1 files changed, 267 insertions, 61 deletions

diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index d145ce804620..b082210df9c8 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -1069,6 +1069,12 @@ static int iterate_dir_item(struct btrfs_root *root, struct btrfs_path *path,
 			}
 		}
 
+		ret = btrfs_is_name_len_valid(eb, path->slots[0],
+			  (unsigned long)(di + 1), name_len + data_len);
+		if (!ret) {
+			ret = -EIO;
+			goto out;
+		}
 		if (name_len + data_len > buf_len) {
 			buf_len = name_len + data_len;
 			if (is_vmalloc_addr(buf)) {
@@ -1083,7 +1089,7 @@ static int iterate_dir_item(struct btrfs_root *root, struct btrfs_path *path,
 				buf = tmp;
 			}
 			if (!buf) {
-				buf = vmalloc(buf_len);
+				buf = kvmalloc(buf_len, GFP_KERNEL);
 				if (!buf) {
 					ret = -ENOMEM;
 					goto out;
@@ -1681,6 +1687,9 @@ static int is_inode_existent(struct send_ctx *sctx, u64 ino, u64 gen)
 {
 	int ret;
 
+	if (ino == BTRFS_FIRST_FREE_OBJECTID)
+		return 1;
+
 	ret = get_cur_inode_state(sctx, ino, gen);
 	if (ret < 0)
 		goto out;
@@ -1847,7 +1856,7 @@ out:
  */
 static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 			      const char *name, int name_len,
-			      u64 *who_ino, u64 *who_gen)
+			      u64 *who_ino, u64 *who_gen, u64 *who_mode)
 {
 	int ret = 0;
 	u64 gen;
@@ -1866,7 +1875,7 @@ static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 	 * not deleted and then re-created, if it was then we have no overwrite
 	 * and we can just unlink this entry.
 	 */
-	if (sctx->parent_root) {
+	if (sctx->parent_root && dir != BTRFS_FIRST_FREE_OBJECTID) {
 		ret = get_inode_info(sctx->parent_root, dir, NULL, &gen, NULL,
 				     NULL, NULL, NULL);
 		if (ret < 0 && ret != -ENOENT)
@@ -1896,7 +1905,7 @@ static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 	if (other_inode > sctx->send_progress ||
 	    is_waiting_for_move(sctx, other_inode)) {
 		ret = get_inode_info(sctx->parent_root, other_inode, NULL,
-				who_gen, NULL, NULL, NULL, NULL);
+				who_gen, who_mode, NULL, NULL, NULL);
 		if (ret < 0)
 			goto out;
 
@@ -1934,6 +1943,19 @@ static int did_overwrite_ref(struct send_ctx *sctx,
 	if (ret <= 0)
 		goto out;
 
+	if (dir != BTRFS_FIRST_FREE_OBJECTID) {
+		ret = get_inode_info(sctx->send_root, dir, NULL, &gen, NULL,
+				     NULL, NULL, NULL);
+		if (ret < 0 && ret != -ENOENT)
+			goto out;
+		if (ret) {
+			ret = 0;
+			goto out;
+		}
+		if (gen != dir_gen)
+			goto out;
+	}
+
 	/* check if the ref was overwritten by another ref */
 	ret = lookup_dir_item_inode(sctx->send_root, dir, name, name_len,
 			&ow_inode, &other_type);
@@ -2753,15 +2775,20 @@ out:
 
 struct recorded_ref {
 	struct list_head list;
-	char *dir_path;
 	char *name;
 	struct fs_path *full_path;
 	u64 dir;
 	u64 dir_gen;
-	int dir_path_len;
 	int name_len;
 };
 
+static void set_ref_path(struct recorded_ref *ref, struct fs_path *path)
+{
+	ref->full_path = path;
+	ref->name = (char *)kbasename(ref->full_path->start);
+	ref->name_len = ref->full_path->end - ref->name;
+}
+
 /*
  * We need to process new refs before deleted refs, but compare_tree gives us
  * everything mixed. So we first record all refs and later process them.
@@ -2778,17 +2805,7 @@ static int __record_ref(struct list_head *head, u64 dir,
 
 	ref->dir = dir;
 	ref->dir_gen = dir_gen;
-	ref->full_path = path;
-
-	ref->name = (char *)kbasename(ref->full_path->start);
-	ref->name_len = ref->full_path->end - ref->name;
-	ref->dir_path = ref->full_path->start;
-	if (ref->name == ref->full_path->start)
-		ref->dir_path_len = 0;
-	else
-		ref->dir_path_len = ref->full_path->end -
-				ref->full_path->start - 1 - ref->name_len;
-
+	set_ref_path(ref, path);
 	list_add_tail(&ref->list, head);
 	return 0;
 }
@@ -3530,9 +3547,17 @@ static int is_ancestor(struct btrfs_root *root,
 		       struct fs_path *fs_path)
 {
 	u64 ino = ino2;
+	bool free_path = false;
+	int ret = 0;
+
+	if (!fs_path) {
+		fs_path = fs_path_alloc();
+		if (!fs_path)
+			return -ENOMEM;
+		free_path = true;
+	}
 
 	while (ino > BTRFS_FIRST_FREE_OBJECTID) {
-		int ret;
 		u64 parent;
 		u64 parent_gen;
 
@@ -3541,13 +3566,18 @@ static int is_ancestor(struct btrfs_root *root,
 		if (ret < 0) {
 			if (ret == -ENOENT && ino == ino2)
 				ret = 0;
-			return ret;
+			goto out;
+		}
+		if (parent == ino1) {
+			ret = parent_gen == ino1_gen ? 1 : 0;
+			goto out;
 		}
-		if (parent == ino1)
-			return parent_gen == ino1_gen ? 1 : 0;
 		ino = parent;
 	}
-	return 0;
+ out:
+	if (free_path)
+		fs_path_free(fs_path);
+	return ret;
 }
 
 static int wait_for_parent_move(struct send_ctx *sctx,
@@ -3556,6 +3586,7 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 {
 	int ret = 0;
 	u64 ino = parent_ref->dir;
+	u64 ino_gen = parent_ref->dir_gen;
 	u64 parent_ino_before, parent_ino_after;
 	struct fs_path *path_before = NULL;
 	struct fs_path *path_after = NULL;
@@ -3576,6 +3607,8 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 	 * at get_cur_path()).
 	 */
 	while (ino > BTRFS_FIRST_FREE_OBJECTID) {
+		u64 parent_ino_after_gen;
+
 		if (is_waiting_for_move(sctx, ino)) {
 			/*
 			 * If the current inode is an ancestor of ino in the
@@ -3598,7 +3631,7 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 		fs_path_reset(path_after);
 
 		ret = get_first_ref(sctx->send_root, ino, &parent_ino_after,
-				    NULL, path_after);
+				    &parent_ino_after_gen, path_after);
 		if (ret < 0)
 			goto out;
 		ret = get_first_ref(sctx->parent_root, ino, &parent_ino_before,
@@ -3615,10 +3648,20 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 		if (ino > sctx->cur_ino &&
 		    (parent_ino_before != parent_ino_after || len1 != len2 ||
 		     memcmp(path_before->start, path_after->start, len1))) {
-			ret = 1;
-			break;
+			u64 parent_ino_gen;
+
+			ret = get_inode_info(sctx->parent_root, ino, NULL,
+					     &parent_ino_gen, NULL, NULL, NULL,
+					     NULL);
+			if (ret < 0)
+				goto out;
+			if (ino_gen == parent_ino_gen) {
+				ret = 1;
+				break;
+			}
 		}
 		ino = parent_ino_after;
+		ino_gen = parent_ino_after_gen;
 	}
 
 out:
@@ -3640,6 +3683,36 @@ out:
 	return ret;
 }
 
+static int update_ref_path(struct send_ctx *sctx, struct recorded_ref *ref)
+{
+	int ret;
+	struct fs_path *new_path;
+
+	/*
+	 * Our reference's name member points to its full_path member string, so
+	 * we use here a new path.
+	 */
+	new_path = fs_path_alloc();
+	if (!new_path)
+		return -ENOMEM;
+
+	ret = get_cur_path(sctx, ref->dir, ref->dir_gen, new_path);
+	if (ret < 0) {
+		fs_path_free(new_path);
+		return ret;
+	}
+	ret = fs_path_add(new_path, ref->name, ref->name_len);
+	if (ret < 0) {
+		fs_path_free(new_path);
+		return ret;
+	}
+
+	fs_path_free(ref->full_path);
+	set_ref_path(ref, new_path);
+
+	return 0;
+}
+
 /*
  * This does all the move/link/unlink/rmdir magic.
  */
@@ -3653,10 +3726,13 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 	struct fs_path *valid_path = NULL;
 	u64 ow_inode = 0;
 	u64 ow_gen;
+	u64 ow_mode;
 	int did_overwrite = 0;
 	int is_orphan = 0;
 	u64 last_dir_ino_rm = 0;
 	bool can_rename = true;
+	bool orphanized_dir = false;
+	bool orphanized_ancestor = false;
 
 	btrfs_debug(fs_info, "process_recorded_refs %llu", sctx->cur_ino);
 
@@ -3754,7 +3830,7 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 		 */
 		ret = will_overwrite_ref(sctx, cur->dir, cur->dir_gen,
 				cur->name, cur->name_len,
-				&ow_inode, &ow_gen);
+				&ow_inode, &ow_gen, &ow_mode);
 		if (ret < 0)
 			goto out;
 		if (ret) {
@@ -3771,6 +3847,8 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 						cur->full_path);
 				if (ret < 0)
 					goto out;
+				if (S_ISDIR(ow_mode))
+					orphanized_dir = true;
 
 				/*
 				 * If ow_inode has its rename operation delayed
@@ -3808,9 +3886,16 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 				 * might contain the pre-orphanization name of
 				 * ow_inode, which is no longer valid.
 				 */
-				fs_path_reset(valid_path);
-				ret = get_cur_path(sctx, sctx->cur_ino,
-					   sctx->cur_inode_gen, valid_path);
+				ret = is_ancestor(sctx->parent_root,
+						  ow_inode, ow_gen,
+						  sctx->cur_ino, NULL);
+				if (ret > 0) {
+					orphanized_ancestor = true;
+					fs_path_reset(valid_path);
+					ret = get_cur_path(sctx, sctx->cur_ino,
+							   sctx->cur_inode_gen,
+							   valid_path);
+				}
 				if (ret < 0)
 					goto out;
 			} else {
@@ -3869,6 +3954,18 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 				if (ret < 0)
 					goto out;
 			} else {
+				/*
+				 * We might have previously orphanized an inode
+				 * which is an ancestor of our current inode,
+				 * so our reference's full path, which was
+				 * computed before any such orphanizations, must
+				 * be updated.
+				 */
+				if (orphanized_dir) {
+					ret = update_ref_path(sctx, cur);
+					if (ret < 0)
+						goto out;
+				}
 				ret = send_link(sctx, cur->full_path,
 						valid_path);
 				if (ret < 0)
@@ -3931,6 +4028,18 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 			if (ret < 0)
 				goto out;
 			if (!ret) {
+				/*
+				 * If we orphanized any ancestor before, we need
+				 * to recompute the full path for deleted names,
+				 * since any such path was computed before we
+				 * processed any references and orphanized any
+				 * ancestor inode.
+				 */
+				if (orphanized_ancestor) {
+					ret = update_ref_path(sctx, cur);
+					if (ret < 0)
+						goto out;
+				}
 				ret = send_unlink(sctx, cur->full_path);
 				if (ret < 0)
 					goto out;
@@ -5155,15 +5264,18 @@ static int is_extent_unchanged(struct send_ctx *sctx,
 	while (key.offset < ekey->offset + left_len) {
 		ei = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
 		right_type = btrfs_file_extent_type(eb, ei);
-		if (right_type != BTRFS_FILE_EXTENT_REG) {
+		if (right_type != BTRFS_FILE_EXTENT_REG &&
+		    right_type != BTRFS_FILE_EXTENT_INLINE) {
 			ret = 0;
 			goto out;
 		}
 
-		right_disknr = btrfs_file_extent_disk_bytenr(eb, ei);
-		right_len = btrfs_file_extent_num_bytes(eb, ei);
-		right_offset = btrfs_file_extent_offset(eb, ei);
-		right_gen = btrfs_file_extent_generation(eb, ei);
+		if (right_type == BTRFS_FILE_EXTENT_INLINE) {
+			right_len = btrfs_file_extent_inline_len(eb, slot, ei);
+			right_len = PAGE_ALIGN(right_len);
+		} else {
+			right_len = btrfs_file_extent_num_bytes(eb, ei);
+		}
 
 		/*
 		 * Are we at extent 8? If yes, we know the extent is changed.
@@ -5175,6 +5287,23 @@ static int is_extent_unchanged(struct send_ctx *sctx,
 			goto out;
 		}
 
+		/*
+		 * We just wanted to see if when we have an inline extent, what
+		 * follows it is a regular extent (wanted to check the above
+		 * condition for inline extents too). This should normally not
+		 * happen but it's possible for example when we have an inline
+		 * compressed extent representing data with a size matching
+		 * the page size (currently the same as sector size).
+		 */
+		if (right_type == BTRFS_FILE_EXTENT_INLINE) {
+			ret = 0;
+			goto out;
+		}
+
+		right_disknr = btrfs_file_extent_disk_bytenr(eb, ei);
+		right_offset = btrfs_file_extent_offset(eb, ei);
+		right_gen = btrfs_file_extent_generation(eb, ei);
+
 		left_offset_fixed = left_offset;
 		if (key.offset < ekey->offset) {
 			/* Fix the right offset for 2a and 7. */
@@ -5277,6 +5406,81 @@ out:
 	return ret;
 }
 
+static int range_is_hole_in_parent(struct send_ctx *sctx,
+				   const u64 start,
+				   const u64 end)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_root *root = sctx->parent_root;
+	u64 search_start = start;
+	int ret;
+
+	path = alloc_path_for_send();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = sctx->cur_ino;
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = search_start;
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	if (ret > 0 && path->slots[0] > 0)
+		path->slots[0]--;
+
+	while (search_start < end) {
+		struct extent_buffer *leaf = path->nodes[0];
+		int slot = path->slots[0];
+		struct btrfs_file_extent_item *fi;
+		u64 extent_end;
+
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			else if (ret > 0)
+				break;
+			continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (key.objectid < sctx->cur_ino ||
+		    key.type < BTRFS_EXTENT_DATA_KEY)
+			goto next;
+		if (key.objectid > sctx->cur_ino ||
+		    key.type > BTRFS_EXTENT_DATA_KEY ||
+		    key.offset >= end)
+			break;
+
+		fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+		if (btrfs_file_extent_type(leaf, fi) ==
+		    BTRFS_FILE_EXTENT_INLINE) {
+			u64 size = btrfs_file_extent_inline_len(leaf, slot, fi);
+
+			extent_end = ALIGN(key.offset + size,
+					   root->fs_info->sectorsize);
+		} else {
+			extent_end = key.offset +
+				btrfs_file_extent_num_bytes(leaf, fi);
+		}
+		if (extent_end <= start)
+			goto next;
+		if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0) {
+			search_start = extent_end;
+			goto next;
+		}
+		ret = 0;
+		goto out;
+next:
+		path->slots[0]++;
+	}
+	ret = 1;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
 static int maybe_send_hole(struct send_ctx *sctx, struct btrfs_path *path,
 			   struct btrfs_key *key)
 {
@@ -5321,8 +5525,17 @@ static int maybe_send_hole(struct send_ctx *sctx, struct btrfs_path *path,
 			return ret;
 	}
 
-	if (sctx->cur_inode_last_extent < key->offset)
-		ret = send_hole(sctx, key->offset);
+	if (sctx->cur_inode_last_extent < key->offset) {
+		ret = range_is_hole_in_parent(sctx,
+					      sctx->cur_inode_last_extent,
+					      key->offset);
+		if (ret < 0)
+			return ret;
+		else if (ret == 0)
+			ret = send_hole(sctx, key->offset);
+		else
+			ret = 0;
+	}
 	sctx->cur_inode_last_extent = extent_end;
 	return ret;
 }
@@ -6192,8 +6405,13 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
 		goto out;
 	}
 
+	/*
+	 * Check that we don't overflow at later allocations, we request
+	 * clone_sources_count + 1 items, and compare to unsigned long inside
+	 * access_ok.
+	 */
 	if (arg->clone_sources_count >
-	    ULLONG_MAX / sizeof(*arg->clone_sources)) {
+	    ULONG_MAX / sizeof(struct clone_root) - 1) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -6242,22 +6460,16 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
 	sctx->clone_roots_cnt = arg->clone_sources_count;
 
 	sctx->send_max_size = BTRFS_SEND_BUF_SIZE;
-	sctx->send_buf = kmalloc(sctx->send_max_size, GFP_KERNEL | __GFP_NOWARN);
+	sctx->send_buf = kvmalloc(sctx->send_max_size, GFP_KERNEL);
 	if (!sctx->send_buf) {
-		sctx->send_buf = vmalloc(sctx->send_max_size);
-		if (!sctx->send_buf) {
-			ret = -ENOMEM;
-			goto out;
-		}
+		ret = -ENOMEM;
+		goto out;
 	}
 
-	sctx->read_buf = kmalloc(BTRFS_SEND_READ_SIZE, GFP_KERNEL | __GFP_NOWARN);
+	sctx->read_buf = kvmalloc(BTRFS_SEND_READ_SIZE, GFP_KERNEL);
 	if (!sctx->read_buf) {
-		sctx->read_buf = vmalloc(BTRFS_SEND_READ_SIZE);
-		if (!sctx->read_buf) {
-			ret = -ENOMEM;
-			goto out;
-		}
+		ret = -ENOMEM;
+		goto out;
 	}
 
 	sctx->pending_dir_moves = RB_ROOT;
@@ -6266,25 +6478,19 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
 
 	alloc_size = sizeof(struct clone_root) * (arg->clone_sources_count + 1);
 
-	sctx->clone_roots = kzalloc(alloc_size, GFP_KERNEL | __GFP_NOWARN);
+	sctx->clone_roots = kzalloc(alloc_size, GFP_KERNEL);
 	if (!sctx->clone_roots) {
-		sctx->clone_roots = vzalloc(alloc_size);
-		if (!sctx->clone_roots) {
-			ret = -ENOMEM;
-			goto out;
-		}
+		ret = -ENOMEM;
+		goto out;
 	}
 
 	alloc_size = arg->clone_sources_count * sizeof(*arg->clone_sources);
 
 	if (arg->clone_sources_count) {
-		clone_sources_tmp = kmalloc(alloc_size, GFP_KERNEL | __GFP_NOWARN);
+		clone_sources_tmp = kvmalloc(alloc_size, GFP_KERNEL);
 		if (!clone_sources_tmp) {
-			clone_sources_tmp = vmalloc(alloc_size);
-			if (!clone_sources_tmp) {
-				ret = -ENOMEM;
-				goto out;
-			}
+			ret = -ENOMEM;
+			goto out;
 		}
 
 		ret = copy_from_user(clone_sources_tmp, arg->clone_sources,