Yet another critical vulnerability in systemd, this time involving snapd. Ubuntu folk are affected.

"A serious security issue has been discovered in Ubuntu, and it is gaining attention in the cybersecurity community. The vulnerability is identified as CVE-2026-3888 and mainly affects Ubuntu Desktop systems from version 24.04 onwards. This flaw is dangerous because it allows an attacker with limited access to gain full root privileges. Root access means complete control over the entire system."

all 28 comments

sorted by: hot top controversial new old
[–] 22 points 7 months ago (3 children)

Not a very good article. The original write-up (not linked anywhere in the article) is here: https://blog.qualys.com/vulnerabilities-threat-research/2026/03/17/cve-2026-3888-important-snap-flaw-enables-local-privilege-escalation-to-root

They also mention something else that's interesting at the bottom of the write-up:

Secondary Finding: Vulnerability in Ubuntu 25.10 uutils Coreutils

In a proactive security effort prior to the release of Ubuntu Desktop 25.10, the Qualys Threat Research Unit assisted the Ubuntu Security Team in reviewing the uutils coreutils package (a Rust rewrite of standard GNU utilities).

A race condition in the rm utility allowed an unprivileged local attacker to replace directory entries with symlinks during root-owned cron executions (specifically /etc/cron.daily/apport). Successful exploitation could lead to arbitrary file deletion as root or further privilege escalation by targeting snap sandbox directories.

The vulnerability was reported and mitigated prior to the public release of Ubuntu 25.10. The default rm command in Ubuntu 25.10 was reverted to GNU coreutils to mitigate this risk immediately. Upstream fixes have since been applied to the uutils repository.

  • source
  • hideshow 3 child comments
  • [–] 40 points 7 months ago* (5 children)

    Yet another critical vulnerability in systemd

    This is a critical vulnerability in snapd, not systemd. It sounds like it could also be exploited if something other than systemd deleted the files in /tmp/. Or if /tmp/ was not mounted.

  • source
  • hideshow 5 child comments
  • [+] [S] -14 points 7 months ago (2 children)

    Yet another critical vulnerability in the much vaunted systemd has been exposed by a misbehaving app - in this case snapd.

    Both need patching.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 19 points 7 months ago (1 child)

    Nothing in the Qualys report nor the Ubuntu page for the CVE indicate that there is something to be fixed in systemd, only that you can create systemd-tmpfiles rules that will expose the vulnerability in snapd.

    What do you think systemd-tmpfiles needs to do differently?

  • source
  • parent
  • hideshow 1 child comment
  • [–] [S] -2 points 7 months ago

    Read the bug in the systemd repo. meanwhile the systemd cultists will defend it to the death. Look, if you value the positive aspects of systemd but dislike the cancerous assimilation of the entire userland, dinit is a perfectly good option.

  • source
  • parent
  • [–] 17 points 7 months ago

    snap

    well there's your problem

  • source
  • [–] 16 points 7 months ago* (4 children)

    This is not a systemd flaw. This is a snap bug.

    e:f;b

  • source
  • hideshow 4 child comments
  • [–] [S] 0 points 7 months ago (3 children)

    Any errant application can expose this glaring systemd flaw

  • source
  • parent
  • hideshow 3 child comments
  • [–] 2 points 7 months ago* (2 children)

    There is no systemd flaw here.

    snap-confine creates /tmp/.snap owned by root.

    systemd-tmpfiles can delete this directory because it also has root privileges. It will do so if the directory is inactive for, by default, 30 days. Files can be excluded from this by adding a .conf file to /etc/tmpfiles.d/, snap-confine does not do this.

    Because the files are not excluded they will be deleted. systemd-tmpfiles can do this because it is running as root.

    Once they are deleted a USER can recreate /tmp/.snap with malicious code.

    snap-confine never verifies that the directory is owned by root, and performs its security checks before its privileged file operations, creating a race window. Because snap-confine is setuid root, it then bind-mounts files from the attacker-controlled /tmp/.snap into the snap sandbox's filesystem, allowing an attacker to execute arbitrary code as root.

    What is the systemd-tmpfiles flaw? It does exactly what it is supposed to do, and it provides a means to exclude directories from its process. snap doesn't configure systemd to ignore the directories and it doesn't perform appropriate checks on the directory's ownership.

  • source
  • parent
  • hideshow 2 child comments
  • [–] [S] -1 points 7 months ago* (1 child)

    Exactly. systemd has a glaring security hole that had to be kludged. As the OP I posted the article to warn non technical users of the danger but systemd defenders league are predictably blind to any possible flaw in their golden calf and cannot resist the temptation to rush to battle. yawn

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 7 months ago*

    What is the security hole in systemd?

    You haven't answered this.

    systemd-tmpfiles exists to delete inactive files in /tmp. That's not a security flaw, that's system maintenance. It's the documented purpose of systemd-tmpfiles and it performs exactly was documented.

    The security hole is in snap-confine which does not verify that its own directory is owned by root before mounting it AS ROOT. That's the security hole.

    So, again, what is the security hole in systemd?

    If you don't have an answer then just say so. Resorting to name calling and trying to frame this as if I'm the irrational one is absurd.

  • source
  • parent
  • [–] 7 points 7 months ago (5 children)

    When I need to create scratch files I usually operate in /tmp. Almost all directories there that I saw were using randomized paths (e.g. UUIDs). I guess this is to prevent problems mentioned in the article. So, I believe this would be a vulnerability of snap, not systemd.

    I use Fedora where /tmp is created as tmpfs, which lives in RAM and is cleared when the system is shut down. I wonder what's the benefit of Ubuntu's approach.

  • source
  • hideshow 5 child comments
  • [–] [S] -4 points 7 months ago (4 children)

    If you think about it for even a minute this is still a glaring cve in systemd, exposed in this case, by misbehaving snapd. systemd still needed to be patched and so did snapd.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 10 points 7 months ago (2 children)

    No Dylan, don't bother fixing this shit, go straight for the boot licking commit.

  • source
  • hideshow 2 child comments
  • [–] 2 points 7 months ago* (1 child)

    how would i know if Kubuntu 25.10 is affected (based on ubuntu)?
    i guess this means yes?
    command 'snap' from deb snapd (2.73+ubuntu25.10.1)
    as it is lower than the version mentioned in the article "Upstream snapd: versions prior to 2.75"

    now the question is how do i force an update on that thing?
    sudo apt upgrade did not include an update for snapd:

    Upgrading:
    bpftool linux-headers-generic linux-libc-dev linux-tools-common
    linux-generic linux-image-generic linux-perf

    Installing dependencies:
    linux-headers-6.17.0-20 linux-image-6.17.0-20-generic linux-tools-6.17.0-20
    linux-headers-6.17.0-20-generic linux-modules-6.17.0-20-generic linux-tools-6.17.0-20-generic

    Suggested packages:
    linux-tools

    Not upgrading yet due to phasing:
    fwupd libfwupd3

    Summary:
    Upgrading: 7, Installing: 6, Removing: 0, Not Upgrading: 2
    Download size: 212 MB
    Space needed: 421 MB / 417 GB available

    edit:
    i tried sudo apt install snapd but it returned:

    snapd is already the newest version (2.73+ubuntu25.10.1).
    snapd set to manually installed.

    edit2:
    or am i save because of this?:

    Ubuntu 25.10 LTS: snapd versions prior to 2.73+ubuntu25.10.1

  • source
  • hideshow 1 child comment
  • [–] 1 point 7 months ago*

    "Systemd is built badly by weaponized dunning-kruger" -- pros

    exploit [happens]

    World: surprised pikachu

  • source