Arch Linux

Jacek Poplawski

Member

From: Poland

Registered: 2006-01-10

Posts: 736

Website

2.6.24.1

Will 2.6.24.1 or exploit patch be in [core] as soon as possible, or should we compile our own kernel today?

shining

Pacman Developer

Registered: 2006-05-10

Posts: 2,043

Re: 2.6.24.1

The information was already put everywhere :
http://bbs.archlinux.org/viewtopic.php?id=43701
http://archlinux.org/news/383/
http://www.archlinux.org/pipermail/arch … 16797.html

The bug is only really fixed by 2.6.24.2 , but arch package includes a pre version of that release, so it's fine.

---

pacman roulette : pacman -S $(pacman -Slq | LANG=C sort -R | head -n $((RANDOM % 10)))

Jacek Poplawski

Member

From: Poland

Registered: 2006-01-10

Posts: 736

Website

Re: 2.6.24.1

Great! Thanks.

broch

Banned

From: L.A. California

Registered: 2006-11-13

Posts: 975

Re: 2.6.24.1

not so fast..
http://bugs.debian.org/cgi-bin/bugrepor … =464953#14

pre-2.6.24.2
hotfix seems to cause problems

there are two fixes and only the latter seems to be safe:

Just for the record: Do not use the "hotfix" named disable-vmsplice-if-
exploitable.c. The hotfix first tries to run the exploit (which would be
totally unnecessary for the actual "fix" by the way and is therefore a
very dumb thing to do), and this still leads to kernel memory corruption
which will render the system unstable. You can imagine what might come
from corrupted kernel beside a simple crash (e.g. data loss).

I don't know which is applied in Arch kernel?

jacko

Member

Registered: 2007-11-23

Posts: 840

Re: 2.6.24.1

http://cvs.archlinux.org/cgi-bin/viewcv … text/plain

Here is the patch, doesn't look like they added the disable-vmsplice-if-
exploitable.c hotfix. I do not see it listed anywhere in the patch.

Last edited by jacko (2008-02-11 17:49:17)

shining

Pacman Developer

Registered: 2006-05-10

Posts: 2,043

Re: 2.6.24.1

Uhm, this thing is not a kernel patch, it's an external program (just the exploit slightly modified to fix the issue rather than exploit it).
That's ugly, and it's not something that could have been applied to the arch kernel. It's a program you could have run manually after each boot.

The real fix is the the first change in the patch linked above by jacko.
It's very similar to the one that apparently got applied in 2.6.24.2 :
http://git.kernel.org/?p=linux/kernel/g … 42272fd063

Both are likely equivalent.

---

pacman roulette : pacman -S $(pacman -Slq | LANG=C sort -R | head -n $((RANDOM % 10)))