FatGid: FreeBSD 14.x kernel local privilege escalation
98 points by WhyNotHugo a day ago | 41 comments
  • socphoenix a day ago |
    Not sure why this is saying it isn’t patched, they released the notice including fix for 14.4 yesterday?
    • irishcoffee 21 hours ago |
      Maybe they're not up to snuff on yesterday? They published this yesterday.

      > The bug was silently fixed in the main branch on 2025-11-27 (commit 000d5b52c19ff3858a6f0cbb405d47713c4267a4) as a side effect of a broader function refactoring. The fix has not been backported to stable/14 or releng/14.4. FreeBSD 14.4-RELEASE remains vulnerable.

      > FreeBSD 15.0 still carries the sizeof(*groups) typo and is therefore vulnerable, but the surrounding code differs enough from 14.4 that the chain primitives developed here do not lift the overflow into a working LPE on that branch. On 15.0 the bug remains a kernel panic triggered by any unprivileged user.

  • turkeyboi a day ago |
    Why does this need to be a whole ass website
    • dragontamer 21 hours ago |
      What?

      Is there something in this website that feels unnecessary? It seems like a good format of sharing high quality information.

      This looks like a full bug into a complete root escalation of a kernel. That's hard to do and deserving of praise. The fact that we have a writeup organized like this is awesome.

      -------

      This is sort of the expert level stuff that I thought HackerNews would most enjoy.

    • cryo32 21 hours ago |
      You're not going to get anywhere in the security sector unless you gain notoriety i.e. are noticed.

      This appears to come from dressing up like Elton John in a feather suit and hiring a marketing team.

      • tptacek 21 hours ago |
        It's a wall of text about a kernel stack overflow. I'm not sure where the "Elton John" part is. Is it... that they used an accent color?
        • 866-RON-0-FEZ 19 hours ago |
          Maybe the researcher was wearing windshield-wiper spectacles when he discovered the vulnerability.

          I don't understand why you're being so defensive about this.

          • tptacek 19 hours ago |
            Because it's a tiresome, tropey, and ultimately invalid complaint. Look downthread at the person who said the FreeBSD commit log was better than this page, despite being inscrutable to security practitioners who don't work in the kernel and not saying a word about proven exploit vectors.

            These complaints aren't about what's better or worse for the user community; they're about people trying to put vulnerability researchers in their place.

            • 866-RON-0-FEZ 19 hours ago |
              While I believe whimsical names will always be silly, I do concede that commit log is effectively useless to 99% of eyeballs.
              • tptacek 19 hours ago |
                It's not even a complete description of the vulnerability. It's what the kernel maintainers need to know to understand and fix the bug in the code. The claim that it's superior to the branded vulnerability page gives away the whole game.
            • cryo32 14 hours ago |
              As a member of the user community, I only care when it appears on the security list and then that it's patched. And I want to see that on the changelog. And that's about it.

              I don't want or need fanfare, marketing or any of that stuff.

              It's a bug for fucks sake. There will be people having web pages for Gnome user interface bugs next.

        • GoblinSlayer 18 hours ago |
          Buffer overflow.
    • tptacek 21 hours ago |
      Why not? This weird complaint has been happening since ~2010 and it has never made any sense. You are strictly better off with the website than without it. When it was vulnerability researchers getting all peevish about the status competition they were running, I at least understood where the complaint was coming from, but even among practitioners, branded vulnerabilities are so much the norm at this point that there's no status implication anymore.
      • themafia 19 hours ago |
        > You are strictly better off with the website than without it.

        Why? This is a better resource in every way: https://cgit.freebsd.org/src/commit/?id=000d5b52c19ff3858a6f...

        It details the actual problem instead of showing off tired stack exploit tricks.

        • tptacek 19 hours ago |
          No, that commit log is obviously not better than the page explaining the vulnerability and the exploit vectors.

          Case in point: what's "tired" about the stack exploitation techniques they're using here?

          And, while you're not right, even stipulating that you were, what would that matter? How is anyone better off with less explanation of a vulnerability?

          • themafia 18 hours ago |
            The website explains an exploit. I've seen exploits before. The commit explains the unique issue.

            I'm more interested in the why than the how.

            I suppose people with different overall goals will see that differently.

            • tptacek 17 hours ago |
              You didn't answer my question. What's "tired" about the exploit technique here?
        • tom_ 18 hours ago |
          Is that even the fix though? The problem sizeof*groups expression has already been removed by that point. This fixes something but it's not obviously related to the vulnerability description.

          git log -S suggests 4cd93df95e697942adf0ff038fc8f357cbb07cf9, which looks more likely: https://cgit.freebsd.org/src/commit/?id=4cd93df95e697942adf0... - though not to say you don't want the later commit too. I'm sure you do.

    • rs_rs_rs_rs_rs 20 hours ago |
      Have you not done anything remotely interesting for you that you want to build a website so the whole world can see it?
    • 866-RON-0-FEZ 19 hours ago |
      It gives legitimacy to whatever whimsical name was given to a vulnerability by registering the domain.

      CVE numbers are for boring professionals.

      • elkrapo 19 hours ago |
        The cool kiddies wait and time their disclosures on the cool numbers.
  • djha-skin 21 hours ago |
    TrueNAS is on FreeBSD, as well as lots of network equipment. This does affect us more than we think as operators.
    • ActionHank 21 hours ago |
      Possibly Playstation as well.
      • sbankowi 20 hours ago |
        Also Netgate's devices running PFSense.
        • anygivnthursday 19 hours ago |
          And OPNSense boxes
      • loeg 18 hours ago |
        PlayStation 4 was a fork of FreeBSD 9, and is immune to this bug introduced in 14. Sony also changes a LOT, I'm not sure anything dealing with unix credentials even exists in this fork. It's not clear how much FreeBSD is even used in PlayStation 5 (2020), but it would be based off 12 or earlier (also immune to this bug from 14) (13 was released in 2021).
        • HDBaseT 11 hours ago |
          What about PlayStation 3?

          Whilst all are "soft-moddable" via HEN, a large number of the Slim and Superslims are not compatible with full custom firmware. Recently a hardware based exploit 'badWDSD' released which allow CFW, although even still a small number of Superslims are not compatible.

          • loeg 10 hours ago |
            I thought it went without saying that PS3 (2006) firmware isn't based on FreeBSD 14 (2023).
    • yjftsjthsd-h 21 hours ago |
      I would think that pure-storage NAS or network equipment was effectively completely immune to local privilege escalation. I'll give you the NAS where it might be running untrusted containers or such, but that's it.
    • 1over137 20 hours ago |
      Alas, TrueNAS actually switched to Linux a couple of years ago.
      • sbankowi 20 hours ago |
        FreeBSD was the reason I chose TrueNAS Core. Unfortunately, you are right, TrueNAS Scale (Linux) is where they are focusing all their attention. At this point I will not purchase additional TrueNAS equipment as I feel I was "rug pulled." I get that they are going after more of the Docker container/app market, but I just want a solid ZFS w/excellent networking NAS device. Linux is close to this ideal, but it isn't as "Set and Forget" as FreeBSD (IMO).
        • BadBadJellyBean 19 hours ago |
          You usually don't really interact with the OS underneath at all so I don't think it makes much of a difference unless you are very fond of Jails.

          I mean that is the whole point of a NAS OS. It gives you a GUI and you don't have to worry about the rest.

        • NexRebular 16 hours ago |
          > solid ZFS w/excellent networking NAS device.

          illumos distros might be a good alternative. I have OmniOS[0] as a filer and SmartOS[1] running hypervisor duties on zones and bhyve.

          [0] https://omnios.org [1] https://docs.smartos.org

    • 866-RON-0-FEZ 19 hours ago |
      Juniper JunOS is based on FreeBSD IIRC.
      • sophacles 17 hours ago |
        They've been moving their NOSes to a linux based platform.
        • NexRebular 16 hours ago |
          It really is a tragedy how everything is being monoculturized to linux.
    • doublerabbit 18 hours ago |
      TrueNAS was, but they now use Linux as the main distribution.
    • ksec 27 minutes ago |
      > This does affect us more than we think as operators.

      I hope this is true but as other comments have suggested, Juniper and TrueNAS are moving or no longer on FreeBSD.

      I am now wondering if Netflix may one day give up FreeBSD for their cache server as well.

  • badgersnake 17 hours ago |
    LPEs are really not impressive enough to warrant names and websites.
  • CodeCompost 16 hours ago |
    I'm curious to see how many issues an LLM can find for OpenBSD. Not knocking OpenBSD at all, genuinely curious to see if all the careful development can stand up to LLMs.