Btrfs: incompatible format change to remove hole extents

Btrfs has always had these filler extent data items for holes in inodes.  This
has made somethings very easy, like logging hole punches and sending hole
punches.  However for large holey files these extent data items are pure
overhead.  So add an incompatible feature to no longer add hole extents to
reduce the amount of metadata used by these sort of files.  This has a few
changes for logging and send obviously since they will need to detect holes and
log/send the holes if there are any.  I've tested this thoroughly with xfstests
and it doesn't cause any issues with and without the incompat format set.
Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <clm@fb.com>

1 files changed, 149 insertions, 9 deletions

diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index 945d1db98f26..29803b4129fc 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -111,6 +111,7 @@ struct send_ctx {
 	int cur_inode_deleted;
 	u64 cur_inode_size;
 	u64 cur_inode_mode;
+	u64 cur_inode_last_extent;
 
 	u64 send_progress;
 
@@ -145,6 +146,13 @@ struct name_cache_entry {
 	char name[];
 };
 
+static int need_send_hole(struct send_ctx *sctx)
+{
+	return (sctx->parent_root && !sctx->cur_inode_new &&
+		!sctx->cur_inode_new_gen && !sctx->cur_inode_deleted &&
+		S_ISREG(sctx->cur_inode_mode));
+}
+
 static void fs_path_reset(struct fs_path *p)
 {
 	if (p->reversed) {
@@ -3752,6 +3760,39 @@ out:
 	return ret;
 }
 
+static int send_hole(struct send_ctx *sctx, u64 end)
+{
+	struct fs_path *p = NULL;
+	u64 offset = sctx->cur_inode_last_extent;
+	u64 len;
+	int ret = 0;
+
+	p = fs_path_alloc();
+	if (!p)
+		return -ENOMEM;
+	memset(sctx->read_buf, 0, BTRFS_SEND_READ_SIZE);
+	while (offset < end) {
+		len = min_t(u64, end - offset, BTRFS_SEND_READ_SIZE);
+
+		ret = begin_cmd(sctx, BTRFS_SEND_C_WRITE);
+		if (ret < 0)
+			break;
+		ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
+		if (ret < 0)
+			break;
+		TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
+		TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
+		TLV_PUT(sctx, BTRFS_SEND_A_DATA, sctx->read_buf, len);
+		ret = send_cmd(sctx);
+		if (ret < 0)
+			break;
+		offset += len;
+	}
+tlv_put_failure:
+	fs_path_free(p);
+	return ret;
+}
+
 static int send_write_or_clone(struct send_ctx *sctx,
 			       struct btrfs_path *path,
 			       struct btrfs_key *key,
@@ -3979,6 +4020,84 @@ out:
 	return ret;
 }
 
+static int get_last_extent(struct send_ctx *sctx, u64 offset)
+{
+	struct btrfs_path *path;
+	struct btrfs_root *root = sctx->send_root;
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_key key;
+	u64 extent_end;
+	u8 type;
+	int ret;
+
+	path = alloc_path_for_send();
+	if (!path)
+		return -ENOMEM;
+
+	sctx->cur_inode_last_extent = 0;
+
+	key.objectid = sctx->cur_ino;
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = offset;
+	ret = btrfs_search_slot_for_read(root, &key, path, 0, 1);
+	if (ret < 0)
+		goto out;
+	ret = 0;
+	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+	if (key.objectid != sctx->cur_ino || key.type != BTRFS_EXTENT_DATA_KEY)
+		goto out;
+
+	fi = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			    struct btrfs_file_extent_item);
+	type = btrfs_file_extent_type(path->nodes[0], fi);
+	if (type == BTRFS_FILE_EXTENT_INLINE) {
+		u64 size = btrfs_file_extent_inline_len(path->nodes[0], fi);
+		extent_end = ALIGN(key.offset + size,
+				   sctx->send_root->sectorsize);
+	} else {
+		extent_end = key.offset +
+			btrfs_file_extent_num_bytes(path->nodes[0], fi);
+	}
+	sctx->cur_inode_last_extent = extent_end;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int maybe_send_hole(struct send_ctx *sctx, struct btrfs_path *path,
+			   struct btrfs_key *key)
+{
+	struct btrfs_file_extent_item *fi;
+	u64 extent_end;
+	u8 type;
+	int ret = 0;
+
+	if (sctx->cur_ino != key->objectid || !need_send_hole(sctx))
+		return 0;
+
+	if (sctx->cur_inode_last_extent == (u64)-1) {
+		ret = get_last_extent(sctx, key->offset - 1);
+		if (ret)
+			return ret;
+	}
+
+	fi = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			    struct btrfs_file_extent_item);
+	type = btrfs_file_extent_type(path->nodes[0], fi);
+	if (type == BTRFS_FILE_EXTENT_INLINE) {
+		u64 size = btrfs_file_extent_inline_len(path->nodes[0], fi);
+		extent_end = ALIGN(key->offset + size,
+				   sctx->send_root->sectorsize);
+	} else {
+		extent_end = key->offset +
+			btrfs_file_extent_num_bytes(path->nodes[0], fi);
+	}
+	if (sctx->cur_inode_last_extent < key->offset)
+		ret = send_hole(sctx, key->offset);
+	sctx->cur_inode_last_extent = extent_end;
+	return ret;
+}
+
 static int process_extent(struct send_ctx *sctx,
 			  struct btrfs_path *path,
 			  struct btrfs_key *key)
@@ -3995,7 +4114,7 @@ static int process_extent(struct send_ctx *sctx,
 			goto out;
 		if (ret) {
 			ret = 0;
-			goto out;
+			goto out_hole;
 		}
 	} else {
 		struct btrfs_file_extent_item *ei;
@@ -4031,7 +4150,10 @@ static int process_extent(struct send_ctx *sctx,
 		goto out;
 
 	ret = send_write_or_clone(sctx, path, key, found_clone);
-
+	if (ret)
+		goto out;
+out_hole:
+	ret = maybe_send_hole(sctx, path, key);
 out:
 	return ret;
 }
@@ -4157,6 +4279,19 @@ static int finish_inode_if_needed(struct send_ctx *sctx, int at_end)
 	}
 
 	if (S_ISREG(sctx->cur_inode_mode)) {
+		if (need_send_hole(sctx)) {
+			if (sctx->cur_inode_last_extent == (u64)-1) {
+				ret = get_last_extent(sctx, (u64)-1);
+				if (ret)
+					goto out;
+			}
+			if (sctx->cur_inode_last_extent <
+			    sctx->cur_inode_size) {
+				ret = send_hole(sctx, sctx->cur_inode_size);
+				if (ret)
+					goto out;
+			}
+		}
 		ret = send_truncate(sctx, sctx->cur_ino, sctx->cur_inode_gen,
 				sctx->cur_inode_size);
 		if (ret < 0)
@@ -4200,6 +4335,7 @@ static int changed_inode(struct send_ctx *sctx,
 
 	sctx->cur_ino = key->objectid;
 	sctx->cur_inode_new_gen = 0;
+	sctx->cur_inode_last_extent = (u64)-1;
 
 	/*
 	 * Set send_progress to current inode. This will tell all get_cur_xxx
@@ -4480,14 +4616,18 @@ static int changed_cb(struct btrfs_root *left_root,
 	struct send_ctx *sctx = ctx;
 
 	if (result == BTRFS_COMPARE_TREE_SAME) {
-		if (key->type != BTRFS_INODE_REF_KEY &&
-		    key->type != BTRFS_INODE_EXTREF_KEY)
-			return 0;
-		ret = compare_refs(sctx, left_path, key);
-		if (!ret)
+		if (key->type == BTRFS_INODE_REF_KEY ||
+		    key->type == BTRFS_INODE_EXTREF_KEY) {
+			ret = compare_refs(sctx, left_path, key);
+			if (!ret)
+				return 0;
+			if (ret < 0)
+				return ret;
+		} else if (key->type == BTRFS_EXTENT_DATA_KEY) {
+			return maybe_send_hole(sctx, left_path, key);
+		} else {
 			return 0;
-		if (ret < 0)
-			return ret;
+		}
 		result = BTRFS_COMPARE_TREE_CHANGED;
 		ret = 0;
 	}