btrfs: remove wrong comment for can_nocow_extent()

The comment for can_nocow_extent() says that the function will flush
ordered extents, however that never happens and was never true before the
comment was added in commit e4ecaf9 ("btrfs: add comments for
btrfs_check_can_nocow() and can_nocow_extent()"). This is true only for
the function btrfs_check_can_nocow(), which after that commit was renamed
to check_can_nocow(). So just remove that part of the comment.

Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

fs/btrfs/inode.c

@@ -7105,9 +7105,6 @@ static struct extent_map *btrfs_new_extent_direct(struct btrfs_inode *inode,
  * @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.
- *
  * Return:
  * >0	and update @len if we can do nocow write
  *  0	if we can't do nocow write