Merge branch 'kdave-part1-enospc' into for-linus-4.8

fs/btrfs/ctree.h

@@ -439,6 +439,8 @@ struct btrfs_space_info {
 	struct list_head list;
 	/* Protected by the spinlock 'lock'. */
 	struct list_head ro_bgs;
+	struct list_head priority_tickets;
+	struct list_head tickets;
 
 	struct rw_semaphore groups_sem;
 	/* for block groups in our same type */
@@ -2624,6 +2626,15 @@ enum btrfs_reserve_flush_enum {
 	BTRFS_RESERVE_FLUSH_ALL,
 };
 
+enum btrfs_flush_state {
+	FLUSH_DELAYED_ITEMS_NR	=	1,
+	FLUSH_DELAYED_ITEMS	=	2,
+	FLUSH_DELALLOC		=	3,
+	FLUSH_DELALLOC_WAIT	=	4,
+	ALLOC_CHUNK		=	5,
+	COMMIT_TRANS		=	6,
+};
+
 int btrfs_check_data_free_space(struct inode *inode, u64 start, u64 len);
 int btrfs_alloc_data_chunk_ondemand(struct inode *inode, u64 bytes);
 void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len);
@@ -2661,8 +2672,8 @@ int btrfs_block_rsv_refill(struct btrfs_root *root,
 			   struct btrfs_block_rsv *block_rsv, u64 min_reserved,
 			   enum btrfs_reserve_flush_enum flush);
 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
-			    struct btrfs_block_rsv *dst_rsv,
+			    struct btrfs_block_rsv *dst_rsv, u64 num_bytes,
-			    u64 num_bytes);
+			    int update_size);
 int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info,
 			     struct btrfs_block_rsv *dest, u64 num_bytes,
 			     int min_factor);

fs/btrfs/delayed-inode.c

@@ -553,7 +553,7 @@ static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans,
 	dst_rsv = &root->fs_info->delayed_block_rsv;
 
 	num_bytes = btrfs_calc_trans_metadata_size(root, 1);
-	ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
+	ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1);
 	if (!ret) {
 		trace_btrfs_space_reservation(root->fs_info, "delayed_item",
 					      item->key.objectid,
@@ -597,6 +597,29 @@ static int btrfs_delayed_inode_reserve_metadata(
 
 	num_bytes = btrfs_calc_trans_metadata_size(root, 1);
 
+	/*
+	 * If our block_rsv is the delalloc block reserve then check and see if
+	 * we have our extra reservation for updating the inode.  If not fall
+	 * through and try to reserve space quickly.
+	 *
+	 * We used to try and steal from the delalloc block rsv or the global
+	 * reserve, but we'd steal a full reservation, which isn't kind.  We are
+	 * here through delalloc which means we've likely just cowed down close
+	 * to the leaf that contains the inode, so we would steal less just
+	 * doing the fallback inode update, so if we do end up having to steal
+	 * from the global block rsv we hopefully only steal one or two blocks
+	 * worth which is less likely to hurt us.
+	 */
+	if (src_rsv && src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) {
+		spin_lock(&BTRFS_I(inode)->lock);
+		if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED,
+				       &BTRFS_I(inode)->runtime_flags))
+			release = true;
+		else
+			src_rsv = NULL;
+		spin_unlock(&BTRFS_I(inode)->lock);
+	}
+
 	/*
 	 * btrfs_dirty_inode will update the inode under btrfs_join_transaction
 	 * which doesn't reserve space for speed.  This is a problem since we
@@ -626,51 +649,10 @@ static int btrfs_delayed_inode_reserve_metadata(
 						      num_bytes, 1);
 		}
 		return ret;
-	} else if (src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) {
-		spin_lock(&BTRFS_I(inode)->lock);
-		if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED,
-				       &BTRFS_I(inode)->runtime_flags)) {
-			spin_unlock(&BTRFS_I(inode)->lock);
-			release = true;
-			goto migrate;
-		}
-		spin_unlock(&BTRFS_I(inode)->lock);
-
-		/* Ok we didn't have space pre-reserved.  This shouldn't happen
-		 * too often but it can happen if we do delalloc to an existing
-		 * inode which gets dirtied because of the time update, and then
-		 * isn't touched again until after the transaction commits and
-		 * then we try to write out the data.  First try to be nice and
-		 * reserve something strictly for us.  If not be a pain and try
-		 * to steal from the delalloc block rsv.
-		 */
-		ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes,
-					  BTRFS_RESERVE_NO_FLUSH);
-		if (!ret)
-			goto out;
-
-		ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
-		if (!ret)
-			goto out;
-
-		if (btrfs_test_opt(root, ENOSPC_DEBUG)) {
-			btrfs_debug(root->fs_info,
-				    "block rsv migrate returned %d", ret);
-			WARN_ON(1);
-		}
-		/*
-		 * Ok this is a problem, let's just steal from the global rsv
-		 * since this really shouldn't happen that often.
-		 */
-		ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv,
-					      dst_rsv, num_bytes);
-		goto out;
 	}
 
-migrate:
+	ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1);
-	ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
 
-out:
 	/*
 	 * Migrate only takes a reservation, it doesn't touch the size of the
 	 * block_rsv.  This is to simplify people who don't normally have things

fs/btrfs/file.c

@@ -2479,7 +2479,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
 	}
 
 	ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv,
-				      min_size);
+				      min_size, 0);
 	BUG_ON(ret);
 	trans->block_rsv = rsv;
 
@@ -2522,7 +2522,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
 		}
 
 		ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv,
-					      rsv, min_size);
+					      rsv, min_size, 0);
 		BUG_ON(ret);	/* shouldn't happen */
 		trans->block_rsv = rsv;
 

fs/btrfs/inode.c

@@ -5264,7 +5264,7 @@ void btrfs_evict_inode(struct inode *inode)
 		if (steal_from_global) {
 			if (!btrfs_check_space_for_delayed_refs(trans, root))
 				ret = btrfs_block_rsv_migrate(global_rsv, rsv,
-							      min_size);
+							      min_size, 0);
 			else
 				ret = -ENOSPC;
 		}
@@ -9116,7 +9116,7 @@ static int btrfs_truncate(struct inode *inode)
 
 	/* Migrate the slack space for the truncate to our reserve */
 	ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv,
-				      min_size);
+				      min_size, 0);
 	BUG_ON(ret);
 
 	/*
@@ -9156,7 +9156,7 @@ static int btrfs_truncate(struct inode *inode)
 		}
 
 		ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv,
-					      rsv, min_size);
+					      rsv, min_size, 0);
 		BUG_ON(ret);	/* shouldn't happen */
 		trans->block_rsv = rsv;
 	}
@@ -9177,7 +9177,6 @@ static int btrfs_truncate(struct inode *inode)
 		ret = btrfs_end_transaction(trans, root);
 		btrfs_btree_balance_dirty(root);
 	}
-
 out:
 	btrfs_free_block_rsv(root, rsv);
 

fs/btrfs/relocation.c

@@ -2604,25 +2604,28 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans,
 
 	trans->block_rsv = rc->block_rsv;
 	rc->reserved_bytes += num_bytes;
+
+	/*
+	 * We are under a transaction here so we can only do limited flushing.
+	 * If we get an enospc just kick back -EAGAIN so we know to drop the
+	 * transaction and try to refill when we can flush all the things.
+	 */
 	ret = btrfs_block_rsv_refill(root, rc->block_rsv, num_bytes,
-				BTRFS_RESERVE_FLUSH_ALL);
+				BTRFS_RESERVE_FLUSH_LIMIT);
 	if (ret) {
-		if (ret == -EAGAIN) {
+		tmp = rc->extent_root->nodesize * RELOCATION_RESERVED_NODES;
-			tmp = rc->extent_root->nodesize *
-				RELOCATION_RESERVED_NODES;
 		while (tmp <= rc->reserved_bytes)
 			tmp <<= 1;
 		/*
 		 * only one thread can access block_rsv at this point,
 		 * so we don't need hold lock to protect block_rsv.
 		 * we expand more reservation size here to allow enough
-			 * space for relocation and we will return earlier in
+		 * space for relocation and we will return eailer in
 		 * enospc case.
 		 */
 		rc->block_rsv->size = tmp + rc->extent_root->nodesize *
 			RELOCATION_RESERVED_NODES;
-		}
+		return -EAGAIN;
-		return ret;
 	}
 
 	return 0;
@@ -3871,6 +3874,7 @@ static noinline_for_stack
 int prepare_to_relocate(struct reloc_control *rc)
 {
 	struct btrfs_trans_handle *trans;
+	int ret;
 
 	rc->block_rsv = btrfs_alloc_block_rsv(rc->extent_root,
 					      BTRFS_BLOCK_RSV_TEMP);
@@ -3885,6 +3889,11 @@ int prepare_to_relocate(struct reloc_control *rc)
 	rc->reserved_bytes = 0;
 	rc->block_rsv->size = rc->extent_root->nodesize *
 			      RELOCATION_RESERVED_NODES;
+	ret = btrfs_block_rsv_refill(rc->extent_root,
+				     rc->block_rsv, rc->block_rsv->size,
+				     BTRFS_RESERVE_FLUSH_ALL);
+	if (ret)
+		return ret;
 
 	rc->create_reloc_tree = 1;
 	set_reloc_control(rc);
@@ -4643,7 +4652,7 @@ int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 	if (rc->merge_reloc_tree) {
 		ret = btrfs_block_rsv_migrate(&pending->block_rsv,
 					      rc->block_rsv,
-					      rc->nodes_relocated);
+					      rc->nodes_relocated, 1);
 		if (ret)
 			return ret;
 	}

include/trace/events/btrfs.h

@@ -440,6 +440,46 @@ TRACE_EVENT(btrfs_sync_fs,
 	TP_printk("wait = %d", __entry->wait)
 );
 
+TRACE_EVENT(btrfs_add_block_group,
+
+	TP_PROTO(struct btrfs_fs_info *fs_info,
+		 struct btrfs_block_group_cache *block_group, int create),
+
+	TP_ARGS(fs_info, block_group, create),
+
+	TP_STRUCT__entry(
+		__array(	u8,	fsid,	BTRFS_UUID_SIZE	)
+		__field(	u64,	offset			)
+		__field(	u64,	size			)
+		__field(	u64,	flags			)
+		__field(	u64,	bytes_used		)
+		__field(	u64,	bytes_super		)
+		__field(	int,	create			)
+	),
+
+	TP_fast_assign(
+		memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE);
+		__entry->offset		= block_group->key.objectid;
+		__entry->size		= block_group->key.offset;
+		__entry->flags		= block_group->flags;
+		__entry->bytes_used	=
+			btrfs_block_group_used(&block_group->item);
+		__entry->bytes_super	= block_group->bytes_super;
+		__entry->create		= create;
+	),
+
+	TP_printk("%pU: block_group offset = %llu, size = %llu, "
+		  "flags = %llu(%s), bytes_used = %llu, bytes_super = %llu, "
+		  "create = %d", __entry->fsid,
+		  (unsigned long long)__entry->offset,
+		  (unsigned long long)__entry->size,
+		  (unsigned long long)__entry->flags,
+		  __print_flags((unsigned long)__entry->flags, "|",
+				BTRFS_GROUP_FLAGS),
+		  (unsigned long long)__entry->bytes_used,
+		  (unsigned long long)__entry->bytes_super, __entry->create)
+);
+
 #define show_ref_action(action)						\
 	__print_symbolic(action,					\
 		{ BTRFS_ADD_DELAYED_REF,    "ADD_DELAYED_REF" },	\
@@ -744,6 +784,88 @@ TRACE_EVENT(btrfs_space_reservation,
 		  __entry->bytes)
 );
 
+#define show_flush_action(action)						\
+	__print_symbolic(action,						\
+		{ BTRFS_RESERVE_NO_FLUSH,	"BTRFS_RESERVE_NO_FLUSH"},	\
+		{ BTRFS_RESERVE_FLUSH_LIMIT,	"BTRFS_RESERVE_FLUSH_LIMIT"},	\
+		{ BTRFS_RESERVE_FLUSH_ALL,	"BTRFS_RESERVE_FLUSH_ALL"})
+
+TRACE_EVENT(btrfs_trigger_flush,
+
+	TP_PROTO(struct btrfs_fs_info *fs_info, u64 flags, u64 bytes,
+		 int flush, char *reason),
+
+	TP_ARGS(fs_info, flags, bytes, flush, reason),
+
+	TP_STRUCT__entry(
+		__array(	u8,	fsid,	BTRFS_UUID_SIZE	)
+		__field(	u64,	flags			)
+		__field(	u64,	bytes			)
+		__field(	int,	flush			)
+		__string(	reason,	reason			)
+	),
+
+	TP_fast_assign(
+		memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE);
+		__entry->flags	= flags;
+		__entry->bytes	= bytes;
+		__entry->flush	= flush;
+		__assign_str(reason, reason)
+	),
+
+	TP_printk("%pU: %s: flush = %d(%s), flags = %llu(%s), bytes = %llu",
+		  __entry->fsid, __get_str(reason), __entry->flush,
+		  show_flush_action(__entry->flush),
+		  (unsigned long long)__entry->flags,
+		  __print_flags((unsigned long)__entry->flags, "|",
+				BTRFS_GROUP_FLAGS),
+		  (unsigned long long)__entry->bytes)
+);
+
+#define show_flush_state(state)							\
+	__print_symbolic(state,							\
+		{ FLUSH_DELAYED_ITEMS_NR,	"FLUSH_DELAYED_ITEMS_NR"},	\
+		{ FLUSH_DELAYED_ITEMS,		"FLUSH_DELAYED_ITEMS"},		\
+		{ FLUSH_DELALLOC,		"FLUSH_DELALLOC"},		\
+		{ FLUSH_DELALLOC_WAIT,		"FLUSH_DELALLOC_WAIT"},		\
+		{ ALLOC_CHUNK,			"ALLOC_CHUNK"},			\
+		{ COMMIT_TRANS,			"COMMIT_TRANS"})
+
+TRACE_EVENT(btrfs_flush_space,
+
+	TP_PROTO(struct btrfs_fs_info *fs_info, u64 flags, u64 num_bytes,
+		 u64 orig_bytes, int state, int ret),
+
+	TP_ARGS(fs_info, flags, num_bytes, orig_bytes, state, ret),
+
+	TP_STRUCT__entry(
+		__array(	u8,	fsid,	BTRFS_UUID_SIZE	)
+		__field(	u64,	flags			)
+		__field(	u64,	num_bytes		)
+		__field(	u64,	orig_bytes		)
+		__field(	int,	state			)
+		__field(	int,	ret			)
+	),
+
+	TP_fast_assign(
+		memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE);
+		__entry->flags		=	flags;
+		__entry->num_bytes	=	num_bytes;
+		__entry->orig_bytes	=	orig_bytes;
+		__entry->state		=	state;
+		__entry->ret		=	ret;
+	),
+
+	TP_printk("%pU: state = %d(%s), flags = %llu(%s), num_bytes = %llu, "
+		  "orig_bytes = %llu, ret = %d", __entry->fsid, __entry->state,
+		  show_flush_state(__entry->state),
+		  (unsigned long long)__entry->flags,
+		  __print_flags((unsigned long)__entry->flags, "|",
+				BTRFS_GROUP_FLAGS),
+		  (unsigned long long)__entry->num_bytes,
+		  (unsigned long long)__entry->orig_bytes, __entry->ret)
+);
+
 DECLARE_EVENT_CLASS(btrfs__reserved_extent,
 
 	TP_PROTO(struct btrfs_root *root, u64 start, u64 len),
@@ -751,18 +873,21 @@ DECLARE_EVENT_CLASS(btrfs__reserved_extent,
 	TP_ARGS(root, start, len),
 
 	TP_STRUCT__entry(
+		__array(	u8,	fsid,	BTRFS_UUID_SIZE	)
 		__field(	u64,	root_objectid		)
 		__field(	u64,	start			)
 		__field(	u64,	len			)
 	),
 
 	TP_fast_assign(
+		memcpy(__entry->fsid, root->fs_info->fsid, BTRFS_UUID_SIZE);
 		__entry->root_objectid	= root->root_key.objectid;
 		__entry->start		= start;
 		__entry->len		= len;
 	),
 
-	TP_printk("root = %llu(%s), start = %llu, len = %llu",
+	TP_printk("%pU: root = %llu(%s), start = %llu, len = %llu",
+		  __entry->fsid,
 		  show_root_type(__entry->root_objectid),
 		  (unsigned long long)__entry->start,
 		  (unsigned long long)__entry->len)
@@ -819,6 +944,7 @@ DECLARE_EVENT_CLASS(btrfs__reserve_extent,
 	TP_ARGS(root, block_group, start, len),
 
 	TP_STRUCT__entry(
+		__array(	u8,	fsid,	BTRFS_UUID_SIZE	)
 		__field(	u64,	root_objectid		)
 		__field(	u64,	bg_objectid		)
 		__field(	u64,	flags			)
@@ -827,6 +953,7 @@ DECLARE_EVENT_CLASS(btrfs__reserve_extent,
 	),
 
 	TP_fast_assign(
+		memcpy(__entry->fsid, root->fs_info->fsid, BTRFS_UUID_SIZE);
 		__entry->root_objectid	= root->root_key.objectid;
 		__entry->bg_objectid	= block_group->key.objectid;
 		__entry->flags		= block_group->flags;
@@ -834,8 +961,8 @@ DECLARE_EVENT_CLASS(btrfs__reserve_extent,
 		__entry->len		= len;
 	),
 
-	TP_printk("root = %Lu(%s), block_group = %Lu, flags = %Lu(%s), "
+	TP_printk("%pU: root = %Lu(%s), block_group = %Lu, flags = %Lu(%s), "
-		  "start = %Lu, len = %Lu",
+		  "start = %Lu, len = %Lu", __entry->fsid,
 		  show_root_type(__entry->root_objectid), __entry->bg_objectid,
 		  __entry->flags, __print_flags((unsigned long)__entry->flags,
 						"|", BTRFS_GROUP_FLAGS),