14 files changed, 247 insertions, 142 deletions
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index 0b278b117cbe..14fce27b4780 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -22,8 +22,9 @@
 #include "disk-io.h"
 #include "transaction.h"
-#define BTRFS_DELAYED_WRITEBACK         400
+#define BTRFS_DELAYED_WRITEBACK         512
-#define BTRFS_DELAYED_BACKGROUND        100
+#define BTRFS_DELAYED_BACKGROUND        128
+#define BTRFS_DELAYED_BATCH             16
 static struct kmem_cache *delayed_node_cache;
@@ -494,6 +495,15 @@ static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node,
                                        BTRFS_DELAYED_DELETION_ITEM);
 }
+static void finish_one_item(struct btrfs_delayed_root *delayed_root)
+{
+        int seq = atomic_inc_return(&delayed_root->items_seq);
+        if ((atomic_dec_return(&delayed_root->items) <
+            BTRFS_DELAYED_BACKGROUND || seq % BTRFS_DELAYED_BATCH == 0) &&
+            waitqueue_active(&delayed_root->wait))
+                wake_up(&delayed_root->wait);
+}
 static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item)
 {
        struct rb_root *root;
@@ -512,10 +522,8 @@ static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item)
        rb_erase(&delayed_item->rb_node, root);
        delayed_item->delayed_node->count--;
-        if (atomic_dec_return(&delayed_root->items) <
-            BTRFS_DELAYED_BACKGROUND &&
+        finish_one_item(delayed_root);
-            waitqueue_active(&delayed_root->wait))
-                wake_up(&delayed_root->wait);
 }
 static void btrfs_release_delayed_item(struct btrfs_delayed_item *item)
@@ -1056,10 +1064,7 @@ static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node)
                delayed_node->count--;
                delayed_root = delayed_node->root->fs_info->delayed_root;
-                if (atomic_dec_return(&delayed_root->items) <
+                finish_one_item(delayed_root);
-                    BTRFS_DELAYED_BACKGROUND &&
-                    waitqueue_active(&delayed_root->wait))
-                        wake_up(&delayed_root->wait);
        }
 }
@@ -1304,35 +1309,44 @@ void btrfs_remove_delayed_node(struct inode *inode)
        btrfs_release_delayed_node(delayed_node);
 }
-struct btrfs_async_delayed_node {
+struct btrfs_async_delayed_work {
-        struct btrfs_root *root;
+        struct btrfs_delayed_root *delayed_root;
-        struct btrfs_delayed_node *delayed_node;
+        int nr;
        struct btrfs_work work;
 };
-static void btrfs_async_run_delayed_node_done(struct btrfs_work *work)
+static void btrfs_async_run_delayed_root(struct btrfs_work *work)
 {
-        struct btrfs_async_delayed_node *async_node;
+        struct btrfs_async_delayed_work *async_work;
+        struct btrfs_delayed_root *delayed_root;
        struct btrfs_trans_handle *trans;
        struct btrfs_path *path;
        struct btrfs_delayed_node *delayed_node = NULL;
        struct btrfs_root *root;
        struct btrfs_block_rsv *block_rsv;
-        int need_requeue = 0;
+        int total_done = 0;
-        async_node = container_of(work, struct btrfs_async_delayed_node, work);
+        async_work = container_of(work, struct btrfs_async_delayed_work, work);
+        delayed_root = async_work->delayed_root;
        path = btrfs_alloc_path();
        if (!path)
                goto out;
-        path->leave_spinning = 1;
-        delayed_node = async_node->delayed_node;
+again:
+        if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND / 2)
+                goto free_path;
+        delayed_node = btrfs_first_prepared_delayed_node(delayed_root);
+        if (!delayed_node)
+                goto free_path;
+        path->leave_spinning = 1;
        root = delayed_node->root;
        trans = btrfs_join_transaction(root);
        if (IS_ERR(trans))
-                goto free_path;
+                goto release_path;
        block_rsv = trans->block_rsv;
        trans->block_rsv = &root->fs_info->delayed_block_rsv;
@@ -1363,57 +1377,47 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work)
         * Task1 will sleep until the transaction is commited.
         */
        mutex_lock(&delayed_node->mutex);
-        if (delayed_node->count)
+        btrfs_dequeue_delayed_node(root->fs_info->delayed_root, delayed_node);
-                need_requeue = 1;
-        else
-                btrfs_dequeue_delayed_node(root->fs_info->delayed_root,
-                                           delayed_node);
        mutex_unlock(&delayed_node->mutex);
        trans->block_rsv = block_rsv;
        btrfs_end_transaction_dmeta(trans, root);
        btrfs_btree_balance_dirty_nodelay(root);
+release_path:
+        btrfs_release_path(path);
+        total_done++;
+        btrfs_release_prepared_delayed_node(delayed_node);
+        if (async_work->nr == 0 || total_done < async_work->nr)
+                goto again;
 free_path:
        btrfs_free_path(path);
 out:
-        if (need_requeue)
+        wake_up(&delayed_root->wait);
-                btrfs_requeue_work(&async_node->work);
+        kfree(async_work);
-        else {
-                btrfs_release_prepared_delayed_node(delayed_node);
-                kfree(async_node);
-        }
 }
 static int btrfs_wq_run_delayed_node(struct btrfs_delayed_root *delayed_root,
-                                     struct btrfs_root *root, int all)
+                                     struct btrfs_root *root, int nr)
 {
-        struct btrfs_async_delayed_node *async_node;
+        struct btrfs_async_delayed_work *async_work;
-        struct btrfs_delayed_node *curr;
-        int count = 0;
-again:
+        if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND)
-        curr = btrfs_first_prepared_delayed_node(delayed_root);
-        if (!curr)
                return 0;
-        async_node = kmalloc(sizeof(*async_node), GFP_NOFS);
+        async_work = kmalloc(sizeof(*async_work), GFP_NOFS);
-        if (!async_node) {
+        if (!async_work)
-                btrfs_release_prepared_delayed_node(curr);
                return -ENOMEM;
-        }
-        async_node->root = root;
-        async_node->delayed_node = curr;
-        async_node->work.func = btrfs_async_run_delayed_node_done;
-        async_node->work.flags = 0;
-        btrfs_queue_worker(&root->fs_info->delayed_workers, &async_node->work);
+        async_work->delayed_root = delayed_root;
-        count++;
+        async_work->work.func = btrfs_async_run_delayed_root;
+        async_work->work.flags = 0;
-        if (all || count < 4)
+        async_work->nr = nr;
-                goto again;
+        btrfs_queue_worker(&root->fs_info->delayed_workers, &async_work->work);
        return 0;
 }
@@ -1424,30 +1428,55 @@ void btrfs_assert_delayed_root_empty(struct btrfs_root *root)
        WARN_ON(btrfs_first_delayed_node(delayed_root));
 }
+static int refs_newer(struct btrfs_delayed_root *delayed_root,
+                      int seq, int count)
+{
+        int val = atomic_read(&delayed_root->items_seq);
+        if (val < seq || val >= seq + count)
+                return 1;
+        return 0;
+}
 void btrfs_balance_delayed_items(struct btrfs_root *root)
 {
        struct btrfs_delayed_root *delayed_root;
+        int seq;
        delayed_root = btrfs_get_delayed_root(root);
        if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND)
                return;
+        seq = atomic_read(&delayed_root->items_seq);
        if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) {
                int ret;
-                ret = btrfs_wq_run_delayed_node(delayed_root, root, 1);
+                DEFINE_WAIT(__wait);
+                ret = btrfs_wq_run_delayed_node(delayed_root, root, 0);
                if (ret)
                        return;
-                wait_event_interruptible_timeout(
+                while (1) {
-                                delayed_root->wait,
+                        prepare_to_wait(&delayed_root->wait, &__wait,
-                                (atomic_read(&delayed_root->items) <
+                                        TASK_INTERRUPTIBLE);
-                                 BTRFS_DELAYED_BACKGROUND),
-                                HZ);
+                        if (refs_newer(delayed_root, seq,
-                return;
+                                       BTRFS_DELAYED_BATCH) ||
+                            atomic_read(&delayed_root->items) <
+                            BTRFS_DELAYED_BACKGROUND) {
+                                break;
+                        }
+                        if (!signal_pending(current))
+                                schedule();
+                        else
+                                break;
+                }
+                finish_wait(&delayed_root->wait, &__wait);
        }
-        btrfs_wq_run_delayed_node(delayed_root, root, 0);
+        btrfs_wq_run_delayed_node(delayed_root, root, BTRFS_DELAYED_BATCH);
 }
 /* Will return 0 or -ENOMEM */
diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h
index 78b6ad0fc669..1d5c5f7abe3e 100644
--- a/fs/btrfs/delayed-inode.h
+++ b/fs/btrfs/delayed-inode.h
@@ -43,6 +43,7 @@ struct btrfs_delayed_root {
         */
        struct list_head prepare_list;
        atomic_t items;         /* for delayed items */
+        atomic_t items_seq;     /* for delayed items */
        int nodes;              /* for delayed nodes */
        wait_queue_head_t wait;
 };
@@ -86,6 +87,7 @@ static inline void btrfs_init_delayed_root(
                                struct btrfs_delayed_root *delayed_root)
 {
        atomic_set(&delayed_root->items, 0);
+        atomic_set(&delayed_root->items_seq, 0);
        delayed_root->nodes = 0;
        spin_lock_init(&delayed_root->lock);
        init_waitqueue_head(&delayed_root->wait);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 02369a3c162e..7d84651e850b 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -62,7 +62,7 @@ static void btrfs_destroy_ordered_operations(struct btrfs_transaction *t,
 static void btrfs_destroy_ordered_extents(struct btrfs_root *root);
 static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
                                      struct btrfs_root *root);
-static void btrfs_destroy_pending_snapshots(struct btrfs_transaction *t);
+static void btrfs_evict_pending_snapshots(struct btrfs_transaction *t);
 static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root);
 static int btrfs_destroy_marked_extents(struct btrfs_root *root,
                                        struct extent_io_tree *dirty_pages,
@@ -3687,7 +3687,7 @@ int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
        return ret;
 }
-static void btrfs_destroy_pending_snapshots(struct btrfs_transaction *t)
+static void btrfs_evict_pending_snapshots(struct btrfs_transaction *t)
 {
        struct btrfs_pending_snapshot *snapshot;
        struct list_head splice;
@@ -3700,10 +3700,8 @@ static void btrfs_destroy_pending_snapshots(struct btrfs_transaction *t)
                snapshot = list_entry(splice.next,
                                      struct btrfs_pending_snapshot,
                                      list);
+                snapshot->error = -ECANCELED;
                list_del_init(&snapshot->list);
-                kfree(snapshot);
        }
 }
@@ -3840,6 +3838,8 @@ void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
        cur_trans->blocked = 1;
        wake_up(&root->fs_info->transaction_blocked_wait);
+        btrfs_evict_pending_snapshots(cur_trans);
        cur_trans->blocked = 0;
        wake_up(&root->fs_info->transaction_wait);
@@ -3849,8 +3849,6 @@ void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
        btrfs_destroy_delayed_inodes(root);
        btrfs_assert_delayed_root_empty(root);
-        btrfs_destroy_pending_snapshots(cur_trans);
        btrfs_destroy_marked_extents(root, &cur_trans->dirty_pages,
                                     EXTENT_DIRTY);
        btrfs_destroy_pinned_extent(root,
@@ -3894,6 +3892,8 @@ int btrfs_cleanup_transaction(struct btrfs_root *root)
                if (waitqueue_active(&root->fs_info->transaction_blocked_wait))
                        wake_up(&root->fs_info->transaction_blocked_wait);
+                btrfs_evict_pending_snapshots(t);
                t->blocked = 0;
                smp_mb();
                if (waitqueue_active(&root->fs_info->transaction_wait))
@@ -3907,8 +3907,6 @@ int btrfs_cleanup_transaction(struct btrfs_root *root)
                btrfs_destroy_delayed_inodes(root);
                btrfs_assert_delayed_root_empty(root);
-                btrfs_destroy_pending_snapshots(t);
                btrfs_destroy_delalloc_inodes(root);
                spin_lock(&root->fs_info->trans_lock);
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 3e074dab2d57..9ac2eca681eb 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -1467,8 +1467,11 @@ int lookup_inline_extent_backref(struct btrfs_trans_handle *trans,
        if (ret && !insert) {
                err = -ENOENT;
                goto out;
+        } else if (ret) {
+                err = -EIO;
+                WARN_ON(1);
+                goto out;
        }
-        BUG_ON(ret); /* Corruption */
        leaf = path->nodes[0];
        item_size = btrfs_item_size_nr(leaf, path->slots[0]);
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index af1d0605a5c1..5b4ea5f55b8f 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -591,6 +591,7 @@ void btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
                }
                compressed = test_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
                clear_bit(EXTENT_FLAG_PINNED, &em->flags);
+                clear_bit(EXTENT_FLAG_LOGGING, &flags);
                remove_extent_mapping(em_tree, em);
                if (no_splits)
                        goto next;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index c226daefd65d..ca1b767d51f7 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -2312,6 +2312,7 @@ again:
        key.type = BTRFS_EXTENT_DATA_KEY;
        key.offset = start;
+        path->leave_spinning = 1;
        if (merge) {
                struct btrfs_file_extent_item *fi;
                u64 extent_len;
@@ -2368,6 +2369,7 @@ again:
        btrfs_mark_buffer_dirty(leaf);
        inode_add_bytes(inode, len);
+        btrfs_release_path(path);
        ret = btrfs_inc_extent_ref(trans, root, new->bytenr,
                        new->disk_len, 0,
@@ -2381,6 +2383,7 @@ again:
        ret = 1;
 out_free_path:
        btrfs_release_path(path);
+        path->leave_spinning = 0;
        btrfs_end_transaction(trans, root);
 out_unlock:
        unlock_extent_cached(&BTRFS_I(inode)->io_tree, lock_start, lock_end,
@@ -8502,6 +8505,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
        struct btrfs_key ins;
        u64 cur_offset = start;
        u64 i_size;
+        u64 cur_bytes;
        int ret = 0;
        bool own_trans = true;
@@ -8516,8 +8520,9 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
                        }
                }
-                ret = btrfs_reserve_extent(trans, root,
+                cur_bytes = min(num_bytes, 256ULL * 1024 * 1024);
-                                           min(num_bytes, 256ULL * 1024 * 1024),
+                cur_bytes = max(cur_bytes, min_size);
+                ret = btrfs_reserve_extent(trans, root, cur_bytes,
                                           min_size, 0, *alloc_hint, &ins, 1);
                if (ret) {
                        if (own_trans)
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index c83086fdda05..2c02310ff2d9 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -527,6 +527,8 @@ fail:
        if (async_transid) {
                *async_transid = trans->transid;
                err = btrfs_commit_transaction_async(trans, root, 1);
+                if (err)
+                        err = btrfs_commit_transaction(trans, root);
        } else {
                err = btrfs_commit_transaction(trans, root);
        }
@@ -592,16 +594,14 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
                *async_transid = trans->transid;
                ret = btrfs_commit_transaction_async(trans,
                                     root->fs_info->extent_root, 1);
+                if (ret)
+                        ret = btrfs_commit_transaction(trans, root);
        } else {
                ret = btrfs_commit_transaction(trans,
                                               root->fs_info->extent_root);
        }
-        if (ret) {
+        if (ret)
-                /* cleanup_transaction has freed this for us */
-                if (trans->aborted)
-                        pending_snapshot = NULL;
                goto fail;
-        }
        ret = pending_snapshot->error;
        if (ret)
@@ -2245,13 +2245,6 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
        if (ret)
                return ret;
-        if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
-                        1)) {
-                pr_info("btrfs: dev add/delete/balance/replace/resize operation in progress\n");
-                mnt_drop_write_file(file);
-                return -EINVAL;
-        }
        if (btrfs_root_readonly(root)) {
                ret = -EROFS;
                goto out;
@@ -2306,7 +2299,6 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
                ret = -EINVAL;
        }
 out:
-        atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0);
        mnt_drop_write_file(file);
        return ret;
 }
diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h
index ca52681e5f40..b81e0e9a4894 100644
--- a/fs/btrfs/locking.h
+++ b/fs/btrfs/locking.h
@@ -26,7 +26,6 @@
 void btrfs_tree_lock(struct extent_buffer *eb);
 void btrfs_tree_unlock(struct extent_buffer *eb);
-int btrfs_try_spin_lock(struct extent_buffer *eb);
 void btrfs_tree_read_lock(struct extent_buffer *eb);
 void btrfs_tree_read_unlock(struct extent_buffer *eb);
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index aee4b1cc3d98..5471e47d6559 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -1525,21 +1525,23 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
                if ((qg->lim_flags & BTRFS_QGROUP_LIMIT_MAX_RFER) &&
                    qg->reserved + qg->rfer + num_bytes >
-                    qg->max_rfer)
+                    qg->max_rfer) {
                        ret = -EDQUOT;
+                        goto out;
+                }
                if ((qg->lim_flags & BTRFS_QGROUP_LIMIT_MAX_EXCL) &&
                    qg->reserved + qg->excl + num_bytes >
-                    qg->max_excl)
+                    qg->max_excl) {
                        ret = -EDQUOT;
+                        goto out;
+                }
                list_for_each_entry(glist, &qg->groups, next_group) {
                        ulist_add(ulist, glist->group->qgroupid,
                                  (uintptr_t)glist->group, GFP_ATOMIC);
                }
        }
-        if (ret)
-                goto out;
        /*
         * no limits exceeded, now record the reservation into all qgroups
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 50695dc5e2ab..b67171e6d688 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -1269,6 +1269,8 @@ static int __update_reloc_root(struct btrfs_root *root, int del)
        }
        spin_unlock(&rc->reloc_root_tree.lock);
+        if (!node)
+                return 0;
        BUG_ON((struct btrfs_root *)node->data != root);
        if (!del) {
@@ -2238,13 +2240,28 @@ again:
 }
 static noinline_for_stack
+void free_reloc_roots(struct list_head *list)
+{
+        struct btrfs_root *reloc_root;
+        while (!list_empty(list)) {
+                reloc_root = list_entry(list->next, struct btrfs_root,
+                                        root_list);
+                __update_reloc_root(reloc_root, 1);
+                free_extent_buffer(reloc_root->node);
+                free_extent_buffer(reloc_root->commit_root);
+                kfree(reloc_root);
+        }
+}
+static noinline_for_stack
 int merge_reloc_roots(struct reloc_control *rc)
 {
        struct btrfs_root *root;
        struct btrfs_root *reloc_root;
        LIST_HEAD(reloc_roots);
        int found = 0;
-        int ret;
+        int ret = 0;
 again:
        root = rc->extent_root;
@@ -2270,20 +2287,33 @@ again:
                        BUG_ON(root->reloc_root != reloc_root);
                        ret = merge_reloc_root(rc, root);
-                        BUG_ON(ret);
+                        if (ret)
+                                goto out;
                } else {
                        list_del_init(&reloc_root->root_list);
                }
                ret = btrfs_drop_snapshot(reloc_root, rc->block_rsv, 0, 1);
-                BUG_ON(ret < 0);
+                if (ret < 0) {
+                        if (list_empty(&reloc_root->root_list))
+                                list_add_tail(&reloc_root->root_list,
+                                              &reloc_roots);
+                        goto out;
+                }
        }
        if (found) {
                found = 0;
                goto again;
        }
+out:
+        if (ret) {
+                btrfs_std_error(root->fs_info, ret);
+                if (!list_empty(&reloc_roots))
+                        free_reloc_roots(&reloc_roots);
+        }
        BUG_ON(!RB_EMPTY_ROOT(&rc->reloc_root_tree.rb_root));
-        return 0;
+        return ret;
 }
 static void free_block_list(struct rb_root *blocks)
@@ -2818,8 +2848,10 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
        int err = 0;
        path = btrfs_alloc_path();
-        if (!path)
+        if (!path) {
-                return -ENOMEM;
+                err = -ENOMEM;
+                goto out_path;
+        }
        rb_node = rb_first(blocks);
        while (rb_node) {
@@ -2858,10 +2890,11 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
                rb_node = rb_next(rb_node);
        }
 out:
-        free_block_list(blocks);
        err = finish_pending_nodes(trans, rc, path, err);
        btrfs_free_path(path);
+out_path:
+        free_block_list(blocks);
        return err;
 }
@@ -3698,7 +3731,15 @@ int prepare_to_relocate(struct reloc_control *rc)
        set_reloc_control(rc);
        trans = btrfs_join_transaction(rc->extent_root);
-        BUG_ON(IS_ERR(trans));
+        if (IS_ERR(trans)) {
+                unset_reloc_control(rc);
+                /*
+                 * extent tree is not a ref_cow tree and has no reloc_root to
+                 * cleanup.  And callers are responsible to free the above
+                 * block rsv.
+                 */
+                return PTR_ERR(trans);
+        }
        btrfs_commit_transaction(trans, rc->extent_root);
        return 0;
 }
@@ -3730,7 +3771,11 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
        while (1) {
                progress++;
                trans = btrfs_start_transaction(rc->extent_root, 0);
-                BUG_ON(IS_ERR(trans));
+                if (IS_ERR(trans)) {
+                        err = PTR_ERR(trans);
+                        trans = NULL;
+                        break;
+                }
 restart:
                if (update_backref_cache(trans, &rc->backref_cache)) {
                        btrfs_end_transaction(trans, rc->extent_root);
@@ -4264,14 +4309,9 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 out_free:
        kfree(rc);
 out:
-        while (!list_empty(&reloc_roots)) {
+        if (!list_empty(&reloc_roots))
-                reloc_root = list_entry(reloc_roots.next,
+                free_reloc_roots(&reloc_roots);
-                                        struct btrfs_root, root_list);
-                list_del(&reloc_root->root_list);
-                free_extent_buffer(reloc_root->node);
-                free_extent_buffer(reloc_root->commit_root);
-                kfree(reloc_root);
-        }
        btrfs_free_path(path);
        if (err == 0) {
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 68a29a1ea068..f6b88595f858 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -1558,6 +1558,7 @@ static struct file_system_type btrfs_fs_type = {
        .kill_sb        = btrfs_kill_super,
        .fs_flags       = FS_REQUIRES_DEV,
 };
+MODULE_ALIAS_FS("btrfs");
 /*
 * used by btrfsctl to scan devices when no FS is mounted
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index e52da6fb1165..50767bbaad6c 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -625,14 +625,13 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
        btrfs_trans_release_metadata(trans, root);
        trans->block_rsv = NULL;
-        /*
-         * the same root has to be passed to start_transaction and
-         * end_transaction. Subvolume quota depends on this.
-         */
-        WARN_ON(trans->root != root);
        if (trans->qgroup_reserved) {
-                btrfs_qgroup_free(root, trans->qgroup_reserved);
+                /*
+                 * the same root has to be passed here between start_transaction
+                 * and end_transaction. Subvolume quota depends on this.
+                 */
+                btrfs_qgroup_free(trans->root, trans->qgroup_reserved);
                trans->qgroup_reserved = 0;
        }
@@ -1052,7 +1051,12 @@ int btrfs_defrag_root(struct btrfs_root *root)
 /*
 * new snapshots need to be created at a very specific time in the
- * transaction commit.  This does the actual creation
+ * transaction commit.  This does the actual creation.
+ *
+ * Note:
+ * If the error which may affect the commitment of the current transaction
+ * happens, we should return the error number. If the error which just affect
+ * the creation of the pending snapshots, just return 0.
 */
 static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
                                   struct btrfs_fs_info *fs_info,
@@ -1071,7 +1075,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
        struct extent_buffer *tmp;
        struct extent_buffer *old;
        struct timespec cur_time = CURRENT_TIME;
-        int ret;
+        int ret = 0;
        u64 to_reserve = 0;
        u64 index = 0;
        u64 objectid;
@@ -1080,40 +1084,36 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
        path = btrfs_alloc_path();
        if (!path) {
-                ret = pending->error = -ENOMEM;
+                pending->error = -ENOMEM;
-                return ret;
+                return 0;
        }
        new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS);
        if (!new_root_item) {
-                ret = pending->error = -ENOMEM;
+                pending->error = -ENOMEM;
                goto root_item_alloc_fail;
        }
-        ret = btrfs_find_free_objectid(tree_root, &objectid);
+        pending->error = btrfs_find_free_objectid(tree_root, &objectid);
-        if (ret) {
+        if (pending->error)
-                pending->error = ret;
                goto no_free_objectid;
-        }
        btrfs_reloc_pre_snapshot(trans, pending, &to_reserve);
        if (to_reserve > 0) {
-                ret = btrfs_block_rsv_add(root, &pending->block_rsv,
+                pending->error = btrfs_block_rsv_add(root,
-                                          to_reserve,
+                                                     &pending->block_rsv,
-                                          BTRFS_RESERVE_NO_FLUSH);
+                                                     to_reserve,
-                if (ret) {
+                                                     BTRFS_RESERVE_NO_FLUSH);
-                        pending->error = ret;
+                if (pending->error)
                        goto no_free_objectid;
-                }
        }
-        ret = btrfs_qgroup_inherit(trans, fs_info, root->root_key.objectid,
+        pending->error = btrfs_qgroup_inherit(trans, fs_info,
-                                   objectid, pending->inherit);
+                                              root->root_key.objectid,
-        if (ret) {
+                                              objectid, pending->inherit);
-                pending->error = ret;
+        if (pending->error)
                goto no_free_objectid;
-        }
        key.objectid = objectid;
        key.offset = (u64)-1;
@@ -1141,7 +1141,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
                                         dentry->d_name.len, 0);
        if (dir_item != NULL && !IS_ERR(dir_item)) {
                pending->error = -EEXIST;
-                goto fail;
+                goto dir_item_existed;
        } else if (IS_ERR(dir_item)) {
                ret = PTR_ERR(dir_item);
                btrfs_abort_transaction(trans, root, ret);
@@ -1272,6 +1272,8 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
        if (ret)
                btrfs_abort_transaction(trans, root, ret);
 fail:
+        pending->error = ret;
+dir_item_existed:
        trans->block_rsv = rsv;
        trans->bytes_reserved = 0;
 no_free_objectid:
@@ -1287,12 +1289,17 @@ root_item_alloc_fail:
 static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans,
                                             struct btrfs_fs_info *fs_info)
 {
-        struct btrfs_pending_snapshot *pending;
+        struct btrfs_pending_snapshot *pending, *next;
        struct list_head *head = &trans->transaction->pending_snapshots;
+        int ret = 0;
-        list_for_each_entry(pending, head, list)
+        list_for_each_entry_safe(pending, next, head, list) {
-                create_pending_snapshot(trans, fs_info, pending);
+                list_del(&pending->list);
-        return 0;
+                ret = create_pending_snapshot(trans, fs_info, pending);
+                if (ret)
+                        break;
+        }
+        return ret;
 }
 static void update_super_roots(struct btrfs_root *root)
@@ -1448,6 +1455,13 @@ static void cleanup_transaction(struct btrfs_trans_handle *trans,
        btrfs_abort_transaction(trans, root, err);
        spin_lock(&root->fs_info->trans_lock);
+        if (list_empty(&cur_trans->list)) {
+                spin_unlock(&root->fs_info->trans_lock);
+                btrfs_end_transaction(trans, root);
+                return;
+        }
        list_del_init(&cur_trans->list);
        if (cur_trans == root->fs_info->running_transaction) {
                root->fs_info->trans_no_join = 1;
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index c7ef569eb22a..451fad96ecd1 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -1382,7 +1382,10 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
        btrfs_release_path(path);
        if (ret == 0) {
-                btrfs_inc_nlink(inode);
+                if (!inode->i_nlink)
+                        set_nlink(inode, 1);
+                else
+                        btrfs_inc_nlink(inode);
                ret = btrfs_update_inode(trans, root, inode);
        } else if (ret == -EEXIST) {
                ret = 0;
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 35bb2d4ed29f..5989a92236f7 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -684,6 +684,12 @@ int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
                __btrfs_close_devices(fs_devices);
                free_fs_devices(fs_devices);
        }
+        /*
+         * Wait for rcu kworkers under __btrfs_close_devices
+         * to finish all blkdev_puts so device is really
+         * free when umount is done.
+         */
+        rcu_barrier();
        return ret;
 }
@@ -2379,7 +2385,11 @@ static int btrfs_relocate_chunk(struct btrfs_root *root,
                return ret;
        trans = btrfs_start_transaction(root, 0);
-        BUG_ON(IS_ERR(trans));
+        if (IS_ERR(trans)) {
+                ret = PTR_ERR(trans);
+                btrfs_std_error(root->fs_info, ret);
+                return ret;
+        }
        lock_chunks(root);
@@ -3050,7 +3060,8 @@ static void __cancel_balance(struct btrfs_fs_info *fs_info)
        unset_balance_control(fs_info);
        ret = del_balance_item(fs_info->tree_root);
-        BUG_ON(ret);
+        if (ret)
+                btrfs_std_error(fs_info, ret);
        atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
 }
@@ -3230,6 +3241,11 @@ int btrfs_balance(struct btrfs_balance_control *bctl,
                update_ioctl_balance_args(fs_info, 0, bargs);
        }
+        if ((ret && ret != -ECANCELED && ret != -ENOSPC) ||
+            balance_need_close(fs_info)) {
+                __cancel_balance(fs_info);
+        }
        wake_up(&fs_info->balance_wait_q);
        return ret;