Attacking UNIX Systems via CUPS, Part I | CUPS Remote Code Execution
Attacking UNIX Systems via CUPS, Part I | CUPS Remote Code Execution
Hello friends, this is the first of two, possibly three (if and when I have time to finish the Windows research) writeups. We will start with targeting GNU/Linux systems with an RCE. As someone who’s
Worse than the exploit, is hearing the struggles the author faced to report it
Twenty-two days of arguments, condescension, several gaslighting attempts, more or less subtle personal attacks, dozens of emails and messages, more than 100 pages of text in total. Hours and hours and hours and hours and fucking hours. Not to mention somehow being judged by a big chunk of the infosec community with a tendency of talking and judging situations they simply don’t know.
I mean, OK, it's a vulnerability and there are interesting implications, but this is hardly significant in any pracitcal sense of the word.
the potential victim has to run their system without a firewall, has to print to the printer they've never interacted with before and then the attacker can run shit with whatever the printing system's user id is, which shouldn't be an issue on any reasonably modern distro.
I routinely remove cups and friends from any system I run because I have no need for printing and it bothers me to see it constantly during every system upgrade.
Wait, so how do we print now?
Lol what has more of an attack surface: CUPS or a reactos VM?
RPI print server + sshfs?
Udp sent to port 631. Isn't firewall on router going to block that anyway?
I have cups (but not cups-browsed) installed, but I only start the service when I need to print something a few times a year. Until then it is only a binary sitting in a folder, nothing more.
Honestly it isn't a big deal if you just use it on local host. Just make sure cups is sandboxes like it should be. (Systemd)
Yes, but exactly that was/is the issue of this bug. cups-browsed was attaching itself to every available IP on the system. And cups-browsed can't only be bind to localhost, it would defeat the whole purpose of that tool. For it to be able to find other printers in the network it needs to be bound to a non-localhost-IP address. So, not much to sandbox
cups-browsed <= 2.0.1 binds on UDP INADDR_ANY:631 trusting any packet from any source
Well that would explain why I didn't have it installed (although I did have other parts of cups until jwz coincidentally reminded us two days ago that it can all be removed if you don't have a printer.) I clear out anything that opens ports I don't need to be open. A practice I would recommend to anyone.
I mean I guess that’s sound advice if you don’t need to, I don’t know, print?
Indeed. But cups-browsed isn't necessary in order to be able to print things, it's for automatically discovering new printers on the network.
BRB.
sudo apt purge cupsDone. This should not even be part of baseline Ubuntu desktop. Speaking for myself but I have not had a printer for about 15 years. The paperless office really did become reality.
About 25 years for me.
For most people it's a better option to use a local print shop for the odd times that they need something printed.
More options for printing too.
I don't know if I would like to have my personal data that I needed to print out on any system in a print shop. Printers and Copy machines in print shops often have internal HDD where the files are stored for caching reasons, often for months or even years (depending of the size of the HDD and how much the device is used) until some internal cleanup process deletes them.
Oh, hey! Wasn't even a problem for me.
I had updates today that were security related about this.
damn. that's a pretty shit one.