George,

It is an exploit that uses a hole that was inadvertently created in the main 
Linux kernel a decade ago.

Nobody knew the hole was there (at least, anybody that did know about it was 
keeping it to themselves) until a researcher from xinit started investigating.

Xinit used a LLM AKA AI assisted program to find the hole.  Their business 
model is to charge developers to analyze their code for them using their secret 
sauce.  I guess they got tired of spamming LinkedIn for customers and decided 
to do something spectacular to get attention by seeing if they could use their 
product on the Linux kernel to find a hole and they hit the jackpot.

The exploit basically means anyone who has user-level access to a vulnerable 
Linux system can become the root user.  And every Linux system using a kernel 
released in the last decade is vulnerable.

For example, the Apache web server is usually installed to run under a 
non-privileged userID often "www"   So let's say someone has a webserver that 
they allow people to upload stuff to.  All you have to do is upload a simple 
script then cause it to run under the www user via a POST or GET command and 
bang -your now root on that webserver and can run a back door or something you 
can use to remote in and pwn the server (gain control of the server).

When xinit found this they patched it then submitted the patch April 1st.  They 
did not plaster "this patch plugs a massive security hole" in the patch 
description.  Instead they used a non-committal description of it, basically 
saying it was making the code faster and more efficient.  Since Linus, who 
manages patch submittals to the Linux kernel, is known to bite on anything that 
purports to make Linux run faster, the patch was quickly accepted.  It's also 
possible that xinit privately emailed Linus and explained the ramifications of 
what they found and asked him to shut the F up about it.

Debian and Ubuntu's dev teams both have new releases coming out, Debian is 
releasing Debina 13 and Ubuntu is releasing 26.04.  Between them and 
distributions based on Debian, they are the 600 pound gorillas of Linux with 
probably almost half of Linux installs while RedHat and it's derivatives   So, 
xinit knew that their kernel patch would be picked up by these new releases and 
it was.

So once the stage was set, xinit then published proof of concept on a vanity 
website.  The python code they published is real, it does work, I've tested it. 
 They then passed the word to a BUNCH of technical bloggers and commercial news 
sites and articles started appearing.  They can pretend to appear like angels 
because anyone who complains will be told "if you were running current distros 
(that were released last week but let's not quibble) then it wouldn't be a 
problem for you" by most of the Linux fanboys on the Internet who don't know 
anything.

But the story does not end there.

Yesterday afternoon (afternoon PST that is) Canonical's ubuntu.com website came 
under attack and went offline.  I posted about this on this list yesterday but 
didn't know at the time they were under attack.

The attacker is undoubtedly Russian Mafia as they have sent extortion demands 
to Canonical already.  They CLAIM that they are Iranians but this is obviously 
ridiculous as the Iranians are a little busy right now, and the demands are for 
money - which obviously, the Iranians can't use right now anyway to buy 
anything since the US has them blockaded.  Real Iranians would demand food or 
humanitarian aid or a cessation to the shooting war or something like that.

In the meantime, xinit has also threatened - well, actually, PROMISED, to 
release more zero day exploits that they claim to have have discovered.  They 
must really be desperate for attention and customers.

There's a lot of people running Ubuntu desktop and server right now who are 
scared and looking for information.  The likelihood is that they WILL NOT do 
the obvious and simple thing which is run "get update, get upgrade" as root on 
their systems which would download updates and patch their systems, until they 
are sure they are doing the right thing.  They are trying to get more 
authoritative information which they can't since the ubuntu.com website is 
offline.  The Ubuntu update servers are NOT affected by the DDoS attack

The longer that people DO NOT update, the more time that the Russian Mafia or 
whoever is running the attack, will have to develop exploits that use this 
hole, and use them to break into systems that are not patched.  That is IMHO 
the real goal of the Canonical attack - sow confusion and doubt.

Run updates on your system ASAP and then forget about it, you will be protected.

Ted


-----Original Message-----
From: PLUG <[email protected]> On Behalf Of Ben Koenig via PLUG
Sent: Friday, May 1, 2026 10:13 AM
To: Portland Linux/Unix Group <[email protected]>
Cc: Ben Koenig <[email protected]>
Subject: Re: [PLUG] exploit in the wild

Yeah, wallposting can muddy the waters.

The original link seemed sketchy so I just skimmed through the details without 
actually running it. Looks like Russell beat the internet to this one because 
in the past couple days it's been popping up on a lot of blogs/forums.

If anyone wants a real link as opposed to the sales pitch from the company that 
found it, here's a few:

https://nvd.nist.gov/vuln/detail/CVE-2026-31431
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=a664bf3d603d

Note that the NIST page references the copy.fail website people have been 
referring to.
-Ben

On Friday, May 1st, 2026 at 9:47 AM, George <[email protected]> wrote:

> I can't tell what you guys are talking about. I assume you're talking 
> about a virus or an exploit.
> It seems that whatever you're talking about is explained in a rando 
> web link in the original interest group email. Just how many of you 
> guys click links in your emails when you're researching security? (I'm 
> kidding, after about four emails, i just looked it up from an 
> independent source. Still not sure if i use algif_aead but I'm the only user 
> on my network, er,...
> that i know of....)
> 
> 
> On Fri, May 1, 2026, 8:12 AM Ted Mittelstaedt <[email protected]>
> wrote:
> 
> > That may work for now however according to:
> >
> > https://xint.io/blog/copy-fail-linux-distributions
> >
> > "...The scan also identified other high severity vulnerabilities, 
> > including another privilege escalation bug. These other bugs are 
> > still in the responsible disclosure process."
> >
> > And we know now that from xinit's POV responsible disclosure means 
> > insert a patch then wait 30 days and publish a zero day.
> >
> > So this isn't going to be the only one of these rodeos.  It's just 
> > the first.
> >
> > Ted
> >
> > -----Original Message-----
> > From: PLUG <[email protected]> On Behalf Of King 
> > Beowulf
> > Sent: Friday, May 1, 2026 7:46 AM
> > To: [email protected]
> > Subject: Re: [PLUG] exploit in the wild
> >
> > On 4/30/26 17:11, Ted Mittelstaedt wrote:
> > > I can confirm that the latest apt-get update to Ubuntu 24.04 as of 
> > > a few
> > minutes ago is disabling the aead module.
> > >
> > > For an un-updated system, running python3 copy_fail_exp.py gets 
> > > you a
> > root shell.   For an updated system it gets an error.  For Ubuntu 26.04 it
> > merely asks for the root password.
> > >
> > > Ted
> > >
> > >
> >
> > or run
> >
> > find / * -perm -4004 -type f -exec ls -ld {} \; > setuid.txt
> >
> > and remove 'r' flag from user, user group, and other group.
> >
> > On Slackware, most setuid root utilities are not user readable.
> >
> > # ls -l /usr/bin/sudo
> > -rws--x--x 1 root root 289800 Jul 26  2025 /usr/bin/sudo* # ls -l 
> > /bin/su -rws--x--x 1 root root 59552 Feb 13  2021 /bin/su*
> >
> > There are a few that are unfortunately.
> >
> > This will mitigate the exploit until patched.
> >
> > -Ed
> >
> >
> >
> >
> 

Reply via email to