *c, #define DBGKEY(key) dbg_key_str0(c, (key)) #define DBGKEY1(key) dbg_key_str1(c, (key)) +extern spinlock_t dbg_lock; + #define ubifs_dbg_msg(type, fmt, ...) \ pr_debug("UBIFS DBG " type ": " fmt "\n", ##__VA_ARGS__) @@ -163,41 +219,36 @@ const char *dbg_key_str1(const struct ubifs_info *c, /* Additional recovery messages */ #define dbg_rcvry(fmt, ...) ubifs_dbg_msg("rcvry", fmt, ##__VA_ARGS__) -/* - * Debugging check flags. - * - * UBIFS_CHK_GEN: general checks - * UBIFS_CHK_TNC: check TNC - * UBIFS_CHK_IDX_SZ: check index size - * UBIFS_CHK_ORPH: check orphans - * UBIFS_CHK_OLD_IDX: check the old index - * UBIFS_CHK_LPROPS: check lprops - * UBIFS_CHK_FS: check the file-system - */ -enum { - UBIFS_CHK_GEN = 0x1, - UBIFS_CHK_TNC = 0x2, - UBIFS_CHK_IDX_SZ = 0x4, - UBIFS_CHK_ORPH = 0x8, - UBIFS_CHK_OLD_IDX = 0x10, - UBIFS_CHK_LPROPS = 0x20, - UBIFS_CHK_FS = 0x40, -}; +extern struct ubifs_global_debug_info ubifs_dbg; -/* - * Special testing flags. - * - * UBIFS_TST_RCVRY: failure mode for recovery testing - */ -enum { - UBIFS_TST_RCVRY = 0x4, -}; - -extern spinlock_t dbg_lock; - -extern unsigned int ubifs_msg_flags; -extern unsigned int ubifs_chk_flags; -extern unsigned int ubifs_tst_flags; +static inline int dbg_is_chk_gen(const struct ubifs_info *c) +{ + return !!(ubifs_dbg.chk_gen || c->dbg->chk_gen); +} +static inline int dbg_is_chk_index(const struct ubifs_info *c) +{ + return !!(ubifs_dbg.chk_index || c->dbg->chk_index); +} +static inline int dbg_is_chk_orph(const struct ubifs_info *c) +{ + return !!(ubifs_dbg.chk_orph || c->dbg->chk_orph); +} +static inline int dbg_is_chk_lprops(const struct ubifs_info *c) +{ + return !!(ubifs_dbg.chk_lprops || c->dbg->chk_lprops); +} +static inline int dbg_is_chk_fs(const struct ubifs_info *c) +{ + return !!(ubifs_dbg.chk_fs || c->dbg->chk_fs); +} +static inline int dbg_is_tst_rcvry(const struct ubifs_info *c) +{ + return !!(ubifs_dbg.tst_rcvry || c->dbg->tst_rcvry); +} +static inline int dbg_is_power_cut(const struct ubifs_info *c) +{ + return !!c->dbg->pc_happened; +} int ubifs_debugging_init(struct ubifs_info *c); void ubifs_debugging_exit(struct ubifs_info *c); @@ -208,7 +259,7 @@ const char *dbg_cstate(int cmt_state); const char *dbg_jhead(int jhead); const char *dbg_get_key_dump(const struct ubifs_info *c, const union ubifs_key *key); -void dbg_dump_inode(const struct ubifs_info *c, const struct inode *inode); +void dbg_dump_inode(struct ubifs_info *c, const struct inode *inode); void dbg_dump_node(const struct ubifs_info *c, const void *node); void dbg_dump_lpt_node(const struct ubifs_info *c, void *node, int lnum, int offs); @@ -219,6 +270,8 @@ void dbg_dump_lprop(const struct ubifs_info *c, const struct ubifs_lprops *lp); void dbg_dump_lprops(struct ubifs_info *c); void dbg_dump_lpt_info(struct ubifs_info *c); void dbg_dump_leb(const struct ubifs_info *c, int lnum); +void dbg_dump_sleb(const struct ubifs_info *c, + const struct ubifs_scan_leb *sleb, int offs); void dbg_dump_znode(const struct ubifs_info *c, const struct ubifs_znode *znode); void dbg_dump_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat); @@ -241,8 +294,8 @@ int dbg_check_cats(struct ubifs_info *c); int dbg_check_ltab(struct ubifs_info *c); int dbg_chk_lpt_free_spc(struct ubifs_info *c); int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len); -int dbg_check_synced_i_size(struct inode *inode); -int dbg_check_dir_size(struct ubifs_info *c, const struct inode *dir); +int dbg_check_synced_i_size(const struct ubifs_info *c, struct inode *inode); +int dbg_check_dir(struct ubifs_info *c, const struct inode *dir); int dbg_check_tnc(struct ubifs_info *c, int extra); int dbg_check_idx_size(struct ubifs_info *c, long long idx_size); int dbg_check_filesystem(struct ubifs_info *c); @@ -255,54 +308,12 @@ int dbg_check_inode_size(struct ubifs_info *c, const struct inode *inode, int dbg_check_data_nodes_order(struct ubifs_info *c, struct list_head *head); int dbg_check_nondata_nodes_order(struct ubifs_info *c, struct list_head *head); -/* Force the use of in-the-gaps method for testing */ -static inline int dbg_force_in_the_gaps_enabled(void) -{ - return ubifs_chk_flags & UBIFS_CHK_GEN; -} -int dbg_force_in_the_gaps(void); - -/* Failure mode for recovery testing */ -#define dbg_failure_mode (ubifs_tst_flags & UBIFS_TST_RCVRY) - -#ifndef UBIFS_DBG_PRESERVE_UBI -#define ubi_leb_read dbg_leb_read -#define ubi_leb_write dbg_leb_write -#define ubi_leb_change dbg_leb_change -#define ubi_leb_erase dbg_leb_erase -#define ubi_leb_unmap dbg_leb_unmap -#define ubi_is_mapped dbg_is_mapped -#define ubi_leb_map dbg_leb_map -#endif - -int dbg_leb_read(struct ubi_volume_desc *desc, int lnum, char *buf, int offset, - int len, int check); -int dbg_leb_write(struct ubi_volume_desc *desc, int lnum, const void *buf, - int offset, int len, int dtype); -int dbg_leb_change(struct ubi_volume_desc *desc, int lnum, const void *buf, - int len, int dtype); -int dbg_leb_erase(struct ubi_volume_desc *desc, int lnum); -int dbg_leb_unmap(struct ubi_volume_desc *desc, int lnum); -int dbg_is_mapped(struct ubi_volume_desc *desc, int lnum); -int dbg_leb_map(struct ubi_volume_desc *desc, int lnum, int dtype); - -static inline int dbg_read(struct ubi_volume_desc *desc, int lnum, char *buf, - int offset, int len) -{ - return dbg_leb_read(desc, lnum, buf, offset, len, 0); -} - -static inline int dbg_write(struct ubi_volume_desc *desc, int lnum, - const void *buf, int offset, int len) -{ - return dbg_leb_write(desc, lnum, buf, offset, len, UBI_UNKNOWN); -} - -static inline int dbg_change(struct ubi_volume_desc *desc, int lnum, - const void *buf, int len) -{ - return dbg_leb_change(desc, lnum, buf, len, UBI_UNKNOWN); -} +int dbg_leb_write(struct ubifs_info *c, int lnum, const void *buf, int offs, + int len, int dtype); +int dbg_leb_change(struct ubifs_info *c, int lnum, const void *buf, int len, + int dtype); +int dbg_leb_unmap(struct ubifs_info *c, int lnum); +int dbg_leb_map(struct ubifs_info *c, int lnum, int dtype); /* Debugfs-related stuff */ int dbg_debugfs_init(void); @@ -314,7 +325,7 @@ void dbg_debugfs_exit_fs(struct ubifs_info *c); /* Use "if (0)" to make compiler check arguments even if debugging is off */ #define ubifs_assert(expr) do { \ - if (0 && (expr)) \ + if (0) \ printk(KERN_CRIT "UBIFS assert failed in %s at %u (pid %d)\n", \ __func__, __LINE__, current->pid); \ } while (0) @@ -324,9 +335,12 @@ void dbg_debugfs_exit_fs(struct ubifs_info *c); ubifs_err(fmt, ##__VA_ARGS__); \ } while (0) -#define ubifs_dbg_msg(fmt, ...) do { \ - if (0) \ - pr_debug(fmt "\n", ##__VA_ARGS__); \ +#define DBGKEY(key) ((char *)(key)) +#define DBGKEY1(key) ((char *)(key)) + +#define ubifs_dbg_msg(fmt, ...) do { \ + if (0) \ + printk(KERN_DEBUG fmt "\n", ##__VA_ARGS__); \ } while (0) #define dbg_dump_stack() @@ -347,9 +361,6 @@ void dbg_debugfs_exit_fs(struct ubifs_info *c); #define dbg_scan(fmt, ...) ubifs_dbg_msg(fmt, ##__VA_ARGS__) #define dbg_rcvry(fmt, ...) ubifs_dbg_msg(fmt, ##__VA_ARGS__) -#define DBGKEY(key) ((char *)(key)) -#define DBGKEY1(key) ((char *)(key)) - static inline int ubifs_debugging_init(struct ubifs_info *c) { return 0; } static inline void ubifs_debugging_exit(struct ubifs_info *c) { return; } static inline const char *dbg_ntype(int type) { return ""; } @@ -358,7 +369,7 @@ static inline const char *dbg_jhead(int jhead) { return ""; } static inline const char * dbg_get_key_dump(const struct ubifs_info *c, const union ubifs_key *key) { return ""; } -static inline void dbg_dump_inode(const struct ubifs_info *c, +static inline void dbg_dump_inode(struct ubifs_info *c, const struct inode *inode) { return; } static inline void dbg_dump_node(const struct ubifs_info *c, const void *node) { return; } @@ -379,6 +390,9 @@ static inline void dbg_dump_lpt_info(struct ubifs_info *c) { return; } static inline void dbg_dump_leb(const struct ubifs_info *c, int lnum) { return; } static inline void +dbg_dump_sleb(const struct ubifs_info *c, + const struct ubifs_scan_leb *sleb, int offs) { return; } +static inline void dbg_dump_znode(const struct ubifs_info *c, const struct ubifs_znode *znode) { return; } static inline void dbg_dump_heap(struct ubifs_info *c, @@ -410,9 +424,11 @@ static inline int dbg_check_ltab(struct ubifs_info *c) { return 0; } static inline int dbg_chk_lpt_free_spc(struct ubifs_info *c) { return 0; } static inline int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len) { return 0; } -static inline int dbg_check_synced_i_size(struct inode *inode) { return 0; } -static inline int dbg_check_dir_size(struct ubifs_info *c, - const struct inode *dir) { return 0; } +static inline int +dbg_check_synced_i_size(const struct ubifs_info *c, + struct inode *inode) { return 0; } +static inline int dbg_check_dir(struct ubifs_info *c, + const struct inode *dir) { return 0; } static inline int dbg_check_tnc(struct ubifs_info *c, int extra) { return 0; } static inline int dbg_check_idx_size(struct ubifs_info *c, long long idx_size) { return 0; } @@ -432,9 +448,23 @@ static inline int dbg_check_nondata_nodes_order(struct ubifs_info *c, struct list_head *head) { return 0; } -static inline int dbg_force_in_the_gaps(void) { return 0; } -#define dbg_force_in_the_gaps_enabled() 0 -#define dbg_failure_mode 0 +static inline int dbg_leb_write(struct ubifs_info *c, int lnum, + const void *buf, int offset, + int len, int dtype) { return 0; } +static inline int dbg_leb_change(struct ubifs_info *c, int lnum, + const void *buf, int len, + int dtype) { return 0; } +static inline int dbg_leb_unmap(struct ubifs_info *c, int lnum) { return 0; } +static inline int dbg_leb_map(struct ubifs_info *c, int lnum, + int dtype) { return 0; } + +static inline int dbg_is_chk_gen(const struct ubifs_info *c) { return 0; } +static inline int dbg_is_chk_index(const struct ubifs_info *c) { return 0; } +static inline int dbg_is_chk_orph(const struct ubifs_info *c) { return 0; } +static inline int dbg_is_chk_lprops(const struct ubifs_info *c) { return 0; } +static inline int dbg_is_chk_fs(const struct ubifs_info *c) { return 0; } +static inline int dbg_is_tst_rcvry(const struct ubifs_info *c) { return 0; } +static inline int dbg_is_power_cut(const struct ubifs_info *c) { return 0; } static inline int dbg_debugfs_init(void) { return 0; } static inline void dbg_debugfs_exit(void) { return; } diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c index 936a038..aaebf0f 100644 --- a/fs/ubifs/dir.c +++ b/fs/ubifs/dir.c @@ -102,7 +102,7 @@ struct inode *ubifs_new_inode(struct ubifs_info *c, const struct inode *dir, * UBIFS has to fully control "clean <-> dirty" transitions of inodes * to make budgeting work. */ - inode->i_flags |= (S_NOCMTIME); + inode->i_flags |= S_NOCMTIME; inode_init_owner(inode, dir, mode); inode->i_mtime = inode->i_atime = inode->i_ctime = @@ -172,9 +172,11 @@ struct inode *ubifs_new_inode(struct ubifs_info *c, const struct inode *dir, #ifdef CONFIG_UBIFS_FS_DEBUG -static int dbg_check_name(struct ubifs_dent_node *dent, struct qstr *nm) +static int dbg_check_name(const struct ubifs_info *c, + const struct ubifs_dent_node *dent, + const struct qstr *nm) { - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; if (le16_to_cpu(dent->nlen) != nm->len) return -EINVAL; @@ -185,7 +187,7 @@ static int dbg_check_name(struct ubifs_dent_node *dent, struct qstr *nm) #else -#define dbg_check_name(dent, nm) 0 +#define dbg_check_name(c, dent, nm) 0 #endif @@ -219,7 +221,7 @@ static struct dentry *ubifs_lookup(struct inode *dir, struct dentry *dentry, goto out; } - if (dbg_check_name(dent, &dentry->d_name)) { + if (dbg_check_name(c, dent, &dentry->d_name)) { err = -EINVAL; goto out; } @@ -546,7 +548,7 @@ static int ubifs_link(struct dentry *old_dentry, struct inode *dir, ubifs_assert(mutex_is_locked(&dir->i_mutex)); ubifs_assert(mutex_is_locked(&inode->i_mutex)); - err = dbg_check_synced_i_size(inode); + err = dbg_check_synced_i_size(c, inode); if (err) return err; @@ -601,7 +603,7 @@ static int ubifs_unlink(struct inode *dir, struct dentry *dentry) inode->i_nlink, dir->i_ino); ubifs_assert(mutex_is_locked(&dir->i_mutex)); ubifs_assert(mutex_is_locked(&inode->i_mutex)); - err = dbg_check_synced_i_size(inode); + err = dbg_check_synced_i_size(c, inode); if (err) return err; diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c index 5e7fccf..6589006 100644 --- a/fs/ubifs/file.c +++ b/fs/ubifs/file.c @@ -1263,7 +1263,7 @@ int ubifs_setattr(struct dentry *dentry, struct iattr *attr) if (err) return err; - err = dbg_check_synced_i_size(inode); + err = dbg_check_synced_i_size(c, inode); if (err) return err; @@ -1304,7 +1304,7 @@ static void *ubifs_follow_link(struct dentry *dentry, struct nameidata *nd) return NULL; } -int ubifs_fsync(struct file *file, int datasync) +int ubifs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; struct ubifs_info *c = inode->i_sb->s_fs_info; @@ -1319,14 +1319,16 @@ int ubifs_fsync(struct file *file, int datasync) */ return 0; - /* - * VFS has already synchronized dirty pages for this inode. Synchronize - * the inode unless this is a 'datasync()' call. - */ + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + mutex_lock(&inode->i_mutex); + + /* Synchronize the inode unless this is a 'datasync()' call. */ if (!datasync || (inode->i_state & I_DIRTY_DATASYNC)) { err = inode->i_sb->s_op->write_inode(inode, NULL); if (err) - return err; + goto out; } /* @@ -1334,10 +1336,9 @@ int ubifs_fsync(struct file *file, int datasync) * them. */ err = ubifs_sync_wbufs_by_inode(c, inode); - if (err) - return err; - - return 0; +out: + mutex_unlock(&inode->i_mutex); + return err; } /** @@ -1521,8 +1522,7 @@ static int ubifs_vm_page_mkwrite(struct vm_area_struct *vma, ubifs_release_dirty_inode_budget(c, ui); } - unlock_page(page); - return 0; + return VM_FAULT_LOCKED; out_unlock: unlock_page(page); diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c index 3be645e..9228950 100644 --- a/fs/ubifs/io.c +++ b/fs/ubifs/io.c @@ -86,8 +86,125 @@ void ubifs_ro_mode(struct ubifs_info *c, int err) c->no_chk_data_crc = 0; c->vfs_sb->s_flags |= MS_RDONLY; ubifs_warn("switched to read-only mode, error %d", err); + dump_stack(); + } +} + +/* + * Below are simple wrappers over UBI I/O functions which include some + * additional checks and UBIFS debugging stuff. See corresponding UBI function + * for more information. + */ + +int ubifs_leb_read(const struct ubifs_info *c, int lnum, void *buf, int offs, + int len, int even_ebadmsg) +{ + int err; + + err = ubi_read(c->ubi, lnum, buf, offs, len); + /* + * In case of %-EBADMSG print the error message only if the + * @even_ebadmsg is true. + */ + if (err && (err != -EBADMSG || even_ebadmsg)) { + ubifs_err("reading %d bytes from LEB %d:%d failed, error %d", + len, lnum, offs, err); + dbg_dump_stack(); + } + return err; +} + +int ubifs_leb_write(struct ubifs_info *c, int lnum, const void *buf, int offs, + int len, int dtype) +{ + int err; + + ubifs_assert(!c->ro_media && !c->ro_mount); + if (c->ro_error) + return -EROFS; + if (!dbg_is_tst_rcvry(c)) + err = ubi_leb_write(c->ubi, lnum, buf, offs, len, dtype); + else + err = dbg_leb_write(c, lnum, buf, offs, len, dtype); + if (err) { + ubifs_err("writing %d bytes to LEB %d:%d failed, error %d", + len, lnum, offs, err); + ubifs_ro_mode(c, err); + dbg_dump_stack(); + } + return err; +} + +int ubifs_leb_change(struct ubifs_info *c, int lnum, const void *buf, int len, + int dtype) +{ + int err; + + ubifs_assert(!c->ro_media && !c->ro_mount); + if (c->ro_error) + return -EROFS; + if (!dbg_is_tst_rcvry(c)) + err = ubi_leb_change(c->ubi, lnum, buf, len, dtype); + else + err = dbg_leb_change(c, lnum, buf, len, dtype); + if (err) { + ubifs_err("changing %d bytes in LEB %d failed, error %d", + len, lnum, err); + ubifs_ro_mode(c, err); + dbg_dump_stack(); + } + return err; +} + +int ubifs_leb_unmap(struct ubifs_info *c, int lnum) +{ + int err; + + ubifs_assert(!c->ro_media && !c->ro_mount); + if (c->ro_error) + return -EROFS; + if (!dbg_is_tst_rcvry(c)) + err = ubi_leb_unmap(c->ubi, lnum); + else + err = dbg_leb_unmap(c, lnum); + if (err) { + ubifs_err("unmap LEB %d failed, error %d", lnum, err); + ubifs_ro_mode(c, err); + dbg_dump_stack(); + } + return err; +} + +int ubifs_leb_map(struct ubifs_info *c, int lnum, int dtype) +{ + int err; + + ubifs_assert(!c->ro_media && !c->ro_mount); + if (c->ro_error) + return -EROFS; + if (!dbg_is_tst_rcvry(c)) + err = ubi_leb_map(c->ubi, lnum, dtype); + else + err = dbg_leb_map(c, lnum, dtype); + if (err) { + ubifs_err("mapping LEB %d failed, error %d", lnum, err); + ubifs_ro_mode(c, err); + dbg_dump_stack(); + } + return err; +} + +int ubifs_is_mapped(const struct ubifs_info *c, int lnum) +{ + int err; + + err = ubi_is_mapped(c->ubi, lnum); + if (err < 0) { + ubifs_err("ubi_is_mapped failed for LEB %d, error %d", + lnum, err); dbg_dump_stack(); } + return err; } /** @@ -406,14 +523,10 @@ int ubifs_wbuf_sync_nolock(struct ubifs_wbuf *wbuf) dirt = sync_len - wbuf->used; if (dirt) ubifs_pad(c, wbuf->buf + wbuf->used, dirt); - err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs, - sync_len, wbuf->dtype); - if (err) { - ubifs_err("cannot write %d bytes to LEB %d:%d", - sync_len, wbuf->lnum, wbuf->offs); - dbg_dump_stack(); + err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, wbuf->offs, sync_len, + wbuf->dtype); + if (err) return err; - } spin_lock(&wbuf->lock); wbuf->offs += sync_len; @@ -605,9 +718,9 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len) if (aligned_len == wbuf->avail) { dbg_io("flush jhead %s wbuf to LEB %d:%d", dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs); - err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, - wbuf->offs, wbuf->size, - wbuf->dtype); + err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, + wbuf->offs, wbuf->size, + wbuf->dtype); if (err) goto out; @@ -642,8 +755,8 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len) dbg_io("flush jhead %s wbuf to LEB %d:%d", dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs); memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail); - err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs, - wbuf->size, wbuf->dtype); + err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, wbuf->offs, + wbuf->size, wbuf->dtype); if (err) goto out; @@ -661,8 +774,8 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len) */ dbg_io("write %d bytes to LEB %d:%d", wbuf->size, wbuf->lnum, wbuf->offs); - err = ubi_leb_write(c->ubi, wbuf->lnum, buf, wbuf->offs, - wbuf->size, wbuf->dtype); + err = ubifs_leb_write(c, wbuf->lnum, buf, wbuf->offs, + wbuf->size, wbuf->dtype); if (err) goto out; @@ -683,8 +796,8 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len) n <<= c->max_write_shift; dbg_io("write %d bytes to LEB %d:%d", n, wbuf->lnum, wbuf->offs); - err = ubi_leb_write(c->ubi, wbuf->lnum, buf + written, - wbuf->offs, n, wbuf->dtype); + err = ubifs_leb_write(c, wbuf->lnum, buf + written, + wbuf->offs, n, wbuf->dtype); if (err) goto out; wbuf->offs += n; @@ -766,13 +879,9 @@ int ubifs_write_node(struct ubifs_info *c, void *buf, int len, int lnum, return -EROFS; ubifs_prepare_node(c, buf, len, 1); - err = ubi_leb_write(c->ubi, lnum, buf, offs, buf_len, dtype); - if (err) { - ubifs_err("cannot write %d bytes to LEB %d:%d, error %d", - buf_len, lnum, offs, err); + err = ubifs_leb_write(c, lnum, buf, offs, buf_len, dtype); + if (err) dbg_dump_node(c, buf); - dbg_dump_stack(); - } return err; } @@ -824,13 +933,9 @@ int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len, if (rlen > 0) { /* Read everything that goes before write-buffer */ - err = ubi_read(c->ubi, lnum, buf, offs, rlen); - if (err && err != -EBADMSG) { - ubifs_err("failed to read node %d from LEB %d:%d, " - "error %d", type, lnum, offs, err); - dbg_dump_stack(); + err = ubifs_leb_read(c, lnum, buf, offs, rlen, 0); + if (err && err != -EBADMSG) return err; - } } if (type != ch->node_type) { @@ -885,12 +990,9 @@ int ubifs_read_node(const struct ubifs_info *c, void *buf, int type, int len, ubifs_assert(!(offs & 7) && offs < c->leb_size); ubifs_assert(type >= 0 && type < UBIFS_NODE_TYPES_CNT); - err = ubi_read(c->ubi, lnum, buf, offs, len); - if (err && err != -EBADMSG) { - ubifs_err("cannot read node %d from LEB %d:%d, error %d", - type, lnum, offs, err); + err = ubifs_leb_read(c, lnum, buf, offs, len, 0); + if (err && err != -EBADMSG) return err; - } if (type != ch->node_type) { ubifs_err("bad node type (%d but expected %d)", diff --git a/fs/ubifs/log.c b/fs/ubifs/log.c index affea94..843beda 100644 --- a/fs/ubifs/log.c +++ b/fs/ubifs/log.c @@ -110,10 +110,14 @@ static inline long long empty_log_bytes(const struct ubifs_info *c) h = (long long)c->lhead_lnum * c->leb_size + c->lhead_offs; t = (long long)c->ltail_lnum * c->leb_size; - if (h >= t) + if (h > t) return c->log_bytes - h + t; - else + else if (h != t) return t - h; + else if (c->lhead_lnum != c->ltail_lnum) + return 0; + else + return c->log_bytes; } /** @@ -262,7 +266,7 @@ int ubifs_add_bud_to_log(struct ubifs_info *c, int jhead, int lnum, int offs) * an unclean reboot, because the target LEB might have been * unmapped, but not yet physically erased. */ - err = ubi_leb_map(c->ubi, bud->lnum, UBI_SHORTTERM); + err = ubifs_leb_map(c, bud->lnum, UBI_SHORTTERM); if (err) goto out_unlock; } @@ -283,8 +287,6 @@ int ubifs_add_bud_to_log(struct ubifs_info *c, int jhead, int lnum, int offs) return 0; out_unlock: - if (err != -EAGAIN) - ubifs_ro_mode(c, err); mutex_unlock(&c->log_mutex); kfree(ref); kfree(bud); @@ -455,9 +457,9 @@ out: * @ltail_lnum: new log tail LEB number * * This function is called on when the commit operation was finished. It - * moves log tail to new position and unmaps LEBs which contain obsolete data. - * Returns zero in case of success and a negative error code in case of - * failure. + * moves log tail to new position and updates the master node so that it stores + * the new log tail LEB number. Returns zero in case of success and a negative + * error code in case of failure. */ int ubifs_log_end_commit(struct ubifs_info *c, int ltail_lnum) { @@ -485,7 +487,12 @@ int ubifs_log_end_commit(struct ubifs_info *c, int ltail_lnum) spin_unlock(&c->buds_lock); err = dbg_check_bud_bytes(c); + if (err) + goto out; + err = ubifs_write_master(c); + +out: mutex_unlock(&c->log_mutex); return err; } @@ -752,7 +759,7 @@ static int dbg_check_bud_bytes(struct ubifs_info *c) struct ubifs_bud *bud; long long bud_bytes = 0; - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; spin_lock(&c->buds_lock); diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c index ae6f74a..ea9d491 100644 --- a/fs/ubifs/lprops.c +++ b/fs/ubifs/lprops.c @@ -510,7 +510,7 @@ static int is_lprops_dirty(struct ubifs_info *c, struct ubifs_lprops *lprops) pnode = (struct ubifs_pnode *)container_of(lprops - pos, struct ubifs_pnode, lprops[0]); - return !test_bit(COW_ZNODE, &pnode->flags) && + return !test_bit(COW_CNODE, &pnode->flags) && test_bit(DIRTY_CNODE, &pnode->flags); } @@ -866,7 +866,7 @@ int dbg_check_cats(struct ubifs_info *c) struct list_head *pos; int i, cat; - if (!(ubifs_chk_flags & (UBIFS_CHK_GEN | UBIFS_CHK_LPROPS))) + if (!dbg_is_chk_gen(c) && !dbg_is_chk_lprops(c)) return 0; list_for_each_entry(lprops, &c->empty_list, list) { @@ -964,7 +964,7 @@ void dbg_check_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat, { int i = 0, j, err = 0; - if (!(ubifs_chk_flags & (UBIFS_CHK_GEN | UBIFS_CHK_LPROPS))) + if (!dbg_is_chk_gen(c) && !dbg_is_chk_lprops(c)) return; for (i = 0; i < heap->cnt; i++) { @@ -1268,7 +1268,7 @@ int dbg_check_lprops(struct ubifs_info *c) int i, err; struct ubifs_lp_stats lst; - if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS)) + if (!dbg_is_chk_lprops(c)) return 0; /* diff --git a/fs/ubifs/lpt.c b/fs/ubifs/lpt.c index ef5155e..6189c74 100644 --- a/fs/ubifs/lpt.c +++ b/fs/ubifs/lpt.c @@ -701,8 +701,8 @@ int ubifs_create_dflt_lpt(struct ubifs_info *c, int *main_lebs, int lpt_first, alen = ALIGN(len, c->min_io_size); set_ltab(c, lnum, c->leb_size - alen, alen - len); memset(p, 0xff, alen - len); - err = ubi_leb_change(c->ubi, lnum++, buf, alen, - UBI_SHORTTERM); + err = ubifs_leb_change(c, lnum++, buf, alen, + UBI_SHORTTERM); if (err) goto out; p = buf; @@ -732,8 +732,8 @@ int ubifs_create_dflt_lpt(struct ubifs_info *c, int *main_lebs, int lpt_first, set_ltab(c, lnum, c->leb_size - alen, alen - len); memset(p, 0xff, alen - len); - err = ubi_leb_change(c->ubi, lnum++, buf, alen, - UBI_SHORTTERM); + err = ubifs_leb_change(c, lnum++, buf, alen, + UBI_SHORTTERM); if (err) goto out; p = buf; @@ -780,8 +780,8 @@ int ubifs_create_dflt_lpt(struct ubifs_info *c, int *main_lebs, int lpt_first, alen = ALIGN(len, c->min_io_size); set_ltab(c, lnum, c->leb_size - alen, alen - len); memset(p, 0xff, alen - len); - err = ubi_leb_change(c->ubi, lnum++, buf, alen, - UBI_SHORTTERM); + err = ubifs_leb_change(c, lnum++, buf, alen, + UBI_SHORTTERM); if (err) goto out; p = buf; @@ -806,7 +806,7 @@ int ubifs_create_dflt_lpt(struct ubifs_info *c, int *main_lebs, int lpt_first, alen = ALIGN(len, c->min_io_size); set_ltab(c, lnum, c->leb_size - alen, alen - len); memset(p, 0xff, alen - len); - err = ubi_leb_change(c->ubi, lnum++, buf, alen, UBI_SHORTTERM); + err = ubifs_leb_change(c, lnum++, buf, alen, UBI_SHORTTERM); if (err) goto out; p = buf; @@ -826,7 +826,7 @@ int ubifs_create_dflt_lpt(struct ubifs_info *c, int *main_lebs, int lpt_first, /* Write remaining buffer */ memset(p, 0xff, alen - len); - err = ubi_leb_change(c->ubi, lnum, buf, alen, UBI_SHORTTERM); + err = ubifs_leb_change(c, lnum, buf, alen, UBI_SHORTTERM); if (err) goto out; @@ -1222,7 +1222,7 @@ int ubifs_read_nnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip) if (c->big_lpt) nnode->num = calc_nnode_num_from_parent(c, parent, iip); } else { - err = ubi_read(c->ubi, lnum, buf, offs, c->nnode_sz); + err = ubifs_leb_read(c, lnum, buf, offs, c->nnode_sz, 1); if (err) goto out; err = ubifs_unpack_nnode(c, buf, nnode); @@ -1247,6 +1247,7 @@ int ubifs_read_nnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip) out: ubifs_err("error %d reading nnode at %d:%d", err, lnum, offs); + dbg_dump_stack(); kfree(nnode); return err; } @@ -1290,7 +1291,7 @@ static int read_pnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip) lprops->flags = ubifs_categorize_lprops(c, lprops); } } else { - err = ubi_read(c->ubi, lnum, buf, offs, c->pnode_sz); + err = ubifs_leb_read(c, lnum, buf, offs, c->pnode_sz, 1); if (err) goto out; err = unpack_pnode(c, buf, pnode); @@ -1312,6 +1313,7 @@ static int read_pnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip) out: ubifs_err("error %d reading pnode at %d:%d", err, lnum, offs); dbg_dump_pnode(c, pnode, parent, iip); + dbg_dump_stack(); dbg_msg("calc num: %d", calc_pnode_num_from_parent(c, parent, iip)); kfree(pnode); return err; @@ -1331,7 +1333,7 @@ static int read_ltab(struct ubifs_info *c) buf = vmalloc(c->ltab_sz); if (!buf) return -ENOMEM; - err = ubi_read(c->ubi, c->ltab_lnum, buf, c->ltab_offs, c->ltab_sz); + err = ubifs_leb_read(c, c->ltab_lnum, buf, c->ltab_offs, c->ltab_sz, 1); if (err) goto out; err = unpack_ltab(c, buf); @@ -1354,7 +1356,8 @@ static int read_lsave(struct ubifs_info *c) buf = vmalloc(c->lsave_sz); if (!buf) return -ENOMEM; - err = ubi_read(c->ubi, c->lsave_lnum, buf, c->lsave_offs, c->lsave_sz); + err = ubifs_leb_read(c, c->lsave_lnum, buf, c->lsave_offs, + c->lsave_sz, 1); if (err) goto out; err = unpack_lsave(c, buf); @@ -1814,8 +1817,8 @@ static struct ubifs_nnode *scan_get_nnode(struct ubifs_info *c, if (c->big_lpt) nnode->num = calc_nnode_num_from_parent(c, parent, iip); } else { - err = ubi_read(c->ubi, branch->lnum, buf, branch->offs, - c->nnode_sz); + err = ubifs_leb_read(c, branch->lnum, buf, branch->offs, + c->nnode_sz, 1); if (err) return ERR_PTR(err); err = ubifs_unpack_nnode(c, buf, nnode); @@ -1883,8 +1886,8 @@ static struct ubifs_pnode *scan_get_pnode(struct ubifs_info *c, ubifs_assert(branch->lnum >= c->lpt_first && branch->lnum <= c->lpt_last); ubifs_assert(branch->offs >= 0 && branch->offs < c->leb_size); - err = ubi_read(c->ubi, branch->lnum, buf, branch->offs, - c->pnode_sz); + err = ubifs_leb_read(c, branch->lnum, buf, branch->offs, + c->pnode_sz, 1); if (err) return ERR_PTR(err); err = unpack_pnode(c, buf, pnode); @@ -2224,7 +2227,7 @@ int dbg_check_lpt_nodes(struct ubifs_info *c, struct ubifs_cnode *cnode, struct ubifs_cnode *cn; int num, iip = 0, err; - if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS)) + if (!dbg_is_chk_lprops(c)) return 0; while (cnode) { diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c index dfcb574..cddd6bd 100644 --- a/fs/ubifs/lpt_commit.c +++ b/fs/ubifs/lpt_commit.c @@ -27,6 +27,7 @@ #include <linux/crc16.h> #include <linux/slab.h> +#include <linux/random.h> #include "ubifs.h" #ifdef CONFIG_UBIFS_FS_DEBUG @@ -116,8 +117,8 @@ static int get_cnodes_to_commit(struct ubifs_info *c) return 0; cnt += 1; while (1) { - ubifs_assert(!test_bit(COW_ZNODE, &cnode->flags)); - __set_bit(COW_ZNODE, &cnode->flags); + ubifs_assert(!test_bit(COW_CNODE, &cnode->flags)); + __set_bit(COW_CNODE, &cnode->flags); cnext = next_dirty_cnode(cnode); if (!cnext) { cnode->cnext = c->lpt_cnext; @@ -465,7 +466,7 @@ static int write_cnodes(struct ubifs_info *c) */ clear_bit(DIRTY_CNODE, &cnode->flags); smp_mb__before_clear_bit(); - clear_bit(COW_ZNODE, &cnode->flags); + clear_bit(COW_CNODE, &cnode->flags); smp_mb__after_clear_bit(); offs += len; dbg_chk_lpt_sz(c, 1, len); @@ -1160,11 +1161,11 @@ static int lpt_gc_lnum(struct ubifs_info *c, int lnum) void *buf = c->lpt_buf; dbg_lp("LEB %d", lnum); - err = ubi_read(c->ubi, lnum, buf, 0, c->leb_size); - if (err) { - ubifs_err("cannot read LEB %d, error %d", lnum, err); + + err = ubifs_leb_read(c, lnum, buf, 0, c->leb_size, 1); + if (err) return err; - } + while (1) { if (!is_a_node(c, buf, len)) { int pad_len; @@ -1640,7 +1641,7 @@ static int dbg_check_ltab_lnum(struct ubifs_info *c, int lnum) int ret; void *buf, *p; - if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS)) + if (!dbg_is_chk_lprops(c)) return 0; buf = p = __vmalloc(c->leb_size, GFP_NOFS, PAGE_KERNEL); @@ -1650,11 +1651,11 @@ static int dbg_check_ltab_lnum(struct ubifs_info *c, int lnum) } dbg_lp("LEB %d", lnum); - err = ubi_read(c->ubi, lnum, buf, 0, c->leb_size); - if (err) { - dbg_msg("ubi_read failed, LEB %d, error %d", lnum, err); + + err = ubifs_leb_read(c, lnum, buf, 0, c->leb_size, 1); + if (err) goto out; - } + while (1) { if (!is_a_node(c, p, len)) { int i, pad_len; @@ -1711,7 +1712,7 @@ int dbg_check_ltab(struct ubifs_info *c) { int lnum, err, i, cnt; - if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS)) + if (!dbg_is_chk_lprops(c)) return 0; /* Bring the entire tree into memory */ @@ -1754,7 +1755,7 @@ int dbg_chk_lpt_free_spc(struct ubifs_info *c) long long free = 0; int i; - if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS)) + if (!dbg_is_chk_lprops(c)) return 0; for (i = 0; i < c->lpt_lebs; i++) { @@ -1796,7 +1797,7 @@ int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len) long long chk_lpt_sz, lpt_sz; int err = 0; - if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS)) + if (!dbg_is_chk_lprops(c)) return 0; switch (action) { @@ -1901,11 +1902,10 @@ static void dump_lpt_leb(const struct ubifs_info *c, int lnum) return; } - err = ubi_read(c->ubi, lnum, buf, 0, c->leb_size); - if (err) { - ubifs_err("cannot read LEB %d, error %d", lnum, err); + err = ubifs_leb_read(c, lnum, buf, 0, c->leb_size, 1); + if (err) goto out; - } + while (1) { offs = c->leb_size - len; if (!is_a_node(c, p, len)) { @@ -2019,7 +2019,7 @@ static int dbg_populate_lsave(struct ubifs_info *c) struct ubifs_lpt_heap *heap; int i; - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; if (random32() & 3) return 0; diff --git a/fs/ubifs/master.c b/fs/ubifs/master.c index 278c238..bb9f481 100644 --- a/fs/ubifs/master.c +++ b/fs/ubifs/master.c @@ -352,10 +352,9 @@ int ubifs_read_master(struct ubifs_info *c) * ubifs_write_master - write master node. * @c: UBIFS file-system description object * - * This function writes the master node. The caller has to take the - * @c->mst_mutex lock before calling this function. Returns zero in case of - * success and a negative error code in case of failure. The master node is - * written twice to enable recovery. + * This function writes the master node. Returns zero in case of success and a + * negative error code in case of failure. The master node is written twice to + * enable recovery. */ int ubifs_write_master(struct ubifs_info *c) { diff --git a/fs/ubifs/misc.h b/fs/ubifs/misc.h index 0b5296a..ee7cb5e 100644 --- a/fs/ubifs/misc.h +++ b/fs/ubifs/misc.h @@ -39,6 +39,29 @@ static inline int ubifs_zn_dirty(const struct ubifs_znode *znode) } /** + * ubifs_zn_obsolete - check if znode is obsolete. + * @znode: znode to check + * + * This helper function returns %1 if @znode is obsolete and %0 otherwise. + */ +static inline int ubifs_zn_obsolete(const struct ubifs_znode *znode) +{ + return !!test_bit(OBSOLETE_ZNODE, &znode->flags); +} + +/** + * ubifs_zn_cow - check if znode has to be copied on write. + * @znode: znode to check + * + * This helper function returns %1 if @znode is has COW flag set and %0 + * otherwise. + */ +static inline int ubifs_zn_cow(const struct ubifs_znode *znode) +{ + return !!test_bit(COW_ZNODE, &znode->flags); +} + +/** * ubifs_wake_up_bgt - wake up background thread. * @c: UBIFS file-system description object */ @@ -122,86 +145,6 @@ static inline int ubifs_wbuf_sync(struct ubifs_wbuf *wbuf) } /** - * ubifs_leb_unmap - unmap an LEB. - * @c: UBIFS file-system description object - * @lnum: LEB number to unmap - * - * This function returns %0 on success and a negative error code on failure. - */ -static inline int ubifs_leb_unmap(const struct ubifs_info *c, int lnum) -{ - int err; - - ubifs_assert(!c->ro_media && !c->ro_mount); - if (c->ro_error) - return -EROFS; - err = ubi_leb_unmap(c->ubi, lnum); - if (err) { - ubifs_err("unmap LEB %d failed, error %d", lnum, err); - return err; - } - - return 0; -} - -/** - * ubifs_leb_write - write to a LEB. - * @c: UBIFS file-system description object - * @lnum: LEB number to write - * @buf: buffer to write from - * @offs: offset within LEB to write to - * @len: length to write - * @dtype: data type - * - * This function returns %0 on success and a negative error code on failure. - */ -static inline int ubifs_leb_write(const struct ubifs_info *c, int lnum, - const void *buf, int offs, int len, int dtype) -{ - int err; - - ubifs_assert(!c->ro_media && !c->ro_mount); - if (c->ro_error) - return -EROFS; - err = ubi_leb_write(c->ubi, lnum, buf, offs, len, dtype); - if (err) { - ubifs_err("writing %d bytes at %d:%d, error %d", - len, lnum, offs, err); - return err; - } - - return 0; -} - -/** - * ubifs_leb_change - atomic LEB change. - * @c: UBIFS file-system description object - * @lnum: LEB number to write - * @buf: buffer to write from - * @len: length to write - * @dtype: data type - * - * This function returns %0 on success and a negative error code on failure. - */ -static inline int ubifs_leb_change(const struct ubifs_info *c, int lnum, - const void *buf, int len, int dtype) -{ - int err; - - ubifs_assert(!c->ro_media && !c->ro_mount); - if (c->ro_error) - return -EROFS; - err = ubi_leb_change(c->ubi, lnum, buf, len, dtype); - if (err) { - ubifs_err("changing %d bytes in LEB %d, error %d", - len, lnum, err); - return err; - } - - return 0; -} - -/** * ubifs_encode_dev - encode device node IDs. * @dev: UBIFS device node information * @rdev: device IDs to encode diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c index a5422ff..f9c90b5 100644 --- a/fs/ubifs/orphan.c +++ b/fs/ubifs/orphan.c @@ -130,13 +130,14 @@ void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum) else if (inum > o->inum) p = p->rb_right; else { - if (o->dnext) { + if (o->del) { spin_unlock(&c->orphan_lock); dbg_gen("deleted twice ino %lu", (unsigned long)inum); return; } if (o->cnext) { + o->del = 1; o->dnext = c->orph_dnext; c->orph_dnext = o; spin_unlock(&c->orphan_lock); @@ -447,6 +448,7 @@ static void erase_deleted(struct ubifs_info *c) orphan = dnext; dnext = orphan->dnext; ubifs_assert(!orphan->new); + ubifs_assert(orphan->del); rb_erase(&orphan->rb, &c->orph_tree); list_del(&orphan->list); c->tot_orphans -= 1; @@ -536,6 +538,7 @@ static int insert_dead_orphan(struct ubifs_info *c, ino_t inum) rb_link_node(&orphan->rb, parent, p); rb_insert_color(&orphan->rb, &c->orph_tree); list_add_tail(&orphan->list, &c->orph_list); + orphan->del = 1; orphan->dnext = c->orph_dnext; c->orph_dnext = orphan; dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum, @@ -929,7 +932,7 @@ static int dbg_check_orphans(struct ubifs_info *c) struct check_info ci; int err; - if (!(ubifs_chk_flags & UBIFS_CHK_ORPH)) + if (!dbg_is_chk_orph(c)) return 0; ci.last_ino = 0; diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c index 783d8e0..ee4f43f 100644 --- a/fs/ubifs/recovery.c +++ b/fs/ubifs/recovery.c @@ -117,7 +117,7 @@ static int get_master_node(const struct ubifs_info *c, int lnum, void **pbuf, if (!sbuf) return -ENOMEM; - err = ubi_read(c->ubi, lnum, sbuf, 0, c->leb_size); + err = ubifs_leb_read(c, lnum, sbuf, 0, c->leb_size, 0); if (err && err != -EBADMSG) goto out_free; @@ -213,10 +213,10 @@ static int write_rcvrd_mst_node(struct ubifs_info *c, mst->flags |= cpu_to_le32(UBIFS_MST_RCVRY); ubifs_prepare_node(c, mst, UBIFS_MST_NODE_SZ, 1); - err = ubi_leb_change(c->ubi, lnum, mst, sz, UBI_SHORTTERM); + err = ubifs_leb_change(c, lnum, mst, sz, UBI_SHORTTERM); if (err) goto out; - err = ubi_leb_change(c->ubi, lnum + 1, mst, sz, UBI_SHORTTERM); + err = ubifs_leb_change(c, lnum + 1, mst, sz, UBI_SHORTTERM); if (err) goto out; out: @@ -274,7 +274,8 @@ int ubifs_recover_master_node(struct ubifs_info *c) if (cor1) goto out_err; mst = mst1; - } else if (offs1 == 0 && offs2 + sz >= c->leb_size) { + } else if (offs1 == 0 && + c->leb_size - offs2 - sz < sz) { /* 1st LEB was unmapped and written, 2nd not */ if (cor1) goto out_err; @@ -539,8 +540,8 @@ static int fix_unclean_leb(struct ubifs_info *c, struct ubifs_scan_leb *sleb, int len = ALIGN(endpt, c->min_io_size); if (start) { - err = ubi_read(c->ubi, lnum, sleb->buf, 0, - start); + err = ubifs_leb_read(c, lnum, sleb->buf, 0, + start, 1); if (err) return err; } @@ -554,8 +555,8 @@ static int fix_unclean_leb(struct ubifs_info *c, struct ubifs_scan_leb *sleb, ubifs_pad(c, buf, pad_len); } } - err = ubi_leb_change(c->ubi, lnum, sleb->buf, len, - UBI_UNKNOWN); + err = ubifs_leb_change(c, lnum, sleb->buf, len, + UBI_UNKNOWN); if (err) return err; } @@ -819,7 +820,8 @@ static int get_cs_sqnum(struct ubifs_info *c, int lnum, int offs, return -ENOMEM; if (c->leb_size - offs < UBIFS_CS_NODE_SZ) goto out_err; - err = ubi_read(c->ubi, lnum, (void *)cs_node, offs, UBIFS_CS_NODE_SZ); + err = ubifs_leb_read(c, lnum, (void *)cs_node, offs, + UBIFS_CS_NODE_SZ, 0); if (err && err != -EBADMSG) goto out_free; ret = ubifs_scan_a_node(c, cs_node, UBIFS_CS_NODE_SZ, lnum, offs, 0); @@ -919,8 +921,7 @@ struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum, * * This function returns %0 on success and a negative error code on failure. */ -static int recover_head(const struct ubifs_info *c, int lnum, int offs, - void *sbuf) +static int recover_head(struct ubifs_info *c, int lnum, int offs, void *sbuf) { int len = c->max_write_size, err; @@ -931,15 +932,15 @@ static int recover_head(const struct ubifs_info *c, int lnum, int offs, return 0; /* Read at the head location and check it is empty flash */ - err = ubi_read(c->ubi, lnum, sbuf, offs, len); + err = ubifs_leb_read(c, lnum, sbuf, offs, len, 1); if (err || !is_empty(sbuf, len)) { dbg_rcvry("cleaning head at %d:%d", lnum, offs); if (offs == 0) return ubifs_leb_unmap(c, lnum); - err = ubi_read(c->ubi, lnum, sbuf, 0, offs); + err = ubifs_leb_read(c, lnum, sbuf, 0, offs, 1); if (err) return err; - return ubi_leb_change(c->ubi, lnum, sbuf, offs, UBI_UNKNOWN); + return ubifs_leb_change(c, lnum, sbuf, offs, UBI_UNKNOWN); } return 0; @@ -962,7 +963,7 @@ static int recover_head(const struct ubifs_info *c, int lnum, int offs, * * This function returns %0 on success and a negative error code on failure. */ -int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf) +int ubifs_recover_inl_heads(struct ubifs_info *c, void *sbuf) { int err; @@ -982,7 +983,7 @@ int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf) } /** - * clean_an_unclean_leb - read and write a LEB to remove corruption. + * clean_an_unclean_leb - read and write a LEB to remove corruption. * @c: UBIFS file-system description object * @ucleb: unclean LEB information * @sbuf: LEB-sized buffer to use @@ -993,7 +994,7 @@ int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf) * * This function returns %0 on success and a negative error code on failure. */ -static int clean_an_unclean_leb(const struct ubifs_info *c, +static int clean_an_unclean_leb(struct ubifs_info *c, struct ubifs_unclean_leb *ucleb, void *sbuf) { int err, lnum = ucleb->lnum, offs = 0, len = ucleb->endpt, quiet = 1; @@ -1009,7 +1010,7 @@ static int clean_an_unclean_leb(const struct ubifs_info *c, return 0; } - err = ubi_read(c->ubi, lnum, buf, offs, len); + err = ubifs_leb_read(c, lnum, buf, offs, len, 0); if (err && err != -EBADMSG) return err; @@ -1069,7 +1070,7 @@ static int clean_an_unclean_leb(const struct ubifs_info *c, } /* Write back the LEB atomically */ - err = ubi_leb_change(c->ubi, lnum, sbuf, len, UBI_UNKNOWN); + err = ubifs_leb_change(c, lnum, sbuf, len, UBI_UNKNOWN); if (err) return err; @@ -1089,7 +1090,7 @@ static int clean_an_unclean_leb(const struct ubifs_info *c, * * This function returns %0 on success and a negative error code on failure. */ -int ubifs_clean_lebs(const struct ubifs_info *c, void *sbuf) +int ubifs_clean_lebs(struct ubifs_info *c, void *sbuf) { dbg_rcvry("recovery"); while (!list_empty(&c->unclean_leb_list)) { @@ -1454,7 +1455,7 @@ static int fix_size_in_place(struct ubifs_info *c, struct size_entry *e) if (i_size >= e->d_size) return 0; /* Read the LEB */ - err = ubi_read(c->ubi, lnum, c->sbuf, 0, c->leb_size); + err = ubifs_leb_read(c, lnum, c->sbuf, 0, c->leb_size, 1); if (err) goto out; /* Change the size field and recalculate the CRC */ @@ -1470,7 +1471,7 @@ static int fix_size_in_place(struct ubifs_info *c, struct size_entry *e) len -= 1; len = ALIGN(len + 1, c->min_io_size); /* Atomically write the fixed LEB back again */ - err = ubi_leb_change(c->ubi, lnum, c->sbuf, len, UBI_UNKNOWN); + err = ubifs_leb_change(c, lnum, c->sbuf, len, UBI_UNKNOWN); if (err) goto out; dbg_rcvry("inode %lu at %d:%d size %lld -> %lld", diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c index 5e97161..ccabaf1 100644 --- a/fs/ubifs/replay.c +++ b/fs/ubifs/replay.c @@ -523,8 +523,7 @@ static int is_last_bud(struct ubifs_info *c, struct ubifs_bud *bud) if (!list_is_last(&next->list, &jh->buds_list)) return 0; - err = ubi_read(c->ubi, next->lnum, (char *)&data, - next->start, 4); + err = ubifs_leb_read(c, next->lnum, (char *)&data, next->start, 4, 1); if (err) return 0; diff --git a/fs/ubifs/sb.c b/fs/ubifs/sb.c index 1250016..b73ecd8 100644 --- a/fs/ubifs/sb.c +++ b/fs/ubifs/sb.c @@ -247,7 +247,7 @@ static int create_default_filesystem(struct ubifs_info *c) mst->total_dirty = cpu_to_le64(tmp64); /* The indexing LEB does not contribute to dark space */ - tmp64 = (c->main_lebs - 1) * c->dark_wm; + tmp64 = ((long long)(c->main_lebs - 1) * c->dark_wm); mst->total_dark = cpu_to_le64(tmp64); mst->total_used = cpu_to_le64(UBIFS_INO_NODE_SZ); @@ -674,15 +674,15 @@ static int fixup_leb(struct ubifs_info *c, int lnum, int len) if (len == 0) { dbg_mnt("unmap empty LEB %d", lnum); - return ubi_leb_unmap(c->ubi, lnum); + return ubifs_leb_unmap(c, lnum); } dbg_mnt("fixup LEB %d, data len %d", lnum, len); - err = ubi_read(c->ubi, lnum, c->sbuf, 0, len); + err = ubifs_leb_read(c, lnum, c->sbuf, 0, len, 1); if (err) return err; - return ubi_leb_change(c->ubi, lnum, c->sbuf, len, UBI_UNKNOWN); + return ubifs_leb_change(c, lnum, c->sbuf, len, UBI_UNKNOWN); } /** diff --git a/fs/ubifs/scan.c b/fs/ubifs/scan.c index 36216b4..37383e8 100644 --- a/fs/ubifs/scan.c +++ b/fs/ubifs/scan.c @@ -148,7 +148,7 @@ struct ubifs_scan_leb *ubifs_start_scan(const struct ubifs_info *c, int lnum, INIT_LIST_HEAD(&sleb->nodes); sleb->buf = sbuf; - err = ubi_read(c->ubi, lnum, sbuf + offs, offs, c->leb_size - offs); + err = ubifs_leb_read(c, lnum, sbuf + offs, offs, c->leb_size - offs, 0); if (err && err != -EBADMSG) { ubifs_err("cannot read %d bytes from LEB %d:%d," " error %d", c->leb_size - offs, lnum, offs, err); @@ -240,7 +240,7 @@ void ubifs_scanned_corruption(const struct ubifs_info *c, int lnum, int offs, int len; ubifs_err("corruption at LEB %d:%d", lnum, offs); - if (dbg_failure_mode) + if (dbg_is_tst_rcvry(c)) return; len = c->leb_size - offs; if (len > 8192) diff --git a/fs/ubifs/shrinker.c b/fs/ubifs/shrinker.c index 9e1d056..e0a7a76 100644 --- a/fs/ubifs/shrinker.c +++ b/fs/ubifs/shrinker.c @@ -128,7 +128,6 @@ static int shrink_tnc(struct ubifs_info *c, int nr, int age, int *contention) freed = ubifs_destroy_tnc_subtree(znode); atomic_long_sub(freed, &ubifs_clean_zn_cnt); atomic_long_sub(freed, &c->clean_zn_cnt); - ubifs_assert(atomic_long_read(&c->clean_zn_cnt) >= 0); total_freed += freed; znode = zprev; } diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index db04976..201bcfc 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -85,7 +85,7 @@ static int validate_inode(struct ubifs_info *c, const struct inode *inode) if (ui->data_len < 0 || ui->data_len > UBIFS_MAX_INO_DATA) return 4; - if (ui->xattr && (inode->i_mode & S_IFMT) != S_IFREG) + if (ui->xattr && !S_ISREG(inode->i_mode)) return 5; if (!ubifs_compr_present(ui->compr_type)) { @@ -94,7 +94,7 @@ static int validate_inode(struct ubifs_info *c, const struct inode *inode) ubifs_compr_name(ui->compr_type)); } - err = dbg_check_dir_size(c, inode); + err = dbg_check_dir(c, inode); return err; } @@ -129,7 +129,7 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum) goto out_ino; inode->i_flags |= (S_NOCMTIME | S_NOATIME); - inode->i_nlink = le32_to_cpu(ino->nlink); + set_nlink(inode, le32_to_cpu(ino->nlink)); inode->i_uid = le32_to_cpu(ino->uid); inode->i_gid = le32_to_cpu(ino->gid); inode->i_atime.tv_sec = (int64_t)le64_to_cpu(ino->atime_sec); @@ -914,7 +914,7 @@ static int check_volume_empty(struct ubifs_info *c) c->empty = 1; for (lnum = 0; lnum < c->leb_cnt; lnum++) { - err = ubi_is_mapped(c->ubi, lnum); + err = ubifs_is_mapped(c, lnum); if (unlikely(err < 0)) return err; if (err == 1) { @@ -1985,7 +1985,6 @@ static struct ubifs_info *alloc_ubifs_info(struct ubi_volume_desc *ubi) mutex_init(&c->lp_mutex); mutex_init(&c->tnc_mutex); mutex_init(&c->log_mutex); - mutex_init(&c->mst_mutex); mutex_init(&c->umount_mutex); mutex_init(&c->bu_mutex); mutex_init(&c->write_reserve_mutex); @@ -2264,19 +2263,12 @@ static int __init ubifs_init(void) return -EINVAL; } - err = register_filesystem(&ubifs_fs_type); - if (err) { - ubifs_err("cannot register file system, error %d", err); - return err; - } - - err = -ENOMEM; ubifs_inode_slab = kmem_cache_create("ubifs_inode_slab", sizeof(struct ubifs_inode), 0, SLAB_MEM_SPREAD | SLAB_RECLAIM_ACCOUNT, &inode_slab_ctor); if (!ubifs_inode_slab) - goto out_reg; + return -ENOMEM; register_shrinker(&ubifs_shrinker_info); @@ -2288,15 +2280,20 @@ static int __init ubifs_init(void) if (err) goto out_compr; + err = register_filesystem(&ubifs_fs_type); + if (err) { + ubifs_err("cannot register file system, error %d", err); + goto out_dbg; + } return 0; +out_dbg: + dbg_debugfs_exit(); out_compr: ubifs_compressors_exit(); out_shrinker: unregister_shrinker(&ubifs_shrinker_info); kmem_cache_destroy(ubifs_inode_slab); -out_reg: - unregister_filesystem(&ubifs_fs_type); return err; } /* late_initcall to let compressors initialize first */ diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c index 91b4213..0667386 100644 --- a/fs/ubifs/tnc.c +++ b/fs/ubifs/tnc.c @@ -223,7 +223,7 @@ static struct ubifs_znode *copy_znode(struct ubifs_info *c, __set_bit(DIRTY_ZNODE, &zn->flags); __clear_bit(COW_ZNODE, &zn->flags); - ubifs_assert(!test_bit(OBSOLETE_ZNODE, &znode->flags)); + ubifs_assert(!ubifs_zn_obsolete(znode)); __set_bit(OBSOLETE_ZNODE, &znode->flags); if (znode->level != 0) { @@ -271,7 +271,7 @@ static struct ubifs_znode *dirty_cow_znode(struct ubifs_info *c, struct ubifs_znode *zn; int err; - if (!test_bit(COW_ZNODE, &znode->flags)) { + if (!ubifs_zn_cow(znode)) { /* znode is not being committed */ if (!test_and_set_bit(DIRTY_ZNODE, &znode->flags)) { atomic_long_inc(&c->dirty_zn_cnt); @@ -462,7 +462,7 @@ static int try_read_node(const struct ubifs_info *c, void *buf, int type, dbg_io("LEB %d:%d, %s, length %d", lnum, offs, dbg_ntype(type), len); - err = ubi_read(c->ubi, lnum, buf, offs, len); + err = ubifs_leb_read(c, lnum, buf, offs, len, 1); if (err) { ubifs_err("cannot read node type %d from LEB %d:%d, error %d", type, lnum, offs, err); @@ -1666,7 +1666,7 @@ static int read_wbuf(struct ubifs_wbuf *wbuf, void *buf, int len, int lnum, if (!overlap) { /* We may safely unlock the write-buffer and read the data */ spin_unlock(&wbuf->lock); - return ubi_read(c->ubi, lnum, buf, offs, len); + return ubifs_leb_read(c, lnum, buf, offs, len, 0); } /* Don't read under wbuf */ @@ -1680,7 +1680,7 @@ static int read_wbuf(struct ubifs_wbuf *wbuf, void *buf, int len, int lnum, if (rlen > 0) /* Read everything that goes before write-buffer */ - return ubi_read(c->ubi, lnum, buf, offs, rlen); + return ubifs_leb_read(c, lnum, buf, offs, rlen, 0); return 0; } @@ -1767,7 +1767,7 @@ int ubifs_tnc_bulk_read(struct ubifs_info *c, struct bu_info *bu) if (wbuf) err = read_wbuf(wbuf, bu->buf, len, lnum, offs); else - err = ubi_read(c->ubi, lnum, bu->buf, offs, len); + err = ubifs_leb_read(c, lnum, bu->buf, offs, len, 0); /* Check for a race with GC */ if (maybe_leb_gced(c, lnum, bu->gc_seq)) @@ -2423,7 +2423,7 @@ static int tnc_delete(struct ubifs_info *c, struct ubifs_znode *znode, int n) */ do { - ubifs_assert(!test_bit(OBSOLETE_ZNODE, &znode->flags)); + ubifs_assert(!ubifs_zn_obsolete(znode)); ubifs_assert(ubifs_zn_dirty(znode)); zp = znode->parent; @@ -2479,9 +2479,8 @@ static int tnc_delete(struct ubifs_info *c, struct ubifs_znode *znode, int n) c->zroot.offs = zbr->offs; c->zroot.len = zbr->len; c->zroot.znode = znode; - ubifs_assert(!test_bit(OBSOLETE_ZNODE, - &zp->flags)); - ubifs_assert(test_bit(DIRTY_ZNODE, &zp->flags)); + ubifs_assert(!ubifs_zn_obsolete(zp)); + ubifs_assert(ubifs_zn_dirty(zp)); atomic_long_dec(&c->dirty_zn_cnt); if (zp->cnext) { @@ -2865,7 +2864,7 @@ static void tnc_destroy_cnext(struct ubifs_info *c) struct ubifs_znode *znode = cnext; cnext = cnext->cnext; - if (test_bit(OBSOLETE_ZNODE, &znode->flags)) + if (ubifs_zn_obsolete(znode)) kfree(znode); } while (cnext && cnext != c->cnext); } @@ -3301,7 +3300,7 @@ int dbg_check_inode_size(struct ubifs_info *c, const struct inode *inode, if (!S_ISREG(inode->i_mode)) return 0; - if (!(ubifs_chk_flags & UBIFS_CHK_GEN)) + if (!dbg_is_chk_gen(c)) return 0; block = (size + UBIFS_BLOCK_SIZE - 1) >> UBIFS_BLOCK_SHIFT; @@ -3337,9 +3336,10 @@ out_dump: ubifs_err("inode %lu has size %lld, but there are data at offset %lld " "(data key %s)", (unsigned long)inode->i_ino, size, ((loff_t)block) << UBIFS_BLOCK_SHIFT, DBGKEY(key)); + mutex_unlock(&c->tnc_mutex); dbg_dump_inode(c, inode); dbg_dump_stack(); - err = -EINVAL; + return -EINVAL; out_unlock: mutex_unlock(&c->tnc_mutex); diff --git a/fs/ubifs/tnc_commit.c b/fs/ubifs/tnc_commit.c index 41920f3..4c15f07 100644 --- a/fs/ubifs/tnc_commit.c +++ b/fs/ubifs/tnc_commit.c @@ -22,6 +22,7 @@ /* This file implements TNC functions for committing */ +#include <linux/random.h> #include "ubifs.h" /** @@ -87,8 +88,12 @@ static int make_idx_node(struct ubifs_info *c, struct ubifs_idx_node *idx, atomic_long_dec(&c->dirty_zn_cnt); ubifs_assert(ubifs_zn_dirty(znode)); - ubifs_assert(test_bit(COW_ZNODE, &znode->flags)); + ubifs_assert(ubifs_zn_cow(znode)); + /* + * Note, unlike 'write_index()' we do not add memory barriers here + * because this function is called with @c->tnc_mutex locked. + */ __clear_bit(DIRTY_ZNODE, &znode->flags); __clear_bit(COW_ZNODE, &znode->flags); @@ -377,7 +382,7 @@ static int layout_in_gaps(struct ubifs_info *c, int cnt) c->gap_lebs = NULL; return err; } - if (dbg_force_in_the_gaps_enabled()) { + if (!dbg_is_chk_index(c)) { /* * Do not print scary warnings if the debugging * option which forces in-the-gaps is enabled. @@ -491,25 +496,6 @@ static int layout_in_empty_space(struct ubifs_info *c) else next_len = ubifs_idx_node_sz(c, cnext->child_cnt); - if (c->min_io_size == 1) { - buf_offs += ALIGN(len, 8); - if (next_len) { - if (buf_offs + next_len <= c->leb_size) - continue; - err = ubifs_update_one_lp(c, lnum, 0, - c->leb_size - buf_offs, 0, 0); - if (err) - return err; - lnum = -1; - continue; - } - err = ubifs_update_one_lp(c, lnum, - c->leb_size - buf_offs, 0, 0, 0); - if (err) - return err; - break; - } - /* Update buffer positions */ wlen = used + len; used += ALIGN(len, 8); @@ -658,7 +644,7 @@ static int get_znodes_to_commit(struct ubifs_info *c) } cnt += 1; while (1) { - ubifs_assert(!test_bit(COW_ZNODE, &znode->flags)); + ubifs_assert(!ubifs_zn_cow(znode)); __set_bit(COW_ZNODE, &znode->flags); znode->alt = 0; cnext = find_next_dirty(znode); @@ -704,7 +690,7 @@ static int alloc_idx_lebs(struct ubifs_info *c, int cnt) c->ilebs[c->ileb_cnt++] = lnum; dbg_cmt("LEB %d", lnum); } - if (dbg_force_in_the_gaps()) + if (dbg_is_chk_index(c) && !(random32() & 7)) return -ENOSPC; return 0; } @@ -830,7 +816,7 @@ static int write_index(struct ubifs_info *c) struct ubifs_idx_node *idx; struct ubifs_znode *znode, *cnext; int i, lnum, offs, len, next_len, buf_len, buf_offs, used; - int avail, wlen, err, lnum_pos = 0; + int avail, wlen, err, lnum_pos = 0, blen, nxt_offs; cnext = c->enext; if (!cnext) @@ -907,7 +893,7 @@ static int write_index(struct ubifs_info *c) cnext = znode->cnext; ubifs_assert(ubifs_zn_dirty(znode)); - ubifs_assert(test_bit(COW_ZNODE, &znode->flags)); + ubifs_assert(ubifs_zn_cow(znode)); /* * It is important that other threads should see %DIRTY_ZNODE @@ -922,6 +908,28 @@ static int write_index(struct ubifs_info *c) clear_bit(COW_ZNODE, &znode->flags); smp_mb__after_clear_bit(); + /* + * We have marked the znode as clean but have not updated the + * @c->clean_zn_cnt counter. If this znode becomes dirty again + * before 'free_obsolete_znodes()' is called, then + * @c->clean_zn_cnt will be decremented before it gets + * incremented (resulting in 2 decrements for the same znode). + * This means that @c->clean_zn_cnt may become negative for a + * while. + * + * Q: why we cannot increment @c->clean_zn_cnt? + * A: because we do not have the @c->tnc_mutex locked, and the + * following code would be racy and buggy: + * + * if (!ubifs_zn_obsolete(znode)) { + * atomic_long_inc(&c->clean_zn_cnt); + * atomic_long_inc(&ubifs_clean_zn_cnt); + * } + * + * Thus, we just delay the @c->clean_zn_cnt update until we + * have the mutex locked. + */ + /* Do not access znode from this point on */ /* Update buffer positions */ @@ -938,65 +946,38 @@ static int write_index(struct ubifs_info *c) else next_len = ubifs_idx_node_sz(c, cnext->child_cnt); - if (c->min_io_size == 1) { - /* - * Write the prepared index node immediately if there is - * no minimum IO size - */ - err = ubifs_leb_write(c, lnum, c->cbuf, buf_offs, - wlen, UBI_SHORTTERM); - if (err) - return err; - buf_offs += ALIGN(wlen, 8); - if (next_len) { - used = 0; - avail = buf_len; - if (buf_offs + next_len > c->leb_size) { - err = ubifs_update_one_lp(c, lnum, - LPROPS_NC, 0, 0, LPROPS_TAKEN); - if (err) - return err; - lnum = -1; - } + nxt_offs = buf_offs + used + next_len; + if (next_len && nxt_offs <= c->leb_size) { + if (avail > 0) continue; - } + else + blen = buf_len; } else { - int blen, nxt_offs = buf_offs + used + next_len; - - if (next_len && nxt_offs <= c->leb_size) { - if (avail > 0) - continue; - else - blen = buf_len; - } else { - wlen = ALIGN(wlen, 8); - blen = ALIGN(wlen, c->min_io_size); - ubifs_pad(c, c->cbuf + wlen, blen - wlen); - } - /* - * The buffer is full or there are no more znodes - * to do - */ - err = ubifs_leb_write(c, lnum, c->cbuf, buf_offs, - blen, UBI_SHORTTERM); - if (err) - return err; - buf_offs += blen; - if (next_len) { - if (nxt_offs > c->leb_size) { - err = ubifs_update_one_lp(c, lnum, - LPROPS_NC, 0, 0, LPROPS_TAKEN); - if (err) - return err; - lnum = -1; - } - used -= blen; - if (used < 0) - used = 0; - avail = buf_len - used; - memmove(c->cbuf, c->cbuf + blen, used); - continue; + wlen = ALIGN(wlen, 8); + blen = ALIGN(wlen, c->min_io_size); + ubifs_pad(c, c->cbuf + wlen, blen - wlen); + } + + /* The buffer is full or there are no more znodes to do */ + err = ubifs_leb_write(c, lnum, c->cbuf, buf_offs, blen, + UBI_SHORTTERM); + if (err) + return err; + buf_offs += blen; + if (next_len) { + if (nxt_offs > c->leb_size) { + err = ubifs_update_one_lp(c, lnum, LPROPS_NC, 0, + 0, LPROPS_TAKEN); + if (err) + return err; + lnum = -1; } + used -= blen; + if (used < 0) + used = 0; + avail = buf_len - used; + memmove(c->cbuf, c->cbuf + blen, used); + continue; } break; } @@ -1029,7 +1010,7 @@ static void free_obsolete_znodes(struct ubifs_info *c) do { znode = cnext; cnext = znode->cnext; - if (test_bit(OBSOLETE_ZNODE, &znode->flags)) + if (ubifs_zn_obsolete(znode)) kfree(znode); else { znode->cnext = NULL; diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 62d50a9..223dd42 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -230,14 +230,14 @@ enum { * LPT cnode flag bits. * * DIRTY_CNODE: cnode is dirty - * COW_CNODE: cnode is being committed and must be copied before writing * OBSOLETE_CNODE: cnode is being committed and has been copied (or deleted), - * so it can (and must) be freed when the commit is finished + * so it can (and must) be freed when the commit is finished + * COW_CNODE: cnode is being committed and must be copied before writing */ enum { DIRTY_CNODE = 0, - COW_CNODE = 1, - OBSOLETE_CNODE = 2, + OBSOLETE_CNODE = 1, + COW_CNODE = 2, }; /* @@ -908,6 +908,7 @@ struct ubifs_budget_req { * @dnext: next orphan to delete * @inum: inode number * @new: %1 => added since the last commit, otherwise %0 + * @del: %1 => delete pending, otherwise %0 */ struct ubifs_orphan { struct rb_node rb; @@ -917,6 +918,7 @@ struct ubifs_orphan { struct ubifs_orphan *dnext; ino_t inum; int new; + unsigned del:1; }; /** @@ -1042,7 +1044,6 @@ struct ubifs_debug_info; * * @mst_node: master node * @mst_offs: offset of valid master node - * @mst_mutex: protects the master node area, @mst_node, and @mst_offs * * @max_bu_buf_len: maximum bulk-read buffer length * @bu_mutex: protects the pre-allocated bulk-read buffer and @c->bu @@ -1282,7 +1283,6 @@ struct ubifs_info { struct ubifs_mst_node *mst_node; int mst_offs; - struct mutex mst_mutex; int max_bu_buf_len; struct mutex bu_mutex; @@ -1471,6 +1471,15 @@ extern struct ubifs_compressor *ubifs_compressors[UBIFS_COMPR_TYPES_CNT]; /* io.c */ void ubifs_ro_mode(struct ubifs_info *c, int err); +int ubifs_leb_read(const struct ubifs_info *c, int lnum, void *buf, int offs, + int len, int even_ebadmsg); +int ubifs_leb_write(struct ubifs_info *c, int lnum, const void *buf, int offs, + int len, int dtype); +int ubifs_leb_change(struct ubifs_info *c, int lnum, const void *buf, int len, + int dtype); +int ubifs_leb_unmap(struct ubifs_info *c, int lnum); +int ubifs_leb_map(struct ubifs_info *c, int lnum, int dtype); +int ubifs_is_mapped(const struct ubifs_info *c, int lnum); int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len); int ubifs_wbuf_seek_nolock(struct ubifs_wbuf *wbuf, int lnum, int offs, int dtype); @@ -1723,7 +1732,7 @@ const struct ubifs_lprops *ubifs_fast_find_frdi_idx(struct ubifs_info *c); int ubifs_calc_dark(const struct ubifs_info *c, int spc); /* file.c */ -int ubifs_fsync(struct file *file, int datasync); +int ubifs_fsync(struct file *file, loff_t start, loff_t end, int datasync); int ubifs_setattr(struct dentry *dentry, struct iattr *attr); /* dir.c */ @@ -1750,8 +1759,8 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum, int offs, void *sbuf, int jhead); struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum, int offs, void *sbuf); -int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf); -int ubifs_clean_lebs(const struct ubifs_info *c, void *sbuf); +int ubifs_recover_inl_heads(struct ubifs_info *c, void *sbuf); +int ubifs_clean_lebs(struct ubifs_info *c, void *sbuf); int ubifs_rcvry_gc_commit(struct ubifs_info *c); int ubifs_recover_size_accum(struct ubifs_info *c, union ubifs_key *key, int deletion, loff_t new_size); diff --git a/fs/ubifs/xattr.c b/fs/ubifs/xattr.c index 16f19f5..bf18f7a 100644 --- a/fs/ubifs/xattr.c +++ b/fs/ubifs/xattr.c @@ -558,10 +558,10 @@ int ubifs_removexattr(struct dentry *dentry, const char *name) } ubifs_assert(inode->i_nlink == 1); - inode->i_nlink = 0; + clear_nlink(inode); err = remove_xattr(c, host, inode, &nm); if (err) - inode->i_nlink = 1; + set_nlink(inode, 1); /* If @i_nlink is 0, 'iput()' will delete the inode */ iput(inode); diff --git a/fs/udf/balloc.c b/fs/udf/balloc.c index 95518a9..987585b 100644 --- a/fs/udf/balloc.c +++ b/fs/udf/balloc.c @@ -59,8 +59,8 @@ static int __load_block_bitmap(struct super_block *sb, int nr_groups = bitmap->s_nr_groups; if (block_group >= nr_groups) { - udf_debug("block_group (%d) > nr_groups (%d)\n", block_group, - nr_groups); + udf_debug("block_group (%d) > nr_groups (%d)\n", + block_group, nr_groups); } if (bitmap->s_block_bitmap[block_group]) { @@ -126,8 +126,9 @@ static void udf_bitmap_free_blocks(struct super_block *sb, if (bloc->logicalBlockNum + count < count || (bloc->logicalBlockNum + count) > partmap->s_partition_len) { udf_debug("%d < %d || %d + %d > %d\n", - bloc->logicalBlockNum, 0, bloc->logicalBlockNum, - count, partmap->s_partition_len); + bloc->logicalBlockNum, 0, + bloc->logicalBlockNum, count, + partmap->s_partition_len); goto error_return; } @@ -155,7 +156,7 @@ static void udf_bitmap_free_blocks(struct super_block *sb, if (udf_set_bit(bit + i, bh->b_data)) { udf_debug("bit %ld already set\n", bit + i); udf_debug("byte=%2x\n", - ((char *)bh->b_data)[(bit + i) >> 3]); + ((char *)bh->b_data)[(bit + i) >> 3]); } } udf_add_free_space(sb, sbi->s_partition, count); @@ -369,7 +370,8 @@ static void udf_table_free_blocks(struct super_block *sb, if (bloc->logicalBlockNum + count < count || (bloc->logicalBlockNum + count) > partmap->s_partition_len) { udf_debug("%d < %d || %d + %d > %d\n", - bloc->logicalBlockNum, 0, bloc->logicalBlockNum, count, + bloc->logicalBlockNum, 0, + bloc->logicalBlockNum, count, partmap->s_partition_len); goto error_return; } diff --git a/fs/udf/dir.c b/fs/udf/dir.c index eb8bfe2..56341af 100644 --- a/fs/udf/dir.c +++ b/fs/udf/dir.c @@ -163,7 +163,8 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, struct kernel_lb_addr tloc = lelb_to_cpu(cfi.icb.extLocation); iblock = udf_get_lb_pblock(dir->i_sb, &tloc, 0); - flen = udf_get_filename(dir->i_sb, nameptr, fname, lfi); + flen = udf_get_filename(dir->i_sb, nameptr, lfi, fname, + UDF_NAME_LEN); dt_type = DT_UNKNOWN; } diff --git a/fs/udf/directory.c b/fs/udf/directory.c index 2ffdb67..3e44f57 100644 --- a/fs/udf/directory.c +++ b/fs/udf/directory.c @@ -162,8 +162,8 @@ struct fileIdentDesc *udf_get_fileident(void *buffer, int bufsize, int *offset) int padlen; if ((!buffer) || (!offset)) { - udf_debug("invalidparms\n, buffer=%p, offset=%p\n", buffer, - offset); + udf_debug("invalidparms, buffer=%p, offset=%p\n", + buffer, offset); return NULL; } @@ -201,7 +201,7 @@ struct short_ad *udf_get_fileshortad(uint8_t *ptr, int maxoffset, uint32_t *offs struct short_ad *sa; if ((!ptr) || (!offset)) { - printk(KERN_ERR "udf: udf_get_fileshortad() invalidparms\n"); + pr_err("%s: invalidparms\n", __func__); return NULL; } @@ -223,7 +223,7 @@ struct long_ad *udf_get_filelongad(uint8_t *ptr, int maxoffset, uint32_t *offset struct long_ad *la; if ((!ptr) || (!offset)) { - printk(KERN_ERR "udf: udf_get_filelongad() invalidparms\n"); + pr_err("%s: invalidparms\n", __func__); return NULL; } diff --git a/fs/udf/file.c b/fs/udf/file.c index 8eb9628..874c9e3 100644 --- a/fs/udf/file.c +++ b/fs/udf/file.c @@ -173,7 +173,7 @@ long udf_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) long old_block, new_block; int result = -EINVAL; - if (file_permission(filp, MAY_READ) != 0) { + if (inode_permission(inode, MAY_READ) != 0) { udf_debug("no permission to access inode %lu\n", inode->i_ino); result = -EPERM; goto out; diff --git a/fs/udf/inode.c b/fs/udf/inode.c index 957c974..e081440 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -37,6 +37,7 @@ #include <linux/writeback.h> #include <linux/slab.h> #include <linux/crc-itu-t.h> +#include <linux/mpage.h> #include "udf_i.h" #include "udf_sb.h" @@ -83,12 +84,10 @@ void udf_evict_inode(struct inode *inode) end_writeback(inode); if (iinfo->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB && inode->i_size != iinfo->i_lenExtents) { - printk(KERN_WARNING "UDF-fs (%s): Inode %lu (mode %o) has " - "inode size %llu different from extent length %llu. " - "Filesystem need not be standards compliant.\n", - inode->i_sb->s_id, inode->i_ino, inode->i_mode, - (unsigned long long)inode->i_size, - (unsigned long long)iinfo->i_lenExtents); + udf_warn(inode->i_sb, "Inode %lu (mode %o) has inode size %llu different from extent length %llu. Filesystem need not be standards compliant.\n", + inode->i_ino, inode->i_mode, + (unsigned long long)inode->i_size, + (unsigned long long)iinfo->i_lenExtents); } kfree(iinfo->i_ext.i_data); iinfo->i_ext.i_data = NULL; @@ -104,7 +103,13 @@ static int udf_writepage(struct page *page, struct writeback_control *wbc) static int udf_readpage(struct file *file, struct page *page) { - return block_read_full_page(page, udf_get_block); + return mpage_readpage(page, udf_get_block); +} + +static int udf_readpages(struct file *file, struct address_space *mapping, + struct list_head *pages, unsigned nr_pages) +{ + return mpage_readpages(mapping, pages, nr_pages, udf_get_block); } static int udf_write_begin(struct file *file, struct address_space *mapping, @@ -139,6 +144,7 @@ static sector_t udf_bmap(struct address_space *mapping, sector_t block) const struct address_space_operations udf_aops = { .readpage = udf_readpage, + .readpages = udf_readpages, .writepage = udf_writepage, .write_begin = udf_write_begin, .write_end = generic_write_end, @@ -1170,13 +1176,22 @@ update_time: return 0; } +/* + * Maximum length of linked list formed by ICB hierarchy. The chosen number is + * arbitrary - just that we hopefully don't limit any real use of rewritten + * inode on write-once media but avoid looping for too long on corrupted media. + */ +#define UDF_MAX_ICB_NESTING 1024 + static void __udf_read_inode(struct inode *inode) { struct buffer_head *bh = NULL; struct fileEntry *fe; uint16_t ident; struct udf_inode_info *iinfo = UDF_I(inode); + unsigned int indirections = 0; +reread: /* * Set defaults, but the inode is still incomplete! * Note: get_new_inode() sets the following on a new inode: @@ -1191,16 +1206,15 @@ static void __udf_read_inode(struct inode *inode) */ bh = udf_read_ptagged(inode->i_sb, &iinfo->i_location, 0, &ident); if (!bh) { - printk(KERN_ERR "udf: udf_read_inode(ino %ld) failed !bh\n", - inode->i_ino); + udf_err(inode->i_sb, "(ino %ld) failed !bh\n", inode->i_ino); make_bad_inode(inode); return; } if (ident != TAG_IDENT_FE && ident != TAG_IDENT_EFE && ident != TAG_IDENT_USE) { - printk(KERN_ERR "udf: udf_read_inode(ino %ld) " - "failed ident=%d\n", inode->i_ino, ident); + udf_err(inode->i_sb, "(ino %ld) failed ident=%d\n", + inode->i_ino, ident); brelse(bh); make_bad_inode(inode); return; @@ -1214,34 +1228,32 @@ static void __udf_read_inode(struct inode *inode) ibh = udf_read_ptagged(inode->i_sb, &iinfo->i_location, 1, &ident); if (ident == TAG_IDENT_IE && ibh) { - struct buffer_head *nbh = NULL; struct kernel_lb_addr loc; struct indirectEntry *ie; ie = (struct indirectEntry *)ibh->b_data; loc = lelb_to_cpu(ie->indirectICB.extLocation); - if (ie->indirectICB.extLength && - (nbh = udf_read_ptagged(inode->i_sb, &loc, 0, - &ident))) { - if (ident == TAG_IDENT_FE || - ident == TAG_IDENT_EFE) { - memcpy(&iinfo->i_location, - &loc, - sizeof(struct kernel_lb_addr)); - brelse(bh); - brelse(ibh); - brelse(nbh); - __udf_read_inode(inode); + if (ie->indirectICB.extLength) { + brelse(bh); + brelse(ibh); + memcpy(&iinfo->i_location, &loc, + sizeof(struct kernel_lb_addr)); + if (++indirections > UDF_MAX_ICB_NESTING) { + udf_err(inode->i_sb, + "too many ICBs in ICB hierarchy" + " (max %d supported)\n", + UDF_MAX_ICB_NESTING); + make_bad_inode(inode); return; } - brelse(nbh); + goto reread; } } brelse(ibh); } else if (fe->icbTag.strategyType != cpu_to_le16(4)) { - printk(KERN_ERR "udf: unsupported strategy type: %d\n", - le16_to_cpu(fe->icbTag.strategyType)); + udf_err(inode->i_sb, "unsupported strategy type: %d\n", + le16_to_cpu(fe->icbTag.strategyType)); brelse(bh); make_bad_inode(inode); return; @@ -1258,6 +1270,8 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) int offset; struct udf_sb_info *sbi = UDF_SB(inode->i_sb); struct udf_inode_info *iinfo = UDF_I(inode); + unsigned int link_count; + int bs = inode->i_sb->s_blocksize; fe = (struct fileEntry *)bh->b_data; efe = (struct extendedFileEntry *)bh->b_data; @@ -1278,41 +1292,38 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) if (fe->descTag.tagIdent == cpu_to_le16(TAG_IDENT_EFE)) { iinfo->i_efe = 1; iinfo->i_use = 0; - if (udf_alloc_i_data(inode, inode->i_sb->s_blocksize - + if (udf_alloc_i_data(inode, bs - sizeof(struct extendedFileEntry))) { make_bad_inode(inode); return; } memcpy(iinfo->i_ext.i_data, bh->b_data + sizeof(struct extendedFileEntry), - inode->i_sb->s_blocksize - - sizeof(struct extendedFileEntry)); + bs - sizeof(struct extendedFileEntry)); } else if (fe->descTag.tagIdent == cpu_to_le16(TAG_IDENT_FE)) { iinfo->i_efe = 0; iinfo->i_use = 0; - if (udf_alloc_i_data(inode, inode->i_sb->s_blocksize - - sizeof(struct fileEntry))) { + if (udf_alloc_i_data(inode, bs - sizeof(struct fileEntry))) { make_bad_inode(inode); return; } memcpy(iinfo->i_ext.i_data, bh->b_data + sizeof(struct fileEntry), - inode->i_sb->s_blocksize - sizeof(struct fileEntry)); + bs - sizeof(struct fileEntry)); } else if (fe->descTag.tagIdent == cpu_to_le16(TAG_IDENT_USE)) { iinfo->i_efe = 0; iinfo->i_use = 1; iinfo->i_lenAlloc = le32_to_cpu( ((struct unallocSpaceEntry *)bh->b_data)-> lengthAllocDescs); - if (udf_alloc_i_data(inode, inode->i_sb->s_blocksize - + if (udf_alloc_i_data(inode, bs - sizeof(struct unallocSpaceEntry))) { make_bad_inode(inode); return; } memcpy(iinfo->i_ext.i_data, bh->b_data + sizeof(struct unallocSpaceEntry), - inode->i_sb->s_blocksize - - sizeof(struct unallocSpaceEntry)); + bs - sizeof(struct unallocSpaceEntry)); return; } @@ -1340,9 +1351,10 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_mode &= ~sbi->s_umask; read_unlock(&sbi->s_cred_lock); - inode->i_nlink = le16_to_cpu(fe->fileLinkCount); - if (!inode->i_nlink) - inode->i_nlink = 1; + link_count = le16_to_cpu(fe->fileLinkCount); + if (!link_count) + link_count = 1; + set_nlink(inode, link_count); inode->i_size = le64_to_cpu(fe->informationLength); iinfo->i_lenExtents = inode->i_size; @@ -1389,6 +1401,36 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) iinfo->i_lenEAttr; } + /* + * Sanity check length of allocation descriptors and extended attrs to + * avoid integer overflows + */ + if (iinfo->i_lenEAttr > bs || iinfo->i_lenAlloc > bs) { + make_bad_inode(inode); + return; + } + /* Now do exact checks */ + if (udf_file_entry_alloc_offset(inode) + iinfo->i_lenAlloc > bs) { + make_bad_inode(inode); + return; + } + /* Sanity checks for files in ICB so that we don't get confused later */ + if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) { + /* + * For file in ICB data is stored in allocation descriptor + * so sizes should match + */ + if (iinfo->i_lenAlloc != inode->i_size) { + make_bad_inode(inode); + return; + } + /* File in ICB has to fit in there... */ + if (inode->i_size > bs - udf_file_entry_alloc_offset(inode)) { + make_bad_inode(inode); + return; + } + } + switch (fe->icbTag.fileType) { case ICBTAG_FILE_TYPE_DIRECTORY: inode->i_op = &udf_dir_inode_operations; @@ -1435,9 +1477,8 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) udf_debug("METADATA BITMAP FILE-----\n"); break; default: - printk(KERN_ERR "udf: udf_fill_inode(ino %ld) failed unknown " - "file type=%d\n", inode->i_ino, - fe->icbTag.fileType); + udf_err(inode->i_sb, "(ino %ld) failed unknown file type=%d\n", + inode->i_ino, fe->icbTag.fileType); make_bad_inode(inode); return; } @@ -1460,8 +1501,8 @@ static int udf_alloc_i_data(struct inode *inode, size_t size) iinfo->i_ext.i_data = kmalloc(size, GFP_KERNEL); if (!iinfo->i_ext.i_data) { - printk(KERN_ERR "udf:udf_alloc_i_data (ino %ld) " - "no free memory\n", inode->i_ino); + udf_err(inode->i_sb, "(ino %ld) no free memory\n", + inode->i_ino); return -ENOMEM; } @@ -1711,9 +1752,8 @@ out: if (do_sync) { sync_dirty_buffer(bh); if (buffer_write_io_error(bh)) { - printk(KERN_WARNING "IO error syncing udf inode " - "[%s:%08lx]\n", inode->i_sb->s_id, - inode->i_ino); + udf_warn(inode->i_sb, "IO error syncing udf inode [%08lx]\n", + inode->i_ino); err = -EIO; } } @@ -2004,8 +2044,7 @@ int8_t udf_current_aext(struct inode *inode, struct extent_position *epos, *elen = le32_to_cpu(lad->extLength) & UDF_EXTENT_LENGTH_MASK; break; default: - udf_debug("alloc_type = %d unsupported\n", - iinfo->i_alloc_type); + udf_debug("alloc_type = %d unsupported\n", iinfo->i_alloc_type); return -1; } diff --git a/fs/udf/lowlevel.c b/fs/udf/lowlevel.c index 43e24a3..6583fe9 100644 --- a/fs/udf/lowlevel.c +++ b/fs/udf/lowlevel.c @@ -38,7 +38,7 @@ unsigned int udf_get_last_session(struct super_block *sb) if (i == 0) { udf_debug("XA disk: %s, vol_desc_start=%d\n", - (ms_info.xa_flag ? "yes" : "no"), ms_info.addr.lba); + ms_info.xa_flag ? "yes" : "no", ms_info.addr.lba); if (ms_info.xa_flag) /* necessary for a valid ms_info.addr */ vol_desc_start = ms_info.addr.lba; } else { diff --git a/fs/udf/misc.c b/fs/udf/misc.c index 9215700..c175b4d 100644 --- a/fs/udf/misc.c +++ b/fs/udf/misc.c @@ -204,6 +204,7 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, { struct tag *tag_p; struct buffer_head *bh = NULL; + u8 checksum; /* Read the block */ if (block == 0xFFFFFFFF) @@ -211,8 +212,8 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, bh = udf_tread(sb, block); if (!bh) { - udf_debug("block=%d, location=%d: read failed\n", - block, location); + udf_err(sb, "read failed, block=%u, location=%d\n", + block, location); return NULL; } @@ -227,16 +228,18 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, } /* Verify the tag checksum */ - if (udf_tag_checksum(tag_p) != tag_p->tagChecksum) { - printk(KERN_ERR "udf: tag checksum failed block %d\n", block); + checksum = udf_tag_checksum(tag_p); + if (checksum != tag_p->tagChecksum) { + udf_err(sb, "tag checksum failed, block %u: 0x%02x != 0x%02x\n", + block, checksum, tag_p->tagChecksum); goto error_out; } /* Verify the tag version */ if (tag_p->descVersion != cpu_to_le16(0x0002U) && tag_p->descVersion != cpu_to_le16(0x0003U)) { - udf_debug("tag version 0x%04x != 0x0002 || 0x0003 block %d\n", - le16_to_cpu(tag_p->descVersion), block); + udf_err(sb, "tag version 0x%04x != 0x0002 || 0x0003, block %u\n", + le16_to_cpu(tag_p->descVersion), block); goto error_out; } @@ -248,8 +251,8 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, return bh; udf_debug("Crc failure block %d: crc = %d, crclen = %d\n", block, - le16_to_cpu(tag_p->descCRC), le16_to_cpu(tag_p->descCRCLength)); - + le16_to_cpu(tag_p->descCRC), + le16_to_cpu(tag_p->descCRCLength)); error_out: brelse(bh); return NULL; diff --git a/fs/udf/namei.c b/fs/udf/namei.c index d8c1bb5..483d662 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -235,7 +235,8 @@ static struct fileIdentDesc *udf_find_entry(struct inode *dir, if (!lfi) continue; - flen = udf_get_filename(dir->i_sb, nameptr, fname, lfi); + flen = udf_get_filename(dir->i_sb, nameptr, lfi, fname, + UDF_NAME_LEN); if (flen && udf_match(flen, fname, child->len, child->name)) goto out_ok; } @@ -577,8 +578,7 @@ static int udf_create(struct inode *dir, struct dentry *dentry, int mode, fi = udf_add_entry(dir, dentry, &fibh, &cfi, &err); if (!fi) { - inode->i_nlink--; - mark_inode_dirty(inode); + inode_dec_link_count(inode); iput(inode); return err; } @@ -618,8 +618,7 @@ static int udf_mknod(struct inode *dir, struct dentry *dentry, int mode, init_special_inode(inode, mode, rdev); fi = udf_add_entry(dir, dentry, &fibh, &cfi, &err); if (!fi) { - inode->i_nlink--; - mark_inode_dirty(inode); + inode_dec_link_count(inode); iput(inode); return err; } @@ -665,12 +664,11 @@ static int udf_mkdir(struct inode *dir, struct dentry *dentry, int mode) inode->i_fop = &udf_dir_operations; fi = udf_add_entry(inode, NULL, &fibh, &cfi, &err); if (!fi) { - inode->i_nlink--; - mark_inode_dirty(inode); + inode_dec_link_count(inode); iput(inode); goto out; } - inode->i_nlink = 2; + set_nlink(inode, 2); cfi.icb.extLength = cpu_to_le32(inode->i_sb->s_blocksize); cfi.icb.extLocation = cpu_to_lelb(dinfo->i_location); *(__le32 *)((struct allocDescImpUse *)cfi.icb.impUse)->impUse = @@ -683,7 +681,7 @@ static int udf_mkdir(struct inode *dir, struct dentry *dentry, int mode) fi = udf_add_entry(dir, dentry, &fibh, &cfi, &err); if (!fi) { - inode->i_nlink = 0; + clear_nlink(inode); mark_inode_dirty(inode); iput(inode); goto out; @@ -799,9 +797,8 @@ static int udf_rmdir(struct inode *dir, struct dentry *dentry) if (retval) goto end_rmdir; if (inode->i_nlink != 2) - udf_warning(inode->i_sb, "udf_rmdir", - "empty directory has nlink != 2 (%d)", - inode->i_nlink); + udf_warn(inode->i_sb, "empty directory has nlink != 2 (%d)\n", + inode->i_nlink); clear_nlink(inode); inode->i_size = 0; inode_dec_link_count(dir); @@ -840,7 +837,7 @@ static int udf_unlink(struct inode *dir, struct dentry *dentry) if (!inode->i_nlink) { udf_debug("Deleting nonexistent file (%lu), %d\n", inode->i_ino, inode->i_nlink); - inode->i_nlink = 1; + set_nlink(inode, 1); } retval = udf_delete_entry(dir, fi, &fibh, &cfi); if (retval) diff --git a/fs/udf/partition.c b/fs/udf/partition.c index a71090e..d6caf01 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -33,8 +33,8 @@ uint32_t udf_get_pblock(struct super_block *sb, uint32_t block, struct udf_sb_info *sbi = UDF_SB(sb); struct udf_part_map *map; if (partition >= sbi->s_partitions) { - udf_debug("block=%d, partition=%d, offset=%d: " - "invalid partition\n", block, partition, offset); + udf_debug("block=%d, partition=%d, offset=%d: invalid partition\n", + block, partition, offset); return 0xFFFFFFFF; } map = &sbi->s_partmaps[partition]; @@ -60,8 +60,8 @@ uint32_t udf_get_pblock_virt15(struct super_block *sb, uint32_t block, vdata = &map->s_type_specific.s_virtual; if (block > vdata->s_num_entries) { - udf_debug("Trying to access block beyond end of VAT " - "(%d max %d)\n", block, vdata->s_num_entries); + udf_debug("Trying to access block beyond end of VAT (%d max %d)\n", + block, vdata->s_num_entries); return 0xFFFFFFFF; } @@ -321,9 +321,14 @@ uint32_t udf_get_pblock_meta25(struct super_block *sb, uint32_t block, /* We shouldn't mount such media... */ BUG_ON(!inode); retblk = udf_try_read_meta(inode, block, partition, offset); - if (retblk == 0xFFFFFFFF) { - udf_warning(sb, __func__, "error reading from METADATA, " - "trying to read from MIRROR"); + if (retblk == 0xFFFFFFFF && mdata->s_metadata_fe) { + udf_warn(sb, "error reading from METADATA, trying to read from MIRROR\n"); + if (!(mdata->s_flags & MF_MIRROR_FE_LOADED)) { + mdata->s_mirror_fe = udf_find_metadata_inode_efe(sb, + mdata->s_mirror_file_loc, map->s_partition_num); + mdata->s_flags |= MF_MIRROR_FE_LOADED; + } + inode = mdata->s_mirror_fe; if (!inode) return 0xFFFFFFFF; diff --git a/fs/udf/super.c b/fs/udf/super.c index b0c7b53..f66439e 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -76,8 +76,6 @@ #define UDF_DEFAULT_BLOCKSIZE 2048 -static char error_buf[1024]; - /* These are the "meat" - everything else is stuffing */ static int udf_fill_super(struct super_block *, void *, int); static void udf_put_super(struct super_block *); @@ -93,8 +91,6 @@ static void udf_close_lvid(struct super_block *); static unsigned int udf_count_free(struct super_block *); static int udf_statfs(struct dentry *, struct kstatfs *); static int udf_show_options(struct seq_file *, struct vfsmount *); -static void udf_error(struct super_block *sb, const char *function, - const char *fmt, ...); struct logicalVolIntegrityDescImpUse *udf_sb_lvidiu(struct udf_sb_info *sbi) { @@ -245,9 +241,8 @@ static int udf_sb_alloc_partition_maps(struct super_block *sb, u32 count) sbi->s_partmaps = kcalloc(count, sizeof(struct udf_part_map), GFP_KERNEL); if (!sbi->s_partmaps) { - udf_error(sb, __func__, - "Unable to allocate space for %d partition maps", - count); + udf_err(sb, "Unable to allocate space for %d partition maps\n", + count); sbi->s_partitions = 0; return -ENOMEM; } @@ -551,8 +546,7 @@ static int udf_parse_options(char *options, struct udf_options *uopt, uopt->dmode = option & 0777; break; default: - printk(KERN_ERR "udf: bad mount option \"%s\" " - "or missing value\n", p); + pr_err("bad mount option \"%s\" or missing value\n", p); return 0; } } @@ -646,20 +640,16 @@ static loff_t udf_check_vsd(struct super_block *sb) udf_debug("ISO9660 Boot Record found\n"); break; case 1: - udf_debug("ISO9660 Primary Volume Descriptor " - "found\n"); + udf_debug("ISO9660 Primary Volume Descriptor found\n"); break; case 2: - udf_debug("ISO9660 Supplementary Volume " - "Descriptor found\n"); + udf_debug("ISO9660 Supplementary Volume Descriptor found\n"); break; case 3: - udf_debug("ISO9660 Volume Partition Descriptor " - "found\n"); + udf_debug("ISO9660 Volume Partition Descriptor found\n"); break; case 255: - udf_debug("ISO9660 Volume Descriptor Set " - "Terminator found\n"); + udf_debug("ISO9660 Volume Descriptor Set Terminator found\n"); break; default: udf_debug("ISO9660 VRS (%u) found\n", @@ -810,8 +800,7 @@ static int udf_load_pvoldesc(struct super_block *sb, sector_t block) pvoldesc->recordingDateAndTime)) { #ifdef UDFFS_DEBUG struct timestamp *ts = &pvoldesc->recordingDateAndTime; - udf_debug("recording time %04u/%02u/%02u" - " %02u:%02u (%x)\n", + udf_debug("recording time %04u/%02u/%02u %02u:%02u (%x)\n", le16_to_cpu(ts->year), ts->month, ts->day, ts->hour, ts->minute, le16_to_cpu(ts->typeAndTimezone)); #endif @@ -822,7 +811,7 @@ static int udf_load_pvoldesc(struct super_block *sb, sector_t block) strncpy(UDF_SB(sb)->s_volume_ident, outstr->u_name, outstr->u_len > 31 ? 31 : outstr->u_len); udf_debug("volIdent[] = '%s'\n", - UDF_SB(sb)->s_volume_ident); + UDF_SB(sb)->s_volume_ident); } if (!udf_build_ustr(instr, pvoldesc->volSetIdent, 128)) @@ -838,64 +827,57 @@ out1: return ret; } +struct inode *udf_find_metadata_inode_efe(struct super_block *sb, + u32 meta_file_loc, u32 partition_num) +{ + struct kernel_lb_addr addr; + struct inode *metadata_fe; + + addr.logicalBlockNum = meta_file_loc; + addr.partitionReferenceNum = partition_num; + + metadata_fe = udf_iget(sb, &addr); + + if (metadata_fe == NULL) + udf_warn(sb, "metadata inode efe not found\n"); + else if (UDF_I(metadata_fe)->i_alloc_type != ICBTAG_FLAG_AD_SHORT) { + udf_warn(sb, "metadata inode efe does not have short allocation descriptors!\n"); + iput(metadata_fe); + metadata_fe = NULL; + } + + return metadata_fe; +} + static int udf_load_metadata_files(struct super_block *sb, int partition) { struct udf_sb_info *sbi = UDF_SB(sb); struct udf_part_map *map; struct udf_meta_data *mdata; struct kernel_lb_addr addr; - int fe_error = 0; map = &sbi->s_partmaps[partition]; mdata = &map->s_type_specific.s_metadata; /* metadata address */ - addr.logicalBlockNum = mdata->s_meta_file_loc; - addr.partitionReferenceNum = map->s_partition_num; - udf_debug("Metadata file location: block = %d part = %d\n", - addr.logicalBlockNum, addr.partitionReferenceNum); + mdata->s_meta_file_loc, map->s_partition_num); - mdata->s_metadata_fe = udf_iget(sb, &addr); + mdata->s_metadata_fe = udf_find_metadata_inode_efe(sb, + mdata->s_meta_file_loc, map->s_partition_num); if (mdata->s_metadata_fe == NULL) { - udf_warning(sb, __func__, "metadata inode efe not found, " - "will try mirror inode."); - fe_error = 1; - } else if (UDF_I(mdata->s_metadata_fe)->i_alloc_type != - ICBTAG_FLAG_AD_SHORT) { - udf_warning(sb, __func__, "metadata inode efe does not have " - "short allocation descriptors!"); - fe_error = 1; - iput(mdata->s_metadata_fe); - mdata->s_metadata_fe = NULL; - } + /* mirror file entry */ + udf_debug("Mirror metadata file location: block = %d part = %d\n", + mdata->s_mirror_file_loc, map->s_partition_num); - /* mirror file entry */ - addr.logicalBlockNum = mdata->s_mirror_file_loc; - addr.partitionReferenceNum = map->s_partition_num; - - udf_debug("Mirror metadata file location: block = %d part = %d\n", - addr.logicalBlockNum, addr.partitionReferenceNum); + mdata->s_mirror_fe = udf_find_metadata_inode_efe(sb, + mdata->s_mirror_file_loc, map->s_partition_num); - mdata->s_mirror_fe = udf_iget(sb, &addr); - - if (mdata->s_mirror_fe == NULL) { - if (fe_error) { - udf_error(sb, __func__, "mirror inode efe not found " - "and metadata inode is missing too, exiting..."); - goto error_exit; - } else - udf_warning(sb, __func__, "mirror inode efe not found," - " but metadata inode is OK"); - } else if (UDF_I(mdata->s_mirror_fe)->i_alloc_type != - ICBTAG_FLAG_AD_SHORT) { - udf_warning(sb, __func__, "mirror inode efe does not have " - "short allocation descriptors!"); - iput(mdata->s_mirror_fe); - mdata->s_mirror_fe = NULL; - if (fe_error) + if (mdata->s_mirror_fe == NULL) { + udf_err(sb, "Both metadata and mirror metadata inode efe can not found\n"); goto error_exit; + } } /* @@ -908,18 +890,15 @@ static int udf_load_metadata_files(struct super_block *sb, int partition) addr.partitionReferenceNum = map->s_partition_num; udf_debug("Bitmap file location: block = %d part = %d\n", - addr.logicalBlockNum, addr.partitionReferenceNum); + addr.logicalBlockNum, addr.partitionReferenceNum); mdata->s_bitmap_fe = udf_iget(sb, &addr); if (mdata->s_bitmap_fe == NULL) { if (sb->s_flags & MS_RDONLY) - udf_warning(sb, __func__, "bitmap inode efe " - "not found but it's ok since the disc" - " is mounted read-only"); + udf_warn(sb, "bitmap inode efe not found but it's ok since the disc is mounted read-only\n"); else { - udf_error(sb, __func__, "bitmap inode efe not " - "found and attempted read-write mount"); + udf_err(sb, "bitmap inode efe not found and attempted read-write mount\n"); goto error_exit; } } @@ -972,9 +951,8 @@ static struct udf_bitmap *udf_sb_alloc_bitmap(struct super_block *sb, u32 index) bitmap = vzalloc(size); /* TODO: get rid of vzalloc */ if (bitmap == NULL) { - udf_error(sb, __func__, - "Unable to allocate space for bitmap " - "and %d buffer_head pointers", nr_groups); + udf_err(sb, "Unable to allocate space for bitmap and %d buffer_head pointers\n", + nr_groups); return NULL; } @@ -1004,10 +982,9 @@ static int udf_fill_partdesc_info(struct super_block *sb, if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_OVERWRITABLE)) map->s_partition_flags |= UDF_PART_FLAG_OVERWRITABLE; - udf_debug("Partition (%d type %x) starts at physical %d, " - "block length %d\n", p_index, - map->s_partition_type, map->s_partition_root, - map->s_partition_len); + udf_debug("Partition (%d type %x) starts at physical %d, block length %d\n", + p_index, map->s_partition_type, + map->s_partition_root, map->s_partition_len); if (strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR02) && strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR03)) @@ -1024,12 +1001,12 @@ static int udf_fill_partdesc_info(struct super_block *sb, map->s_uspace.s_table = udf_iget(sb, &loc); if (!map->s_uspace.s_table) { udf_debug("cannot load unallocSpaceTable (part %d)\n", - p_index); + p_index); return 1; } map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_TABLE; udf_debug("unallocSpaceTable (part %d) @ %ld\n", - p_index, map->s_uspace.s_table->i_ino); + p_index, map->s_uspace.s_table->i_ino); } if (phd->unallocSpaceBitmap.extLength) { @@ -1042,8 +1019,8 @@ static int udf_fill_partdesc_info(struct super_block *sb, bitmap->s_extPosition = le32_to_cpu( phd->unallocSpaceBitmap.extPosition); map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_BITMAP; - udf_debug("unallocSpaceBitmap (part %d) @ %d\n", p_index, - bitmap->s_extPosition); + udf_debug("unallocSpaceBitmap (part %d) @ %d\n", + p_index, bitmap->s_extPosition); } if (phd->partitionIntegrityTable.extLength) @@ -1059,13 +1036,13 @@ static int udf_fill_partdesc_info(struct super_block *sb, map->s_fspace.s_table = udf_iget(sb, &loc); if (!map->s_fspace.s_table) { udf_debug("cannot load freedSpaceTable (part %d)\n", - p_index); + p_index); return 1; } map->s_partition_flags |= UDF_PART_FLAG_FREED_TABLE; udf_debug("freedSpaceTable (part %d) @ %ld\n", - p_index, map->s_fspace.s_table->i_ino); + p_index, map->s_fspace.s_table->i_ino); } if (phd->freedSpaceBitmap.extLength) { @@ -1078,8 +1055,8 @@ static int udf_fill_partdesc_info(struct super_block *sb, bitmap->s_extPosition = le32_to_cpu( phd->freedSpaceBitmap.extPosition); map->s_partition_flags |= UDF_PART_FLAG_FREED_BITMAP; - udf_debug("freedSpaceBitmap (part %d) @ %d\n", p_index, - bitmap->s_extPosition); + udf_debug("freedSpaceBitmap (part %d) @ %d\n", + p_index, bitmap->s_extPosition); } return 0; } @@ -1119,11 +1096,9 @@ static int udf_load_vat(struct super_block *sb, int p_index, int type1_index) udf_find_vat_block(sb, p_index, type1_index, sbi->s_last_block); if (!sbi->s_vat_inode && sbi->s_last_block != blocks - 1) { - printk(KERN_NOTICE "UDF-fs: Failed to read VAT inode from the" - " last recorded block (%lu), retrying with the last " - "block of the device (%lu).\n", - (unsigned long)sbi->s_last_block, - (unsigned long)blocks - 1); + pr_notice("Failed to read VAT inode from the last recorded block (%lu), retrying with the last block of the device (%lu).\n", + (unsigned long)sbi->s_last_block, + (unsigned long)blocks - 1); udf_find_vat_block(sb, p_index, type1_index, blocks - 1); } if (!sbi->s_vat_inode) @@ -1221,8 +1196,8 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) if (map->s_partition_type == UDF_METADATA_MAP25) { ret = udf_load_metadata_files(sb, i); if (ret) { - printk(KERN_ERR "UDF-fs: error loading MetaData " - "partition map %d\n", i); + udf_err(sb, "error loading MetaData partition map %d\n", + i); goto out_bh; } } else { @@ -1235,9 +1210,7 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) * overwrite blocks instead of relocating them). */ sb->s_flags |= MS_RDONLY; - printk(KERN_NOTICE "UDF-fs: Filesystem marked read-only " - "because writing to pseudooverwrite partition is " - "not implemented.\n"); + pr_notice("Filesystem marked read-only because writing to pseudooverwrite partition is not implemented\n"); } out_bh: /* In case loading failed, we handle cleanup in udf_fill_super */ @@ -1259,13 +1232,13 @@ static int udf_load_sparable_map(struct super_block *sb, map->s_partition_type = UDF_SPARABLE_MAP15; sdata->s_packet_len = le16_to_cpu(spm->packetLength); if (!is_power_of_2(sdata->s_packet_len)) { - udf_error(sb, __func__, "error loading logical volume descriptor: " + udf_err(sb, "error loading logical volume descriptor: " "Invalid packet length %u\n", (unsigned)sdata->s_packet_len); return -EIO; } if (spm->numSparingTables > 4) { - udf_error(sb, __func__, "error loading logical volume descriptor: " + udf_err(sb, "error loading logical volume descriptor: " "Too many sparing tables (%d)\n", (int)spm->numSparingTables); return -EIO; @@ -1312,8 +1285,8 @@ static int udf_load_logicalvol(struct super_block *sb, sector_t block, BUG_ON(ident != TAG_IDENT_LVD); lvd = (struct logicalVolDesc *)bh->b_data; table_len = le32_to_cpu(lvd->mapTableLength); - if (sizeof(*lvd) + table_len > sb->s_blocksize) { - udf_error(sb, __func__, "error loading logical volume descriptor: " + if (table_len > sb->s_blocksize - sizeof(*lvd)) { + udf_err(sb, "error loading logical volume descriptor: " "Partition table too long (%u > %lu)\n", table_len, sb->s_blocksize - sizeof(*lvd)); ret = 1; @@ -1373,9 +1346,8 @@ static int udf_load_logicalvol(struct super_block *sb, sector_t block, struct metadataPartitionMap *mdm = (struct metadataPartitionMap *) &(lvd->partitionMaps[offset]); - udf_debug("Parsing Logical vol part %d " - "type %d id=%s\n", i, type, - UDF_ID_METADATA); + udf_debug("Parsing Logical vol part %d type %d id=%s\n", + i, type, UDF_ID_METADATA); map->s_partition_type = UDF_METADATA_MAP25; map->s_partition_func = udf_get_pblock_meta25; @@ -1390,25 +1362,24 @@ static int udf_load_logicalvol(struct super_block *sb, sector_t block, le32_to_cpu(mdm->allocUnitSize); mdata->s_align_unit_size = le16_to_cpu(mdm->alignUnitSize); - mdata->s_dup_md_flag = - mdm->flags & 0x01; + if (mdm->flags & 0x01) + mdata->s_flags |= MF_DUPLICATE_MD; udf_debug("Metadata Ident suffix=0x%x\n", - (le16_to_cpu( - ((__le16 *) - mdm->partIdent.identSuffix)[0]))); + le16_to_cpu(*(__le16 *) + mdm->partIdent.identSuffix)); udf_debug("Metadata part num=%d\n", - le16_to_cpu(mdm->partitionNum)); + le16_to_cpu(mdm->partitionNum)); udf_debug("Metadata part alloc unit size=%d\n", - le32_to_cpu(mdm->allocUnitSize)); + le32_to_cpu(mdm->allocUnitSize)); udf_debug("Metadata file loc=%d\n", - le32_to_cpu(mdm->metadataFileLoc)); + le32_to_cpu(mdm->metadataFileLoc)); udf_debug("Mirror file loc=%d\n", - le32_to_cpu(mdm->metadataMirrorFileLoc)); + le32_to_cpu(mdm->metadataMirrorFileLoc)); udf_debug("Bitmap file loc=%d\n", - le32_to_cpu(mdm->metadataBitmapFileLoc)); - udf_debug("Duplicate Flag: %d %d\n", - mdata->s_dup_md_flag, mdm->flags); + le32_to_cpu(mdm->metadataBitmapFileLoc)); + udf_debug("Flags: %d %d\n", + mdata->s_flags, mdm->flags); } else { udf_debug("Unknown ident: %s\n", upm2->partIdent.ident); @@ -1418,16 +1389,15 @@ static int udf_load_logicalvol(struct super_block *sb, sector_t block, map->s_partition_num = le16_to_cpu(upm2->partitionNum); } udf_debug("Partition (%d:%d) type %d on volume %d\n", - i, map->s_partition_num, type, - map->s_volumeseqnum); + i, map->s_partition_num, type, map->s_volumeseqnum); } if (fileset) { struct long_ad *la = (struct long_ad *)&(lvd->logicalVolContentsUse[0]); *fileset = lelb_to_cpu(la->extLocation); - udf_debug("FileSet found in LogicalVolDesc at block=%d, " - "partition=%d\n", fileset->logicalBlockNum, + udf_debug("FileSet found in LogicalVolDesc at block=%d, partition=%d\n", + fileset->logicalBlockNum, fileset->partitionReferenceNum); } if (lvd->integritySeqExt.extLength) @@ -1507,9 +1477,9 @@ static noinline int udf_process_sequence(struct super_block *sb, long block, bh = udf_read_tagged(sb, block, block, &ident); if (!bh) { - printk(KERN_ERR "udf: Block %Lu of volume descriptor " - "sequence is corrupted or we could not read " - "it.\n", (unsigned long long)block); + udf_err(sb, + "Block %llu of volume descriptor sequence is corrupted or we could not read it\n", + (unsigned long long)block); return 1; } @@ -1582,7 +1552,7 @@ static noinline int udf_process_sequence(struct super_block *sb, long block, * in a suitable order */ if (!vds[VDS_POS_PRIMARY_VOL_DESC].block) { - printk(KERN_ERR "udf: Primary Volume Descriptor not found!\n"); + udf_err(sb, "Primary Volume Descriptor not found!\n"); return 1; } if (udf_load_pvoldesc(sb, vds[VDS_POS_PRIMARY_VOL_DESC].block)) @@ -1769,7 +1739,7 @@ static int udf_load_vrs(struct super_block *sb, struct udf_options *uopt, if (!sb_set_blocksize(sb, uopt->blocksize)) { if (!silent) - printk(KERN_WARNING "UDF-fs: Bad block size\n"); + udf_warn(sb, "Bad block size\n"); return 0; } sbi->s_last_block = uopt->lastblock; @@ -1778,12 +1748,11 @@ static int udf_load_vrs(struct super_block *sb, struct udf_options *uopt, nsr_off = udf_check_vsd(sb); if (!nsr_off) { if (!silent) - printk(KERN_WARNING "UDF-fs: No VRS found\n"); + udf_warn(sb, "No VRS found\n"); return 0; } if (nsr_off == -1) - udf_debug("Failed to read byte 32768. Assuming open " - "disc. Skipping validity check\n"); + udf_debug("Failed to read byte 32768. Assuming open disc. Skipping validity check\n"); if (!sbi->s_last_block) sbi->s_last_block = udf_get_last_block(sb); } else { @@ -1794,7 +1763,7 @@ static int udf_load_vrs(struct super_block *sb, struct udf_options *uopt, sbi->s_anchor = uopt->anchor; if (!udf_find_anchor(sb, fileset)) { if (!silent) - printk(KERN_WARNING "UDF-fs: No anchor found\n"); + udf_warn(sb, "No anchor found\n"); return 0; } return 1; @@ -1972,8 +1941,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) if (uopt.flags & (1 << UDF_FLAG_UTF8) && uopt.flags & (1 << UDF_FLAG_NLS_MAP)) { - udf_error(sb, "udf_read_super", - "utf8 cannot be combined with iocharset\n"); + udf_err(sb, "utf8 cannot be combined with iocharset\n"); goto error_out; } #ifdef CONFIG_UDF_NLS @@ -2022,15 +1990,14 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) ret = udf_load_vrs(sb, &uopt, silent, &fileset); if (!ret && uopt.blocksize != UDF_DEFAULT_BLOCKSIZE) { if (!silent) - printk(KERN_NOTICE - "UDF-fs: Rescanning with blocksize " - "%d\n", UDF_DEFAULT_BLOCKSIZE); + pr_notice("Rescanning with blocksize %d\n", + UDF_DEFAULT_BLOCKSIZE); uopt.blocksize = UDF_DEFAULT_BLOCKSIZE; ret = udf_load_vrs(sb, &uopt, silent, &fileset); } } if (!ret) { - printk(KERN_WARNING "UDF-fs: No partition found (1)\n"); + udf_warn(sb, "No partition found (1)\n"); goto error_out; } @@ -2045,10 +2012,9 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) le16_to_cpu(lvidiu->maxUDFWriteRev); */ if (minUDFReadRev > UDF_MAX_READ_VERSION) { - printk(KERN_ERR "UDF-fs: minUDFReadRev=%x " - "(max is %x)\n", - le16_to_cpu(lvidiu->minUDFReadRev), - UDF_MAX_READ_VERSION); + udf_err(sb, "minUDFReadRev=%x (max is %x)\n", + le16_to_cpu(lvidiu->minUDFReadRev), + UDF_MAX_READ_VERSION); goto error_out; } else if (minUDFWriteRev > UDF_MAX_WRITE_VERSION) sb->s_flags |= MS_RDONLY; @@ -2062,28 +2028,27 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) } if (!sbi->s_partitions) { - printk(KERN_WARNING "UDF-fs: No partition found (2)\n"); + udf_warn(sb, "No partition found (2)\n"); goto error_out; } if (sbi->s_partmaps[sbi->s_partition].s_partition_flags & UDF_PART_FLAG_READ_ONLY) { - printk(KERN_NOTICE "UDF-fs: Partition marked readonly; " - "forcing readonly mount\n"); + pr_notice("Partition marked readonly; forcing readonly mount\n"); sb->s_flags |= MS_RDONLY; } if (udf_find_fileset(sb, &fileset, &rootdir)) { - printk(KERN_WARNING "UDF-fs: No fileset found\n"); + udf_warn(sb, "No fileset found\n"); goto error_out; } if (!silent) { struct timestamp ts; udf_time_to_disk_stamp(&ts, sbi->s_record_time); - udf_info("UDF: Mounting volume '%s', " - "timestamp %04u/%02u/%02u %02u:%02u (%x)\n", - sbi->s_volume_ident, le16_to_cpu(ts.year), ts.month, ts.day, + udf_info("Mounting volume '%s', timestamp %04u/%02u/%02u %02u:%02u (%x)\n", + sbi->s_volume_ident, + le16_to_cpu(ts.year), ts.month, ts.day, ts.hour, ts.minute, le16_to_cpu(ts.typeAndTimezone)); } if (!(sb->s_flags & MS_RDONLY)) @@ -2094,8 +2059,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) /* perhaps it's not extensible enough, but for now ... */ inode = udf_iget(sb, &rootdir); if (!inode) { - printk(KERN_ERR "UDF-fs: Error in udf_iget, block=%d, " - "partition=%d\n", + udf_err(sb, "Error in udf_iget, block=%d, partition=%d\n", rootdir.logicalBlockNum, rootdir.partitionReferenceNum); goto error_out; } @@ -2103,7 +2067,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) /* Allocate a dentry for the root inode */ sb->s_root = d_alloc_root(inode); if (!sb->s_root) { - printk(KERN_ERR "UDF-fs: Couldn't allocate root dentry\n"); + udf_err(sb, "Couldn't allocate root dentry\n"); iput(inode); goto error_out; } @@ -2131,32 +2095,40 @@ error_out: return -EINVAL; } -static void udf_error(struct super_block *sb, const char *function, - const char *fmt, ...) +void _udf_err(struct super_block *sb, const char *function, + const char *fmt, ...) { + struct va_format vaf; va_list args; - if (!(sb->s_flags & MS_RDONLY)) { - /* mark sb error */ + /* mark sb error */ + if (!(sb->s_flags & MS_RDONLY)) sb->s_dirt = 1; - } + va_start(args, fmt); - vsnprintf(error_buf, sizeof(error_buf), fmt, args); + + vaf.fmt = fmt; + vaf.va = &args; + + pr_err("error (device %s): %s: %pV", sb->s_id, function, &vaf); + va_end(args); - printk(KERN_CRIT "UDF-fs error (device %s): %s: %s\n", - sb->s_id, function, error_buf); } -void udf_warning(struct super_block *sb, const char *function, - const char *fmt, ...) +void _udf_warn(struct super_block *sb, const char *function, + const char *fmt, ...) { + struct va_format vaf; va_list args; va_start(args, fmt); - vsnprintf(error_buf, sizeof(error_buf), fmt, args); + + vaf.fmt = fmt; + vaf.va = &args; + + pr_warn("warning (device %s): %s: %pV", sb->s_id, function, &vaf); + va_end(args); - printk(KERN_WARNING "UDF-fs warning (device %s): %s: %s\n", - sb->s_id, function, error_buf); } static void udf_put_super(struct super_block *sb) @@ -2248,11 +2220,11 @@ static unsigned int udf_count_free_bitmap(struct super_block *sb, bh = udf_read_ptagged(sb, &loc, 0, &ident); if (!bh) { - printk(KERN_ERR "udf: udf_count_free failed\n"); + udf_err(sb, "udf_count_free failed\n"); goto out; } else if (ident != TAG_IDENT_SBD) { brelse(bh); - printk(KERN_ERR "udf: udf_count_free failed\n"); + udf_err(sb, "udf_count_free failed\n"); goto out; } diff --git a/fs/udf/symlink.c b/fs/udf/symlink.c index b1d4488..0422b7b 100644 --- a/fs/udf/symlink.c +++ b/fs/udf/symlink.c @@ -30,43 +30,73 @@ #include <linux/buffer_head.h> #include "udf_i.h" -static void udf_pc_to_char(struct super_block *sb, unsigned char *from, - int fromlen, unsigned char *to) +static int udf_pc_to_char(struct super_block *sb, unsigned char *from, + int fromlen, unsigned char *to, int tolen) { struct pathComponent *pc; int elen = 0; + int comp_len; unsigned char *p = to; + /* Reserve one byte for terminating \0 */ + tolen--; while (elen < fromlen) { pc = (struct pathComponent *)(from + elen); + elen += sizeof(struct pathComponent); switch (pc->componentType) { case 1: - if (pc->lengthComponentIdent == 0) { - p = to; - *p++ = '/'; + /* + * Symlink points to some place which should be agreed + * upon between originator and receiver of the media. Ignore. + */ + if (pc->lengthComponentIdent > 0) { + elen += pc->lengthComponentIdent; + break; } + /* Fall through */ + case 2: + if (tolen == 0) + return -ENAMETOOLONG; + p = to; + *p++ = '/'; + tolen--; break; case 3: + if (tolen < 3) + return -ENAMETOOLONG; memcpy(p, "../", 3); p += 3; + tolen -= 3; break; case 4: + if (tolen < 2) + return -ENAMETOOLONG; memcpy(p, "./", 2); p += 2; + tolen -= 2; /* that would be . - just ignore */ break; case 5: - p += udf_get_filename(sb, pc->componentIdent, p, - pc->lengthComponentIdent); + elen += pc->lengthComponentIdent; + if (elen > fromlen) + return -EIO; + comp_len = udf_get_filename(sb, pc->componentIdent, + pc->lengthComponentIdent, + p, tolen); + p += comp_len; + tolen -= comp_len; + if (tolen == 0) + return -ENAMETOOLONG; *p++ = '/'; + tolen--; break; } - elen += sizeof(struct pathComponent) + pc->lengthComponentIdent; } if (p > to + 1) p[-1] = '\0'; else p[0] = '\0'; + return 0; } static int udf_symlink_filler(struct file *file, struct page *page) @@ -74,11 +104,17 @@ static int udf_symlink_filler(struct file *file, struct page *page) struct inode *inode = page->mapping->host; struct buffer_head *bh = NULL; unsigned char *symlink; - int err = -EIO; + int err; unsigned char *p = kmap(page); struct udf_inode_info *iinfo; uint32_t pos; + /* We don't support symlinks longer than one block */ + if (inode->i_size > inode->i_sb->s_blocksize) { + err = -ENAMETOOLONG; + goto out_unmap; + } + iinfo = UDF_I(inode); pos = udf_block_map(inode, 0); @@ -88,14 +124,18 @@ static int udf_symlink_filler(struct file *file, struct page *page) } else { bh = sb_bread(inode->i_sb, pos); - if (!bh) - goto out; + if (!bh) { + err = -EIO; + goto out_unlock_inode; + } symlink = bh->b_data; } - udf_pc_to_char(inode->i_sb, symlink, inode->i_size, p); + err = udf_pc_to_char(inode->i_sb, symlink, inode->i_size, p, PAGE_SIZE); brelse(bh); + if (err) + goto out_unlock_inode; up_read(&iinfo->i_data_sem); SetPageUptodate(page); @@ -103,9 +143,10 @@ static int udf_symlink_filler(struct file *file, struct page *page) unlock_page(page); return 0; -out: +out_unlock_inode: up_read(&iinfo->i_data_sem); SetPageError(page); +out_unmap: kunmap(page); unlock_page(page); return err; diff --git a/fs/udf/truncate.c b/fs/udf/truncate.c index 8424308..4b98fee 100644 --- a/fs/udf/truncate.c +++ b/fs/udf/truncate.c @@ -95,23 +95,21 @@ void udf_truncate_tail_extent(struct inode *inode) lbcount += elen; if (lbcount > inode->i_size) { if (lbcount - inode->i_size >= inode->i_sb->s_blocksize) - printk(KERN_WARNING - "udf_truncate_tail_extent(): Too long " - "extent after EOF in inode %u: i_size: " - "%Ld lbcount: %Ld extent %u+%u\n", - (unsigned)inode->i_ino, - (long long)inode->i_size, - (long long)lbcount, - (unsigned)eloc.logicalBlockNum, - (unsigned)elen); + udf_warn(inode->i_sb, + "Too long extent after EOF in inode %u: i_size: %lld lbcount: %lld extent %u+%u\n", + (unsigned)inode->i_ino, + (long long)inode->i_size, + (long long)lbcount, + (unsigned)eloc.logicalBlockNum, + (unsigned)elen); nelen = elen - (lbcount - inode->i_size); epos.offset -= adsize; extent_trunc(inode, &epos, &eloc, etype, elen, nelen); epos.offset += adsize; if (udf_next_aext(inode, &epos, &eloc, &elen, 1) != -1) - printk(KERN_ERR "udf_truncate_tail_extent(): " - "Extent after EOF in inode %u.\n", - (unsigned)inode->i_ino); + udf_err(inode->i_sb, + "Extent after EOF in inode %u\n", + (unsigned)inode->i_ino); break; } } diff --git a/fs/udf/udf_sb.h b/fs/udf/udf_sb.h index 54706dc..604f5fc 100644 --- a/fs/udf/udf_sb.h +++ b/fs/udf/udf_sb.h @@ -54,13 +54,16 @@ #pragma pack(1) /* XXX(hch): Why? This file just defines in-core structures */ +#define MF_DUPLICATE_MD 0x01 +#define MF_MIRROR_FE_LOADED 0x02 + struct udf_meta_data { __u32 s_meta_file_loc; __u32 s_mirror_file_loc; __u32 s_bitmap_file_loc; __u32 s_alloc_unit_size; __u16 s_align_unit_size; - __u8 s_dup_md_flag; + int s_flags; struct inode *s_metadata_fe; struct inode *s_mirror_fe; struct inode *s_bitmap_fe; diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index dbd52d4b..8775ab2 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -1,6 +1,8 @@ #ifndef __UDF_DECL_H #define __UDF_DECL_H +#define pr_fmt(fmt) "UDF-fs: " fmt + #include "ecma_167.h" #include "osta_udf.h" @@ -16,23 +18,30 @@ #define UDF_PREALLOCATE #define UDF_DEFAULT_PREALLOC_BLOCKS 8 +extern __printf(3, 4) void _udf_err(struct super_block *sb, + const char *function, const char *fmt, ...); +#define udf_err(sb, fmt, ...) \ + _udf_err(sb, __func__, fmt, ##__VA_ARGS__) + +extern __printf(3, 4) void _udf_warn(struct super_block *sb, + const char *function, const char *fmt, ...); +#define udf_warn(sb, fmt, ...) \ + _udf_warn(sb, __func__, fmt, ##__VA_ARGS__) + +#define udf_info(fmt, ...) \ + pr_info("INFO " fmt, ##__VA_ARGS__) + #undef UDFFS_DEBUG #ifdef UDFFS_DEBUG -#define udf_debug(f, a...) \ -do { \ - printk(KERN_DEBUG "UDF-fs DEBUG %s:%d:%s: ", \ - __FILE__, __LINE__, __func__); \ - printk(f, ##a); \ -} while (0) +#define udf_debug(fmt, ...) \ + printk(KERN_DEBUG pr_fmt("%s:%d:%s: " fmt), \ + __FILE__, __LINE__, __func__, ##__VA_ARGS__) #else -#define udf_debug(f, a...) /**/ +#define udf_debug(fmt, ...) \ + no_printk(fmt, ##__VA_ARGS__) #endif -#define udf_info(f, a...) \ - printk(KERN_INFO "UDF-fs INFO " f, ##a); - - #define udf_fixed_to_variable(x) ( ( ( (x) >> 5 ) * 39 ) + ( (x) & 0x0000001F ) ) #define udf_variable_to_fixed(x) ( ( ( (x) / 39 ) << 5 ) + ( (x) % 39 ) ) @@ -112,8 +121,6 @@ struct extent_position { /* super.c */ -__attribute__((format(printf, 3, 4))) -extern void udf_warning(struct super_block *, const char *, const char *, ...); static inline void udf_updated_lvid(struct super_block *sb) { struct buffer_head *bh = UDF_SB(sb)->s_lvid_bh; @@ -126,6 +133,8 @@ static inline void udf_updated_lvid(struct super_block *sb) UDF_SB(sb)->s_lvid_dirty = 1; } extern u64 lvid_get_unique_id(struct super_block *sb); +struct inode *udf_find_metadata_inode_efe(struct super_block *sb, + u32 meta_file_loc, u32 partition_num); /* namei.c */ extern int udf_write_fi(struct inode *inode, struct fileIdentDesc *, @@ -198,7 +207,8 @@ udf_get_lb_pblock(struct super_block *sb, struct kernel_lb_addr *loc, } /* unicode.c */ -extern int udf_get_filename(struct super_block *, uint8_t *, uint8_t *, int); +extern int udf_get_filename(struct super_block *, uint8_t *, int, uint8_t *, + int); extern int udf_put_filename(struct super_block *, const uint8_t *, uint8_t *, int); extern int udf_build_ustr(struct ustr *, dstring *, int); diff --git a/fs/udf/udftime.c b/fs/udf/udftime.c index b8c828c..1f11483 100644 --- a/fs/udf/udftime.c +++ b/fs/udf/udftime.c @@ -34,9 +34,10 @@ * http://www.boulder.nist.gov/timefreq/pubs/bulletin/leapsecond.htm */ +#include "udfdecl.h" + #include <linux/types.h> #include <linux/kernel.h> -#include "udfdecl.h" #define EPOCH_YEAR 1970 diff --git a/fs/udf/unicode.c b/fs/udf/unicode.c index d03a90b..d29c06f 100644 --- a/fs/udf/unicode.c +++ b/fs/udf/unicode.c @@ -28,7 +28,8 @@ #include "udf_sb.h" -static int udf_translate_to_linux(uint8_t *, uint8_t *, int, uint8_t *, int); +static int udf_translate_to_linux(uint8_t *, int, uint8_t *, int, uint8_t *, + int); static int udf_char_to_ustr(struct ustr *dest, const uint8_t *src, int strlen) { @@ -114,7 +115,7 @@ int udf_CS0toUTF8(struct ustr *utf_o, const struct ustr *ocu_i) cmp_id = ocu_i->u_cmpID; if (cmp_id != 8 && cmp_id != 16) { memset(utf_o, 0, sizeof(struct ustr)); - printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", + pr_err("unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name); return 0; } @@ -242,7 +243,7 @@ try_again: if (utf_cnt) { error_out: ocu[++u_len] = '?'; - printk(KERN_DEBUG "udf: bad UTF-8 character\n"); + printk(KERN_DEBUG pr_fmt("bad UTF-8 character\n")); } ocu[length - 1] = (uint8_t)u_len + 1; @@ -267,7 +268,7 @@ static int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o, cmp_id = ocu_i->u_cmpID; if (cmp_id != 8 && cmp_id != 16) { memset(utf_o, 0, sizeof(struct ustr)); - printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", + pr_err("unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name); return 0; } @@ -333,8 +334,8 @@ try_again: return u_len + 1; } -int udf_get_filename(struct super_block *sb, uint8_t *sname, uint8_t *dname, - int flen) +int udf_get_filename(struct super_block *sb, uint8_t *sname, int slen, + uint8_t *dname, int dlen) { struct ustr *filename, *unifilename; int len = 0; @@ -347,7 +348,7 @@ int udf_get_filename(struct super_block *sb, uint8_t *sname, uint8_t *dname, if (!unifilename) goto out1; - if (udf_build_ustr_exact(unifilename, sname, flen)) + if (udf_build_ustr_exact(unifilename, sname, slen)) goto out2; if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) { @@ -366,7 +367,8 @@ int udf_get_filename(struct super_block *sb, uint8_t *sname, uint8_t *dname, } else goto out2; - len = udf_translate_to_linux(dname, filename->u_name, filename->u_len, + len = udf_translate_to_linux(dname, dlen, + filename->u_name, filename->u_len, unifilename->u_name, unifilename->u_len); out2: kfree(unifilename); @@ -403,10 +405,12 @@ int udf_put_filename(struct super_block *sb, const uint8_t *sname, #define EXT_MARK '.' #define CRC_MARK '#' #define EXT_SIZE 5 +/* Number of chars we need to store generated CRC to make filename unique */ +#define CRC_LEN 5 -static int udf_translate_to_linux(uint8_t *newName, uint8_t *udfName, - int udfLen, uint8_t *fidName, - int fidNameLen) +static int udf_translate_to_linux(uint8_t *newName, int newLen, + uint8_t *udfName, int udfLen, + uint8_t *fidName, int fidNameLen) { int index, newIndex = 0, needsCRC = 0; int extIndex = 0, newExtIndex = 0, hasExt = 0; @@ -440,7 +444,7 @@ static int udf_translate_to_linux(uint8_t *newName, uint8_t *udfName, newExtIndex = newIndex; } } - if (newIndex < 256) + if (newIndex < newLen) newName[newIndex++] = curr; else needsCRC = 1; @@ -468,13 +472,13 @@ static int udf_translate_to_linux(uint8_t *newName, uint8_t *udfName, } ext[localExtIndex++] = curr; } - maxFilenameLen = 250 - localExtIndex; + maxFilenameLen = newLen - CRC_LEN - localExtIndex; if (newIndex > maxFilenameLen) newIndex = maxFilenameLen; else newIndex = newExtIndex; - } else if (newIndex > 250) - newIndex = 250; + } else if (newIndex > newLen - CRC_LEN) + newIndex = newLen - CRC_LEN; newName[newIndex++] = CRC_MARK; valueCRC = crc_itu_t(0, fidName, fidNameLen); newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12]; diff --git a/fs/ufs/ialloc.c b/fs/ufs/ialloc.c index 2eabf04..78a4c70 100644 --- a/fs/ufs/ialloc.c +++ b/fs/ufs/ialloc.c @@ -341,7 +341,7 @@ cg_found: fail_remove_inode: unlock_super(sb); - inode->i_nlink = 0; + clear_nlink(inode); iput(inode); UFSD("EXIT (FAILED): err %d\n", err); return ERR_PTR(err); diff --git a/fs/ufs/inode.c b/fs/ufs/inode.c index b4d791a..879b134 100644 --- a/fs/ufs/inode.c +++ b/fs/ufs/inode.c @@ -589,7 +589,7 @@ static int ufs1_read_inode(struct inode *inode, struct ufs_inode *ufs_inode) * Copy data to the in-core inode. */ inode->i_mode = mode = fs16_to_cpu(sb, ufs_inode->ui_mode); - inode->i_nlink = fs16_to_cpu(sb, ufs_inode->ui_nlink); + set_nlink(inode, fs16_to_cpu(sb, ufs_inode->ui_nlink)); if (inode->i_nlink == 0) { ufs_error (sb, "ufs_read_inode", "inode %lu has zero nlink\n", inode->i_ino); return -1; @@ -637,7 +637,7 @@ static int ufs2_read_inode(struct inode *inode, struct ufs2_inode *ufs2_inode) * Copy data to the in-core inode. */ inode->i_mode = mode = fs16_to_cpu(sb, ufs2_inode->ui_mode); - inode->i_nlink = fs16_to_cpu(sb, ufs2_inode->ui_nlink); + set_nlink(inode, fs16_to_cpu(sb, ufs2_inode->ui_nlink)); if (inode->i_nlink == 0) { ufs_error (sb, "ufs_read_inode", "inode %lu has zero nlink\n", inode->i_ino); return -1; diff --git a/fs/ufs/namei.c b/fs/ufs/namei.c index b57aab9..639d491 100644 --- a/fs/ufs/namei.c +++ b/fs/ufs/namei.c @@ -59,8 +59,6 @@ static struct dentry *ufs_lookup(struct inode * dir, struct dentry *dentry, stru if (ino) inode = ufs_iget(dir->i_sb, ino); unlock_ufs(dir->i_sb); - if (IS_ERR(inode)) - return ERR_CAST(inode); return d_splice_alias(inode, dentry); } diff --git a/fs/ufs/ufs.h b/fs/ufs/ufs.h index 5be2755..c26f2bc 100644 --- a/fs/ufs/ufs.h +++ b/fs/ufs/ufs.h @@ -117,9 +117,12 @@ extern int ufs_getfrag_block (struct inode *inode, sector_t fragment, struct buf extern const struct file_operations ufs_dir_operations; /* super.c */ -extern void ufs_warning (struct super_block *, const char *, const char *, ...) __attribute__ ((format (printf, 3, 4))); -extern void ufs_error (struct super_block *, const char *, const char *, ...) __attribute__ ((format (printf, 3, 4))); -extern void ufs_panic (struct super_block *, const char *, const char *, ...) __attribute__ ((format (printf, 3, 4))); +extern __printf(3, 4) +void ufs_warning(struct super_block *, const char *, const char *, ...); +extern __printf(3, 4) +void ufs_error(struct super_block *, const char *, const char *, ...); +extern __printf(3, 4) +void ufs_panic(struct super_block *, const char *, const char *, ...); /* symlink.c */ extern const struct inode_operations ufs_fast_symlink_inode_operations; |