Btrfs: make the chunk allocator completely tree lockless

When adjusting the enospc rules for relocation I ran into a deadlock because we
were relocating the only system chunk and that forced us to try and allocate a
new system chunk while holding locks in the chunk tree, which caused us to
deadlock.  To fix this I've moved all of the dev extent addition and chunk
addition out to the delayed chunk completion stuff.  We still keep the in-memory
stuff which makes sure everything is consistent.

One change I had to make was to search the commit root of the device tree to
find a free dev extent, and hold onto any chunk em's that we allocated in that
transaction so we do not allocate the same dev extent twice.  This has the side
effect of fixing a bug with balance that has been there ever since balance
existed.  Basically you can free a block group and it's dev extent and then
immediately allocate that dev extent for a new block group and write stuff to
that dev extent, all within the same transaction.  So if you happen to crash
during a balance you could come back to a completely broken file system.  This
patch should keep these sort of things from happening in the future since we
won't be able to allocate free'd dev extents until after the transaction
commits.  This has passed all of the xfstests and my super annoying stress test
followed by a balance.  Thanks,
Signed-off-by: Josef Bacik <jbacik@fusionio.com>

fs/btrfs/extent-tree.c

@@ -7950,6 +7950,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 	struct btrfs_space_info *space_info;
 	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
 	struct btrfs_device *device;
+	struct btrfs_trans_handle *trans;
 	u64 min_free;
 	u64 dev_min = 1;
 	u64 dev_nr = 0;
@@ -8036,6 +8037,13 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 		do_div(min_free, dev_min);
 	}
 
+	/* We need to do this so that we can look at pending chunks */
+	trans = btrfs_join_transaction(root);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		goto out;
+	}
+
 	mutex_lock(&root->fs_info->chunk_mutex);
 	list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) {
 		u64 dev_offset;
@@ -8046,7 +8054,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 		 */
 		if (device->total_bytes > device->bytes_used + min_free &&
 		    !device->is_tgtdev_for_dev_replace) {
-			ret = find_free_dev_extent(device, min_free,
+			ret = find_free_dev_extent(trans, device, min_free,
 						   &dev_offset, NULL);
 			if (!ret)
 				dev_nr++;
@@ -8058,6 +8066,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 		}
 	}
 	mutex_unlock(&root->fs_info->chunk_mutex);
+	btrfs_end_transaction(trans, root);
 out:
 	btrfs_put_block_group(block_group);
 	return ret;
@@ -8423,6 +8432,10 @@ void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans,
 					sizeof(item));
 		if (ret)
 			btrfs_abort_transaction(trans, extent_root, ret);
+		ret = btrfs_finish_chunk_alloc(trans, extent_root,
+					       key.objectid, key.offset);
+		if (ret)
+			btrfs_abort_transaction(trans, extent_root, ret);
 	}
 }
 

fs/btrfs/transaction.c

@@ -63,6 +63,14 @@ static void put_transaction(struct btrfs_transaction *transaction)
 	if (atomic_dec_and_test(&transaction->use_count)) {
 		BUG_ON(!list_empty(&transaction->list));
 		WARN_ON(transaction->delayed_refs.root.rb_node);
+		while (!list_empty(&transaction->pending_chunks)) {
+			struct extent_map *em;
+
+			em = list_first_entry(&transaction->pending_chunks,
+					      struct extent_map, list);
+			list_del_init(&em->list);
+			free_extent_map(em);
+		}
 		kmem_cache_free(btrfs_transaction_cachep, transaction);
 	}
 }
@@ -202,6 +210,7 @@ static noinline int join_transaction(struct btrfs_root *root, unsigned int type)
 
 	INIT_LIST_HEAD(&cur_trans->pending_snapshots);
 	INIT_LIST_HEAD(&cur_trans->ordered_operations);
+	INIT_LIST_HEAD(&cur_trans->pending_chunks);
 	list_add_tail(&cur_trans->list, &fs_info->trans_list);
 	extent_io_tree_init(&cur_trans->dirty_pages,
 			     fs_info->btree_inode->i_mapping);

fs/btrfs/transaction.h

@@ -56,6 +56,7 @@ struct btrfs_transaction {
 	wait_queue_head_t commit_wait;
 	struct list_head pending_snapshots;
 	struct list_head ordered_operations;
+	struct list_head pending_chunks;
 	struct btrfs_delayed_ref_root delayed_refs;
 	int aborted;
 };

fs/btrfs/volumes.h

@@ -316,7 +316,8 @@ int btrfs_recover_balance(struct btrfs_fs_info *fs_info);
 int btrfs_pause_balance(struct btrfs_fs_info *fs_info);
 int btrfs_cancel_balance(struct btrfs_fs_info *fs_info);
 int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset);
-int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
+int find_free_dev_extent(struct btrfs_trans_handle *trans,
+			 struct btrfs_device *device, u64 num_bytes,
 			 u64 *start, u64 *max_avail);
 void btrfs_dev_stat_inc_and_print(struct btrfs_device *dev, int index);
 int btrfs_get_dev_stats(struct btrfs_root *root,
@@ -337,6 +338,9 @@ int btrfs_is_parity_mirror(struct btrfs_mapping_tree *map_tree,
 unsigned long btrfs_full_stripe_len(struct btrfs_root *root,
 				    struct btrfs_mapping_tree *map_tree,
 				    u64 logical);
+int btrfs_finish_chunk_alloc(struct btrfs_trans_handle *trans,
+				struct btrfs_root *extent_root,
+				u64 chunk_offset, u64 chunk_size);
 static inline void btrfs_dev_stat_inc(struct btrfs_device *dev,
 				      int index)
 {