Re: Linux 2.6.30+/SELinux/RHEL5 test kernel 0day, exploiting the unexploitable

[Solar Designer <solar () openwall com>]

Earlier today, I wrote:

> (BTW, I'd be happy to share the mmap_min_addr back-port patch to
> RHEL-5'ish kernels with Red Hat if there's any interest.)

It occurred to me, from a few things I've seen/read lately, that RHEL 5
must indeed already include a back-port of the mmap_min_addr feature,
although somehow it is not in OpenVZ's patch-128.1.1.el5.028stab062.3
(was added after the -128 revision?), which is why I am adding my own
back-port of mmap_min_addr on top of that patch.  I am not using genuine
RHEL 5 (nor CentOS), which is why I don't know.  Maybe Eugene can
comment on this.

> I am going to release 2.4.37.3-ow1 with a CVE-2009-1895 fix in it, and I
> expect it to get into 2.4.37.4.  It's not important for systems with
> "sane" userlands (no crappy SUID-root programs), though.

Released, although I expect it to be replaced with 2.4.37.4-ow1 soon:

http://www.openwall.com/linux/

Alexander