Since he doesn't mention it in his 'fantastic' reporting, OpenSSH 9.6 was released Monday that will patch this attack. Also, since he doesn't mention it, if on the Internet, the MITM would have to be installed at both end points (client side and server side) to be effective without the patch.
Since he doesn’t mention it in his ‘fantastic’ reporting, OpenSSH 9.6 was released Monday that will patch this attack.
I am tempted to delete this post just for the article’s stupid clickbait headline, but it still will probably cause some people to go update their OpenSSH installs, so… meh.
Anyone who actually wants to know details of the vulnerability should read the website about it which is obviously much better than this article.
Also, since he doesn’t mention it, if on the Internet, the MITM would have to be installed at both end points (client side and server side) to be effective without the patch.
Huh? No. The attacker doesn’t need to be in two places or even near either end per se, they could be located at any fully on-path position between the client and server.
Even the researcher who reported this doesn’t go as far as this headline.
“I am an admin, should I drop everything and fix this?”
Probably not.
The attack requires an active Man-in-the-Middle attacker that can intercept and modify the connection’s traffic at the TCP/IP layer. Additionally, we require the negotiation of either ChaCha20-Poly1305, or any CBC cipher in combination with Encrypt-then-MAC as the connection’s encryption mode.
[…]
“So how practical is the attack?”
The Terrapin attack requires an active Man-in-the-Middle attacker, that means some way for an attacker to intercept and modify the data sent from the client or server to the remote peer. This is difficult on the Internet, but can be a plausible attacker model on the local network.
It definitely receives more clicks. I’ve posted this link here a day ago, but arstechnicas title is more engaging. My first thought was whether there’s been another vulnerability found.
That said, this headline isn’t as bad as it could’ve been.
Yeah, if the attacker is in a position to do a MitM attack you have much larger problems than a ssh vulnerability that so far can at most downgrade the encryption of your connection in nearly all cases
Just checked my own sshd configs and I don’t use CBC in them. I’ve based the kex/cipher/Mac configs off of cipherlist.eu and the mozilla docs current standards. Guess it pays to never use default configs for sshd if it’s ever exposed to the Internet.
Edit: I read it wrong. It’s chacha20 OR CBC. I rely heavily on the former with none of the latter.
So you need an MitM situation to even be able to perfom the attack, and the the attack on works on two ciphers? The article says those ciphers are commonly enabled, but are they default or used in relatively modern distributed versions of openssh?
A scan performed by the researchers found that 77 percent of SSH servers exposed to the Internet support at least one of the vulnerable encryption modes, while 57 percent of them list a vulnerable encryption mode as the preferred choice.
That means a client could negotiate one or the other on more than half of all internets exposed openssh daemons.
I haven’t got too whizzed up over this, yet, because I have no ssh daemons exposed without a VPN outer wrapper. However it does look nasty.
I get your point that the exploit existed before it was identified, but an unmitigated exploit that people are aware of is worse than an unmitigated exploit people aren't aware of. Security through obscurity isn't security, of course, but exploiting a vulnerability is easier than finding, then exploiting a vulnerability. There is a reason that notifying the company before publicizing an exploit is the standard for security researchers.
You're right that it's never an OK title, because fuck clickbait, but until it's patched and said patch propagates into the real world, more people being aware of the hole does increase the risk (though it doesn't sound like it's actually a huge show stopper, either).
Weakness and risk are distinct things, though—and while security-through-obscurity is dubious, “strength-through-obscurity” is outright false.
Conflating the two implies that software weaknesses are caused by attackers instead of just exploited by them, and suggests they can be addressed by restricting the external environment rather than by better software audits.
In my opinion Dan Goodin always reports as an alarmist and rarely gives mitigation much focus or in one case I recall, he didn't even mention the vulnerable code never made it to the release branch since they found the vulnerability during testing, until the second to last paragraph (and pretended that paragraph didn't exist in the last paragraph). I can't say in that one case, it wasn't strategic but it sure seemed that way.
For example, he failed to note that the openssh 9.6 patch was released Monday to fix this attack. It would have went perfectly in the section called "Risk assessment" or perhaps in "So what now?" mentioned that people should, I don't know, apply the patch that fixes it.
Another example where he tries scare the reading stating that "researchers found that 77 percent of SSH servers exposed to the Internet support at least one of the vulnerable encryption modes, while 57 percent of them list a vulnerable encryption mode as the preferred choice." which is fine to show how prevalent the algorithms are used but does not mention that the attack would have to be complicated and at both end points to be effective on the Internet or that the attack is defeated with a secure tunnel (IPSec or IKE for example) if still supporting the vulnerable key exchange methods.
He also seems to love to bash FOSS anything as hard as possible, in what to me, feels like a quest to prove proprietary software is more secure than FOSS. When I see his name as an author, I immediately take it with a grain of salt and look for another source of the same information.
There are several ways to exploit LogoFAIL. Remote attacks work by first exploiting an unpatched vulnerability in a browser, media player, or other app and using the administrative control gained to replace the legitimate logo image processed early in the boot process with an identical-looking one that exploits a parser flaw. The other way is to gain brief access to a vulnerable device while it’s unlocked and replace the legitimate image file with a malicious one.
In short, the adversary requires elevated access to replace a file on the EFI partition. In this case, you should consider the machine compromised with or without this flaw.
You weren’t hoping that Secure Boot saves your ass, were you?
Ah, so the next Air Bud movie will be what, Hack Bud?
“There’s nothing in the specifications that says that a dog can’t have admin access.”
“Nothing but 'net!”
Doesn’t this mean that secure boot would save your ass? If you verify that the boot files are signed (secure boot) then you can’t boot these modified files or am I missing something?
If it can execute in ram (as far as I understand, they’ve been talking about fileless attacks, so… Possible?), it can just inject whatever
Addit: also, sucure boot on most systems, well, sucks, unless you remove m$ keys and flash yours, at least. The thing is, they signed shim and whatever was the alternative chainable bootloader (mako or smth?) effectively rendering the whole thing useless; also there was a grub binary distributed as part of some kaspersky’s livecd-s with unlocked config, so, yet again, load whatever tf you want
Last time I enabled secure boot it was with a unified kernel image, there was nothing on the EFI partition that was unsigned.
Idk about the default shim setup but using dracut with uki, rolled keys and luks it’d be secure.
After this you’re protected from offline attacks only though, unless you sign the UKI on a different device any program with root could still sign the modified images itself but no one could do an Evil Maid Attack or similar.
The point with m$ keys was that you should delete them as they’re used to sign stuff that loads literally anything given your maid is insistent enough.
[note: it was mentioned in the arch wiki that sometimes removing m$ keys bricks some (which exactly wasn’t mentioned) devices]
Well, not an expert. We learned now that logos are not signed. I’m not sure the boot menu config file is not either. So on a typical linux setup you can inject a command there.
In many of these cases, however, it’s still possible to run a software tool freely available from the IBV or device vendor website that reflashes the firmware from the OS. To pass security checks, the tool installs the same cryptographically signed UEFI firmware already in use, with only the logo image, which doesn’t require a valid digital signature, changed.
arstechnica.com
Active