56 files changed, 1136 insertions, 770 deletions
diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c
index 87d95a8cddbc..f55ae23b137e 100644
--- a/fs/autofs4/root.c
+++ b/fs/autofs4/root.c
@@ -583,8 +583,6 @@ static int autofs4_dir_unlink(struct inode *dir, struct dentry *dentry)
        if (!autofs4_oz_mode(sbi) && !capable(CAP_SYS_ADMIN))
                return -EACCES;
-        dentry_unhash(dentry);
        if (atomic_dec_and_test(&ino->count)) {
                p_ino = autofs4_dentry_ino(dentry->d_parent);
                if (p_ino && dentry->d_parent != dentry)
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 1f2b19978333..1a2421f908f0 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -1272,8 +1272,8 @@ int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder)
                 * individual writeable reference is too fragile given the
                 * way @mode is used in blkdev_get/put().
                 */
-                if ((disk->flags & GENHD_FL_BLOCK_EVENTS_ON_EXCL_WRITE) &&
+                if (!res && (mode & FMODE_WRITE) && !bdev->bd_write_holder &&
-                    !res && (mode & FMODE_WRITE) && !bdev->bd_write_holder) {
+                    (disk->flags & GENHD_FL_BLOCK_EVENTS_ON_EXCL_WRITE)) {
                        bdev->bd_write_holder = true;
                        disk_block_events(disk);
                }
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 93b1aa932014..52d7eca8c7bf 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -121,9 +121,6 @@ struct btrfs_inode {
         */
        u64 index_cnt;
-        /* the start of block group preferred for allocations. */
-        u64 block_group;
        /* the fsync log has some corner cases that mean we have to check
         * directories to see if any unlinks have been done before
         * the directory was logged.  See tree-log.c for all the
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index b0e18d986e0a..2e667868e0d2 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -43,8 +43,6 @@ struct btrfs_path *btrfs_alloc_path(void)
 {
        struct btrfs_path *path;
        path = kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS);
-        if (path)
-                path->reada = 1;
        return path;
 }
@@ -1224,11 +1222,13 @@ static void reada_for_search(struct btrfs_root *root,
        u64 search;
        u64 target;
        u64 nread = 0;
+        u64 gen;
        int direction = path->reada;
        struct extent_buffer *eb;
        u32 nr;
        u32 blocksize;
        u32 nscan = 0;
+        bool map = true;
        if (level != 1)
                return;
@@ -1250,7 +1250,19 @@ 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;
@@ -1268,14 +1280,23 @@ static void reada_for_search(struct btrfs_root *root,
                search = btrfs_node_blockptr(node, nr);
                if ((search <= target && target - search <= 65536) ||
                    (search > target && search - target <= 65536)) {
-                        readahead_tree_block(root, search, blocksize,
+                        gen = btrfs_node_ptr_generation(node, nr);
-                                     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;
                }
                nscan++;
                if ((nread > 65536 || nscan > 32))
                        break;
        }
+        if (map && node->map_token) {
+                unmap_extent_buffer(node, node->map_token, KM_USER1);
+                node->map_token = NULL;
+        }
 }
 /*
@@ -1648,9 +1669,6 @@ again:
                }
 cow_done:
                BUG_ON(!cow && ins_len);
-                if (level != btrfs_header_level(b))
-                        WARN_ON(1);
-                level = btrfs_header_level(b);
                p->nodes[level] = b;
                if (!p->skip_locking)
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 6c093fa98f61..378b5b4443f3 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -930,7 +930,6 @@ struct btrfs_fs_info {
         * is required instead of the faster short fsync log commits
         */
        u64 last_trans_log_full_commit;
-        u64 open_ioctl_trans;
        unsigned long mount_opt:20;
        unsigned long compress_type:4;
        u64 max_inline;
@@ -947,7 +946,6 @@ struct btrfs_fs_info {
        struct super_block *sb;
        struct inode *btree_inode;
        struct backing_dev_info bdi;
-        struct mutex trans_mutex;
        struct mutex tree_log_mutex;
        struct mutex transaction_kthread_mutex;
        struct mutex cleaner_mutex;
@@ -968,6 +966,7 @@ struct btrfs_fs_info {
        struct rw_semaphore subvol_sem;
        struct srcu_struct subvol_srcu;
+        spinlock_t trans_lock;
        struct list_head trans_list;
        struct list_head hashers;
        struct list_head dead_roots;
@@ -980,6 +979,7 @@ struct btrfs_fs_info {
        atomic_t async_submit_draining;
        atomic_t nr_async_bios;
        atomic_t async_delalloc_pages;
+        atomic_t open_ioctl_trans;
        /*
         * this is used by the balancing code to wait for all the pending
@@ -1044,6 +1044,7 @@ struct btrfs_fs_info {
        int closing;
        int log_root_recovering;
        int enospc_unlink;
+        int trans_no_join;
        u64 total_pinned;
@@ -1065,7 +1066,6 @@ struct btrfs_fs_info {
        struct reloc_control *reloc_ctl;
        spinlock_t delalloc_lock;
-        spinlock_t new_trans_lock;
        u64 delalloc_bytes;
        /* data_alloc_cluster is only used in ssd mode */
@@ -1340,6 +1340,7 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_MOUNT_USER_SUBVOL_RM_ALLOWED (1 << 14)
 #define BTRFS_MOUNT_ENOSPC_DEBUG         (1 << 15)
 #define BTRFS_MOUNT_AUTO_DEFRAG         (1 << 16)
+#define BTRFS_MOUNT_INODE_MAP_CACHE     (1 << 17)
 #define btrfs_clear_opt(o, opt)         ((o) &= ~BTRFS_MOUNT_##opt)
 #define btrfs_set_opt(o, opt)           ((o) |= BTRFS_MOUNT_##opt)
@@ -2238,6 +2239,9 @@ int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
 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,
@@ -2350,6 +2354,15 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
                        struct btrfs_root *root,
                        struct extent_buffer *node,
                        struct extent_buffer *parent);
+static inline int btrfs_fs_closing(struct btrfs_fs_info *fs_info)
+{
+        /*
+         * Get synced with close_ctree()
+         */
+        smp_mb();
+        return fs_info->closing;
+}
 /* root-item.c */
 int btrfs_find_root_ref(struct btrfs_root *tree_root,
                        struct btrfs_path *path,
@@ -2512,8 +2525,7 @@ int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
 int btrfs_writepages(struct address_space *mapping,
                     struct writeback_control *wbc);
 int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
-                             struct btrfs_root *new_root,
+                             struct btrfs_root *new_root, u64 new_dirid);
-                             u64 new_dirid, u64 alloc_hint);
 int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
                         size_t size, struct bio *bio, unsigned long bio_flags);
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index 01e29503a54b..6462c29d2d37 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -678,6 +678,7 @@ static int btrfs_batch_insert_items(struct btrfs_trans_handle *trans,
        INIT_LIST_HEAD(&head);
        next = item;
+        nitems = 0;
        /*
         * count the number of the continuous items that we can insert in batch
@@ -1129,7 +1130,7 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work)
        delayed_node = async_node->delayed_node;
        root = delayed_node->root;
-        trans = btrfs_join_transaction(root, 0);
+        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans))
                goto free_path;
@@ -1572,8 +1573,7 @@ static void fill_stack_inode_item(struct btrfs_trans_handle *trans,
        btrfs_set_stack_inode_transid(inode_item, trans->transid);
        btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev);
        btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags);
-        btrfs_set_stack_inode_block_group(inode_item,
+        btrfs_set_stack_inode_block_group(inode_item, 0);
-                                          BTRFS_I(inode)->block_group);
        btrfs_set_stack_timespec_sec(btrfs_inode_atime(inode_item),
                                     inode->i_atime.tv_sec);
@@ -1595,7 +1595,7 @@ int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans,
                               struct btrfs_root *root, struct inode *inode)
 {
        struct btrfs_delayed_node *delayed_node;
-        int ret;
+        int ret = 0;
        delayed_node = btrfs_get_or_create_delayed_node(inode);
        if (IS_ERR(delayed_node))
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 98b6a71decba..9f68c6898653 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1505,24 +1505,24 @@ static int transaction_kthread(void *arg)
                vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE);
                mutex_lock(&root->fs_info->transaction_kthread_mutex);
-                spin_lock(&root->fs_info->new_trans_lock);
+                spin_lock(&root->fs_info->trans_lock);
                cur = root->fs_info->running_transaction;
                if (!cur) {
-                        spin_unlock(&root->fs_info->new_trans_lock);
+                        spin_unlock(&root->fs_info->trans_lock);
                        goto sleep;
                }
                now = get_seconds();
                if (!cur->blocked &&
                    (now < cur->start_time || now - cur->start_time < 30)) {
-                        spin_unlock(&root->fs_info->new_trans_lock);
+                        spin_unlock(&root->fs_info->trans_lock);
                        delay = HZ * 5;
                        goto sleep;
                }
                transid = cur->transid;
-                spin_unlock(&root->fs_info->new_trans_lock);
+                spin_unlock(&root->fs_info->trans_lock);
-                trans = btrfs_join_transaction(root, 1);
+                trans = btrfs_join_transaction(root);
                BUG_ON(IS_ERR(trans));
                if (transid == trans->transid) {
                        ret = btrfs_commit_transaction(trans, root);
@@ -1613,7 +1613,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
        INIT_LIST_HEAD(&fs_info->ordered_operations);
        INIT_LIST_HEAD(&fs_info->caching_block_groups);
        spin_lock_init(&fs_info->delalloc_lock);
-        spin_lock_init(&fs_info->new_trans_lock);
+        spin_lock_init(&fs_info->trans_lock);
        spin_lock_init(&fs_info->ref_cache_lock);
        spin_lock_init(&fs_info->fs_roots_radix_lock);
        spin_lock_init(&fs_info->delayed_iput_lock);
@@ -1645,6 +1645,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
        fs_info->max_inline = 8192 * 1024;
        fs_info->metadata_ratio = 0;
        fs_info->defrag_inodes = RB_ROOT;
+        fs_info->trans_no_join = 0;
        fs_info->thread_pool_size = min_t(unsigned long,
                                          num_online_cpus() + 2, 8);
@@ -1667,8 +1668,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
        init_waitqueue_head(&fs_info->scrub_pause_wait);
        init_rwsem(&fs_info->scrub_super_lock);
        fs_info->scrub_workers_refcnt = 0;
-        btrfs_init_workers(&fs_info->scrub_workers, "scrub",
-                           fs_info->thread_pool_size, &fs_info->generic_worker);
        sb->s_blocksize = 4096;
        sb->s_blocksize_bits = blksize_bits(4096);
@@ -1709,7 +1708,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
        fs_info->do_barriers = 1;
-        mutex_init(&fs_info->trans_mutex);
        mutex_init(&fs_info->ordered_operations_mutex);
        mutex_init(&fs_info->tree_log_mutex);
        mutex_init(&fs_info->chunk_mutex);
@@ -2479,13 +2477,13 @@ int btrfs_commit_super(struct btrfs_root *root)
        down_write(&root->fs_info->cleanup_work_sem);
        up_write(&root->fs_info->cleanup_work_sem);
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans))
                return PTR_ERR(trans);
        ret = btrfs_commit_transaction(trans, root);
        BUG_ON(ret);
        /* run commit again to drop the original snapshot */
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans))
                return PTR_ERR(trans);
        btrfs_commit_transaction(trans, root);
@@ -2911,9 +2909,8 @@ static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
        INIT_LIST_HEAD(&splice);
-        list_splice_init(&root->fs_info->delalloc_inodes, &splice);
        spin_lock(&root->fs_info->delalloc_lock);
+        list_splice_init(&root->fs_info->delalloc_inodes, &splice);
        while (!list_empty(&splice)) {
                btrfs_inode = list_entry(splice.next, struct btrfs_inode,
@@ -3024,10 +3021,13 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
        WARN_ON(1);
-        mutex_lock(&root->fs_info->trans_mutex);
        mutex_lock(&root->fs_info->transaction_kthread_mutex);
+        spin_lock(&root->fs_info->trans_lock);
        list_splice_init(&root->fs_info->trans_list, &list);
+        root->fs_info->trans_no_join = 1;
+        spin_unlock(&root->fs_info->trans_lock);
        while (!list_empty(&list)) {
                t = list_entry(list.next, struct btrfs_transaction, list);
                if (!t)
@@ -3052,23 +3052,18 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
                t->blocked = 0;
                if (waitqueue_active(&root->fs_info->transaction_wait))
                        wake_up(&root->fs_info->transaction_wait);
-                mutex_unlock(&root->fs_info->trans_mutex);
-                mutex_lock(&root->fs_info->trans_mutex);
                t->commit_done = 1;
                if (waitqueue_active(&t->commit_wait))
                        wake_up(&t->commit_wait);
-                mutex_unlock(&root->fs_info->trans_mutex);
-                mutex_lock(&root->fs_info->trans_mutex);
                btrfs_destroy_pending_snapshots(t);
                btrfs_destroy_delalloc_inodes(root);
-                spin_lock(&root->fs_info->new_trans_lock);
+                spin_lock(&root->fs_info->trans_lock);
                root->fs_info->running_transaction = NULL;
-                spin_unlock(&root->fs_info->new_trans_lock);
+                spin_unlock(&root->fs_info->trans_lock);
                btrfs_destroy_marked_extents(root, &t->dirty_pages,
                                             EXTENT_DIRTY);
@@ -3082,8 +3077,10 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
                kmem_cache_free(btrfs_transaction_cachep, t);
        }
+        spin_lock(&root->fs_info->trans_lock);
+        root->fs_info->trans_no_join = 0;
+        spin_unlock(&root->fs_info->trans_lock);
        mutex_unlock(&root->fs_info->transaction_kthread_mutex);
-        mutex_unlock(&root->fs_info->trans_mutex);
        return 0;
 }
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 169bd62ce776..b42efc2ded51 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -348,7 +348,7 @@ static int caching_kthread(void *data)
         */
        path->skip_locking = 1;
        path->search_commit_root = 1;
-        path->reada = 2;
+        path->reada = 1;
        key.objectid = last;
        key.offset = 0;
@@ -366,8 +366,7 @@ again:
        nritems = btrfs_header_nritems(leaf);
        while (1) {
-                smp_mb();
+                if (btrfs_fs_closing(fs_info) > 1) {
-                if (fs_info->closing > 1) {
                        last = (u64)-1;
                        break;
                }
@@ -379,15 +378,18 @@ again:
                        if (ret)
                                break;
-                        caching_ctl->progress = last;
+                        if (need_resched() ||
-                        btrfs_release_path(path);
+                            btrfs_next_leaf(extent_root, path)) {
-                        up_read(&fs_info->extent_commit_sem);
+                                caching_ctl->progress = last;
-                        mutex_unlock(&caching_ctl->mutex);
+                                btrfs_release_path(path);
-                        if (btrfs_transaction_in_commit(fs_info))
+                                up_read(&fs_info->extent_commit_sem);
-                                schedule_timeout(1);
+                                mutex_unlock(&caching_ctl->mutex);
-                        else
                                cond_resched();
-                        goto again;
+                                goto again;
+                        }
+                        leaf = path->nodes[0];
+                        nritems = btrfs_header_nritems(leaf);
+                        continue;
                }
                if (key.objectid < block_group->key.objectid) {
@@ -3065,7 +3067,7 @@ again:
                        spin_unlock(&data_sinfo->lock);
 alloc:
                        alloc_target = btrfs_get_alloc_profile(root, 1);
-                        trans = btrfs_join_transaction(root, 1);
+                        trans = btrfs_join_transaction(root);
                        if (IS_ERR(trans))
                                return PTR_ERR(trans);
@@ -3087,13 +3089,21 @@ alloc:
                        }
                        goto again;
                }
+                /*
+                 * If we have less pinned bytes than we want to allocate then
+                 * don't bother committing the transaction, it won't help us.
+                 */
+                if (data_sinfo->bytes_pinned < bytes)
+                        committed = 1;
                spin_unlock(&data_sinfo->lock);
                /* commit the current transaction and try again */
 commit_trans:
-                if (!committed && !root->fs_info->open_ioctl_trans) {
+                if (!committed &&
+                    !atomic_read(&root->fs_info->open_ioctl_trans)) {
                        committed = 1;
-                        trans = btrfs_join_transaction(root, 1);
+                        trans = btrfs_join_transaction(root);
                        if (IS_ERR(trans))
                                return PTR_ERR(trans);
                        ret = btrfs_commit_transaction(trans, root);
@@ -3472,7 +3482,7 @@ again:
                goto out;
        ret = -ENOSPC;
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans))
                goto out;
        ret = btrfs_commit_transaction(trans, root);
@@ -3699,7 +3709,7 @@ int btrfs_block_rsv_check(struct btrfs_trans_handle *trans,
                if (trans)
                        return -EAGAIN;
-                trans = btrfs_join_transaction(root, 1);
+                trans = btrfs_join_transaction(root);
                BUG_ON(IS_ERR(trans));
                ret = btrfs_commit_transaction(trans, root);
                return 0;
@@ -3837,6 +3847,37 @@ static void release_global_block_rsv(struct btrfs_fs_info *fs_info)
        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)
@@ -3877,23 +3918,18 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
        struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv;
        /*
-         * one for deleting orphan item, one for updating inode and
+         * We need to hold space in order to delete our orphan item once we've
-         * two for calling btrfs_truncate_inode_items.
+         * added it, so this takes the reservation so we can release it later
-         *
+         * when we are truly done with the orphan item.
-         * btrfs_truncate_inode_items is a delete operation, it frees
-         * more space than it uses in most cases. So two units of
-         * metadata space should be enough for calling it many times.
-         * If all of the metadata space is used, we can commit
-         * transaction and use space it freed.
         */
-        u64 num_bytes = btrfs_calc_trans_metadata_size(root, 4);
+        u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1);
        return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
 }
 void btrfs_orphan_release_metadata(struct inode *inode)
 {
        struct btrfs_root *root = BTRFS_I(inode)->root;
-        u64 num_bytes = btrfs_calc_trans_metadata_size(root, 4);
+        u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1);
        btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes);
 }
@@ -4987,6 +5023,15 @@ have_block_group:
                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) {
+                        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
@@ -5150,6 +5195,7 @@ checks:
                        btrfs_add_free_space(block_group, offset,
                                             search_start - offset);
                BUG_ON(offset > search_start);
+                btrfs_put_block_group(block_group);
                break;
 loop:
                failed_cluster_refill = false;
@@ -5172,9 +5218,7 @@ loop:
         * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try
         *                      again
         */
-        if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE &&
+        if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) {
-            (found_uncached_bg || empty_size || empty_cluster ||
-             allowed_chunk_alloc)) {
                index = 0;
                if (loop == LOOP_FIND_IDEAL && found_uncached_bg) {
                        found_uncached_bg = false;
@@ -5214,42 +5258,39 @@ loop:
                        goto search;
                }
-                if (loop < LOOP_CACHING_WAIT) {
+                loop++;
-                        loop++;
-                        goto search;
-                }
                if (loop == LOOP_ALLOC_CHUNK) {
-                        empty_size = 0;
+                       if (allowed_chunk_alloc) {
-                        empty_cluster = 0;
+                                ret = do_chunk_alloc(trans, root, num_bytes +
-                }
+                                                     2 * 1024 * 1024, data,
+                                                     CHUNK_ALLOC_LIMITED);
+                                allowed_chunk_alloc = 0;
+                                if (ret == 1)
+                                        done_chunk_alloc = 1;
+                        } else if (!done_chunk_alloc &&
+                                   space_info->force_alloc ==
+                                   CHUNK_ALLOC_NO_FORCE) {
+                                space_info->force_alloc = CHUNK_ALLOC_LIMITED;
+                        }
-                if (allowed_chunk_alloc) {
+                       /*
-                        ret = do_chunk_alloc(trans, root, num_bytes +
+                        * We didn't allocate a chunk, go ahead and drop the
-                                             2 * 1024 * 1024, data,
+                        * empty size and loop again.
-                                             CHUNK_ALLOC_LIMITED);
+                        */
-                        allowed_chunk_alloc = 0;
+                       if (!done_chunk_alloc)
-                        done_chunk_alloc = 1;
+                               loop = LOOP_NO_EMPTY_SIZE;
-                } else if (!done_chunk_alloc &&
-                           space_info->force_alloc == CHUNK_ALLOC_NO_FORCE) {
-                        space_info->force_alloc = CHUNK_ALLOC_LIMITED;
                }
-                if (loop < LOOP_NO_EMPTY_SIZE) {
+                if (loop == LOOP_NO_EMPTY_SIZE) {
-                        loop++;
+                        empty_size = 0;
-                        goto search;
+                        empty_cluster = 0;
                }
-                ret = -ENOSPC;
+                goto search;
        } else if (!ins->objectid) {
                ret = -ENOSPC;
-        }
+        } else if (ins->objectid) {
-        /* we found what we needed */
-        if (ins->objectid) {
-                if (!(data & BTRFS_BLOCK_GROUP_DATA))
-                        trans->block_group = block_group->key.objectid;
-                btrfs_put_block_group(block_group);
                ret = 0;
        }
@@ -6526,7 +6567,7 @@ int btrfs_set_block_group_ro(struct btrfs_root *root,
        BUG_ON(cache->ro);
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        BUG_ON(IS_ERR(trans));
        alloc_flags = update_block_group_flags(root, cache->flags);
@@ -6882,6 +6923,7 @@ int btrfs_read_block_groups(struct btrfs_root *root)
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
+        path->reada = 1;
        cache_gen = btrfs_super_cache_generation(&root->fs_info->super_copy);
        if (cache_gen != 0 &&
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index c5d9fbb92bc3..7055d11c1efd 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -1476,7 +1476,7 @@ u64 count_range_bits(struct extent_io_tree *tree,
                        if (total_bytes >= max_bytes)
                                break;
                        if (!found) {
-                                *start = state->start;
+                                *start = max(cur_start, state->start);
                                found = 1;
                        }
                        last = state->end;
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index 4e8445a4757c..a11a92ee2d30 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -126,9 +126,9 @@ struct extent_buffer {
        unsigned long map_len;
        struct page *first_page;
        unsigned long bflags;
-        atomic_t refs;
        struct list_head leak_list;
        struct rcu_head rcu_head;
+        atomic_t refs;
        /* the spinlock is used to protect most operations */
        spinlock_t lock;
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index c6a22d783c35..fa4ef18b66b1 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -129,7 +129,7 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
        if (!btrfs_test_opt(root, AUTO_DEFRAG))
                return 0;
-        if (root->fs_info->closing)
+        if (btrfs_fs_closing(root->fs_info))
                return 0;
        if (BTRFS_I(inode)->in_defrag)
@@ -144,7 +144,7 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
        if (!defrag)
                return -ENOMEM;
-        defrag->ino = inode->i_ino;
+        defrag->ino = btrfs_ino(inode);
        defrag->transid = transid;
        defrag->root = root->root_key.objectid;
@@ -229,7 +229,7 @@ int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info)
                first_ino = defrag->ino + 1;
                rb_erase(&defrag->rb_node, &fs_info->defrag_inodes);
-                if (fs_info->closing)
+                if (btrfs_fs_closing(fs_info))
                        goto next_free;
                spin_unlock(&fs_info->defrag_inodes_lock);
@@ -1480,14 +1480,12 @@ int btrfs_sync_file(struct file *file, int datasync)
         * the current transaction, we can bail out now without any
         * syncing
         */
-        mutex_lock(&root->fs_info->trans_mutex);
+        smp_mb();
        if (BTRFS_I(inode)->last_trans <=
            root->fs_info->last_trans_committed) {
                BTRFS_I(inode)->last_trans = 0;
-                mutex_unlock(&root->fs_info->trans_mutex);
                goto out;
        }
-        mutex_unlock(&root->fs_info->trans_mutex);
        /*
         * ok we haven't committed the transaction yet, lets do a commit
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 70d45795d758..9f985a429877 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -98,7 +98,7 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
                return inode;
        spin_lock(&block_group->lock);
-        if (!root->fs_info->closing) {
+        if (!btrfs_fs_closing(root->fs_info)) {
                block_group->inode = igrab(inode);
                block_group->iref = 1;
        }
@@ -250,7 +250,7 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
        pgoff_t index = 0;
        unsigned long first_page_offset;
        int num_checksums;
-        int ret = 0, ret2;
+        int ret = 0;
        INIT_LIST_HEAD(&bitmaps);
@@ -402,7 +402,14 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
                                spin_lock(&ctl->tree_lock);
                                ret = link_free_space(ctl, e);
                                spin_unlock(&ctl->tree_lock);
-                                BUG_ON(ret);
+                                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) {
@@ -414,10 +421,18 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
                                        goto free_cache;
                                }
                                spin_lock(&ctl->tree_lock);
-                                ret2 = link_free_space(ctl, e);
+                                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);
                        }
@@ -478,8 +493,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
         * If we're unmounting then just return, since this does a search on the
         * normal root and not the commit root and we could deadlock.
         */
-        smp_mb();
+        if (btrfs_fs_closing(fs_info))
-        if (fs_info->closing)
                return 0;
        /*
@@ -575,10 +589,25 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
        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)
@@ -590,12 +619,6 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
                return -1;
        }
-        /* Since the first page has all of our checksums and our generation we
-         * need to calculate the offset into the page that we can start writing
-         * our entries.
-         */
-        first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64);
        /* Get the cluster for this block_group if it exists */
        if (block_group && !list_empty(&block_group->cluster_list))
                cluster = list_entry(block_group->cluster_list.next,
@@ -857,12 +880,14 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
        ret = 1;
 out_free:
+        kfree(checksums);
+        kfree(pages);
+out_update:
        if (ret != 1) {
                invalidate_inode_pages2_range(inode->i_mapping, 0, index);
                BTRFS_I(inode)->generation = 0;
        }
-        kfree(checksums);
-        kfree(pages);
        btrfs_update_inode(trans, root, inode);
        return ret;
 }
@@ -963,10 +988,16 @@ static int tree_insert_offset(struct rb_root *root, u64 offset,
                         * logically.
                         */
                        if (bitmap) {
-                                WARN_ON(info->bitmap);
+                                if (info->bitmap) {
+                                        WARN_ON_ONCE(1);
+                                        return -EEXIST;
+                                }
                                p = &(*p)->rb_right;
                        } else {
-                                WARN_ON(!info->bitmap);
+                                if (!info->bitmap) {
+                                        WARN_ON_ONCE(1);
+                                        return -EEXIST;
+                                }
                                p = &(*p)->rb_left;
                        }
                }
@@ -1386,6 +1417,23 @@ again:
        return 0;
 }
+static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl,
+                               struct btrfs_free_space *info, u64 offset,
+                               u64 bytes)
+{
+        u64 bytes_to_set = 0;
+        u64 end;
+        end = info->offset + (u64)(BITS_PER_BITMAP * ctl->unit);
+        bytes_to_set = min(end - offset, bytes);
+        bitmap_set_bits(ctl, info, offset, bytes_to_set);
+        return bytes_to_set;
+}
 static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
                      struct btrfs_free_space *info)
 {
@@ -1422,12 +1470,18 @@ static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
        return true;
 }
+static struct btrfs_free_space_op free_space_op = {
+        .recalc_thresholds      = recalculate_thresholds,
+        .use_bitmap             = use_bitmap,
+};
 static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
                              struct btrfs_free_space *info)
 {
        struct btrfs_free_space *bitmap_info;
+        struct btrfs_block_group_cache *block_group = NULL;
        int added = 0;
-        u64 bytes, offset, end;
+        u64 bytes, offset, bytes_added;
        int ret;
        bytes = info->bytes;
@@ -1436,7 +1490,49 @@ static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
        if (!ctl->op->use_bitmap(ctl, info))
                return 0;
+        if (ctl->op == &free_space_op)
+                block_group = ctl->private;
 again:
+        /*
+         * Since we link bitmaps right into the cluster we need to see if we
+         * have a cluster here, and if so and it has our bitmap we need to add
+         * the free space to that bitmap.
+         */
+        if (block_group && !list_empty(&block_group->cluster_list)) {
+                struct btrfs_free_cluster *cluster;
+                struct rb_node *node;
+                struct btrfs_free_space *entry;
+                cluster = list_entry(block_group->cluster_list.next,
+                                     struct btrfs_free_cluster,
+                                     block_group_list);
+                spin_lock(&cluster->lock);
+                node = rb_first(&cluster->root);
+                if (!node) {
+                        spin_unlock(&cluster->lock);
+                        goto no_cluster_bitmap;
+                }
+                entry = rb_entry(node, struct btrfs_free_space, offset_index);
+                if (!entry->bitmap) {
+                        spin_unlock(&cluster->lock);
+                        goto no_cluster_bitmap;
+                }
+                if (entry->offset == offset_to_bitmap(ctl, offset)) {
+                        bytes_added = add_bytes_to_bitmap(ctl, entry,
+                                                          offset, bytes);
+                        bytes -= bytes_added;
+                        offset += bytes_added;
+                }
+                spin_unlock(&cluster->lock);
+                if (!bytes) {
+                        ret = 1;
+                        goto out;
+                }
+        }
+no_cluster_bitmap:
        bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
                                         1, 0);
        if (!bitmap_info) {
@@ -1444,19 +1540,10 @@ again:
                goto new_bitmap;
        }
-        end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit);
+        bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
+        bytes -= bytes_added;
-        if (offset >= bitmap_info->offset && offset + bytes > end) {
+        offset += bytes_added;
-                bitmap_set_bits(ctl, bitmap_info, offset, end - offset);
+        added = 0;
-                bytes -= end - offset;
-                offset = end;
-                added = 0;
-        } else if (offset >= bitmap_info->offset && offset + bytes <= end) {
-                bitmap_set_bits(ctl, bitmap_info, offset, bytes);
-                bytes = 0;
-        } else {
-                BUG();
-        }
        if (!bytes) {
                ret = 1;
@@ -1735,11 +1822,6 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
               "\n", count);
 }
-static struct btrfs_free_space_op free_space_op = {
-        .recalc_thresholds      = recalculate_thresholds,
-        .use_bitmap             = use_bitmap,
-};
 void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group)
 {
        struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
@@ -2111,9 +2193,11 @@ again:
 /*
 * This searches the block group for just extents to fill the cluster with.
 */
-static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
+static noinline int
-                                   struct btrfs_free_cluster *cluster,
+setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
-                                   u64 offset, u64 bytes, u64 min_bytes)
+                        struct btrfs_free_cluster *cluster,
+                        struct list_head *bitmaps, u64 offset, u64 bytes,
+                        u64 min_bytes)
 {
        struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
        struct btrfs_free_space *first = NULL;
@@ -2135,6 +2219,8 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
         * extent entry.
         */
        while (entry->bitmap) {
+                if (list_empty(&entry->list))
+                        list_add_tail(&entry->list, bitmaps);
                node = rb_next(&entry->offset_index);
                if (!node)
                        return -ENOSPC;
@@ -2154,8 +2240,12 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
                        return -ENOSPC;
                entry = rb_entry(node, struct btrfs_free_space, offset_index);
-                if (entry->bitmap)
+                if (entry->bitmap) {
+                        if (list_empty(&entry->list))
+                                list_add_tail(&entry->list, bitmaps);
                        continue;
+                }
                /*
                 * we haven't filled the empty size and the window is
                 * very large.  reset and try again
@@ -2207,9 +2297,11 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
 * This specifically looks for bitmaps that may work in the cluster, we assume
 * that we have already failed to find extents that will work.
 */
-static int setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
+static noinline int
-                                struct btrfs_free_cluster *cluster,
+setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
-                                u64 offset, u64 bytes, u64 min_bytes)
+                     struct btrfs_free_cluster *cluster,
+                     struct list_head *bitmaps, u64 offset, u64 bytes,
+                     u64 min_bytes)
 {
        struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
        struct btrfs_free_space *entry;
@@ -2219,10 +2311,39 @@ static int setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
        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.
+         */
+        list_for_each_entry(entry, bitmaps, list) {
+                if (entry->bytes < min_bytes)
+                        continue;
+                ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset,
+                                           bytes, min_bytes);
+                if (!ret)
+                        return 0;
+        }
+        /*
+         * 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.
+         */
+        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);
@@ -2253,6 +2374,8 @@ 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;
        u64 min_bytes;
        int ret;
@@ -2291,11 +2414,16 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
                goto out;
        }
-        ret = setup_cluster_no_bitmap(block_group, cluster, offset, bytes,
+        INIT_LIST_HEAD(&bitmaps);
-                                      min_bytes);
+        ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset,
+                                      bytes, min_bytes);
        if (ret)
-                ret = setup_cluster_bitmap(block_group, cluster, offset,
+                ret = setup_cluster_bitmap(block_group, cluster, &bitmaps,
-                                           bytes, min_bytes);
+                                           offset, bytes, min_bytes);
+        /* Clear our temporary list */
+        list_for_each_entry_safe(entry, tmp, &bitmaps, list)
+                list_del_init(&entry->list);
        if (!ret) {
                atomic_inc(&block_group->count);
@@ -2481,7 +2609,7 @@ struct inode *lookup_free_ino_inode(struct btrfs_root *root,
                return inode;
        spin_lock(&root->cache_lock);
-        if (!root->fs_info->closing)
+        if (!btrfs_fs_closing(root->fs_info))
                root->cache_inode = igrab(inode);
        spin_unlock(&root->cache_lock);
@@ -2504,12 +2632,14 @@ int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
        int ret = 0;
        u64 root_gen = btrfs_root_generation(&root->root_item);
+        if (!btrfs_test_opt(root, INODE_MAP_CACHE))
+                return 0;
        /*
         * If we're unmounting then just return, since this does a search on the
         * normal root and not the commit root and we could deadlock.
         */
-        smp_mb();
+        if (btrfs_fs_closing(fs_info))
-        if (fs_info->closing)
                return 0;
        path = btrfs_alloc_path();
@@ -2543,6 +2673,9 @@ int btrfs_write_out_ino_cache(struct btrfs_root *root,
        struct inode *inode;
        int ret;
+        if (!btrfs_test_opt(root, INODE_MAP_CACHE))
+                return 0;
        inode = lookup_free_ino_inode(root, path);
        if (IS_ERR(inode))
                return 0;
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index 3262cd17a12f..b4087e0fa871 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -38,6 +38,9 @@ static int caching_kthread(void *data)
        int slot;
        int ret;
+        if (!btrfs_test_opt(root, INODE_MAP_CACHE))
+                return 0;
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
@@ -59,8 +62,7 @@ again:
                goto out;
        while (1) {
-                smp_mb();
+                if (btrfs_fs_closing(fs_info))
-                if (fs_info->closing)
                        goto out;
                leaf = path->nodes[0];
@@ -141,6 +143,9 @@ static void start_caching(struct btrfs_root *root)
        int ret;
        u64 objectid;
+        if (!btrfs_test_opt(root, INODE_MAP_CACHE))
+                return;
        spin_lock(&root->cache_lock);
        if (root->cached != BTRFS_CACHE_NO) {
                spin_unlock(&root->cache_lock);
@@ -178,6 +183,9 @@ static void start_caching(struct btrfs_root *root)
 int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid)
 {
+        if (!btrfs_test_opt(root, INODE_MAP_CACHE))
+                return btrfs_find_free_objectid(root, objectid);
 again:
        *objectid = btrfs_find_ino_for_alloc(root);
@@ -201,6 +209,10 @@ 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);
@@ -250,6 +262,9 @@ void btrfs_unpin_free_ino(struct btrfs_root *root)
        struct rb_node *n;
        u64 count;
+        if (!btrfs_test_opt(root, INODE_MAP_CACHE))
+                return;
        while (1) {
                n = rb_first(rbroot);
                if (!n)
@@ -388,9 +403,24 @@ int btrfs_save_ino_cache(struct btrfs_root *root,
        int prealloc;
        bool retry = false;
+        /* only fs tree and subvol/snap needs ino cache */
+        if (root->root_key.objectid != BTRFS_FS_TREE_OBJECTID &&
+            (root->root_key.objectid < BTRFS_FIRST_FREE_OBJECTID ||
+             root->root_key.objectid > BTRFS_LAST_FREE_OBJECTID))
+                return 0;
+        /* Don't save inode cache if we are deleting this root */
+        if (btrfs_root_refs(&root->root_item) == 0 &&
+            root != root->fs_info->tree_root)
+                return 0;
+        if (!btrfs_test_opt(root, INODE_MAP_CACHE))
+                return 0;
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
 again:
        inode = lookup_free_ino_inode(root, path);
        if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) {
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 39a9d5750efd..751ddf8fc58a 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -138,7 +138,6 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
                return -ENOMEM;
        path->leave_spinning = 1;
-        btrfs_set_trans_block_group(trans, inode);
        key.objectid = btrfs_ino(inode);
        key.offset = start;
@@ -426,9 +425,8 @@ again:
                }
        }
        if (start == 0) {
-                trans = btrfs_join_transaction(root, 1);
+                trans = btrfs_join_transaction(root);
                BUG_ON(IS_ERR(trans));
-                btrfs_set_trans_block_group(trans, inode);
                trans->block_rsv = &root->fs_info->delalloc_block_rsv;
                /* lets try to make an inline extent */
@@ -623,8 +621,9 @@ retry:
                            async_extent->start + async_extent->ram_size - 1,
                            GFP_NOFS);
-                trans = btrfs_join_transaction(root, 1);
+                trans = btrfs_join_transaction(root);
                BUG_ON(IS_ERR(trans));
+                trans->block_rsv = &root->fs_info->delalloc_block_rsv;
                ret = btrfs_reserve_extent(trans, root,
                                           async_extent->compressed_size,
                                           async_extent->compressed_size,
@@ -793,9 +792,8 @@ static noinline int cow_file_range(struct inode *inode,
        int ret = 0;
        BUG_ON(is_free_space_inode(root, inode));
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        BUG_ON(IS_ERR(trans));
-        btrfs_set_trans_block_group(trans, inode);
        trans->block_rsv = &root->fs_info->delalloc_block_rsv;
        num_bytes = (end - start + blocksize) & ~(blocksize - 1);
@@ -1077,10 +1075,12 @@ static noinline int run_delalloc_nocow(struct inode *inode,
        nolock = is_free_space_inode(root, inode);
        if (nolock)
-                trans = btrfs_join_transaction_nolock(root, 1);
+                trans = btrfs_join_transaction_nolock(root);
        else
-                trans = btrfs_join_transaction(root, 1);
+                trans = btrfs_join_transaction(root);
        BUG_ON(IS_ERR(trans));
+        trans->block_rsv = &root->fs_info->delalloc_block_rsv;
        cow_start = (u64)-1;
        cur_offset = start;
@@ -1519,8 +1519,6 @@ static noinline int add_pending_csums(struct btrfs_trans_handle *trans,
 {
        struct btrfs_ordered_sum *sum;
-        btrfs_set_trans_block_group(trans, inode);
        list_for_each_entry(sum, list, list) {
                btrfs_csum_file_blocks(trans,
                       BTRFS_I(inode)->root->fs_info->csum_root, sum);
@@ -1735,11 +1733,10 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
                ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent);
                if (!ret) {
                        if (nolock)
-                                trans = btrfs_join_transaction_nolock(root, 1);
+                                trans = btrfs_join_transaction_nolock(root);
                        else
-                                trans = btrfs_join_transaction(root, 1);
+                                trans = btrfs_join_transaction(root);
                        BUG_ON(IS_ERR(trans));
-                        btrfs_set_trans_block_group(trans, inode);
                        trans->block_rsv = &root->fs_info->delalloc_block_rsv;
                        ret = btrfs_update_inode(trans, root, inode);
                        BUG_ON(ret);
@@ -1752,11 +1749,10 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
                         0, &cached_state, GFP_NOFS);
        if (nolock)
-                trans = btrfs_join_transaction_nolock(root, 1);
+                trans = btrfs_join_transaction_nolock(root);
        else
-                trans = btrfs_join_transaction(root, 1);
+                trans = btrfs_join_transaction(root);
        BUG_ON(IS_ERR(trans));
-        btrfs_set_trans_block_group(trans, inode);
        trans->block_rsv = &root->fs_info->delalloc_block_rsv;
        if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
@@ -1990,7 +1986,7 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
        }
        if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)
-                return 0;
+                goto good;
        if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID &&
            test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1, NULL)) {
@@ -2431,7 +2427,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root)
                                        (u64)-1);
        if (root->orphan_block_rsv || root->orphan_item_inserted) {
-                trans = btrfs_join_transaction(root, 1);
+                trans = btrfs_join_transaction(root);
                if (!IS_ERR(trans))
                        btrfs_end_transaction(trans, root);
        }
@@ -2511,12 +2507,12 @@ static void btrfs_read_locked_inode(struct inode *inode)
        struct btrfs_root *root = BTRFS_I(inode)->root;
        struct btrfs_key location;
        int maybe_acls;
-        u64 alloc_group_block;
        u32 rdev;
        int ret;
        path = btrfs_alloc_path();
        BUG_ON(!path);
+        path->leave_spinning = 1;
        memcpy(&location, &BTRFS_I(inode)->location, sizeof(location));
        ret = btrfs_lookup_inode(NULL, root, path, &location, 0);
@@ -2526,6 +2522,12 @@ static void btrfs_read_locked_inode(struct inode *inode)
        leaf = path->nodes[0];
        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);
@@ -2555,8 +2557,6 @@ static void btrfs_read_locked_inode(struct inode *inode)
        BTRFS_I(inode)->index_cnt = (u64)-1;
        BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item);
-        alloc_group_block = btrfs_inode_block_group(leaf, inode_item);
        /*
         * try to precache a NULL acl entry for files that don't have
         * any xattrs or acls
@@ -2566,8 +2566,11 @@ static void btrfs_read_locked_inode(struct inode *inode)
        if (!maybe_acls)
                cache_no_acl(inode);
-        BTRFS_I(inode)->block_group = btrfs_find_block_group(root, 0,
+        if (leaf->map_token) {
-                                                alloc_group_block, 0);
+                unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+                leaf->map_token = NULL;
+        }
        btrfs_free_path(path);
        inode_item = NULL;
@@ -2647,7 +2650,7 @@ static void fill_inode_item(struct btrfs_trans_handle *trans,
        btrfs_set_inode_transid(leaf, item, trans->transid);
        btrfs_set_inode_rdev(leaf, item, inode->i_rdev);
        btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags);
-        btrfs_set_inode_block_group(leaf, item, BTRFS_I(inode)->block_group);
+        btrfs_set_inode_block_group(leaf, item, 0);
        if (leaf->map_token) {
                unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
@@ -3004,8 +3007,6 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
        if (IS_ERR(trans))
                return PTR_ERR(trans);
-        btrfs_set_trans_block_group(trans, dir);
        btrfs_record_unlink_dir(trans, dir, dentry->d_inode, 0);
        ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
@@ -3094,8 +3095,6 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
        if (IS_ERR(trans))
                return PTR_ERR(trans);
-        btrfs_set_trans_block_group(trans, dir);
        if (unlikely(btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
                err = btrfs_unlink_subvol(trans, root, dir,
                                          BTRFS_I(inode)->location.objectid,
@@ -3514,7 +3513,6 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
                                err = PTR_ERR(trans);
                                break;
                        }
-                        btrfs_set_trans_block_group(trans, inode);
                        err = btrfs_drop_extents(trans, inode, cur_offset,
                                                 cur_offset + hole_size,
@@ -3648,9 +3646,8 @@ void btrfs_evict_inode(struct inode *inode)
        btrfs_i_size_write(inode, 0);
        while (1) {
-                trans = btrfs_start_transaction(root, 0);
+                trans = btrfs_join_transaction(root);
                BUG_ON(IS_ERR(trans));
-                btrfs_set_trans_block_group(trans, inode);
                trans->block_rsv = root->orphan_block_rsv;
                ret = btrfs_block_rsv_check(trans, root,
@@ -4133,7 +4130,8 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
-        path->reada = 2;
+        path->reada = 1;
        if (key_type == BTRFS_DIR_INDEX_KEY) {
                INIT_LIST_HEAD(&ins_list);
@@ -4268,18 +4266,16 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc)
        if (BTRFS_I(inode)->dummy_inode)
                return 0;
-        smp_mb();
+        if (btrfs_fs_closing(root->fs_info) && is_free_space_inode(root, inode))
-        if (root->fs_info->closing && is_free_space_inode(root, inode))
                nolock = true;
        if (wbc->sync_mode == WB_SYNC_ALL) {
                if (nolock)
-                        trans = btrfs_join_transaction_nolock(root, 1);
+                        trans = btrfs_join_transaction_nolock(root);
                else
-                        trans = btrfs_join_transaction(root, 1);
+                        trans = btrfs_join_transaction(root);
                if (IS_ERR(trans))
                        return PTR_ERR(trans);
-                btrfs_set_trans_block_group(trans, inode);
                if (nolock)
                        ret = btrfs_end_transaction_nolock(trans, root);
                else
@@ -4303,9 +4299,8 @@ void btrfs_dirty_inode(struct inode *inode, int flags)
        if (BTRFS_I(inode)->dummy_inode)
                return;
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        BUG_ON(IS_ERR(trans));
-        btrfs_set_trans_block_group(trans, inode);
        ret = btrfs_update_inode(trans, root, inode);
        if (ret && ret == -ENOSPC) {
@@ -4319,7 +4314,6 @@ void btrfs_dirty_inode(struct inode *inode, int flags)
                                       PTR_ERR(trans));
                        return;
                }
-                btrfs_set_trans_block_group(trans, inode);
                ret = btrfs_update_inode(trans, root, inode);
                if (ret) {
@@ -4418,8 +4412,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
                                     struct btrfs_root *root,
                                     struct inode *dir,
                                     const char *name, int name_len,
-                                     u64 ref_objectid, u64 objectid,
+                                     u64 ref_objectid, u64 objectid, int mode,
-                                     u64 alloc_hint, int mode, u64 *index)
+                                     u64 *index)
 {
        struct inode *inode;
        struct btrfs_inode_item *inode_item;
@@ -4472,8 +4466,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
                owner = 0;
        else
                owner = 1;
-        BTRFS_I(inode)->block_group =
-                        btrfs_find_block_group(root, 0, alloc_hint, owner);
        key[0].objectid = objectid;
        btrfs_set_key_type(&key[0], BTRFS_INODE_ITEM_KEY);
@@ -4629,15 +4621,13 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
        if (IS_ERR(trans))
                return PTR_ERR(trans);
-        btrfs_set_trans_block_group(trans, dir);
        err = btrfs_find_free_ino(root, &objectid);
        if (err)
                goto out_unlock;
        inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
                                dentry->d_name.len, btrfs_ino(dir), objectid,
-                                BTRFS_I(dir)->block_group, mode, &index);
+                                mode, &index);
        if (IS_ERR(inode)) {
                err = PTR_ERR(inode);
                goto out_unlock;
@@ -4649,7 +4639,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
                goto out_unlock;
        }
-        btrfs_set_trans_block_group(trans, inode);
        err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index);
        if (err)
                drop_inode = 1;
@@ -4658,8 +4647,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
                init_special_inode(inode, inode->i_mode, rdev);
                btrfs_update_inode(trans, root, inode);
        }
-        btrfs_update_inode_block_group(trans, inode);
-        btrfs_update_inode_block_group(trans, dir);
 out_unlock:
        nr = trans->blocks_used;
        btrfs_end_transaction_throttle(trans, root);
@@ -4692,15 +4679,13 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
        if (IS_ERR(trans))
                return PTR_ERR(trans);
-        btrfs_set_trans_block_group(trans, dir);
        err = btrfs_find_free_ino(root, &objectid);
        if (err)
                goto out_unlock;
        inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
                                dentry->d_name.len, btrfs_ino(dir), objectid,
-                                BTRFS_I(dir)->block_group, mode, &index);
+                                mode, &index);
        if (IS_ERR(inode)) {
                err = PTR_ERR(inode);
                goto out_unlock;
@@ -4712,7 +4697,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
                goto out_unlock;
        }
-        btrfs_set_trans_block_group(trans, inode);
        err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index);
        if (err)
                drop_inode = 1;
@@ -4723,8 +4707,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
                inode->i_op = &btrfs_file_inode_operations;
                BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
        }
-        btrfs_update_inode_block_group(trans, inode);
-        btrfs_update_inode_block_group(trans, dir);
 out_unlock:
        nr = trans->blocks_used;
        btrfs_end_transaction_throttle(trans, root);
@@ -4771,8 +4753,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
        btrfs_inc_nlink(inode);
        inode->i_ctime = CURRENT_TIME;
-        btrfs_set_trans_block_group(trans, dir);
        ihold(inode);
        err = btrfs_add_nondir(trans, dir, dentry, inode, 1, index);
@@ -4781,7 +4761,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
                drop_inode = 1;
        } else {
                struct dentry *parent = dget_parent(dentry);
-                btrfs_update_inode_block_group(trans, dir);
                err = btrfs_update_inode(trans, root, inode);
                BUG_ON(err);
                btrfs_log_new_name(trans, inode, NULL, parent);
@@ -4818,7 +4797,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
        trans = btrfs_start_transaction(root, 5);
        if (IS_ERR(trans))
                return PTR_ERR(trans);
-        btrfs_set_trans_block_group(trans, dir);
        err = btrfs_find_free_ino(root, &objectid);
        if (err)
@@ -4826,8 +4804,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
        inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
                                dentry->d_name.len, btrfs_ino(dir), objectid,
-                                BTRFS_I(dir)->block_group, S_IFDIR | mode,
+                                S_IFDIR | mode, &index);
-                                &index);
        if (IS_ERR(inode)) {
                err = PTR_ERR(inode);
                goto out_fail;
@@ -4841,7 +4818,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
        inode->i_op = &btrfs_dir_inode_operations;
        inode->i_fop = &btrfs_dir_file_operations;
-        btrfs_set_trans_block_group(trans, inode);
        btrfs_i_size_write(inode, 0);
        err = btrfs_update_inode(trans, root, inode);
@@ -4855,8 +4831,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
        d_instantiate(dentry, inode);
        drop_on_err = 0;
-        btrfs_update_inode_block_group(trans, inode);
-        btrfs_update_inode_block_group(trans, dir);
 out_fail:
        nr = trans->blocks_used;
@@ -4989,7 +4963,15 @@ again:
        if (!path) {
                path = btrfs_alloc_path();
-                BUG_ON(!path);
+                if (!path) {
+                        err = -ENOMEM;
+                        goto out;
+                }
+                /*
+                 * Chances are we'll be called again, so go ahead and do
+                 * readahead
+                 */
+                path->reada = 1;
        }
        ret = btrfs_lookup_file_extent(trans, root, path,
@@ -5130,8 +5112,10 @@ again:
                                kunmap(page);
                                free_extent_map(em);
                                em = NULL;
                                btrfs_release_path(path);
-                                trans = btrfs_join_transaction(root, 1);
+                                trans = btrfs_join_transaction(root);
                                if (IS_ERR(trans))
                                        return ERR_CAST(trans);
                                goto again;
@@ -5375,7 +5359,7 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode,
                btrfs_drop_extent_cache(inode, start, start + len - 1, 0);
        }
-        trans = btrfs_join_transaction(root, 0);
+        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans))
                return ERR_CAST(trans);
@@ -5611,7 +5595,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
                 * to make sure the current transaction stays open
                 * while we look for nocow cross refs
                 */
-                trans = btrfs_join_transaction(root, 0);
+                trans = btrfs_join_transaction(root);
                if (IS_ERR(trans))
                        goto must_cow;
@@ -5750,7 +5734,7 @@ again:
        BUG_ON(!ordered);
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans)) {
                err = -ENOMEM;
                goto out;
@@ -6500,6 +6484,7 @@ out:
 static int btrfs_truncate(struct inode *inode)
 {
        struct btrfs_root *root = BTRFS_I(inode)->root;
+        struct btrfs_block_rsv *rsv;
        int ret;
        int err = 0;
        struct btrfs_trans_handle *trans;
@@ -6513,28 +6498,80 @@ static int btrfs_truncate(struct inode *inode)
        btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
        btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
-        trans = btrfs_start_transaction(root, 5);
+        /*
-        if (IS_ERR(trans))
+         * Yes ladies and gentelment, this is indeed ugly.  The fact is we have
-                return PTR_ERR(trans);
+         * 3 things going on here
+         *
+         * 1) We need to reserve space for our orphan item and the space to
+         * delete our orphan item.  Lord knows we don't want to have a dangling
+         * orphan item because we didn't reserve space to remove it.
+         *
+         * 2) We need to reserve space to update our inode.
+         *
+         * 3) We need to have something to cache all the space that is going to
+         * be free'd up by the truncate operation, but also have some slack
+         * space reserved in case it uses space during the truncate (thank you
+         * very much snapshotting).
+         *
+         * And we need these to all be seperate.  The fact is we can use alot of
+         * space doing the truncate, and we have no earthly idea how much space
+         * we will use, so we need the truncate reservation to be seperate so it
+         * doesn't end up using space reserved for updating the inode or
+         * removing the orphan item.  We also need to be able to stop the
+         * transaction and start a new one, which means we need to be able to
+         * update the inode several times, and we have no idea of knowing how
+         * many times that will be, so we can't just reserve 1 item for the
+         * entirety of the opration, so that has to be done seperately as well.
+         * Then there is the orphan item, which does indeed need to be held on
+         * to for the whole operation, and we need nobody to touch this reserved
+         * space except the orphan code.
+         *
+         * So that leaves us with
+         *
+         * 1) root->orphan_block_rsv - for the orphan deletion.
+         * 2) rsv - for the truncate reservation, which we will steal from the
+         * transaction reservation.
+         * 3) fs_info->trans_block_rsv - this will have 1 items worth left for
+         * updating the inode.
+         */
+        rsv = btrfs_alloc_block_rsv(root);
+        if (!rsv)
+                return -ENOMEM;
+        btrfs_add_durable_block_rsv(root->fs_info, rsv);
+        trans = btrfs_start_transaction(root, 4);
+        if (IS_ERR(trans)) {
+                err = PTR_ERR(trans);
+                goto out;
+        }
-        btrfs_set_trans_block_group(trans, inode);
+        /*
+         * 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);
+        BUG_ON(ret);
        ret = btrfs_orphan_add(trans, inode);
        if (ret) {
                btrfs_end_transaction(trans, root);
-                return ret;
+                goto out;
        }
        nr = trans->blocks_used;
        btrfs_end_transaction(trans, root);
        btrfs_btree_balance_dirty(root, nr);
-        /* Now start a transaction for the truncate */
+        /*
-        trans = btrfs_start_transaction(root, 0);
+         * Ok so we've already migrated our bytes over for the truncate, so here
-        if (IS_ERR(trans))
+         * just reserve the one slot we need for updating the inode.
-                return PTR_ERR(trans);
+         */
-        btrfs_set_trans_block_group(trans, inode);
+        trans = btrfs_start_transaction(root, 1);
-        trans->block_rsv = root->orphan_block_rsv;
+        if (IS_ERR(trans)) {
+                err = PTR_ERR(trans);
+                goto out;
+        }
+        trans->block_rsv = rsv;
        /*
         * setattr is responsible for setting the ordered_data_close flag,
@@ -6558,24 +6595,17 @@ static int btrfs_truncate(struct inode *inode)
        while (1) {
                if (!trans) {
-                        trans = btrfs_start_transaction(root, 0);
+                        trans = btrfs_start_transaction(root, 3);
-                        if (IS_ERR(trans))
+                        if (IS_ERR(trans)) {
-                                return PTR_ERR(trans);
+                                err = PTR_ERR(trans);
-                        btrfs_set_trans_block_group(trans, inode);
+                                goto out;
-                        trans->block_rsv = root->orphan_block_rsv;
+                        }
-                }
-                ret = btrfs_block_rsv_check(trans, root,
+                        ret = btrfs_truncate_reserve_metadata(trans, root,
-                                            root->orphan_block_rsv, 0, 5);
+                                                              rsv);
-                if (ret == -EAGAIN) {
+                        BUG_ON(ret);
-                        ret = btrfs_commit_transaction(trans, root);
-                        if (ret)
+                        trans->block_rsv = rsv;
-                                return ret;
-                        trans = NULL;
-                        continue;
-                } else if (ret) {
-                        err = ret;
-                        break;
                }
                ret = btrfs_truncate_inode_items(trans, root, inode,
@@ -6586,6 +6616,7 @@ static int btrfs_truncate(struct inode *inode)
                        break;
                }
+                trans->block_rsv = &root->fs_info->trans_block_rsv;
                ret = btrfs_update_inode(trans, root, inode);
                if (ret) {
                        err = ret;
@@ -6599,6 +6630,7 @@ static int btrfs_truncate(struct inode *inode)
        }
        if (ret == 0 && inode->i_nlink > 0) {
+                trans->block_rsv = root->orphan_block_rsv;
                ret = btrfs_orphan_del(trans, inode);
                if (ret)
                        err = ret;
@@ -6610,15 +6642,20 @@ 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;
        nr = trans->blocks_used;
        ret = btrfs_end_transaction_throttle(trans, root);
+        btrfs_btree_balance_dirty(root, nr);
+out:
+        btrfs_free_block_rsv(root, rsv);
        if (ret && !err)
                err = ret;
-        btrfs_btree_balance_dirty(root, nr);
        return err;
 }
@@ -6627,15 +6664,14 @@ static int btrfs_truncate(struct inode *inode)
 * create a new subvolume directory/inode (helper for the ioctl).
 */
 int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
-                             struct btrfs_root *new_root,
+                             struct btrfs_root *new_root, u64 new_dirid)
-                             u64 new_dirid, u64 alloc_hint)
 {
        struct inode *inode;
        int err;
        u64 index = 0;
        inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, new_dirid,
-                                new_dirid, alloc_hint, S_IFDIR | 0700, &index);
+                                new_dirid, S_IFDIR | 0700, &index);
        if (IS_ERR(inode))
                return PTR_ERR(inode);
        inode->i_op = &btrfs_dir_inode_operations;
@@ -6748,21 +6784,6 @@ void btrfs_destroy_inode(struct inode *inode)
                spin_unlock(&root->fs_info->ordered_extent_lock);
        }
-        if (root == root->fs_info->tree_root) {
-                struct btrfs_block_group_cache *block_group;
-                block_group = btrfs_lookup_block_group(root->fs_info,
-                                                BTRFS_I(inode)->block_group);
-                if (block_group && block_group->inode == inode) {
-                        spin_lock(&block_group->lock);
-                        block_group->inode = NULL;
-                        spin_unlock(&block_group->lock);
-                        btrfs_put_block_group(block_group);
-                } else if (block_group) {
-                        btrfs_put_block_group(block_group);
-                }
-        }
        spin_lock(&root->orphan_lock);
        if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
                printk(KERN_INFO "BTRFS: inode %llu still on the orphan list\n",
@@ -6948,8 +6969,6 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
                goto out_notrans;
        }
-        btrfs_set_trans_block_group(trans, new_dir);
        if (dest != root)
                btrfs_record_root_in_trans(trans, dest);
@@ -7131,16 +7150,13 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
        if (IS_ERR(trans))
                return PTR_ERR(trans);
-        btrfs_set_trans_block_group(trans, dir);
        err = btrfs_find_free_ino(root, &objectid);
        if (err)
                goto out_unlock;
        inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
                                dentry->d_name.len, btrfs_ino(dir), objectid,
-                                BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO,
+                                S_IFLNK|S_IRWXUGO, &index);
-                                &index);
        if (IS_ERR(inode)) {
                err = PTR_ERR(inode);
                goto out_unlock;
@@ -7152,7 +7168,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
                goto out_unlock;
        }
-        btrfs_set_trans_block_group(trans, inode);
        err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index);
        if (err)
                drop_inode = 1;
@@ -7163,8 +7178,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
                inode->i_op = &btrfs_file_inode_operations;
                BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
        }
-        btrfs_update_inode_block_group(trans, inode);
-        btrfs_update_inode_block_group(trans, dir);
        if (drop_inode)
                goto out_unlock;
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 85e818ce00c5..b793d112d1f6 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -243,7 +243,7 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
                ip->flags &= ~(BTRFS_INODE_COMPRESS | BTRFS_INODE_NOCOMPRESS);
        }
-        trans = btrfs_join_transaction(root, 1);
+        trans = btrfs_join_transaction(root);
        BUG_ON(IS_ERR(trans));
        ret = btrfs_update_inode(trans, root, inode);
@@ -414,8 +414,7 @@ static noinline int create_subvol(struct btrfs_root *root,
        btrfs_record_root_in_trans(trans, new_root);
-        ret = btrfs_create_subvol_root(trans, new_root, new_dirid,
+        ret = btrfs_create_subvol_root(trans, new_root, new_dirid);
-                                       BTRFS_I(dir)->block_group);
        /*
         * insert the directory item
         */
@@ -707,16 +706,17 @@ static int find_new_extents(struct btrfs_root *root,
        struct btrfs_file_extent_item *extent;
        int type;
        int ret;
+        u64 ino = btrfs_ino(inode);
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
-        min_key.objectid = inode->i_ino;
+        min_key.objectid = ino;
        min_key.type = BTRFS_EXTENT_DATA_KEY;
        min_key.offset = *off;
-        max_key.objectid = inode->i_ino;
+        max_key.objectid = ino;
        max_key.type = (u8)-1;
        max_key.offset = (u64)-1;
@@ -727,7 +727,7 @@ static int find_new_extents(struct btrfs_root *root,
                                           path, 0, newer_than);
                if (ret != 0)
                        goto none;
-                if (min_key.objectid != inode->i_ino)
+                if (min_key.objectid != ino)
                        goto none;
                if (min_key.type != BTRFS_EXTENT_DATA_KEY)
                        goto none;
@@ -2054,29 +2054,34 @@ static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg)
 static long btrfs_ioctl_fs_info(struct btrfs_root *root, void __user *arg)
 {
-        struct btrfs_ioctl_fs_info_args fi_args;
+        struct btrfs_ioctl_fs_info_args *fi_args;
        struct btrfs_device *device;
        struct btrfs_device *next;
        struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
+        int ret = 0;
        if (!capable(CAP_SYS_ADMIN))
                return -EPERM;
-        fi_args.num_devices = fs_devices->num_devices;
+        fi_args = kzalloc(sizeof(*fi_args), GFP_KERNEL);
-        fi_args.max_id = 0;
+        if (!fi_args)
-        memcpy(&fi_args.fsid, root->fs_info->fsid, sizeof(fi_args.fsid));
+                return -ENOMEM;
+        fi_args->num_devices = fs_devices->num_devices;
+        memcpy(&fi_args->fsid, root->fs_info->fsid, sizeof(fi_args->fsid));
        mutex_lock(&fs_devices->device_list_mutex);
        list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
-                if (device->devid > fi_args.max_id)
+                if (device->devid > fi_args->max_id)
-                        fi_args.max_id = device->devid;
+                        fi_args->max_id = device->devid;
        }
        mutex_unlock(&fs_devices->device_list_mutex);
-        if (copy_to_user(arg, &fi_args, sizeof(fi_args)))
+        if (copy_to_user(arg, fi_args, sizeof(*fi_args)))
-                return -EFAULT;
+                ret = -EFAULT;
-        return 0;
+        kfree(fi_args);
+        return ret;
 }
 static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg)
@@ -2489,12 +2494,10 @@ static long btrfs_ioctl_trans_start(struct file *file)
        if (ret)
                goto out;
-        mutex_lock(&root->fs_info->trans_mutex);
+        atomic_inc(&root->fs_info->open_ioctl_trans);
-        root->fs_info->open_ioctl_trans++;
-        mutex_unlock(&root->fs_info->trans_mutex);
        ret = -ENOMEM;
-        trans = btrfs_start_ioctl_transaction(root, 0);
+        trans = btrfs_start_ioctl_transaction(root);
        if (IS_ERR(trans))
                goto out_drop;
@@ -2502,9 +2505,7 @@ static long btrfs_ioctl_trans_start(struct file *file)
        return 0;
 out_drop:
-        mutex_lock(&root->fs_info->trans_mutex);
+        atomic_dec(&root->fs_info->open_ioctl_trans);
-        root->fs_info->open_ioctl_trans--;
-        mutex_unlock(&root->fs_info->trans_mutex);
        mnt_drop_write(file->f_path.mnt);
 out:
        return ret;
@@ -2738,9 +2739,7 @@ long btrfs_ioctl_trans_end(struct file *file)
        btrfs_end_transaction(trans, root);
-        mutex_lock(&root->fs_info->trans_mutex);
+        atomic_dec(&root->fs_info->open_ioctl_trans);
-        root->fs_info->open_ioctl_trans--;
-        mutex_unlock(&root->fs_info->trans_mutex);
        mnt_drop_write(file->f_path.mnt);
        return 0;
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index ca38eca70af0..b1ef27cc673b 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -677,6 +677,8 @@ struct backref_node *build_backref_tree(struct reloc_control *rc,
                err = -ENOMEM;
                goto out;
        }
+        path1->reada = 1;
+        path2->reada = 2;
        node = alloc_backref_node(cache);
        if (!node) {
@@ -1999,6 +2001,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
+        path->reada = 1;
        reloc_root = root->reloc_root;
        root_item = &reloc_root->root_item;
@@ -2139,10 +2142,10 @@ int prepare_to_merge(struct reloc_control *rc, int err)
        u64 num_bytes = 0;
        int ret;
-        mutex_lock(&root->fs_info->trans_mutex);
+        spin_lock(&root->fs_info->trans_lock);
        rc->merging_rsv_size += root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2;
        rc->merging_rsv_size += rc->nodes_relocated * 2;
-        mutex_unlock(&root->fs_info->trans_mutex);
+        spin_unlock(&root->fs_info->trans_lock);
 again:
        if (!err) {
                num_bytes = rc->merging_rsv_size;
@@ -2152,7 +2155,7 @@ again:
                        err = ret;
        }
-        trans = btrfs_join_transaction(rc->extent_root, 1);
+        trans = btrfs_join_transaction(rc->extent_root);
        if (IS_ERR(trans)) {
                if (!err)
                        btrfs_block_rsv_release(rc->extent_root,
@@ -2211,9 +2214,9 @@ int merge_reloc_roots(struct reloc_control *rc)
        int ret;
 again:
        root = rc->extent_root;
-        mutex_lock(&root->fs_info->trans_mutex);
+        spin_lock(&root->fs_info->trans_lock);
        list_splice_init(&rc->reloc_roots, &reloc_roots);
-        mutex_unlock(&root->fs_info->trans_mutex);
+        spin_unlock(&root->fs_info->trans_lock);
        while (!list_empty(&reloc_roots)) {
                found = 1;
@@ -3236,7 +3239,7 @@ truncate:
                goto out;
        }
-        trans = btrfs_join_transaction(root, 0);
+        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans)) {
                btrfs_free_path(path);
                ret = PTR_ERR(trans);
@@ -3300,6 +3303,7 @@ static int find_data_references(struct reloc_control *rc,
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
+        path->reada = 1;
        root = read_fs_root(rc->extent_root->fs_info, ref_root);
        if (IS_ERR(root)) {
@@ -3586,17 +3590,17 @@ next:
 static void set_reloc_control(struct reloc_control *rc)
 {
        struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
-        mutex_lock(&fs_info->trans_mutex);
+        spin_lock(&fs_info->trans_lock);
        fs_info->reloc_ctl = rc;
-        mutex_unlock(&fs_info->trans_mutex);
+        spin_unlock(&fs_info->trans_lock);
 }
 static void unset_reloc_control(struct reloc_control *rc)
 {
        struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
-        mutex_lock(&fs_info->trans_mutex);
+        spin_lock(&fs_info->trans_lock);
        fs_info->reloc_ctl = NULL;
-        mutex_unlock(&fs_info->trans_mutex);
+        spin_unlock(&fs_info->trans_lock);
 }
 static int check_extent_flags(u64 flags)
@@ -3645,7 +3649,7 @@ int prepare_to_relocate(struct reloc_control *rc)
        rc->create_reloc_tree = 1;
        set_reloc_control(rc);
-        trans = btrfs_join_transaction(rc->extent_root, 1);
+        trans = btrfs_join_transaction(rc->extent_root);
        BUG_ON(IS_ERR(trans));
        btrfs_commit_transaction(trans, rc->extent_root);
        return 0;
@@ -3668,6 +3672,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
+        path->reada = 1;
        ret = prepare_to_relocate(rc);
        if (ret) {
@@ -3834,7 +3839,7 @@ restart:
        btrfs_block_rsv_release(rc->extent_root, rc->block_rsv, (u64)-1);
        /* get rid of pinned extents */
-        trans = btrfs_join_transaction(rc->extent_root, 1);
+        trans = btrfs_join_transaction(rc->extent_root);
        if (IS_ERR(trans))
                err = PTR_ERR(trans);
        else
@@ -4093,6 +4098,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
+        path->reada = -1;
        key.objectid = BTRFS_TREE_RELOC_OBJECTID;
        key.type = BTRFS_ROOT_ITEM_KEY;
@@ -4159,7 +4165,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
        set_reloc_control(rc);
-        trans = btrfs_join_transaction(rc->extent_root, 1);
+        trans = btrfs_join_transaction(rc->extent_root);
        if (IS_ERR(trans)) {
                unset_reloc_control(rc);
                err = PTR_ERR(trans);
@@ -4193,7 +4199,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
        unset_reloc_control(rc);
-        trans = btrfs_join_transaction(rc->extent_root, 1);
+        trans = btrfs_join_transaction(rc->extent_root);
        if (IS_ERR(trans))
                err = PTR_ERR(trans);
        else
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index 6dfed0c27ac3..a8d03d5efb5d 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -16,13 +16,7 @@
 * Boston, MA 021110-1307, USA.
 */
-#include <linux/sched.h>
-#include <linux/pagemap.h>
-#include <linux/writeback.h>
 #include <linux/blkdev.h>
-#include <linux/rbtree.h>
-#include <linux/slab.h>
-#include <linux/workqueue.h>
 #include "ctree.h"
 #include "volumes.h"
 #include "disk-io.h"
@@ -117,33 +111,37 @@ static void scrub_free_csums(struct scrub_dev *sdev)
        }
 }
+static void scrub_free_bio(struct bio *bio)
+{
+        int i;
+        struct page *last_page = NULL;
+        if (!bio)
+                return;
+        for (i = 0; i < bio->bi_vcnt; ++i) {
+                if (bio->bi_io_vec[i].bv_page == last_page)
+                        continue;
+                last_page = bio->bi_io_vec[i].bv_page;
+                __free_page(last_page);
+        }
+        bio_put(bio);
+}
 static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev)
 {
        int i;
-        int j;
-        struct page *last_page;
        if (!sdev)
                return;
        for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
                struct scrub_bio *sbio = sdev->bios[i];
-                struct bio *bio;
                if (!sbio)
                        break;
-                bio = sbio->bio;
+                scrub_free_bio(sbio->bio);
-                if (bio) {
-                        last_page = NULL;
-                        for (j = 0; j < bio->bi_vcnt; ++j) {
-                                if (bio->bi_io_vec[j].bv_page == last_page)
-                                        continue;
-                                last_page = bio->bi_io_vec[j].bv_page;
-                                __free_page(last_page);
-                        }
-                        bio_put(bio);
-                }
                kfree(sbio);
        }
@@ -156,8 +154,6 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
 {
        struct scrub_dev *sdev;
        int             i;
-        int             j;
-        int             ret;
        struct btrfs_fs_info *fs_info = dev->dev_root->fs_info;
        sdev = kzalloc(sizeof(*sdev), GFP_NOFS);
@@ -165,7 +161,6 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
                goto nomem;
        sdev->dev = dev;
        for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
-                struct bio *bio;
                struct scrub_bio *sbio;
                sbio = kzalloc(sizeof(*sbio), GFP_NOFS);
@@ -173,32 +168,10 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
                        goto nomem;
                sdev->bios[i] = sbio;
-                bio = bio_kmalloc(GFP_NOFS, SCRUB_PAGES_PER_BIO);
-                if (!bio)
-                        goto nomem;
                sbio->index = i;
                sbio->sdev = sdev;
-                sbio->bio = bio;
                sbio->count = 0;
                sbio->work.func = scrub_checksum;
-                bio->bi_private = sdev->bios[i];
-                bio->bi_end_io = scrub_bio_end_io;
-                bio->bi_sector = 0;
-                bio->bi_bdev = dev->bdev;
-                bio->bi_size = 0;
-                for (j = 0; j < SCRUB_PAGES_PER_BIO; ++j) {
-                        struct page *page;
-                        page = alloc_page(GFP_NOFS);
-                        if (!page)
-                                goto nomem;
-                        ret = bio_add_page(bio, page, PAGE_SIZE, 0);
-                        if (!ret)
-                                goto nomem;
-                }
-                WARN_ON(bio->bi_vcnt != SCRUB_PAGES_PER_BIO);
                if (i != SCRUB_BIOS_PER_DEV-1)
                        sdev->bios[i]->next_free = i + 1;
@@ -369,9 +342,6 @@ static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
        int ret;
        DECLARE_COMPLETION_ONSTACK(complete);
-        /* we are going to wait on this IO */
-        rw |= REQ_SYNC;
        bio = bio_alloc(GFP_NOFS, 1);
        bio->bi_bdev = bdev;
        bio->bi_sector = sector;
@@ -380,6 +350,7 @@ static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
        bio->bi_private = &complete;
        submit_bio(rw, bio);
+        /* this will also unplug the queue */
        wait_for_completion(&complete);
        ret = !test_bit(BIO_UPTODATE, &bio->bi_flags);
@@ -394,6 +365,7 @@ static void scrub_bio_end_io(struct bio *bio, int err)
        struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
        sbio->err = err;
+        sbio->bio = bio;
        btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work);
 }
@@ -453,6 +425,8 @@ static void scrub_checksum(struct btrfs_work *work)
        }
 out:
+        scrub_free_bio(sbio->bio);
+        sbio->bio = NULL;
        spin_lock(&sdev->list_lock);
        sbio->next_free = sdev->first_free;
        sdev->first_free = sbio->index;
@@ -583,25 +557,50 @@ 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];
-        sbio->bio->bi_sector = sbio->physical >> 9;
+        bio = bio_alloc(GFP_NOFS, sbio->count);
-        sbio->bio->bi_size = sbio->count * PAGE_SIZE;
+        if (!bio)
-        sbio->bio->bi_next = NULL;
+                goto nomem;
-        sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
-        sbio->bio->bi_comp_cpu = -1;
+        bio->bi_private = sbio;
-        sbio->bio->bi_bdev = sdev->dev->bdev;
+        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(0, sbio->bio);
+        submit_bio(READ, bio);
        return 0;
+nomem:
+        scrub_free_bio(bio);
+        return -ENOMEM;
 }
 static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len,
@@ -633,7 +632,11 @@ again:
                sbio->logical = logical;
        } else if (sbio->physical + sbio->count * PAGE_SIZE != physical ||
                   sbio->logical + sbio->count * PAGE_SIZE != logical) {
-                scrub_submit(sdev);
+                int ret;
+                ret = scrub_submit(sdev);
+                if (ret)
+                        return ret;
                goto again;
        }
        sbio->spag[sbio->count].flags = flags;
@@ -645,8 +648,13 @@ again:
                memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size);
        }
        ++sbio->count;
-        if (sbio->count == SCRUB_PAGES_PER_BIO || force)
+        if (sbio->count == SCRUB_PAGES_PER_BIO || force) {
-                scrub_submit(sdev);
+                int ret;
+                ret = scrub_submit(sdev);
+                if (ret)
+                        return ret;
+        }
        return 0;
 }
@@ -727,6 +735,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
        struct btrfs_root *root = fs_info->extent_root;
        struct btrfs_root *csum_root = fs_info->csum_root;
        struct btrfs_extent_item *extent;
+        struct blk_plug plug;
        u64 flags;
        int ret;
        int slot;
@@ -789,18 +798,12 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
                ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
                if (ret < 0)
-                        goto out;
+                        goto out_noplug;
-                l = path->nodes[0];
-                slot = path->slots[0];
-                btrfs_item_key_to_cpu(l, &key, slot);
-                if (key.objectid != logical) {
-                        ret = btrfs_previous_item(root, path, 0,
-                                                  BTRFS_EXTENT_ITEM_KEY);
-                        if (ret < 0)
-                                goto out;
-                }
+                /*
+                 * 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];
@@ -809,7 +812,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
                                if (ret == 0)
                                        continue;
                                if (ret < 0)
-                                        goto out;
+                                        goto out_noplug;
                                break;
                        }
@@ -831,6 +834,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
         * 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) {
@@ -890,15 +894,20 @@ again:
                ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
                if (ret < 0)
                        goto out;
+                if (ret > 0) {
-                l = path->nodes[0];
-                slot = path->slots[0];
-                btrfs_item_key_to_cpu(l, &key, slot);
-                if (key.objectid != logical) {
                        ret = btrfs_previous_item(root, path, 0,
                                                  BTRFS_EXTENT_ITEM_KEY);
                        if (ret < 0)
                                goto out;
+                        if (ret > 0) {
+                                /* there's no smaller item, so stick with the
+                                 * larger one */
+                                btrfs_release_path(path);
+                                ret = btrfs_search_slot(NULL, root, &key,
+                                                        path, 0, 0);
+                                if (ret < 0)
+                                        goto out;
+                        }
                }
                while (1) {
@@ -972,6 +981,8 @@ next:
        scrub_submit(sdev);
 out:
+        blk_finish_plug(&plug);
+out_noplug:
        btrfs_free_path(path);
        return ret < 0 ? ret : 0;
 }
@@ -1047,8 +1058,15 @@ int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end)
        while (1) {
                ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
                if (ret < 0)
-                        goto out;
+                        break;
-                ret = 0;
+                if (ret > 0) {
+                        if (path->slots[0] >=
+                            btrfs_header_nritems(path->nodes[0])) {
+                                ret = btrfs_next_leaf(root, path);
+                                if (ret)
+                                        break;
+                        }
+                }
                l = path->nodes[0];
                slot = path->slots[0];
@@ -1058,7 +1076,7 @@ int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end)
                if (found_key.objectid != sdev->dev->devid)
                        break;
-                if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
+                if (btrfs_key_type(&found_key) != BTRFS_DEV_EXTENT_KEY)
                        break;
                if (found_key.offset >= end)
@@ -1087,7 +1105,7 @@ int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end)
                cache = btrfs_lookup_block_group(fs_info, chunk_offset);
                if (!cache) {
                        ret = -ENOENT;
-                        goto out;
+                        break;
                }
                ret = scrub_chunk(sdev, chunk_tree, chunk_objectid,
                                  chunk_offset, length);
@@ -1099,9 +1117,13 @@ int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end)
                btrfs_release_path(path);
        }
-out:
        btrfs_free_path(path);
-        return ret;
+        /*
+         * ret can still be 1 from search_slot or next_leaf,
+         * that's not an error
+         */
+        return ret < 0 ? ret : 0;
 }
 static noinline_for_stack int scrub_supers(struct scrub_dev *sdev)
@@ -1138,8 +1160,12 @@ static noinline_for_stack int scrub_workers_get(struct btrfs_root *root)
        struct btrfs_fs_info *fs_info = root->fs_info;
        mutex_lock(&fs_info->scrub_lock);
-        if (fs_info->scrub_workers_refcnt == 0)
+        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);
+        }
        ++fs_info->scrub_workers_refcnt;
        mutex_unlock(&fs_info->scrub_lock);
@@ -1166,7 +1192,7 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
        int ret;
        struct btrfs_device *dev;
-        if (root->fs_info->closing)
+        if (btrfs_fs_closing(root->fs_info))
                return -EINVAL;
        /*
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 9b2e7e5bc3ef..0bb4ebbb71b7 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -161,7 +161,8 @@ enum {
        Opt_compress_type, Opt_compress_force, Opt_compress_force_type,
        Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
        Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
-        Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, Opt_err,
+        Opt_enospc_debug, Opt_subvolrootid, Opt_defrag,
+        Opt_inode_cache, Opt_err,
 };
 static match_table_t tokens = {
@@ -193,6 +194,7 @@ static match_table_t tokens = {
        {Opt_enospc_debug, "enospc_debug"},
        {Opt_subvolrootid, "subvolrootid=%d"},
        {Opt_defrag, "autodefrag"},
+        {Opt_inode_cache, "inode_cache"},
        {Opt_err, NULL},
 };
@@ -361,6 +363,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
                        printk(KERN_INFO "btrfs: enabling disk space caching\n");
                        btrfs_set_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);
+                        break;
                case Opt_clear_cache:
                        printk(KERN_INFO "btrfs: force clearing of disk cache\n");
                        btrfs_set_opt(info->mount_opt, CLEAR_CACHE);
@@ -819,7 +825,7 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
        } else {
                char b[BDEVNAME_SIZE];
-                s->s_flags = flags;
+                s->s_flags = flags | MS_NOSEC;
                strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
                error = btrfs_fill_super(s, fs_devices, data,
                                         flags & MS_SILENT ? 1 : 0);
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index dc80f7156923..2b3590b9fe98 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -35,6 +35,7 @@ static noinline void put_transaction(struct btrfs_transaction *transaction)
 {
        WARN_ON(atomic_read(&transaction->use_count) == 0);
        if (atomic_dec_and_test(&transaction->use_count)) {
+                BUG_ON(!list_empty(&transaction->list));
                memset(transaction, 0, sizeof(*transaction));
                kmem_cache_free(btrfs_transaction_cachep, transaction);
        }
@@ -49,46 +50,72 @@ static noinline void switch_commit_root(struct btrfs_root *root)
 /*
 * either allocate a new transaction or hop into the existing one
 */
-static noinline int join_transaction(struct btrfs_root *root)
+static noinline int join_transaction(struct btrfs_root *root, int nofail)
 {
        struct btrfs_transaction *cur_trans;
+        spin_lock(&root->fs_info->trans_lock);
+        if (root->fs_info->trans_no_join) {
+                if (!nofail) {
+                        spin_unlock(&root->fs_info->trans_lock);
+                        return -EBUSY;
+                }
+        }
        cur_trans = root->fs_info->running_transaction;
-        if (!cur_trans) {
+        if (cur_trans) {
-                cur_trans = kmem_cache_alloc(btrfs_transaction_cachep,
+                atomic_inc(&cur_trans->use_count);
-                                             GFP_NOFS);
-                if (!cur_trans)
-                        return -ENOMEM;
-                root->fs_info->generation++;
-                atomic_set(&cur_trans->num_writers, 1);
-                cur_trans->num_joined = 0;
-                cur_trans->transid = root->fs_info->generation;
-                init_waitqueue_head(&cur_trans->writer_wait);
-                init_waitqueue_head(&cur_trans->commit_wait);
-                cur_trans->in_commit = 0;
-                cur_trans->blocked = 0;
-                atomic_set(&cur_trans->use_count, 1);
-                cur_trans->commit_done = 0;
-                cur_trans->start_time = get_seconds();
-                cur_trans->delayed_refs.root = RB_ROOT;
-                cur_trans->delayed_refs.num_entries = 0;
-                cur_trans->delayed_refs.num_heads_ready = 0;
-                cur_trans->delayed_refs.num_heads = 0;
-                cur_trans->delayed_refs.flushing = 0;
-                cur_trans->delayed_refs.run_delayed_start = 0;
-                spin_lock_init(&cur_trans->delayed_refs.lock);
-                INIT_LIST_HEAD(&cur_trans->pending_snapshots);
-                list_add_tail(&cur_trans->list, &root->fs_info->trans_list);
-                extent_io_tree_init(&cur_trans->dirty_pages,
-                                     root->fs_info->btree_inode->i_mapping);
-                spin_lock(&root->fs_info->new_trans_lock);
-                root->fs_info->running_transaction = cur_trans;
-                spin_unlock(&root->fs_info->new_trans_lock);
-        } else {
                atomic_inc(&cur_trans->num_writers);
                cur_trans->num_joined++;
+                spin_unlock(&root->fs_info->trans_lock);
+                return 0;
        }
+        spin_unlock(&root->fs_info->trans_lock);
+        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) {
+                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;
+        }
+        atomic_set(&cur_trans->num_writers, 1);
+        cur_trans->num_joined = 0;
+        init_waitqueue_head(&cur_trans->writer_wait);
+        init_waitqueue_head(&cur_trans->commit_wait);
+        cur_trans->in_commit = 0;
+        cur_trans->blocked = 0;
+        /*
+         * One for this trans handle, one so it will live on until we
+         * commit the transaction.
+         */
+        atomic_set(&cur_trans->use_count, 2);
+        cur_trans->commit_done = 0;
+        cur_trans->start_time = get_seconds();
+        cur_trans->delayed_refs.root = RB_ROOT;
+        cur_trans->delayed_refs.num_entries = 0;
+        cur_trans->delayed_refs.num_heads_ready = 0;
+        cur_trans->delayed_refs.num_heads = 0;
+        cur_trans->delayed_refs.flushing = 0;
+        cur_trans->delayed_refs.run_delayed_start = 0;
+        spin_lock_init(&cur_trans->commit_lock);
+        spin_lock_init(&cur_trans->delayed_refs.lock);
+        INIT_LIST_HEAD(&cur_trans->pending_snapshots);
+        list_add_tail(&cur_trans->list, &root->fs_info->trans_list);
+        extent_io_tree_init(&cur_trans->dirty_pages,
+                             root->fs_info->btree_inode->i_mapping);
+        root->fs_info->generation++;
+        cur_trans->transid = root->fs_info->generation;
+        root->fs_info->running_transaction = cur_trans;
+        spin_unlock(&root->fs_info->trans_lock);
        return 0;
 }
@@ -99,39 +126,28 @@ static noinline int join_transaction(struct btrfs_root *root)
 * to make sure the old root from before we joined the transaction is deleted
 * when the transaction commits
 */
-static noinline int record_root_in_trans(struct btrfs_trans_handle *trans,
+int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
-                                         struct btrfs_root *root)
+                               struct btrfs_root *root)
 {
        if (root->ref_cows && root->last_trans < trans->transid) {
                WARN_ON(root == root->fs_info->extent_root);
                WARN_ON(root->commit_root != root->node);
+                spin_lock(&root->fs_info->fs_roots_radix_lock);
+                if (root->last_trans == trans->transid) {
+                        spin_unlock(&root->fs_info->fs_roots_radix_lock);
+                        return 0;
+                }
+                root->last_trans = trans->transid;
                radix_tree_tag_set(&root->fs_info->fs_roots_radix,
                           (unsigned long)root->root_key.objectid,
                           BTRFS_ROOT_TRANS_TAG);
-                root->last_trans = trans->transid;
+                spin_unlock(&root->fs_info->fs_roots_radix_lock);
                btrfs_init_reloc_root(trans, root);
        }
        return 0;
 }
-int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
-                               struct btrfs_root *root)
-{
-        if (!root->ref_cows)
-                return 0;
-        mutex_lock(&root->fs_info->trans_mutex);
-        if (root->last_trans == trans->transid) {
-                mutex_unlock(&root->fs_info->trans_mutex);
-                return 0;
-        }
-        record_root_in_trans(trans, root);
-        mutex_unlock(&root->fs_info->trans_mutex);
-        return 0;
-}
 /* wait for commit against the current transaction to become unblocked
 * when this is done, it is safe to start a new transaction, but the current
 * transaction might not be fully on disk.
@@ -140,21 +156,23 @@ static void wait_current_trans(struct btrfs_root *root)
 {
        struct btrfs_transaction *cur_trans;
+        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;
-                        mutex_unlock(&root->fs_info->trans_mutex);
                        schedule();
-                        mutex_lock(&root->fs_info->trans_mutex);
                }
                finish_wait(&root->fs_info->transaction_wait, &wait);
                put_transaction(cur_trans);
+        } else {
+                spin_unlock(&root->fs_info->trans_lock);
        }
 }
@@ -167,10 +185,16 @@ enum btrfs_trans_type {
 static int may_wait_transaction(struct btrfs_root *root, int type)
 {
-        if (!root->fs_info->log_root_recovering &&
+        if (root->fs_info->log_root_recovering)
-            ((type == TRANS_START && !root->fs_info->open_ioctl_trans) ||
+                return 0;
-             type == TRANS_USERSPACE))
+        if (type == TRANS_USERSPACE)
                return 1;
+        if (type == TRANS_START &&
+            !atomic_read(&root->fs_info->open_ioctl_trans))
+                return 1;
        return 0;
 }
@@ -184,36 +208,44 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
        if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR)
                return ERR_PTR(-EROFS);
+        if (current->journal_info) {
+                WARN_ON(type != TRANS_JOIN && type != TRANS_JOIN_NOLOCK);
+                h = current->journal_info;
+                h->use_count++;
+                h->orig_rsv = h->block_rsv;
+                h->block_rsv = NULL;
+                goto got_it;
+        }
 again:
        h = kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
        if (!h)
                return ERR_PTR(-ENOMEM);
-        if (type != TRANS_JOIN_NOLOCK)
-                mutex_lock(&root->fs_info->trans_mutex);
        if (may_wait_transaction(root, type))
                wait_current_trans(root);
-        ret = join_transaction(root);
+        do {
+                ret = join_transaction(root, type == TRANS_JOIN_NOLOCK);
+                if (ret == -EBUSY)
+                        wait_current_trans(root);
+        } while (ret == -EBUSY);
        if (ret < 0) {
                kmem_cache_free(btrfs_trans_handle_cachep, h);
-                if (type != TRANS_JOIN_NOLOCK)
-                        mutex_unlock(&root->fs_info->trans_mutex);
                return ERR_PTR(ret);
        }
        cur_trans = root->fs_info->running_transaction;
-        atomic_inc(&cur_trans->use_count);
-        if (type != TRANS_JOIN_NOLOCK)
-                mutex_unlock(&root->fs_info->trans_mutex);
        h->transid = cur_trans->transid;
        h->transaction = cur_trans;
        h->blocks_used = 0;
-        h->block_group = 0;
        h->bytes_reserved = 0;
        h->delayed_ref_updates = 0;
+        h->use_count = 1;
        h->block_rsv = NULL;
+        h->orig_rsv = NULL;
        smp_mb();
        if (cur_trans->blocked && may_wait_transaction(root, type)) {
@@ -241,11 +273,8 @@ again:
                }
        }
-        if (type != TRANS_JOIN_NOLOCK)
+got_it:
-                mutex_lock(&root->fs_info->trans_mutex);
+        btrfs_record_root_in_trans(h, root);
-        record_root_in_trans(h, root);
-        if (type != TRANS_JOIN_NOLOCK)
-                mutex_unlock(&root->fs_info->trans_mutex);
        if (!current->journal_info && type != TRANS_USERSPACE)
                current->journal_info = h;
@@ -257,22 +286,19 @@ struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
 {
        return start_transaction(root, num_items, TRANS_START);
 }
-struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
+struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root)
-                                                   int num_blocks)
 {
        return start_transaction(root, 0, TRANS_JOIN);
 }
-struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root,
+struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root)
-                                                          int num_blocks)
 {
        return start_transaction(root, 0, TRANS_JOIN_NOLOCK);
 }
-struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
+struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *root)
-                                                         int num_blocks)
 {
-        return start_transaction(r, 0, TRANS_USERSPACE);
+        return start_transaction(root, 0, TRANS_USERSPACE);
 }
 /* wait for a transaction commit to be fully complete */
@@ -280,17 +306,13 @@ static noinline int wait_for_commit(struct btrfs_root *root,
                                    struct btrfs_transaction *commit)
 {
        DEFINE_WAIT(wait);
-        mutex_lock(&root->fs_info->trans_mutex);
        while (!commit->commit_done) {
                prepare_to_wait(&commit->commit_wait, &wait,
                                TASK_UNINTERRUPTIBLE);
                if (commit->commit_done)
                        break;
-                mutex_unlock(&root->fs_info->trans_mutex);
                schedule();
-                mutex_lock(&root->fs_info->trans_mutex);
        }
-        mutex_unlock(&root->fs_info->trans_mutex);
        finish_wait(&commit->commit_wait, &wait);
        return 0;
 }
@@ -300,59 +322,56 @@ int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid)
        struct btrfs_transaction *cur_trans = NULL, *t;
        int ret;
-        mutex_lock(&root->fs_info->trans_mutex);
        ret = 0;
        if (transid) {
                if (transid <= root->fs_info->last_trans_committed)
-                        goto out_unlock;
+                        goto out;
                /* find specified transaction */
+                spin_lock(&root->fs_info->trans_lock);
                list_for_each_entry(t, &root->fs_info->trans_list, list) {
                        if (t->transid == transid) {
                                cur_trans = t;
+                                atomic_inc(&cur_trans->use_count);
                                break;
                        }
                        if (t->transid > transid)
                                break;
                }
+                spin_unlock(&root->fs_info->trans_lock);
                ret = -EINVAL;
                if (!cur_trans)
-                        goto out_unlock;  /* bad transid */
+                        goto out;  /* bad transid */
        } else {
                /* find newest transaction that is committing | committed */
+                spin_lock(&root->fs_info->trans_lock);
                list_for_each_entry_reverse(t, &root->fs_info->trans_list,
                                            list) {
                        if (t->in_commit) {
                                if (t->commit_done)
-                                        goto out_unlock;
+                                        break;
                                cur_trans = t;
+                                atomic_inc(&cur_trans->use_count);
                                break;
                        }
                }
+                spin_unlock(&root->fs_info->trans_lock);
                if (!cur_trans)
-                        goto out_unlock;  /* nothing committing|committed */
+                        goto out;  /* nothing committing|committed */
        }
-        atomic_inc(&cur_trans->use_count);
-        mutex_unlock(&root->fs_info->trans_mutex);
        wait_for_commit(root, cur_trans);
-        mutex_lock(&root->fs_info->trans_mutex);
        put_transaction(cur_trans);
        ret = 0;
-out_unlock:
+out:
-        mutex_unlock(&root->fs_info->trans_mutex);
        return ret;
 }
 void btrfs_throttle(struct btrfs_root *root)
 {
-        mutex_lock(&root->fs_info->trans_mutex);
+        if (!atomic_read(&root->fs_info->open_ioctl_trans))
-        if (!root->fs_info->open_ioctl_trans)
                wait_current_trans(root);
-        mutex_unlock(&root->fs_info->trans_mutex);
 }
 static int should_end_transaction(struct btrfs_trans_handle *trans,
@@ -370,6 +389,7 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
        struct btrfs_transaction *cur_trans = trans->transaction;
        int updates;
+        smp_mb();
        if (cur_trans->blocked || cur_trans->delayed_refs.flushing)
                return 1;
@@ -388,6 +408,11 @@ 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) {
+                trans->block_rsv = trans->orig_rsv;
+                return 0;
+        }
        while (count < 4) {
                unsigned long cur = trans->delayed_ref_updates;
                trans->delayed_ref_updates = 0;
@@ -410,9 +435,11 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
        btrfs_trans_release_metadata(trans, root);
-        if (lock && !root->fs_info->open_ioctl_trans &&
+        if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) &&
-            should_end_transaction(trans, root))
+            should_end_transaction(trans, root)) {
                trans->transaction->blocked = 1;
+                smp_wmb();
+        }
        if (lock && cur_trans->blocked && !cur_trans->in_commit) {
                if (throttle)
@@ -703,9 +730,9 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans,
 */
 int btrfs_add_dead_root(struct btrfs_root *root)
 {
-        mutex_lock(&root->fs_info->trans_mutex);
+        spin_lock(&root->fs_info->trans_lock);
        list_add(&root->root_list, &root->fs_info->dead_roots);
-        mutex_unlock(&root->fs_info->trans_mutex);
+        spin_unlock(&root->fs_info->trans_lock);
        return 0;
 }
@@ -721,6 +748,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
        int ret;
        int err = 0;
+        spin_lock(&fs_info->fs_roots_radix_lock);
        while (1) {
                ret = radix_tree_gang_lookup_tag(&fs_info->fs_roots_radix,
                                                 (void **)gang, 0,
@@ -733,6 +761,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
                        radix_tree_tag_clear(&fs_info->fs_roots_radix,
                                        (unsigned long)root->root_key.objectid,
                                        BTRFS_ROOT_TRANS_TAG);
+                        spin_unlock(&fs_info->fs_roots_radix_lock);
                        btrfs_free_log(trans, root);
                        btrfs_update_reloc_root(trans, root);
@@ -753,10 +782,12 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
                        err = btrfs_update_root(trans, fs_info->tree_root,
                                                &root->root_key,
                                                &root->root_item);
+                        spin_lock(&fs_info->fs_roots_radix_lock);
                        if (err)
                                break;
                }
        }
+        spin_unlock(&fs_info->fs_roots_radix_lock);
        return err;
 }
@@ -786,7 +817,7 @@ int btrfs_defrag_root(struct btrfs_root *root, int cacheonly)
                btrfs_btree_balance_dirty(info->tree_root, nr);
                cond_resched();
-                if (root->fs_info->closing || ret != -EAGAIN)
+                if (btrfs_fs_closing(root->fs_info) || ret != -EAGAIN)
                        break;
        }
        root->defrag_running = 0;
@@ -851,7 +882,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
        parent = dget_parent(dentry);
        parent_inode = parent->d_inode;
        parent_root = BTRFS_I(parent_inode)->root;
-        record_root_in_trans(trans, parent_root);
+        btrfs_record_root_in_trans(trans, parent_root);
        /*
         * insert the directory item
@@ -869,7 +900,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
        ret = btrfs_update_inode(trans, parent_root, parent_inode);
        BUG_ON(ret);
-        record_root_in_trans(trans, root);
+        btrfs_record_root_in_trans(trans, root);
        btrfs_set_root_last_snapshot(&root->root_item, trans->transid);
        memcpy(new_root_item, &root->root_item, sizeof(*new_root_item));
        btrfs_check_and_init_root_item(new_root_item);
@@ -967,20 +998,20 @@ static void update_super_roots(struct btrfs_root *root)
 int btrfs_transaction_in_commit(struct btrfs_fs_info *info)
 {
        int ret = 0;
-        spin_lock(&info->new_trans_lock);
+        spin_lock(&info->trans_lock);
        if (info->running_transaction)
                ret = info->running_transaction->in_commit;
-        spin_unlock(&info->new_trans_lock);
+        spin_unlock(&info->trans_lock);
        return ret;
 }
 int btrfs_transaction_blocked(struct btrfs_fs_info *info)
 {
        int ret = 0;
-        spin_lock(&info->new_trans_lock);
+        spin_lock(&info->trans_lock);
        if (info->running_transaction)
                ret = info->running_transaction->blocked;
-        spin_unlock(&info->new_trans_lock);
+        spin_unlock(&info->trans_lock);
        return ret;
 }
@@ -1004,9 +1035,7 @@ static void wait_current_trans_commit_start(struct btrfs_root *root,
                                    &wait);
                        break;
                }
-                mutex_unlock(&root->fs_info->trans_mutex);
                schedule();
-                mutex_lock(&root->fs_info->trans_mutex);
                finish_wait(&root->fs_info->transaction_blocked_wait, &wait);
        }
 }
@@ -1032,9 +1061,7 @@ static void wait_current_trans_commit_start_and_unblock(struct btrfs_root *root,
                                    &wait);
                        break;
                }
-                mutex_unlock(&root->fs_info->trans_mutex);
                schedule();
-                mutex_lock(&root->fs_info->trans_mutex);
                finish_wait(&root->fs_info->transaction_wait,
                            &wait);
        }
@@ -1072,7 +1099,7 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans,
        INIT_DELAYED_WORK(&ac->work, do_async_commit);
        ac->root = root;
-        ac->newtrans = btrfs_join_transaction(root, 0);
+        ac->newtrans = btrfs_join_transaction(root);
        if (IS_ERR(ac->newtrans)) {
                int err = PTR_ERR(ac->newtrans);
                kfree(ac);
@@ -1080,23 +1107,22 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans,
        }
        /* take transaction reference */
-        mutex_lock(&root->fs_info->trans_mutex);
        cur_trans = trans->transaction;
        atomic_inc(&cur_trans->use_count);
-        mutex_unlock(&root->fs_info->trans_mutex);
        btrfs_end_transaction(trans, root);
        schedule_delayed_work(&ac->work, 0);
        /* wait for transaction to start and unblock */
-        mutex_lock(&root->fs_info->trans_mutex);
        if (wait_for_unblock)
                wait_current_trans_commit_start_and_unblock(root, cur_trans);
        else
                wait_current_trans_commit_start(root, cur_trans);
-        put_transaction(cur_trans);
-        mutex_unlock(&root->fs_info->trans_mutex);
+        if (current->journal_info == trans)
+                current->journal_info = NULL;
+        put_transaction(cur_trans);
        return 0;
 }
@@ -1139,38 +1165,41 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
        ret = btrfs_run_delayed_refs(trans, root, 0);
        BUG_ON(ret);
-        mutex_lock(&root->fs_info->trans_mutex);
+        spin_lock(&cur_trans->commit_lock);
        if (cur_trans->in_commit) {
+                spin_unlock(&cur_trans->commit_lock);
                atomic_inc(&cur_trans->use_count);
-                mutex_unlock(&root->fs_info->trans_mutex);
                btrfs_end_transaction(trans, root);
                ret = wait_for_commit(root, cur_trans);
                BUG_ON(ret);
-                mutex_lock(&root->fs_info->trans_mutex);
                put_transaction(cur_trans);
-                mutex_unlock(&root->fs_info->trans_mutex);
                return 0;
        }
        trans->transaction->in_commit = 1;
        trans->transaction->blocked = 1;
+        spin_unlock(&cur_trans->commit_lock);
        wake_up(&root->fs_info->transaction_blocked_wait);
+        spin_lock(&root->fs_info->trans_lock);
        if (cur_trans->list.prev != &root->fs_info->trans_list) {
                prev_trans = list_entry(cur_trans->list.prev,
                                        struct btrfs_transaction, list);
                if (!prev_trans->commit_done) {
                        atomic_inc(&prev_trans->use_count);
-                        mutex_unlock(&root->fs_info->trans_mutex);
+                        spin_unlock(&root->fs_info->trans_lock);
                        wait_for_commit(root, prev_trans);
-                        mutex_lock(&root->fs_info->trans_mutex);
                        put_transaction(prev_trans);
+                } else {
+                        spin_unlock(&root->fs_info->trans_lock);
                }
+        } else {
+                spin_unlock(&root->fs_info->trans_lock);
        }
        if (now < cur_trans->start_time || now - cur_trans->start_time < 1)
@@ -1178,12 +1207,12 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
        do {
                int snap_pending = 0;
                joined = cur_trans->num_joined;
                if (!list_empty(&trans->transaction->pending_snapshots))
                        snap_pending = 1;
                WARN_ON(cur_trans != trans->transaction);
-                mutex_unlock(&root->fs_info->trans_mutex);
                if (flush_on_commit || snap_pending) {
                        btrfs_start_delalloc_inodes(root, 1);
@@ -1206,14 +1235,15 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
                prepare_to_wait(&cur_trans->writer_wait, &wait,
                                TASK_UNINTERRUPTIBLE);
-                smp_mb();
                if (atomic_read(&cur_trans->num_writers) > 1)
                        schedule_timeout(MAX_SCHEDULE_TIMEOUT);
                else if (should_grow)
                        schedule_timeout(1);
-                mutex_lock(&root->fs_info->trans_mutex);
                finish_wait(&cur_trans->writer_wait, &wait);
+                spin_lock(&root->fs_info->trans_lock);
+                root->fs_info->trans_no_join = 1;
+                spin_unlock(&root->fs_info->trans_lock);
        } while (atomic_read(&cur_trans->num_writers) > 1 ||
                 (should_grow && cur_trans->num_joined != joined));
@@ -1258,9 +1288,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
        btrfs_prepare_extent_commit(trans, root);
        cur_trans = root->fs_info->running_transaction;
-        spin_lock(&root->fs_info->new_trans_lock);
-        root->fs_info->running_transaction = NULL;
-        spin_unlock(&root->fs_info->new_trans_lock);
        btrfs_set_root_node(&root->fs_info->tree_root->root_item,
                            root->fs_info->tree_root->node);
@@ -1281,10 +1308,13 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
               sizeof(root->fs_info->super_copy));
        trans->transaction->blocked = 0;
+        spin_lock(&root->fs_info->trans_lock);
+        root->fs_info->running_transaction = NULL;
+        root->fs_info->trans_no_join = 0;
+        spin_unlock(&root->fs_info->trans_lock);
        wake_up(&root->fs_info->transaction_wait);
-        mutex_unlock(&root->fs_info->trans_mutex);
        ret = btrfs_write_and_wait_transaction(trans, root);
        BUG_ON(ret);
        write_ctree_super(trans, root, 0);
@@ -1297,22 +1327,21 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
        btrfs_finish_extent_commit(trans, root);
-        mutex_lock(&root->fs_info->trans_mutex);
        cur_trans->commit_done = 1;
        root->fs_info->last_trans_committed = cur_trans->transid;
        wake_up(&cur_trans->commit_wait);
+        spin_lock(&root->fs_info->trans_lock);
        list_del_init(&cur_trans->list);
+        spin_unlock(&root->fs_info->trans_lock);
        put_transaction(cur_trans);
        put_transaction(cur_trans);
        trace_btrfs_transaction_commit(root);
-        mutex_unlock(&root->fs_info->trans_mutex);
        btrfs_scrub_continue(root);
        if (current->journal_info == trans)
@@ -1334,9 +1363,9 @@ int btrfs_clean_old_snapshots(struct btrfs_root *root)
        LIST_HEAD(list);
        struct btrfs_fs_info *fs_info = root->fs_info;
-        mutex_lock(&fs_info->trans_mutex);
+        spin_lock(&fs_info->trans_lock);
        list_splice_init(&fs_info->dead_roots, &list);
-        mutex_unlock(&fs_info->trans_mutex);
+        spin_unlock(&fs_info->trans_lock);
        while (!list_empty(&list)) {
                root = list_entry(list.next, struct btrfs_root, root_list);
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index 804c88639e5d..02564e6230ac 100644
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -28,10 +28,12 @@ struct btrfs_transaction {
         * transaction can end
         */
        atomic_t num_writers;
+        atomic_t use_count;
        unsigned long num_joined;
+        spinlock_t commit_lock;
        int in_commit;
-        atomic_t use_count;
        int commit_done;
        int blocked;
        struct list_head list;
@@ -45,13 +47,14 @@ struct btrfs_transaction {
 struct btrfs_trans_handle {
        u64 transid;
-        u64 block_group;
        u64 bytes_reserved;
+        unsigned long use_count;
        unsigned long blocks_reserved;
        unsigned long blocks_used;
        unsigned long delayed_ref_updates;
        struct btrfs_transaction *transaction;
        struct btrfs_block_rsv *block_rsv;
+        struct btrfs_block_rsv *orig_rsv;
 };
 struct btrfs_pending_snapshot {
@@ -66,19 +69,6 @@ struct btrfs_pending_snapshot {
        struct list_head list;
 };
-static inline void btrfs_set_trans_block_group(struct btrfs_trans_handle *trans,
-                                               struct inode *inode)
-{
-        trans->block_group = BTRFS_I(inode)->block_group;
-}
-static inline void btrfs_update_inode_block_group(
-                                          struct btrfs_trans_handle *trans,
-                                          struct inode *inode)
-{
-        BTRFS_I(inode)->block_group = trans->block_group;
-}
 static inline void btrfs_set_inode_last_trans(struct btrfs_trans_handle *trans,
                                              struct inode *inode)
 {
@@ -92,12 +82,9 @@ int btrfs_end_transaction_nolock(struct btrfs_trans_handle *trans,
                                 struct btrfs_root *root);
 struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
                                                   int num_items);
-struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
+struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root);
-                                                  int num_blocks);
+struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root);
-struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root,
+struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *root);
-                                                          int num_blocks);
-struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
-                                                         int num_blocks);
 int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid);
 int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
                                     struct btrfs_root *root);
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index c48214ef5c09..1efa56e18f9b 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -504,7 +504,7 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
                BUG_ON(!new_device);
                memcpy(new_device, device, sizeof(*new_device));
                new_device->name = kstrdup(device->name, GFP_NOFS);
-                BUG_ON(!new_device->name);
+                BUG_ON(device->name && !new_device->name);
                new_device->bdev = NULL;
                new_device->writeable = 0;
                new_device->in_fs_metadata = 0;
@@ -689,12 +689,8 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
        transid = btrfs_super_generation(disk_super);
        if (disk_super->label[0])
                printk(KERN_INFO "device label %s ", disk_super->label);
-        else {
+        else
-                /* FIXME, make a readl uuid parser */
+                printk(KERN_INFO "device fsid %pU ", disk_super->fsid);
-                printk(KERN_INFO "device fsid %llx-%llx ",
-                       *(unsigned long long *)disk_super->fsid,
-                       *(unsigned long long *)(disk_super->fsid + 8));
-        }
        printk(KERN_CONT "devid %llu transid %llu %s\n",
               (unsigned long long)devid, (unsigned long long)transid, path);
        ret = device_list_add(path, disk_super, devid, fs_devices_ret);
diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c
index f3107e4b4d56..5366fe452ab0 100644
--- a/fs/btrfs/xattr.c
+++ b/fs/btrfs/xattr.c
@@ -158,8 +158,6 @@ int __btrfs_setxattr(struct btrfs_trans_handle *trans,
        if (IS_ERR(trans))
                return PTR_ERR(trans);
-        btrfs_set_trans_block_group(trans, inode);
        ret = do_setxattr(trans, inode, name, value, size, flags);
        if (ret)
                goto out;
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 33da49dc3cc6..5a3953db8118 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -453,7 +453,7 @@ static int ceph_writepage(struct page *page, struct writeback_control *wbc)
        int err;
        struct inode *inode = page->mapping->host;
        BUG_ON(!inode);
-        igrab(inode);
+        ihold(inode);
        err = writepage_nounlock(page, wbc);
        unlock_page(page);
        iput(inode);
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 1f72b00447c4..f605753c8fe9 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2940,14 +2940,12 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc)
        while (!list_empty(&mdsc->cap_dirty)) {
                ci = list_first_entry(&mdsc->cap_dirty, struct ceph_inode_info,
                                      i_dirty_item);
-                inode = igrab(&ci->vfs_inode);
+                inode = &ci->vfs_inode;
+                ihold(inode);
                dout("flush_dirty_caps %p\n", inode);
                spin_unlock(&mdsc->cap_dirty_lock);
-                if (inode) {
+                ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_FLUSH, NULL);
-                        ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_FLUSH,
+                iput(inode);
-                                        NULL);
-                        iput(inode);
-                }
                spin_lock(&mdsc->cap_dirty_lock);
        }
        spin_unlock(&mdsc->cap_dirty_lock);
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 33729e822bb9..ef8f08c343e8 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -308,7 +308,8 @@ more:
                req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS);
                if (IS_ERR(req))
                        return PTR_ERR(req);
-                req->r_inode = igrab(inode);
+                req->r_inode = inode;
+                ihold(inode);
                req->r_dentry = dget(filp->f_dentry);
                /* hints to request -> mds selection code */
                req->r_direct_mode = USE_AUTH_MDS;
@@ -787,10 +788,12 @@ static int ceph_link(struct dentry *old_dentry, struct inode *dir,
        req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
        req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
        err = ceph_mdsc_do_request(mdsc, dir, req);
-        if (err)
+        if (err) {
                d_drop(dentry);
-        else if (!req->r_reply_info.head->is_dentry)
+        } else if (!req->r_reply_info.head->is_dentry) {
-                d_instantiate(dentry, igrab(old_dentry->d_inode));
+                ihold(old_dentry->d_inode);
+                d_instantiate(dentry, old_dentry->d_inode);
+        }
        ceph_mdsc_put_request(req);
        return err;
 }
diff --git a/fs/ceph/export.c b/fs/ceph/export.c
index a610d3d67488..f67b687550de 100644
--- a/fs/ceph/export.c
+++ b/fs/ceph/export.c
@@ -109,7 +109,7 @@ static struct dentry *__fh_to_dentry(struct super_block *sb,
                err = ceph_mdsc_do_request(mdsc, NULL, req);
                inode = req->r_target_inode;
                if (inode)
-                        igrab(inode);
+                        ihold(inode);
                ceph_mdsc_put_request(req);
                if (!inode)
                        return ERR_PTR(-ESTALE);
@@ -167,7 +167,7 @@ static struct dentry *__cfh_to_dentry(struct super_block *sb,
                err = ceph_mdsc_do_request(mdsc, NULL, req);
                inode = req->r_target_inode;
                if (inode)
-                        igrab(inode);
+                        ihold(inode);
                ceph_mdsc_put_request(req);
                if (!inode)
                        return ERR_PTR(err ? err : -ESTALE);
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 203252d88d9f..9542f07d0b93 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -191,7 +191,8 @@ int ceph_open(struct inode *inode, struct file *file)
                err = PTR_ERR(req);
                goto out;
        }
-        req->r_inode = igrab(inode);
+        req->r_inode = inode;
+        ihold(inode);
        req->r_num_caps = 1;
        err = ceph_mdsc_do_request(mdsc, parent_inode, req);
        if (!err)
@@ -282,7 +283,7 @@ int ceph_release(struct inode *inode, struct file *file)
 static int striped_read(struct inode *inode,
                        u64 off, u64 len,
                        struct page **pages, int num_pages,
-                        int *checkeof, bool align_to_pages,
+                        int *checkeof, bool o_direct,
                        unsigned long buf_align)
 {
        struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
@@ -307,7 +308,7 @@ static int striped_read(struct inode *inode,
        io_align = off & ~PAGE_MASK;
 more:
-        if (align_to_pages)
+        if (o_direct)
                page_align = (pos - io_align + buf_align) & ~PAGE_MASK;
        else
                page_align = pos & ~PAGE_MASK;
@@ -317,10 +318,10 @@ more:
                                  ci->i_truncate_seq,
                                  ci->i_truncate_size,
                                  page_pos, pages_left, page_align);
-        hit_stripe = this_len < left;
-        was_short = ret >= 0 && ret < this_len;
        if (ret == -ENOENT)
                ret = 0;
+        hit_stripe = this_len < left;
+        was_short = ret >= 0 && ret < this_len;
        dout("striped_read %llu~%u (read %u) got %d%s%s\n", pos, left, read,
             ret, hit_stripe ? " HITSTRIPE" : "", was_short ? " SHORT" : "");
@@ -345,20 +346,22 @@ more:
        }
        if (was_short) {
-                /* was original extent fully inside i_size? */
+                /* did we bounce off eof? */
-                if (pos + left <= inode->i_size) {
+                if (pos + left > inode->i_size)
-                        dout("zero tail\n");
+                        *checkeof = 1;
-                        ceph_zero_page_vector_range(page_off + read, len - read,
+                /* zero trailing bytes (inside i_size) */
+                if (left > 0 && pos < inode->i_size) {
+                        if (pos + left > inode->i_size)
+                                left = inode->i_size - pos;
+                        dout("zero tail %d\n", left);
+                        ceph_zero_page_vector_range(page_off + read, left,
                                                    pages);
-                        read = len;
+                        read += left;
-                        goto out;
                }
-                /* check i_size */
-                *checkeof = 1;
        }
-out:
        if (ret >= 0)
                ret = read;
        dout("striped_read returns %d\n", ret);
@@ -658,7 +661,7 @@ out:
                /* hit EOF or hole? */
                if (statret == 0 && *ppos < inode->i_size) {
-                        dout("aio_read sync_read hit hole, reading more\n");
+                        dout("aio_read sync_read hit hole, ppos %lld < size %lld, reading more\n", *ppos, inode->i_size);
                        read += ret;
                        base += ret;
                        len -= ret;
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 70b6a4839c38..d8858e96ab18 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -1101,10 +1101,10 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
                                goto done;
                        }
                        req->r_dentry = dn;  /* may have spliced */
-                        igrab(in);
+                        ihold(in);
                } else if (ceph_ino(in) == vino.ino &&
                           ceph_snap(in) == vino.snap) {
-                        igrab(in);
+                        ihold(in);
                } else {
                        dout(" %p links to %p %llx.%llx, not %llx.%llx\n",
                             dn, in, ceph_ino(in), ceph_snap(in),
@@ -1144,7 +1144,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
                        goto done;
                }
                req->r_dentry = dn;  /* may have spliced */
-                igrab(in);
+                ihold(in);
                rinfo->head->is_dentry = 1;  /* fool notrace handlers */
        }
@@ -1328,7 +1328,7 @@ void ceph_queue_writeback(struct inode *inode)
        if (queue_work(ceph_inode_to_client(inode)->wb_wq,
                       &ceph_inode(inode)->i_wb_work)) {
                dout("ceph_queue_writeback %p\n", inode);
-                igrab(inode);
+                ihold(inode);
        } else {
                dout("ceph_queue_writeback %p failed\n", inode);
        }
@@ -1353,7 +1353,7 @@ void ceph_queue_invalidate(struct inode *inode)
        if (queue_work(ceph_inode_to_client(inode)->pg_inv_wq,
                       &ceph_inode(inode)->i_pg_inv_work)) {
                dout("ceph_queue_invalidate %p\n", inode);
-                igrab(inode);
+                ihold(inode);
        } else {
                dout("ceph_queue_invalidate %p failed\n", inode);
        }
@@ -1477,7 +1477,7 @@ void ceph_queue_vmtruncate(struct inode *inode)
        if (queue_work(ceph_sb_to_client(inode->i_sb)->trunc_wq,
                       &ci->i_vmtruncate_work)) {
                dout("ceph_queue_vmtruncate %p\n", inode);
-                igrab(inode);
+                ihold(inode);
        } else {
                dout("ceph_queue_vmtruncate %p failed, pending=%d\n",
                     inode, ci->i_truncate_pending);
@@ -1738,7 +1738,8 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr)
                __mark_inode_dirty(inode, inode_dirty_flags);
        if (mask) {
-                req->r_inode = igrab(inode);
+                req->r_inode = inode;
+                ihold(inode);
                req->r_inode_drop = release;
                req->r_args.setattr.mask = cpu_to_le32(mask);
                req->r_num_caps = 1;
@@ -1779,7 +1780,8 @@ int ceph_do_getattr(struct inode *inode, int mask)
        req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETATTR, USE_ANY_MDS);
        if (IS_ERR(req))
                return PTR_ERR(req);
-        req->r_inode = igrab(inode);
+        req->r_inode = inode;
+        ihold(inode);
        req->r_num_caps = 1;
        req->r_args.getattr.mask = cpu_to_le32(mask);
        err = ceph_mdsc_do_request(mdsc, NULL, req);
diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index 8888c9ba68db..ef0b5f48e13a 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -73,7 +73,8 @@ static long ceph_ioctl_set_layout(struct file *file, void __user *arg)
                                       USE_AUTH_MDS);
        if (IS_ERR(req))
                return PTR_ERR(req);
-        req->r_inode = igrab(inode);
+        req->r_inode = inode;
+        ihold(inode);
        req->r_inode_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_FILE_EXCL;
        req->r_args.setlayout.layout.fl_stripe_unit =
@@ -135,7 +136,8 @@ static long ceph_ioctl_set_layout_policy (struct file *file, void __user *arg)
        if (IS_ERR(req))
                return PTR_ERR(req);
-        req->r_inode = igrab(inode);
+        req->r_inode = inode;
+        ihold(inode);
        req->r_args.setlayout.layout.fl_stripe_unit =
                        cpu_to_le32(l.stripe_unit);
diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c
index 476b329867d4..80576d05d687 100644
--- a/fs/ceph/locks.c
+++ b/fs/ceph/locks.c
@@ -23,7 +23,8 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file,
        req = ceph_mdsc_create_request(mdsc, operation, USE_AUTH_MDS);
        if (IS_ERR(req))
                return PTR_ERR(req);
-        req->r_inode = igrab(inode);
+        req->r_inode = inode;
+        ihold(inode);
        /* mds requires start and length rather than start and end */
        if (LLONG_MAX == fl->fl_end)
@@ -32,11 +33,10 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file,
                length = fl->fl_end - fl->fl_start + 1;
        dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, "
-             "length: %llu, wait: %d, type`: %d", (int)lock_type,
+             "length: %llu, wait: %d, type: %d", (int)lock_type,
             (int)operation, (u64)fl->fl_pid, fl->fl_start,
             length, wait, fl->fl_type);
        req->r_args.filelock_change.rule = lock_type;
        req->r_args.filelock_change.type = cmd;
        req->r_args.filelock_change.pid = cpu_to_le64((u64)fl->fl_pid);
@@ -70,7 +70,7 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file,
        }
        ceph_mdsc_put_request(req);
        dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, "
-             "length: %llu, wait: %d, type`: %d, err code %d", (int)lock_type,
+             "length: %llu, wait: %d, type: %d, err code %d", (int)lock_type,
             (int)operation, (u64)fl->fl_pid, fl->fl_start,
             length, wait, fl->fl_type, err);
        return err;
@@ -109,16 +109,20 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl)
                        dout("mds locked, locking locally");
                        err = posix_lock_file(file, fl, NULL);
                        if (err && (CEPH_MDS_OP_SETFILELOCK == op)) {
-                                /* undo! This should only happen if the kernel detects
+                                /* undo! This should only happen if
-                                 * local deadlock. */
+                                 * the kernel detects local
+                                 * deadlock. */
                                ceph_lock_message(CEPH_LOCK_FCNTL, op, file,
                                                  CEPH_LOCK_UNLOCK, 0, fl);
-                                dout("got %d on posix_lock_file, undid lock", err);
+                                dout("got %d on posix_lock_file, undid lock",
+                                     err);
                        }
                }
-        } else {
+        } else if (err == -ERESTARTSYS) {
-                dout("mds returned error code %d", err);
+                dout("undoing lock\n");
+                ceph_lock_message(CEPH_LOCK_FCNTL, op, file,
+                                  CEPH_LOCK_UNLOCK, 0, fl);
        }
        return err;
 }
@@ -155,8 +159,11 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
                                          file, CEPH_LOCK_UNLOCK, 0, fl);
                        dout("got %d on flock_lock_file_wait, undid lock", err);
                }
-        } else {
+        } else if (err == -ERESTARTSYS) {
-                dout("mds error code %d", err);
+                dout("undoing lock\n");
+                ceph_lock_message(CEPH_LOCK_FLOCK,
+                                  CEPH_MDS_OP_SETFILELOCK,
+                                  file, CEPH_LOCK_UNLOCK, 0, fl);
        }
        return err;
 }
diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c
index 24067d68a554..54b14de2e729 100644
--- a/fs/ceph/snap.c
+++ b/fs/ceph/snap.c
@@ -722,7 +722,7 @@ static void flush_snaps(struct ceph_mds_client *mdsc)
                ci = list_first_entry(&mdsc->snap_flush_list,
                                struct ceph_inode_info, i_snap_flush_item);
                inode = &ci->vfs_inode;
-                igrab(inode);
+                ihold(inode);
                spin_unlock(&mdsc->snap_flush_lock);
                spin_lock(&inode->i_lock);
                __ceph_flush_snaps(ci, &session, 0);
diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c
index f2b628696180..f42d730f1b66 100644
--- a/fs/ceph/xattr.c
+++ b/fs/ceph/xattr.c
@@ -665,7 +665,8 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name,
                err = PTR_ERR(req);
                goto out;
        }
-        req->r_inode = igrab(inode);
+        req->r_inode = inode;
+        ihold(inode);
        req->r_inode_drop = CEPH_CAP_XATTR_SHARED;
        req->r_num_caps = 1;
        req->r_args.setxattr.flags = cpu_to_le32(flags);
@@ -795,7 +796,8 @@ static int ceph_send_removexattr(struct dentry *dentry, const char *name)
                                       USE_AUTH_MDS);
        if (IS_ERR(req))
                return PTR_ERR(req);
-        req->r_inode = igrab(inode);
+        req->r_inode = inode;
+        ihold(inode);
        req->r_inode_drop = CEPH_CAP_XATTR_SHARED;
        req->r_num_caps = 1;
        req->r_path2 = kstrdup(name, GFP_NOFS);
diff --git a/fs/cifs/Kconfig b/fs/cifs/Kconfig
index 1cd4c3a1862d..53ed1ad2c112 100644
--- a/fs/cifs/Kconfig
+++ b/fs/cifs/Kconfig
@@ -7,6 +7,7 @@ config CIFS
        select CRYPTO_MD5
        select CRYPTO_HMAC
        select CRYPTO_ARC4
+        select CRYPTO_ECB
        select CRYPTO_DES
        help
          This is the client VFS module for the Common Internet File System
@@ -148,7 +149,7 @@ config CIFS_FSCACHE
 config CIFS_ACL
          bool "Provide CIFS ACL support (EXPERIMENTAL)"
-          depends on EXPERIMENTAL && CIFS_XATTR
+          depends on EXPERIMENTAL && CIFS_XATTR && KEYS
          help
            Allows to fetch CIFS/NTFS ACL from the server.  The DACL blob
            is handed over to the application/caller.
diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
index dfbd9f1f373d..5a0ee7f2af06 100644
--- a/fs/cifs/cifsencrypt.c
+++ b/fs/cifs/cifsencrypt.c
@@ -184,7 +184,7 @@ int cifs_verify_signature(struct smb_hdr *cifs_pdu,
        if (cifs_pdu == NULL || server == NULL)
                return -EINVAL;
-        if (cifs_pdu->Command == SMB_COM_NEGOTIATE)
+        if (!server->session_estab)
                return 0;
        if (cifs_pdu->Command == SMB_COM_LOCKING_ANDX) {
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 6d88b82537c3..bb659eb73810 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -784,7 +784,7 @@ cifs_parse_mount_options(const char *mountdata, const char *devname,
                         struct smb_vol *vol)
 {
        char *value, *data, *end;
-        char *mountdata_copy, *options;
+        char *mountdata_copy = NULL, *options;
        unsigned int  temp_len, i, j;
        char separator[2];
        short int override_uid = -1;
@@ -1391,7 +1391,7 @@ cifs_parse_mount_options(const char *mountdata, const char *devname,
                                "/proc/fs/cifs/LookupCacheEnabled to 0\n");
                } else if (strnicmp(data, "fsc", 3) == 0) {
 #ifndef CONFIG_CIFS_FSCACHE
-                        cERROR(1, "FS-Cache support needs CONFIG_CIFS_FSCACHE"
+                        cERROR(1, "FS-Cache support needs CONFIG_CIFS_FSCACHE "
                                  "kernel config option set");
                        goto cifs_parse_mount_err;
 #endif
@@ -1976,7 +1976,7 @@ cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb_vol *volume_info)
                warned_on_ntlm = true;
                cERROR(1, "default security mechanism requested.  The default "
                        "security mechanism will be upgraded from ntlm to "
-                        "ntlmv2 in kernel release 2.6.41");
+                        "ntlmv2 in kernel release 3.1");
        }
        ses->overrideSecFlg = volume_info->secFlg;
diff --git a/fs/dcookies.c b/fs/dcookies.c
index a21cabdbd87b..dda0dc702d1b 100644
--- a/fs/dcookies.c
+++ b/fs/dcookies.c
@@ -178,6 +178,8 @@ SYSCALL_DEFINE(lookup_dcookie)(u64 cookie64, char __user * buf, size_t len)
        /* FIXME: (deleted) ? */
        path = d_path(&dcs->path, kbuf, PAGE_SIZE);
+        mutex_unlock(&dcookie_mutex);
        if (IS_ERR(path)) {
                err = PTR_ERR(path);
                goto out_free;
@@ -194,6 +196,7 @@ SYSCALL_DEFINE(lookup_dcookie)(u64 cookie64, char __user * buf, size_t len)
 out_free:
        kfree(kbuf);
+        return err;
 out:
        mutex_unlock(&dcookie_mutex);
        return err;
diff --git a/fs/exec.c b/fs/exec.c
index ea5f748906a8..97e0d52d72fd 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -1093,6 +1093,7 @@ int flush_old_exec(struct linux_binprm * bprm)
        bprm->mm = NULL;                /* We're using it now */
+        set_fs(USER_DS);
        current->flags &= ~(PF_RANDOMIZE | PF_KTHREAD);
        flush_thread();
        current->personality &= ~bprm->per_clear;
@@ -1357,10 +1358,6 @@ int search_binary_handler(struct linux_binprm *bprm,struct pt_regs *regs)
        if (retval)
                return retval;
-        /* kernel module loader fixup */
-        /* so we don't try to load run modprobe in kernel space. */
-        set_fs(USER_DS);
        retval = audit_bprm(bprm);
        if (retval)
                return retval;
diff --git a/fs/fat/file.c b/fs/fat/file.c
index 7257752b6d5d..7018e1d8902d 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -102,7 +102,7 @@ static int fat_ioctl_set_attributes(struct file *file, u32 __user *user_attr)
                if (attr & ATTR_SYS)
                        inode->i_flags |= S_IMMUTABLE;
                else
-                        inode->i_flags &= S_IMMUTABLE;
+                        inode->i_flags &= ~S_IMMUTABLE;
        }
        fat_save_attrs(inode, attr);
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index cc6ec4b2f0ff..38f84cd48b67 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -921,6 +921,8 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
        if (sb->s_flags & MS_MANDLOCK)
                goto err;
+        sb->s_flags &= ~MS_NOSEC;
        if (!parse_fuse_opt((char *) data, &d, is_bdev))
                goto err;
diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 2792a790e50b..1c1336e7b3b2 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -663,14 +663,19 @@ static void glock_work_func(struct work_struct *work)
                drop_ref = 1;
        }
        spin_lock(&gl->gl_spin);
-        if (test_and_clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
+        if (test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
            gl->gl_state != LM_ST_UNLOCKED &&
            gl->gl_demote_state != LM_ST_EXCLUSIVE) {
                unsigned long holdtime, now = jiffies;
                holdtime = gl->gl_tchange + gl->gl_ops->go_min_hold_time;
                if (time_before(now, holdtime))
                        delay = holdtime - now;
-                set_bit(delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE, &gl->gl_flags);
+                if (!delay) {
+                        clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags);
+                        set_bit(GLF_DEMOTE, &gl->gl_flags);
+                }
        }
        run_queue(gl, 0);
        spin_unlock(&gl->gl_spin);
diff --git a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c
index 278e3fb40b71..583636f745e5 100644
--- a/fs/jfs/jfs_logmgr.c
+++ b/fs/jfs/jfs_logmgr.c
@@ -1123,7 +1123,7 @@ int lmLogOpen(struct super_block *sb)
        bdev = blkdev_get_by_dev(sbi->logdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL,
                                 log);
        if (IS_ERR(bdev)) {
-                rc = -PTR_ERR(bdev);
+                rc = PTR_ERR(bdev);
                goto free;
        }
diff --git a/fs/namei.c b/fs/namei.c
index 1ab641f2e78e..9802345df5e7 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -2579,6 +2579,7 @@ int vfs_rmdir(struct inode *dir, struct dentry *dentry)
        if (error)
                goto out;
+        shrink_dcache_parent(dentry);
        error = dir->i_op->rmdir(dir, dentry);
        if (error)
                goto out;
@@ -2623,6 +2624,10 @@ static long do_rmdir(int dfd, const char __user *pathname)
        error = PTR_ERR(dentry);
        if (IS_ERR(dentry))
                goto exit2;
+        if (!dentry->d_inode) {
+                error = -ENOENT;
+                goto exit3;
+        }
        error = mnt_want_write(nd.path.mnt);
        if (error)
                goto exit3;
@@ -2708,11 +2713,10 @@ static long do_unlinkat(int dfd, const char __user *pathname)
        error = PTR_ERR(dentry);
        if (!IS_ERR(dentry)) {
                /* Why not before? Because we want correct error value */
-                if (nd.last.name[nd.last.len])
-                        goto slashes;
                inode = dentry->d_inode;
-                if (inode)
+                if (nd.last.name[nd.last.len] || !inode)
-                        ihold(inode);
+                        goto slashes;
+                ihold(inode);
                error = mnt_want_write(nd.path.mnt);
                if (error)
                        goto exit2;
@@ -2993,6 +2997,8 @@ static int vfs_rename_dir(struct inode *old_dir, struct dentry *old_dentry,
        if (d_mountpoint(old_dentry) || d_mountpoint(new_dentry))
                goto out;
+        if (target)
+                shrink_dcache_parent(new_dentry);
        error = old_dir->i_op->rename(old_dir, old_dentry, new_dir, new_dentry);
        if (error)
                goto out;
diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c
index 7eafe468a29c..b2e3ff347620 100644
--- a/fs/nilfs2/btree.c
+++ b/fs/nilfs2/btree.c
@@ -1346,6 +1346,11 @@ static void nilfs_btree_shrink(struct nilfs_bmap *btree,
        path[level].bp_bh = NULL;
 }
+static void nilfs_btree_nop(struct nilfs_bmap *btree,
+                            struct nilfs_btree_path *path,
+                            int level, __u64 *keyp, __u64 *ptrp)
+{
+}
 static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree,
                                      struct nilfs_btree_path *path,
@@ -1356,20 +1361,19 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree,
        struct buffer_head *bh;
        struct nilfs_btree_node *node, *parent, *sib;
        __u64 sibptr;
-        int pindex, level, ncmin, ncmax, ncblk, ret;
+        int pindex, dindex, level, ncmin, ncmax, ncblk, ret;
        ret = 0;
        stats->bs_nblocks = 0;
        ncmin = NILFS_BTREE_NODE_NCHILDREN_MIN(nilfs_btree_node_size(btree));
        ncblk = nilfs_btree_nchildren_per_block(btree);
-        for (level = NILFS_BTREE_LEVEL_NODE_MIN;
+        for (level = NILFS_BTREE_LEVEL_NODE_MIN, dindex = path[level].bp_index;
             level < nilfs_btree_height(btree) - 1;
             level++) {
                node = nilfs_btree_get_nonroot_node(path, level);
                path[level].bp_oldreq.bpr_ptr =
-                        nilfs_btree_node_get_ptr(node, path[level].bp_index,
+                        nilfs_btree_node_get_ptr(node, dindex, ncblk);
-                                                 ncblk);
                ret = nilfs_bmap_prepare_end_ptr(btree,
                                                 &path[level].bp_oldreq, dat);
                if (ret < 0)
@@ -1383,6 +1387,7 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree,
                parent = nilfs_btree_get_node(btree, path, level + 1, &ncmax);
                pindex = path[level + 1].bp_index;
+                dindex = pindex;
                if (pindex > 0) {
                        /* left sibling */
@@ -1421,6 +1426,14 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree,
                                path[level].bp_sib_bh = bh;
                                path[level].bp_op = nilfs_btree_concat_right;
                                stats->bs_nblocks++;
+                                /*
+                                 * When merging right sibling node
+                                 * into the current node, pointer to
+                                 * the right sibling node must be
+                                 * terminated instead.  The adjustment
+                                 * below is required for that.
+                                 */
+                                dindex = pindex + 1;
                                /* continue; */
                        }
                } else {
@@ -1431,29 +1444,31 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree,
                            NILFS_BTREE_ROOT_NCHILDREN_MAX) {
                                path[level].bp_op = nilfs_btree_shrink;
                                stats->bs_nblocks += 2;
+                                level++;
+                                path[level].bp_op = nilfs_btree_nop;
+                                goto shrink_root_child;
                        } else {
                                path[level].bp_op = nilfs_btree_do_delete;
                                stats->bs_nblocks++;
+                                goto out;
                        }
-                        goto out;
                }
        }
+        /* child of the root node is deleted */
+        path[level].bp_op = nilfs_btree_do_delete;
+        stats->bs_nblocks++;
+shrink_root_child:
        node = nilfs_btree_get_root(btree);
        path[level].bp_oldreq.bpr_ptr =
-                nilfs_btree_node_get_ptr(node, path[level].bp_index,
+                nilfs_btree_node_get_ptr(node, dindex,
                                         NILFS_BTREE_ROOT_NCHILDREN_MAX);
        ret = nilfs_bmap_prepare_end_ptr(btree, &path[level].bp_oldreq, dat);
        if (ret < 0)
                goto err_out_child_node;
-        /* child of the root node is deleted */
-        path[level].bp_op = nilfs_btree_do_delete;
-        stats->bs_nblocks++;
        /* success */
 out:
        *levelp = level;
diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 141646e88fb5..bb24ab6c282f 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -2573,7 +2573,7 @@ static struct nilfs_sc_info *nilfs_segctor_new(struct super_block *sb,
        sci->sc_watermark = NILFS_SC_DEFAULT_WATERMARK;
        if (nilfs->ns_interval)
-                sci->sc_interval = nilfs->ns_interval;
+                sci->sc_interval = HZ * nilfs->ns_interval;
        if (nilfs->ns_watermark)
                sci->sc_watermark = nilfs->ns_watermark;
        return sci;
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index cdbaf5e97308..56f61027236b 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -1072,7 +1072,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
        sb->s_magic = OCFS2_SUPER_MAGIC;
-        sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
+        sb->s_flags = (sb->s_flags & ~(MS_POSIXACL | MS_NOSEC)) |
                ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
        /* Hard readonly mode only if: bdev_read_only, MS_RDONLY,
diff --git a/fs/partitions/check.c b/fs/partitions/check.c
index f82e762eeca2..d545e97d99c3 100644
--- a/fs/partitions/check.c
+++ b/fs/partitions/check.c
@@ -255,13 +255,7 @@ ssize_t part_discard_alignment_show(struct device *dev,
                                   struct device_attribute *attr, char *buf)
 {
        struct hd_struct *p = dev_to_part(dev);
-        struct gendisk *disk = dev_to_disk(dev);
+        return sprintf(buf, "%u\n", p->discard_alignment);
-        unsigned int alignment = 0;
-        if (disk->queue)
-                alignment = queue_limit_discard_alignment(&disk->queue->limits,
-                                                                p->start_sect);
-        return sprintf(buf, "%u\n", alignment);
 }
 ssize_t part_stat_show(struct device *dev,
@@ -455,6 +449,8 @@ struct hd_struct *add_partition(struct gendisk *disk, int partno,
        p->start_sect = start;
        p->alignment_offset =
                queue_limit_alignment_offset(&disk->queue->limits, start);
+        p->discard_alignment =
+                queue_limit_discard_alignment(&disk->queue->limits, start);
        p->nr_sects = len;
        p->partno = partno;
        p->policy = get_disk_ro(disk);
diff --git a/fs/super.c b/fs/super.c
index c75593953c52..ab3d672db0de 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -822,7 +822,7 @@ struct dentry *mount_bdev(struct file_system_type *fs_type,
        } else {
                char b[BDEVNAME_SIZE];
-                s->s_flags = flags;
+                s->s_flags = flags | MS_NOSEC;
                s->s_mode = mode;
                strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
                sb_set_blocksize(s, block_size(bdev));
diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index 166951e0dcd3..3be645e012c9 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -581,6 +581,7 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
        ubifs_assert(wbuf->size % c->min_io_size == 0);
        ubifs_assert(mutex_is_locked(&wbuf->io_mutex));
        ubifs_assert(!c->ro_media && !c->ro_mount);
+        ubifs_assert(!c->space_fixup);
        if (c->leb_size - wbuf->offs >= c->max_write_size)
                ubifs_assert(!((wbuf->offs + wbuf->size) % c->max_write_size));
@@ -759,6 +760,7 @@ int ubifs_write_node(struct ubifs_info *c, void *buf, int len, int lnum,
        ubifs_assert(lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
        ubifs_assert(offs % c->min_io_size == 0 && offs < c->leb_size);
        ubifs_assert(!c->ro_media && !c->ro_mount);
+        ubifs_assert(!c->space_fixup);
        if (c->ro_error)
                return -EROFS;
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c
index 34b1679e6e3a..cef0460f4c54 100644
--- a/fs/ubifs/journal.c
+++ b/fs/ubifs/journal.c
@@ -669,6 +669,7 @@ out_free:
 out_release:
        release_head(c, BASEHD);
+        kfree(dent);
 out_ro:
        ubifs_ro_mode(c, err);
        if (last_reference)
diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c
index bd644bf587a8..a5422fffbd69 100644
--- a/fs/ubifs/orphan.c
+++ b/fs/ubifs/orphan.c
@@ -674,7 +674,7 @@ static int kill_orphans(struct ubifs_info *c)
                if (IS_ERR(sleb)) {
                        if (PTR_ERR(sleb) == -EUCLEAN)
                                sleb = ubifs_recover_leb(c, lnum, 0,
-                                                         c->sbuf, 0);
+                                                         c->sbuf, -1);
                        if (IS_ERR(sleb)) {
                                err = PTR_ERR(sleb);
                                break;
diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c
index 731d9e2e7b50..783d8e0beb76 100644
--- a/fs/ubifs/recovery.c
+++ b/fs/ubifs/recovery.c
@@ -564,19 +564,15 @@ static int fix_unclean_leb(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
 }
 /**
- * drop_last_node - drop the last node or group of nodes.
+ * drop_last_group - drop the last group of nodes.
 * @sleb: scanned LEB information
 * @offs: offset of dropped nodes is returned here
- * @grouped: non-zero if whole group of nodes have to be dropped
 *
 * This is a helper function for 'ubifs_recover_leb()' which drops the last
- * node of the scanned LEB or the last group of nodes if @grouped is not zero.
+ * group of nodes of the scanned LEB.
- * This function returns %1 if a node was dropped and %0 otherwise.
 */
-static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
+static void drop_last_group(struct ubifs_scan_leb *sleb, int *offs)
 {
-        int dropped = 0;
        while (!list_empty(&sleb->nodes)) {
                struct ubifs_scan_node *snod;
                struct ubifs_ch *ch;
@@ -585,17 +581,40 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
                                  list);
                ch = snod->node;
                if (ch->group_type != UBIFS_IN_NODE_GROUP)
-                        return dropped;
+                        break;
-                dbg_rcvry("dropping node at %d:%d", sleb->lnum, snod->offs);
+                dbg_rcvry("dropping grouped node at %d:%d",
+                          sleb->lnum, snod->offs);
+                *offs = snod->offs;
+                list_del(&snod->list);
+                kfree(snod);
+                sleb->nodes_cnt -= 1;
+        }
+}
+/**
+ * drop_last_node - drop the last node.
+ * @sleb: scanned LEB information
+ * @offs: offset of dropped nodes is returned here
+ * @grouped: non-zero if whole group of nodes have to be dropped
+ *
+ * This is a helper function for 'ubifs_recover_leb()' which drops the last
+ * node of the scanned LEB.
+ */
+static void drop_last_node(struct ubifs_scan_leb *sleb, int *offs)
+{
+        struct ubifs_scan_node *snod;
+        if (!list_empty(&sleb->nodes)) {
+                snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node,
+                                  list);
+                dbg_rcvry("dropping last node at %d:%d", sleb->lnum, snod->offs);
                *offs = snod->offs;
                list_del(&snod->list);
                kfree(snod);
                sleb->nodes_cnt -= 1;
-                dropped = 1;
-                if (!grouped)
-                        break;
        }
-        return dropped;
 }
 /**
@@ -604,7 +623,8 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
 * @lnum: LEB number
 * @offs: offset
 * @sbuf: LEB-sized buffer to use
- * @grouped: nodes may be grouped for recovery
+ * @jhead: journal head number this LEB belongs to (%-1 if the LEB does not
+ *         belong to any journal head)
 *
 * This function does a scan of a LEB, but caters for errors that might have
 * been caused by the unclean unmount from which we are attempting to recover.
@@ -612,13 +632,14 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped)
 * found, and a negative error code in case of failure.
 */
 struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
-                                         int offs, void *sbuf, int grouped)
+                                         int offs, void *sbuf, int jhead)
 {
        int ret = 0, err, len = c->leb_size - offs, start = offs, min_io_unit;
+        int grouped = jhead == -1 ? 0 : c->jheads[jhead].grouped;
        struct ubifs_scan_leb *sleb;
        void *buf = sbuf + offs;
-        dbg_rcvry("%d:%d", lnum, offs);
+        dbg_rcvry("%d:%d, jhead %d, grouped %d", lnum, offs, jhead, grouped);
        sleb = ubifs_start_scan(c, lnum, offs, sbuf);
        if (IS_ERR(sleb))
@@ -635,7 +656,7 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
                 * Scan quietly until there is an error from which we cannot
                 * recover
                 */
-                ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 0);
+                ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 1);
                if (ret == SCANNED_A_NODE) {
                        /* A valid node, and not a padding node */
                        struct ubifs_ch *ch = buf;
@@ -695,59 +716,62 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
                 * If nodes are grouped, always drop the incomplete group at
                 * the end.
                 */
-                drop_last_node(sleb, &offs, 1);
+                drop_last_group(sleb, &offs);
-        /*
+        if (jhead == GCHD) {
-         * While we are in the middle of the same min. I/O unit keep dropping
+                /*
-         * nodes. So basically, what we want is to make sure that the last min.
+                 * If this LEB belongs to the GC head then while we are in the
-         * I/O unit where we saw the corruption is dropped completely with all
+                 * middle of the same min. I/O unit keep dropping nodes. So
-         * the uncorrupted node which may possibly sit there.
+                 * basically, what we want is to make sure that the last min.
-         *
+                 * I/O unit where we saw the corruption is dropped completely
-         * In other words, let's name the min. I/O unit where the corruption
+                 * with all the uncorrupted nodes which may possibly sit there.
-         * starts B, and the previous min. I/O unit A. The below code tries to
+                 *
-         * deal with a situation when half of B contains valid nodes or the end
+                 * In other words, let's name the min. I/O unit where the
-         * of a valid node, and the second half of B contains corrupted data or
+                 * corruption starts B, and the previous min. I/O unit A. The
-         * garbage. This means that UBIFS had been writing to B just before the
+                 * below code tries to deal with a situation when half of B
-         * power cut happened. I do not know how realistic is this scenario
+                 * contains valid nodes or the end of a valid node, and the
-         * that half of the min. I/O unit had been written successfully and the
+                 * second half of B contains corrupted data or garbage. This
-         * other half not, but this is possible in our 'failure mode emulation'
+                 * means that UBIFS had been writing to B just before the power
-         * infrastructure at least.
+                 * cut happened. I do not know how realistic is this scenario
-         *
+                 * that half of the min. I/O unit had been written successfully
-         * So what is the problem, why we need to drop those nodes? Whey can't
+                 * and the other half not, but this is possible in our 'failure
-         * we just clean-up the second half of B by putting a padding node
+                 * mode emulation' infrastructure at least.
-         * there? We can, and this works fine with one exception which was
+                 *
-         * reproduced with power cut emulation testing and happens extremely
+                 * So what is the problem, why we need to drop those nodes? Why
-         * rarely. The description follows, but it is worth noting that that is
+                 * can't we just clean-up the second half of B by putting a
-         * only about the GC head, so we could do this trick only if the bud
+                 * padding node there? We can, and this works fine with one
-         * belongs to the GC head, but it does not seem to be worth an
+                 * exception which was reproduced with power cut emulation
-         * additional "if" statement.
+                 * testing and happens extremely rarely.
-         *
+                 *
-         * So, imagine the file-system is full, we run GC which is moving valid
+                 * Imagine the file-system is full, we run GC which starts
-         * nodes from LEB X to LEB Y (obviously, LEB Y is the current GC head
+                 * moving valid nodes from LEB X to LEB Y (obviously, LEB Y is
-         * LEB). The @c->gc_lnum is -1, which means that GC will retain LEB X
+                 * the current GC head LEB). The @c->gc_lnum is -1, which means
-         * and will try to continue. Imagine that LEB X is currently the
+                 * that GC will retain LEB X and will try to continue. Imagine
-         * dirtiest LEB, and the amount of used space in LEB Y is exactly the
+                 * that LEB X is currently the dirtiest LEB, and the amount of
-         * same as amount of free space in LEB X.
+                 * used space in LEB Y is exactly the same as amount of free
-         *
+                 * space in LEB X.
-         * And a power cut happens when nodes are moved from LEB X to LEB Y. We
+                 *
-         * are here trying to recover LEB Y which is the GC head LEB. We find
+                 * And a power cut happens when nodes are moved from LEB X to
-         * the min. I/O unit B as described above. Then we clean-up LEB Y by
+                 * LEB Y. We are here trying to recover LEB Y which is the GC
-         * padding min. I/O unit. And later 'ubifs_rcvry_gc_commit()' function
+                 * head LEB. We find the min. I/O unit B as described above.
-         * fails, because it cannot find a dirty LEB which could be GC'd into
+                 * Then we clean-up LEB Y by padding min. I/O unit. And later
-         * LEB Y! Even LEB X does not match because the amount of valid nodes
+                 * 'ubifs_rcvry_gc_commit()' function fails, because it cannot
-         * there does not fit the free space in LEB Y any more! And this is
+                 * find a dirty LEB which could be GC'd into LEB Y! Even LEB X
-         * because of the padding node which we added to LEB Y. The
+                 * does not match because the amount of valid nodes there does
-         * user-visible effect of this which I once observed and analysed is
+                 * not fit the free space in LEB Y any more! And this is
-         * that we cannot mount the file-system with -ENOSPC error.
+                 * because of the padding node which we added to LEB Y. The
-         *
+                 * user-visible effect of this which I once observed and
-         * So obviously, to make sure that situation does not happen we should
+                 * analysed is that we cannot mount the file-system with
-         * free min. I/O unit B in LEB Y completely and the last used min. I/O
+                 * -ENOSPC error.
-         * unit in LEB Y should be A. This is basically what the below code
+                 *
-         * tries to do.
+                 * So obviously, to make sure that situation does not happen we
-         */
+                 * should free min. I/O unit B in LEB Y completely and the last
-        while (min_io_unit == round_down(offs, c->min_io_size) &&
+                 * used min. I/O unit in LEB Y should be A. This is basically
-               min_io_unit != offs &&
+                 * what the below code tries to do.
-               drop_last_node(sleb, &offs, grouped));
+                 */
+                while (offs > min_io_unit)
+                        drop_last_node(sleb, &offs);
+        }
        buf = sbuf + offs;
        len = c->leb_size - offs;
@@ -881,7 +905,7 @@ struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum,
                }
                ubifs_scan_destroy(sleb);
        }
-        return ubifs_recover_leb(c, lnum, offs, sbuf, 0);
+        return ubifs_recover_leb(c, lnum, offs, sbuf, -1);
 }
 /**
diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c
index 6617280d1679..5e97161ce4d3 100644
--- a/fs/ubifs/replay.c
+++ b/fs/ubifs/replay.c
@@ -557,8 +557,7 @@ static int replay_bud(struct ubifs_info *c, struct bud_entry *b)
                 * these LEBs could possibly be written to at the power cut
                 * time.
                 */
-                sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf,
+                sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf, b->bud->jhead);
-                                         b->bud->jhead != GCHD);
        else
                sleb = ubifs_scan(c, lnum, offs, c->sbuf, 0);
        if (IS_ERR(sleb))
diff --git a/fs/ubifs/shrinker.c b/fs/ubifs/shrinker.c
index ca953a945029..9e1d05666fed 100644
--- a/fs/ubifs/shrinker.c
+++ b/fs/ubifs/shrinker.c
@@ -284,7 +284,11 @@ int ubifs_shrinker(struct shrinker *shrink, struct shrink_control *sc)
        long clean_zn_cnt = atomic_long_read(&ubifs_clean_zn_cnt);
        if (nr == 0)
-                return clean_zn_cnt;
+                /*
+                 * Due to the way UBIFS updates the clean znode counter it may
+                 * temporarily be negative.
+                 */
+                return clean_zn_cnt >= 0 ? clean_zn_cnt : 1;
        if (!clean_zn_cnt) {
                /*
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 1ab0d22e4c94..b5aeb5a8ebed 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -811,15 +811,18 @@ static int alloc_wbufs(struct ubifs_info *c)
                c->jheads[i].wbuf.sync_callback = &bud_wbuf_callback;
                c->jheads[i].wbuf.jhead = i;
+                c->jheads[i].grouped = 1;
        }
        c->jheads[BASEHD].wbuf.dtype = UBI_SHORTTERM;
        /*
         * Garbage Collector head likely contains long-term data and
-         * does not need to be synchronized by timer.
+         * does not need to be synchronized by timer. Also GC head nodes are
+         * not grouped.
         */
        c->jheads[GCHD].wbuf.dtype = UBI_LONGTERM;
        c->jheads[GCHD].wbuf.no_timer = 1;
+        c->jheads[GCHD].grouped = 0;
        return 0;
 }
@@ -1284,12 +1287,25 @@ static int mount_ubifs(struct ubifs_info *c)
        if ((c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY)) != 0) {
                ubifs_msg("recovery needed");
                c->need_recovery = 1;
-                if (!c->ro_mount) {
+        }
-                        err = ubifs_recover_inl_heads(c, c->sbuf);
-                        if (err)
+        if (c->need_recovery && !c->ro_mount) {
-                                goto out_master;
+                err = ubifs_recover_inl_heads(c, c->sbuf);
-                }
+                if (err)
-        } else if (!c->ro_mount) {
+                        goto out_master;
+        }
+        err = ubifs_lpt_init(c, 1, !c->ro_mount);
+        if (err)
+                goto out_master;
+        if (!c->ro_mount && c->space_fixup) {
+                err = ubifs_fixup_free_space(c);
+                if (err)
+                        goto out_master;
+        }
+        if (!c->ro_mount) {
                /*
                 * Set the "dirty" flag so that if we reboot uncleanly we
                 * will notice this immediately on the next mount.
@@ -1297,13 +1313,9 @@ static int mount_ubifs(struct ubifs_info *c)
                c->mst_node->flags |= cpu_to_le32(UBIFS_MST_DIRTY);
                err = ubifs_write_master(c);
                if (err)
-                        goto out_master;
+                        goto out_lpt;
        }
-        err = ubifs_lpt_init(c, 1, !c->ro_mount);
-        if (err)
-                goto out_lpt;
        err = dbg_check_idx_size(c, c->bi.old_idx_sz);
        if (err)
                goto out_lpt;
@@ -1396,12 +1408,6 @@ static int mount_ubifs(struct ubifs_info *c)
        } else
                ubifs_assert(c->lst.taken_empty_lebs > 0);
-        if (!c->ro_mount && c->space_fixup) {
-                err = ubifs_fixup_free_space(c);
-                if (err)
-                        goto out_infos;
-        }
        err = dbg_check_filesystem(c);
        if (err)
                goto out_infos;
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index 8119b1fd8d94..91b4213dde84 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -2876,12 +2876,13 @@ static void tnc_destroy_cnext(struct ubifs_info *c)
 */
 void ubifs_tnc_close(struct ubifs_info *c)
 {
-        long clean_freed;
        tnc_destroy_cnext(c);
        if (c->zroot.znode) {
-                clean_freed = ubifs_destroy_tnc_subtree(c->zroot.znode);
+                long n;
-                atomic_long_sub(clean_freed, &ubifs_clean_zn_cnt);
+                ubifs_destroy_tnc_subtree(c->zroot.znode);
+                n = atomic_long_read(&c->clean_zn_cnt);
+                atomic_long_sub(n, &ubifs_clean_zn_cnt);
        }
        kfree(c->gap_lebs);
        kfree(c->ilebs);
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index a70d7b4ffb25..f79983d6f860 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -722,12 +722,14 @@ struct ubifs_bud {
 * struct ubifs_jhead - journal head.
 * @wbuf: head's write-buffer
 * @buds_list: list of bud LEBs belonging to this journal head
+ * @grouped: non-zero if UBIFS groups nodes when writing to this journal head
 *
 * Note, the @buds list is protected by the @c->buds_lock.
 */
 struct ubifs_jhead {
        struct ubifs_wbuf wbuf;
        struct list_head buds_list;
+        unsigned int grouped:1;
 };
 /**
@@ -1742,7 +1744,7 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum);
 int ubifs_recover_master_node(struct ubifs_info *c);
 int ubifs_write_rcvrd_mst_node(struct ubifs_info *c);
 struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
-                                         int offs, void *sbuf, int grouped);
+                                         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);