From Wikipedia, the free encyclopedia
Linux is a family of open source Unix-like operating systems based on the Linux kernel, an operating system kernel first released on September 17, 1991 by Linus Torvalds. Linux is typically packaged in a Linux distribution (or distro for short).
Distributions include the Linux kernel and supporting system software and libraries, many of which are provided by the GNU Project. Many Linux distributions use the word "Linux" in their name, but the Free Software Foundation uses the name GNU/Linux to emphasize the importance of GNU software, causing some controversy.
Community icon by Alpár-Etele Méder, licensed under CC BY 3.0
It's an LPE, and doesn't allow full root access to anyone who isn't already a user.
Are you saying LPEs aren't a security hazard?
Nope. Just pointing out an alarmist headline.
Ah. Yah, well, I just went with the article's own headline since so many comms insist on that.
Yup. It wasn't a criticism of you.
I don't think that's really a criticism, more like the reality of modern "journalism"...
I know, they need the clickbait title for the click money but yeah -as usual- I'm mostly shrugging this off
By chaining legitimate services such as udisks loop-mounts and PAM/environment quirks, attackers who own any active GUI or SSH session can vault across polkit's allow_active trust zone and emerge as root in seconds.
I recognize a few of those words.
Basically it's two vulns chained; first one gives a remote user privileges that a physically present user would get, in order to do things like put a thumbdrive in and have it mount. Then that udisks privilege can be subverted to escalate that level to root. So as long as you can start a remote session, you can pull root and it doesn't even look that hard.
So how would a bad actor start a remote session on my Linux pc?
Edited to add, downvoted for trying to learn is a new one for me.
They probably can not. Unless you've setup your router such that anyone can connect to an ssh instance running on your PC, and then also use a bad password. Public wifi + having something like ssh running + having a bad password.
Your PC probably doesn't satisfy these requirements (yay!), but some servers might.
I do run some servers, but use robust passwords.
You probably already do, but if you need SSH, use crowdsec, SSHGuard or fail2ban to help filter bot nets.
I have IPs hitting from all over the world, trying logins all the time. Like several per minute, I can only imagine what it would be like if I wasn't blocking IPs with multiple failed login attempts.
The technique described here is only a concern if the 'bad actor' has access to a user account on your machine in the first place.
Such as username and password?
No, there are other ways to get access to your machine without needing it. In general you can classify vulnerabilities as either code execution or privilege escalation, a code execution vulnerability allows an attacker to execute code on your machine, a privilege escalation allows him to break barriers that you might have imposed on him.
For example, if you're running service X as root, and someone manages to find a way to use something on service X to execute code, they might get a reverse shell to your box and run anything there. So you might set service X to run as your user instead of root, now that vulnerability is less important because it only compromises your user, but the attacker could use this one in conjunction with the other one to gain control of your user, then escalate to become root.
If this is something you're interested in, there's a cool website called hackthebox where you have to do these sort of things for real. If you want to have an idea on how it looks, there are some excellent videos here showing walkthroughs for many of them he boxes, I recommend checking something labeled easy since these boxes can get quite complex, but it will give you a good idea of the steps attackers need to take to compromise your system
Yes, or SSH keys or any other means of user authentication. The cool thing in this technique is that it's twofold and you (as an attacker) can cherry-pick the info given. If you walk up locally to someone's running system, you could skip the first half and go with the 'hey, can you resize this XFS image for me' bit.
While this is a risk, it is only a real risk if the system is already exploited for regular user access. Or if there is an untrustworthy user of the system. So for most, it is not a major concern.
Is it new or is it newly discovered?
Since it is open source... I guess we can rule out an intentional back door.
Since it is open source… I guess we can rule out an intentional back door.
Well, once upon a time I would have agreed with you but the xz backdoor changed my mind on that.
I dunno, I'd slow your roll on that. Hanlon's razor came to notoriety in the field of computer science for a reason. I've done software dev professionally for over ten years now and you wouldn't believe the stupid shit I've seen people write. The only thing that sucks more than a computer is the human writing software for it.
For those unfamiliar, here's Hanlon's razor:
Never attribute to malice that which is adequately explained by stupidity.
EDIT: After a quick look at the CVEs, this definitely sounds like a big ol' fuckup. It sounds like there might be some unsafe defaults in polkit as well?
EDIT: Here's the report from the actual researchers which is MUCH more cogent than OP's article: https://www.openwall.com/lists/oss-security/2025/06/17/4
It's chaining two separate oopsies together. This overview on GitHub also provides more details about the libblockdev side of things: https://github.com/advisories/GHSA-mpgj-hch9-5rvx
Specifically, this section:
However, a local attacker can create a specially crafted XFS image containing a SUID-root shell, then trick udisks into resizing it. This mounts their malicious filesystem with root privileges, allowing them to execute their SUID-root shell and gain complete control of the system.
That really doesn't sound like something intentional to me. That sounds like a HUGE oopsy-woopsy fucky-wucky, to get technical about it.
I'm not saying I think it was malicious, just that I'd no longer assume it wasn't
Newly discovered it appears.
Or ditch udisks in favour of pmount (or udevil?), which shouldn't be affected as far as I can tell. That will get you a few months' grace before a similar problem pops up there.