On Wed, Jun 18, 2025 at 11:30:48AM +0800, Edward Adam Davis wrote: > The reproducer uses a file0 on a ntfs3 file system with a corrupted i_link. > When renaming, the file0's inode is marked as a bad inode because the file > name cannot be deleted. However, before renaming, file0 is a directory. > After the renaming fails, it is marked as a bad inode, which makes it a > regular file. In any case, when opening it after creating a hard link, > pick_link() should not be entered because it is not a symbolic link from > beginning to end. > > Add a check on the symbolic link before entering pick_link() to avoid > triggering unknown exceptions when performing the i_link acquisition > operation on other types of files. > > Reported-by: syzbot+1aa90f0eb1fc3e77d969@xxxxxxxxxxxxxxxxxxxxxxxxx > Closes: https://syzkaller.appspot.com/bug?extid=1aa90f0eb1fc3e77d969 > Tested-by: syzbot+1aa90f0eb1fc3e77d969@xxxxxxxxxxxxxxxxxxxxxxxxx > Signed-off-by: Edward Adam Davis <eadavis@xxxxxx> NAK. This is not the first time that garbage is suggested and no, we are not going to paper over that shite in fs/namei.c. Not going to happen. You ARE NOT ALLOWED to call make_bad_inode() on a live inode, period. Never, ever to be done. There's a lot of assertions it violates and there's no chance in hell to plaster each with that kind of checks. Fix NTFS. End of story.
