Merge tag 'for-5.9-rc2-tag' of git://git.kernel.org/pub/scm/linux/ker…

…nel/git/kdave/linux

Pull btrfs fixes from David Sterba:

 - fix swapfile activation on subvolumes with deleted snapshots

 - error value mixup when removing directory entries from tree log

 - fix lzo compression level reset after previous level setting

 - fix space cache memory leak after transaction abort

 - fix const function attribute

 - more error handling improvements

* tag 'for-5.9-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: detect nocow for swap after snapshot delete
  btrfs: check the right error variable in btrfs_del_dir_entries_in_log
  btrfs: fix space cache memory leak after transaction abort
  btrfs: use the correct const function attribute for btrfs_get_num_csums
  btrfs: reset compression level for lzo on remount
  btrfs: handle errors from async submission

fs/btrfs/ctree.c

@@ -68,7 +68,7 @@ const char *btrfs_super_csum_driver(u16 csum_type)
 		btrfs_csums[csum_type].name;
 }
 
-size_t __const btrfs_get_num_csums(void)
+size_t __attribute_const__ btrfs_get_num_csums(void)
 {
 	return ARRAY_SIZE(btrfs_csums);
 }

fs/btrfs/ctree.h

@@ -2262,7 +2262,7 @@ BTRFS_SETGET_STACK_FUNCS(super_uuid_tree_generation, struct btrfs_super_block,
 int btrfs_super_csum_size(const struct btrfs_super_block *s);
 const char *btrfs_super_csum_name(u16 csum_type);
 const char *btrfs_super_csum_driver(u16 csum_type);
-size_t __const btrfs_get_num_csums(void);
+size_t __attribute_const__ btrfs_get_num_csums(void);
 
 
 /*
@@ -2518,7 +2518,7 @@ int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans,
 				    u64 bytenr, u64 num_bytes);
 int btrfs_exclude_logged_extents(struct extent_buffer *eb);
 int btrfs_cross_ref_exist(struct btrfs_root *root,
-			  u64 objectid, u64 offset, u64 bytenr);
+			  u64 objectid, u64 offset, u64 bytenr, bool strict);
 struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
 					     struct btrfs_root *root,
 					     u64 parent, u64 root_objectid,
@@ -2934,7 +2934,7 @@ struct extent_map *btrfs_get_extent_fiemap(struct btrfs_inode *inode,
 					   u64 start, u64 len);
 noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
 			      u64 *orig_start, u64 *orig_block_len,
-			      u64 *ram_bytes);
+			      u64 *ram_bytes, bool strict);
 
 void __btrfs_del_delalloc_inode(struct btrfs_root *root,
 				struct btrfs_inode *inode);

fs/btrfs/disk-io.c

@@ -4551,6 +4551,7 @@ static void btrfs_cleanup_bg_io(struct btrfs_block_group *cache)
 		cache->io_ctl.inode = NULL;
 		iput(inode);
 	}
+	ASSERT(cache->io_ctl.pages == NULL);
 	btrfs_put_block_group(cache);
 }
 

fs/btrfs/extent-tree.c

@@ -2306,7 +2306,8 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
 
 static noinline int check_committed_ref(struct btrfs_root *root,
 					struct btrfs_path *path,
-					u64 objectid, u64 offset, u64 bytenr)
+					u64 objectid, u64 offset, u64 bytenr,
+					bool strict)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_root *extent_root = fs_info->extent_root;
@@ -2348,9 +2349,13 @@ static noinline int check_committed_ref(struct btrfs_root *root,
 	    btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY))
 		goto out;
 
-	/* If extent created before last snapshot => it's definitely shared */
-	if (btrfs_extent_generation(leaf, ei) <=
-	    btrfs_root_last_snapshot(&root->root_item))
+	/*
+	 * If extent created before last snapshot => it's shared unless the
+	 * snapshot has been deleted. Use the heuristic if strict is false.
+	 */
+	if (!strict &&
+	    (btrfs_extent_generation(leaf, ei) <=
+	     btrfs_root_last_snapshot(&root->root_item)))
 		goto out;
 
 	iref = (struct btrfs_extent_inline_ref *)(ei + 1);
@@ -2375,7 +2380,7 @@ static noinline int check_committed_ref(struct btrfs_root *root,
 }
 
 int btrfs_cross_ref_exist(struct btrfs_root *root, u64 objectid, u64 offset,
-			  u64 bytenr)
+			  u64 bytenr, bool strict)
 {
 	struct btrfs_path *path;
 	int ret;
@@ -2386,7 +2391,7 @@ int btrfs_cross_ref_exist(struct btrfs_root *root, u64 objectid, u64 offset,
 
 	do {
 		ret = check_committed_ref(root, path, objectid,
-					  offset, bytenr);
+					  offset, bytenr, strict);
 		if (ret && ret != -ENOENT)
 			goto out;
 

fs/btrfs/file.c

@@ -1571,7 +1571,7 @@ static int check_can_nocow(struct btrfs_inode *inode, loff_t pos,
 	}
 
 	ret = can_nocow_extent(&inode->vfs_inode, lockstart, &num_bytes,
-			NULL, NULL, NULL);
+			NULL, NULL, NULL, false);
 	if (ret <= 0) {
 		ret = 0;
 		if (!nowait)

fs/btrfs/free-space-cache.c

@@ -1186,7 +1186,6 @@ static int __btrfs_wait_cache_io(struct btrfs_root *root,
 	ret = update_cache_item(trans, root, inode, path, offset,
 				io_ctl->entries, io_ctl->bitmaps);
 out:
-	io_ctl_free(io_ctl);
 	if (ret) {
 		invalidate_inode_pages2(inode->i_mapping);
 		BTRFS_I(inode)->generation = 0;
@@ -1347,6 +1346,7 @@ static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
 	 * them out later
 	 */
 	io_ctl_drop_pages(io_ctl);
+	io_ctl_free(io_ctl);
 
 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
 			     i_size_read(inode) - 1, &cached_state);

fs/btrfs/inode.c

@@ -1610,7 +1610,7 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 				goto out_check;
 			ret = btrfs_cross_ref_exist(root, ino,
 						    found_key.offset -
-						    extent_offset, disk_bytenr);
+						    extent_offset, disk_bytenr, false);
 			if (ret) {
 				/*
 				 * ret could be -EIO if the above fails to read
@@ -2161,11 +2161,8 @@ static blk_status_t btrfs_submit_bio_start(void *private_data, struct bio *bio,
 				    u64 bio_offset)
 {
 	struct inode *inode = private_data;
-	blk_status_t ret = 0;
 
-	ret = btrfs_csum_one_bio(BTRFS_I(inode), bio, 0, 0);
-	BUG_ON(ret); /* -ENOMEM */
-	return 0;
+	return btrfs_csum_one_bio(BTRFS_I(inode), bio, 0, 0);
 }
 
 /*
@@ -6953,6 +6950,8 @@ static struct extent_map *btrfs_new_extent_direct(struct btrfs_inode *inode,
  * @orig_start:	(optional) Return the original file offset of the file extent
  * @orig_len:	(optional) Return the original on-disk length of the file extent
  * @ram_bytes:	(optional) Return the ram_bytes of the file extent
+ * @strict:	if true, omit optimizations that might force us into unnecessary
+ *		cow. e.g., don't trust generation number.
  *
  * This function will flush ordered extents in the range to ensure proper
  * nocow checks for (nowait == false) case.
@@ -6967,7 +6966,7 @@ static struct extent_map *btrfs_new_extent_direct(struct btrfs_inode *inode,
  */
 noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
 			      u64 *orig_start, u64 *orig_block_len,
-			      u64 *ram_bytes)
+			      u64 *ram_bytes, bool strict)
 {
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
 	struct btrfs_path *path;
@@ -7045,8 +7044,9 @@ noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
 	 * Do the same check as in btrfs_cross_ref_exist but without the
 	 * unnecessary search.
 	 */
-	if (btrfs_file_extent_generation(leaf, fi) <=
-	    btrfs_root_last_snapshot(&root->root_item))
+	if (!strict &&
+	    (btrfs_file_extent_generation(leaf, fi) <=
+	     btrfs_root_last_snapshot(&root->root_item)))
 		goto out;
 
 	backref_offset = btrfs_file_extent_offset(leaf, fi);
@@ -7082,7 +7082,8 @@ noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
 	 */
 
 	ret = btrfs_cross_ref_exist(root, btrfs_ino(BTRFS_I(inode)),
-				    key.offset - backref_offset, disk_bytenr);
+				    key.offset - backref_offset, disk_bytenr,
+				    strict);
 	if (ret) {
 		ret = 0;
 		goto out;
@@ -7303,7 +7304,7 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
 		block_start = em->block_start + (start - em->start);
 
 		if (can_nocow_extent(inode, start, &len, &orig_start,
-				     &orig_block_len, &ram_bytes) == 1 &&
+				     &orig_block_len, &ram_bytes, false) == 1 &&
 		    btrfs_inc_nocow_writers(fs_info, block_start)) {
 			struct extent_map *em2;
 
@@ -7619,10 +7620,8 @@ static blk_status_t btrfs_submit_bio_start_direct_io(void *private_data,
 				    struct bio *bio, u64 offset)
 {
 	struct inode *inode = private_data;
-	blk_status_t ret;
-	ret = btrfs_csum_one_bio(BTRFS_I(inode), bio, offset, 1);
-	BUG_ON(ret); /* -ENOMEM */
-	return 0;
+
+	return btrfs_csum_one_bio(BTRFS_I(inode), bio, offset, 1);
 }
 
 static void btrfs_end_dio_bio(struct bio *bio)
@@ -10136,7 +10135,7 @@ static int btrfs_swap_activate(struct swap_info_struct *sis, struct file *file,
 		free_extent_map(em);
 		em = NULL;
 
-		ret = can_nocow_extent(inode, start, &len, NULL, NULL, NULL);
+		ret = can_nocow_extent(inode, start, &len, NULL, NULL, NULL, true);
 		if (ret < 0) {
 			goto out;
 		} else if (ret) {

fs/btrfs/super.c

@@ -625,6 +625,7 @@ int btrfs_parse_options(struct btrfs_fs_info *info, char *options,
 			} else if (strncmp(args[0].from, "lzo", 3) == 0) {
 				compress_type = "lzo";
 				info->compress_type = BTRFS_COMPRESS_LZO;
+				info->compress_level = 0;
 				btrfs_set_opt(info->mount_opt, COMPRESS);
 				btrfs_clear_opt(info->mount_opt, NODATACOW);
 				btrfs_clear_opt(info->mount_opt, NODATASUM);

fs/btrfs/tree-log.c

@@ -3449,11 +3449,13 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	btrfs_free_path(path);
 out_unlock:
 	mutex_unlock(&dir->log_mutex);
-	if (ret == -ENOSPC) {
+	if (err == -ENOSPC) {
 		btrfs_set_log_full_commit(trans);
-		ret = 0;
-	} else if (ret < 0)
-		btrfs_abort_transaction(trans, ret);
+		err = 0;
+	} else if (err < 0 && err != -ENOENT) {
+		/* ENOENT can be returned if the entry hasn't been fsynced yet */
+		btrfs_abort_transaction(trans, err);
+	}
 
 	btrfs_end_log_trans(root);