squashfs: verify inode mode when loading from disk

The inode mode loaded from corrupted disk might by error contain the file
type bits.  Since the file type bits are set by squashfs_read_inode()
using bitwise OR, the file type bits must not be set by
squashfs_new_inode() from squashfs_read_inode(); otherwise, an invalid
file type bits later confuses may_open().

Link: https://lkml.kernel.org/r/[email protected]
Reported-by: syzbot <[email protected]>
Closes: https://syzkaller.appspot.com/bug?extid=895c23f6917da440ed0d
Signed-off-by: Tetsuo Handa <[email protected]>
Reviewed-by: Phillip Lougher <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>

Author: Tetsuo Handa

fs/squashfs/inode.c

@@ -68,6 +68,10 @@ static int squashfs_new_inode(struct super_block *sb, struct inode *inode,
 	inode->i_mode = le16_to_cpu(sqsh_ino->mode);
 	inode->i_size = 0;
 
+	/* File type must not be set at this moment, for it will later be set by the caller. */
+	if (inode->i_mode & S_IFMT)
+		err = -EIO;
+
 	return err;
 }