Stories
Slash Boxes
Comments

SoylentNews is people

posted by Fnord666 on Monday July 03 2017, @10:12PM   Printer-friendly
from the it's-a-feature dept.

Submitted via IRC for TheMightyBuzzard

A bug in Linux's systemd init system causes root permissions to be given to services associated with invalid usernames, and while this could pose a security risk, exploitation is not an easy task.

A developer who uses the online moniker "mapleray" last week discovered a problem related to systemd unit files, the configuration files used to describe resources and their behavior. Mapleray noticed that a systemd unit file containing an invalid username – one that starts with a digit (e.g. "0day") – will initiate the targeted process with root privileges instead of regular user privileges.

Systemd is designed not to allow usernames that start with a numeric character, but Red Hat, CentOS and other Linux distributions do allow such usernames.

"It's systemd's parsing of the User= parameter that determines the naming doesn't follow a set of conventions, and decides to fall back to its default value, root," explained developer Mattias Geniar.

While this sounds like it could be leveraged to obtain root privileges on any Linux installation using systemd, exploiting the bug in an attack is not an easy task. Geniar pointed out that the attacker needs root privileges in the first place to edit the systemd unit file and use it.

[...] Systemd developers have classified this issue as "not-a-bug" and they apparently don't plan on fixing it. Linux users are divided on the matter – some believe this is a vulnerability that could pose a serious security risk, while others agree that a fix is not necessary.

See, this is why we can't have nice init systems.

Source: http://www.securityweek.com/linux-systemd-gives-root-privileges-invalid-usernames


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 2) by unauthorized on Tuesday July 04 2017, @12:29AM (12 children)

    by unauthorized (3776) on Tuesday July 04 2017, @12:29AM (#534619)

    But I'm still not sure I get this. How does a username normally get written into one of these 'unit files' in the first place?

    Unit files are usually hand-written files installed along with the distribution or a package. In the typical use-case they only change with package updates unless an admin manually edits them.

    Starting Score:    1  point
    Karma-Bonus Modifier   +1  

    Total Score:   2  
  • (Score: 3, Interesting) by Arik on Tuesday July 04 2017, @12:53AM (11 children)

    by Arik (4543) on Tuesday July 04 2017, @12:53AM (#534622) Journal
    OK, so that probably justifies thinking that it's not a super-critical, must-solve-this-second sort of bug.

    But how do they figure it's not a bug? Is there any conceivable, theoretical case where this behavior is expected and desired?

    --
    If laughter is the best medicine, who are the best doctors?
    • (Score: 5, Insightful) by vux984 on Tuesday July 04 2017, @01:02AM (5 children)

      by vux984 (5045) on Tuesday July 04 2017, @01:02AM (#534625)

      This Exactly.

      Throw/log an error and don't start the process should be the default behavior. Starting as root if the user name is invalid is beyond the pale.

      • (Score: 2) by kaszz on Tuesday July 04 2017, @02:59AM (1 child)

        by kaszz (4211) on Tuesday July 04 2017, @02:59AM (#534657) Journal

        No log as it binary.. remember? total clusterfuck. Shit like this is what forces administrators to do late nighter to clean up after the braindamage writes code that scriptkids use. (which usually some pointy haired type decided must be used..)

        • (Score: 2) by butthurt on Tuesday July 04 2017, @08:57PM

          by butthurt (6141) on Tuesday July 04 2017, @08:57PM (#534928) Journal

          > No log as it binary.. remember?

          The binary "journal" is a log.

      • (Score: 2) by SDRefugee on Tuesday July 04 2017, @01:06PM (1 child)

        by SDRefugee (4477) on Tuesday July 04 2017, @01:06PM (#534784)

        Starting as root if the user name is invalid is beyond the pale.

        Sounds like something that bunch of psychos in Redmond, Washington would do...

        --
        America should be proud of Edward Snowden, the hero, whether they know it or not..
        • (Score: 0) by Anonymous Coward on Tuesday July 04 2017, @07:48PM

          by Anonymous Coward on Tuesday July 04 2017, @07:48PM (#534903)

          Software requirements specification met!

      • (Score: 2) by Thexalon on Tuesday July 04 2017, @07:51PM

        by Thexalon (636) on Tuesday July 04 2017, @07:51PM (#534906)

        At the very least, if I were going to default to anything at all, I'd default to "nobody".

        --
        The only thing that stops a bad guy with a compiler is a good guy with a compiler.
    • (Score: 3, Interesting) by Jesus_666 on Tuesday July 04 2017, @05:13PM (3 children)

      by Jesus_666 (3044) on Tuesday July 04 2017, @05:13PM (#534841)

      Simple: The observed behavior is in line with how systemd was designed and therefore obviously correct. Since usernames can't begin with a digit and every major Linux distribution has been breaking the standard for years before it was even conceived it's not systemd's problem but that of everyone else. The fix to that problem is obvious: All distros should ban numeric usernames immediately (this is security-relevant, after all) and everyone who uses them should rename the offending users before installing any package updates. Also, they should be ashamed of themselves for being bad at Unix.

      At least that's how I imagine the systemd team sees it.

      • (Score: 2) by vux984 on Tuesday July 04 2017, @06:20PM (2 children)

        by vux984 (5045) on Tuesday July 04 2017, @06:20PM (#534864)

        At least that's how I imagine the systemd team sees it.

        Probably. But EVEN if that's how they see it they STILL have a bug.

        If the systemd team wants to enforce a policy of no usernames that start with a digit, fine, that would be in keeping with their arrogance and ego... but at the very least they should properly reject those credentials as invalid.

        The failure mode for attempting to run a process with "invalid" credentials should NEVER be "run it as root".

        • (Score: 3, Touché) by maxwell demon on Tuesday July 04 2017, @06:43PM (1 child)

          by maxwell demon (1608) Subscriber Badge on Tuesday July 04 2017, @06:43PM (#534876) Journal

          The failure mode for attempting to run a process with "invalid" credentials should NEVER be "run it as root".

          Nah, if Poettering implements this that way, it's obviously exactly the right thing. To follow his lead, I'm going to patch login with the following code:

          if (valid(user, password))
            login_as(user);
          else
            login_as(USER_ROOT);

          SCNR :-)

          --
          The Tao of math: The numbers you can count are not the real numbers.
          • (Score: 2) by DECbot on Wednesday July 05 2017, @04:40PM

            by DECbot (832) on Wednesday July 05 2017, @04:40PM (#535258) Journal

            I understand it to be a little more complex than that...

            if (system.valid(user, password))
                try{ login_as(systemd.valid(user)); }
                catch(NoSuchSystemdUserException ex){ login_as(USER_ROOT); }
            else
                throw new NoSuchUserException("User not found");

            --
            cats~$ sudo chown -R us /home/base
    • (Score: 3, Interesting) by FatPhil on Wednesday July 05 2017, @08:03AM

      by FatPhil (863) <pc-soylentNO@SPAMasdf.fi> on Wednesday July 05 2017, @08:03AM (#535097) Homepage
      A malicious 3rd party could create a package with a legitimate-looking unit file and at least one dodgy executable, but chose to run as an unprivileged user. You trust it initially, as it's running as nobody or thereabouts. However, the uninstallation script in the package could deliberately leave the unit file in place, and one dodgy executable. The account it used is correctly nuked. However, the dodgy executable now runs as root.

      OK it requires social engineering to get people to install/try/uninstall, but still, I'd put that as a "fix right now" kind of bug. The brute force "unrecognised user name" = "don't run" solution sounds (a) like a fix to the bug; and (b) like what would have been the sensible default in the first place. If that's more than a trivial 10 line change to the code (I can barely imagine it being more than 3), then it's a code base that needs nuking from orbit.
      --
      Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves