Re: 2.6.22-rc6(mm1) Unable to handle kernel NULL pointer dereference - git-bisect result

[Linus Torvalds <torvalds@linux-foundation.org>]

On Sun, 8 Jul 2007, Markus Trippelsdorf wrote:
> 
> I tested this further and it turned out that the Linus tree is also
> affected. So I ran git-bisect, after I found out that version
> 2.6.21.6 was not affected by this bug.

git-bisect is wonderful.

> gentoox2 linux # git bisect bad
> 1c710c896eb461895d3c399e15bb5f20b39c9073 is first bad commit
> commit 1c710c896eb461895d3c399e15bb5f20b39c9073
> Author: Ulrich Drepper <drepper@redhat.com>

Ok, I see what's wrong.

We're doing "vfs_permission()" and passing in "nd" to it when "times" is 
NULL, but Uli didn't even *initialize* it for the case of

	filename == NULL && dfd != AT_FDCWD

so yeah, it's broken.

I was planning on doing 2.6.22 today, so I'm actually inclined to revert 
the commit entirely, even though it looks like it's probably fairly 
trivial to fix.

But if we get a really quick and obvious fix, I guess I'll apply it. I 
have to say, the new do_utimes() code is pretty ugly compared to the old 
one exactly in the added special cases (the ones that caused this bug..)

Uli?

		Linus