1y ago

Is xz 5.6.1-3+ still dangerous?

When the xz backdoor was discovered, I quickly uninstalled my Arch based setup with an infected version of the software and switched to a distro that shipped an older version (5.5 or 5.4 or something). I found an article which said that in 5.6.1-3 the backdoor was "fixed" by just not letting the malware part communicating with the vulnerable ssh related stuff and the actual malware is still there? (I didn't understand 80% of the technical terms and abbreviations in it ok?) Like it still sounds kinda dangerous to me, especially since many experts say that we don't know the other ways this malware can use (except for the ssh supply chain) yet. Is it true? Should I stick with the new distro for now or can I absolutely safely switch back and finally say that I use Arch btw again?

P. S. I do know that nothing is completely safe. Here I'm asking just about xz and libxzlk or whatever the name of that library is

EDIT: 69 upvotes. Nice

50 comments

Arch wasn't affected at all, cause the backdoor trigger was only on deb and rpm distros.
However it still a good practice to update your system and leave this version behind. Anyway, Arch already updated and is no longer distributing the backdoor version, therefore 5.6.1-3 is safe
You can use Arch btw again. Actually, you never had to leave it at first
- To add to your point: The .deb ones are most likely safe, since it would only be on the unstable & experimental branches. Your garden variety production servers & personal computers should be fine. That is unless you're into some unusual setup like with playing around with the upcoming version, or for some reason are pulling your own xz build.
  Can't speak for the .rpm tho.
  
  Fedora 39 and 40 (which is still in beta) uses xz 5.4. Fedora 41/rawhide (essentially the development branch) was affected it seems: https://access.redhat.com/security/cve/CVE-2024-3094. CentOS Stream and RHEL have way more outdated packages than that, so they were never vulnerable to this backdoor.
  openSUSE Tumbleweed (their rolling release) was affected: https://news.opensuse.org/2024/03/29/xz-backdoor/, Enterprise or Leap were unaffected.
  
  Yeah, I checked myself when this was first a thing. Debian 12 and Ubuntu 22.04 latest are on 5.4 and 5.2 respectively.
  
  "Safe" is a strong word to use. It's safe from that specific backdoor, and it seems like the known backdoor was the main goal of the attackers, but we don't know if they're playing 4D-Chess and have already implemented another backdoor which they're actively using.
- I see what you did there.
  
  What I did there? o.O
  I don't even know
- I know that Arch wasn't affected but it's only true for the known ssh backdoor. Afaik that thing can contain 100+ more "viruses" in it that we don't yet know about. And btw I was using a distro that was quite a bit different to Arch (no, not Manjaro) so idk if it was any safer than Debian sid
  
  Well, until someone find a new backdoor, I call it safe again
  I'll not lose my mental health to a potentially and unknown shady backdoor that could be installed or not in a lib
  
  Well, we don’t really know if there are backdoors in the old version as well, applying your logic
You switched distros because of this? 😆
- Yes
  
  Well, live and learn :)
TL;DR: starting with 5.6.1-2, XZ is safe on Arch. Safe as in not affected by this particular vulnerability.
Look here: https://gitlab.archlinux.org/archlinux/packaging/packages/xz/-/commit/881385757abdc39d3cfea1c3e34ec09f637424ad
And here: https://security.archlinux.org/ASA-202403-1
5.6.1-2 is where the package switched from building from the tarball (backdoored) to the upstream git repo (clean). The tarball release contained some extra build instructions (which didn't exist in the git repo) that added the backdoor during the build process. The issue arose from the downstream maintainers' assumption that the contents of the tarball and the git repo were identical.
Subsequent changes, and 5.6.1-3, were mostly administrative, like changing the git repository's URL (since the maintainer's github account was banned) and locking out Jia Tan's PGP key.
an article which said that in 5.6.1-3 the backdoor was “fixed” by just not letting the malware part communicating with the vulnerable ssh related stuff
That article is bullshit, don't believe a thing it says. Arch was not affected by the SSH vulnerability because the sshd binary did not link liblzma where the backdoor existed, so they could never communicate in a way that could be exploited by this particular vulnerability. It was not part of the fix.
- 5.6.1-2 is where the package switched from building from the tarball (backdoored) to the upstream git repo (clean)
  This is what I was looking for. Though if 5.6.1-2 doesn't contain the backdoor, why is it listed as the last version that does contain it everywhere?
  
  why is it listed as the last version that does contain it everywhere?
  I don't know, but the official advisory is most likely to be correct. Everything else is a game of Chinese whispers where the information becomes less reliable the more it is passed on. Maybe it's because -2 still had Jia Tan's signing key, and could have, theoretically, accepted commits signed by them.
  Where is it listed as such? Can you give examples?
The back door was crafted to be used by a very specific encryption key. You are are vulnerable if the attackers are specifically targeting you. If you are a tangent of a nation, you should be worried:)
- Well idk what "a tangent of a nation" means but I have political opinions very different to what my nation wants me to have so it might actually be a problem for me
  
  What they mean is if you are a affiliated with a national government. You might also be a target if you are very very rich.
  If you're an average Joe, they probably won't burn it on you.
  
  I mean, appears a country is responsible for the attack after 2 years of preparation. If they don't like you, probably was easy to send someone knock your door instead. Relax :)
When i first heard of xz backdoor, i updated my arch system and the xz-utils package to 5.6.1-3 which in the version numbers seems to be a patch, and it seems to be, so think you're safe from now
It's over. Nothing to see here. Move along.

50 comments