Bugzilla – Bug 1219397
VUL-0: kernel: use-after-free in cls_tcindex
Last modified: 2024-07-09 13:04:52 UTC
A use-after-free was found and exploited during Real World CTF 2024 in cls_tcindex. The module has been dropped upstream but we might still ship it. References: https://github.com/N1ghtu/RWCTF6th-RIPTC
The writeup is public, so we should consider this public. Maintainers, could you please double check that we ship the module? (I find CONFIG_NET_CLS_TCINDEX=m in all base branches) Furthermore, since upstream dropped the code, they might not be interested in working on a fix.
Environment that was targeted: https://github.com/chaitin/Real-World-CTF-6th-Challenges/tree/main/RIPTC https://github.com/chaitin/Real-World-CTF-6th-Challenges/releases/download/x/riptc_attachment_241a4f7b8921b131e3237af987ad4f82.tar.gz
dup of https://bugzilla.suse.com/show_bug.cgi?id=1210335 ?
(In reply to Marcus Meissner from comment #3) > dup of https://bugzilla.suse.com/show_bug.cgi?id=1210335 ? I don't think so. The code provided was the one just before being dropped. Thus including all previous CVE fixes.
(In reply to Thomas Leroy from comment #4) > (In reply to Marcus Meissner from comment #3) > > dup of https://bugzilla.suse.com/show_bug.cgi?id=1210335 ? > > I don't think so. The code provided was the one just before being dropped. > Thus including all previous CVE fixes. Yes, I agree. And yes, we ship tcindex.
This is going to be "interesting". We ship with CONFIG_NET_CLS_TCINDEX in all our kernels upto SLE15-SP5 (inclusive). Since the driver is not upstream, we cannot really hope for an upstream fix. In bug 1210335 Michal Koutny was mentioning blacklisting cls_tcindex module which seems as the sanest course for the future. Assigning to Michal since he seems to be dealing with blacklisting of cls_tcindex...