Panik: your Debian stable system is so ancient it still contains the heartbleed bug.
I thought Debian does do security patches
Of course it was patched in all affected Debian versions: https://security-tracker.debian.org/tracker/CVE-2014-0160
Is linux 6.1 vulnerable to heartbleed? I’m on lmde6 with linux 6.1 btw) edit: as other comment said debian 12 is good so everything alright
I don’t think heartbleed is a kernel bug
Its a CPU bug only the kernel can fix 🤒. The kernel is responsible for its running hardware.Am I dumb? that’s the spectere and meltdown bug. xz-utils malware is a whole other thing, lol.
Heartbleed isn’t a hardware vulnerability either. It’s a bug in OpenSSL. Are you alright?
Clearly not, lol. Every vulnerability is spectere aparently. (To be fair, the other CVE this week is hardware based)
Heartbleed isn’t new either. It’s from years ago. It’s also unrelated to the xz backdoor. Maybe you should get some rest. Check your carbon monoxide alarms are working. If not see a doctor. It sounds like you are having memory issues.
The xz infiltration is a proof of concept.
Anyone who is comforted by the fact they’re not affected by a particular release is misguided. We just don’t yet know the ways in which we are thoroughly screwed.
This is a huge wake up call to OSS maintainers that they need to review code a lot more thoroughly. This is far from the last time we’re going to see this, and it probably wouldn’t have been caught if the attacker hadn’t been sloppy
https://upvote.au/comment/818245
Nah, I’d say the chap was pretty unsloppy.
Just that we were lucky that someone found it.It’s a good thing that
xzis a type of program that people may want to profile.But this is an eye opener for people saying that Linux is “secure” (not more secure, but just secure .) because the code has many eyes on it. --> jump to digression.
This confirms my suspicion that we may be affected by the bystander effect, so we actually have less eyes than required for this.
digression:
- of course I don’t mean that this makes Linux less secure than Windows. The point that it makes it more secure than Windows/MacOS or other closed source systems is already apparent.
- Just that, we can’t consider Linux to be secure (without comparing it to something less secure) as many ppl would, when evangelising Linux.
My point being, tell the whole truth. The newbie that’s taking your advice will thank you for that later on.
The reason I consider this sloppy is because he altered default behavior. Done properly, an injection like this probably could have been done with no change to default behavior, and we’d be even less likely to have gotten lucky.
Looking back we can see all the signs pointing to it, but it still took a lot of getting lucky to find it.
I’ve always considered the “source is open so people can check for vulnerabilities” saying a bit ironic, because I’d bet 99% of us never look, nor could find it if we were looking. The bystander effect is definitely here as we all just assume someone else has audited it.
Done properly, an injection like this probably could have been done with no change to default behaviour,
Interesting.
So the sloppiness was in the implementation and not the social engineering.
But then of course, people tend to be not good at both, fooling people and fooling programmers/computers at the same time. In this case, the chap turned out to be better at fooling people than programmers/computers.
And I am being sloppy for not trying to learn enough about exploits even though I should have a good enough programming base to start it.
It’s a rough balance when you’re trying to convince people unfamiliar with the internals (let alone non technical people) to make the switch. Saying “Linux is safe, but not bulletproof” may scare them back to the devil they know even if there’s no greater guarantee of safety there.
Of course, maybe I am being too hard on people by expecting everyone to put more thought into everything they make a decision for. But it is in fact the lack of thought that tends to cause problems in all areas we see nowadays.
But that’s a topic for somewhere else.
We can simply go by “Linux is more bulletproof than Windows”; instead of calling it “safe”, which would also be wrong.
Also with, “Windows will shoot you with intent, Linux might just get some stray shrapnel.”
- of course I don’t mean that this makes Linux less secure than Windows. The point that it makes it more secure than Windows/MacOS or other closed source systems is already apparent.
And to have strong and continuous analysis of software bills of materials.
I’m just waiting for the backdoor to be found in Firefox and Chromium or some library shared by most applications.
Like libwebp a few months back? Or Log4j?
The thing about browsers is that there are so many accidental exploits already, it makes little sense to introduce your own on top of it.
I always just assumed all the distros I use have backdoors as a fact of life. I take comfort in not being a person of interest to anyone and just blend in with the crowd. I also don’t use windows because for every backdoor my Debian may have, windows will have 100 more. Servers don’t get hacked all the time because it is not linux->internet, it is linux->bunch of stuff->internet, but I am sure backdoors are there.
Your Debian stable system is so ancient you got bigger vulnerabilities to worry about: Panik!
Also the problem was that Debian’s sshd linked to liblzma for some systemd feature to work. This mod was done by Debian team.
Liblzma balls
But do it in private, don’t let my xz.
Even if you’re using debian 12 bookworm and are fully up to date, you’re still running [5.4.1].
The only debian version actually shipping the vulnerable version of the package was sid, and being a canary for this kind of thing is what sid is for, which it’s users know perfectly well.
There was a comment on Mastodon or Lemmy saying that the bad actor had been working with the project for two years so earlier versions may have malicious code as well already.
Distros like gentoo reverted to 5.4.2 for that reason. If debian stable is on 5.4.1 that should be ok.
They did but the malware wasn’t fully implemented yet. They spent quite a while implementing it, I guess to try and make it less obvious.
Needless to say all his work ever will already be being reviewed.
The linked version in stable was not impacted.
What do you mean bigger vulnerabilitirs to worry about in Debian stable?
Mostly a joke about him calling it “ancient”, but there may be some unpatched vulnerabilities in older software. Though there could also be some new ones in newest versions.
Still, unless it’s Alpha/Beta/RC, it’s probably better to keep it up-to-date.Debian patches security vulnerabilities in stable. They don’t change the version numbers or anything but they do fix security holes.
Debian responds to security issues in stable within a fairly short window. They have a dedicated security team.
The malicious changes were submitted by JiaT75, one of the two main xz Utils developers with years of contributions to the project.
“Given the activity over several weeks, the committer is either directly involved or there was some quite severe compromise of their system,” Freund wrote. “Unfortunately the latter looks like the less likely explanation, given they communicated on various lists about the ‘fixes’” provided in recent updates. Those updates and fixes can be found here, here, here, and here. https://arstechnica.com/security/2024/03/backdoor-found-in-widely-used-linux-utility-breaks-encrypted-ssh-connections/
That really sucks. This kind of thing can make people and companies lose trust in open source. I wonder if we will learn the reason behind that. I would guess the developer was paid a lot of money by some organization to risk ruining his reputation like that.
Like the exact same thing can not happen in a closed source codebase. It probably does daily. Since closed codebases the due dilligence and reviews cost money, and nobody can see the state. They are intentionally neglected.
Open source nor closed source is immune to the 5$ wrench hackCan’t decide which one is more relevant - the $5 wrench hack, or any sort of blackmailing.
Exactly, if you are as big a Microsoft, you can’t tell 100% if one of your developer’s is actually being paid by a foreign government. Even if you say completely check the commits other devs make, there will still be deadlines when a code review is just “looks fine, next”.
No, its the exact opposite.
Supply chain conpromise is a level of risk to manage not unique to FOSS. Ever heard of sunburst? It resulted in a lot of Microsofts cloud customers getting wreaked all because their supply chain was compromised.
Do people continue to buy into 365 and Azure? Yes. Without care.
So will this hurt open source projects? Not at all, in fact it will benefit them, highlight just why source code SHOULD be open source and visible to all! We would have had very little to no visibility and capability to monitor closed source. Let alone learn, improve and harden how projects can protect against this increasingly more common attack.
Yeah, I agree but I know some companies will have stupid thoughts like “a company employee is less likely to do that” or “at least we have an employment contract to back us up legally”.
Until they are attacked…
Not to mention a lot of the time the “attack” is from the company themselves. Just look at the Meta malware as an example
The sony CD rootkit comes to mind https://en.m.wikipedia.org/wiki/Sony_BMG_copy_protection_rootkit_scandal
the Meta malware
What is this?
The VPN that performed a man in the middle attack to get data from other apps
Ugh this reminds me of a guy I worked with, he used to be a trucker but became a software tester (he was also very religious).
Anyway he used to hate on open source software and call it open sores. According to him it was all amateur crap. Ugh I still hate that guy and it has been 15 years…
Could be a state actor too
Certainly, that’s why I said organization to be vague.
Sorry I should have been more clear too. I was trying to convey that the dev could have been paid off/threatened or it could be the work of a state actor or team of state actors under an alias. In one case they could care about their reputation but in the other maybe not.
By the sounds of it it was an organised social engineering attack. Almost certainly “Jia Tan” is not a real person, or if it is a real person then it’s a case of stolen identity. Even if I were being threatened to put a backdoor in some software I wouldn’t do it under my real name.
Still paniking, cause the backdoor was apparently targetting Debian servers, it was discovered just by chance and the “mantainer” made commits for 2 years in the same repo
The fact that this was planned is what makes me nervous. Imagine what else is lurking.
and it was only discovered accidentally, when someone was profiling some stuff, noticed SSH using a bit too much CPU power when receiving connections even for invalid usernames/passwords, and spent the time to investigate it more deeply. A lot of developers aren’t that attentive, and it could have easily snuck through.
hey Dan, why don’t you post blogs now?
I’ve been meaning to start blogging again. It’s just been a lack of free time. Need to think of ideas, too.
understandable mate, have a nice day!
I’ve never read your blogs but I think it would be good for more people to do write ups on the XZ backdoor. There is a lot that can be learned and improved
That’s basically the idea of having a stable branch. Where all packages have 2+ years of testing and revisions.
They are stagnant. They keep getting security/fix patches, just not new features.
The slowness is on purpose.
(OP may know, but I don’t know if everyone does.)
Edit: /u/[email protected] is picking up what I was putting down.
The slowness is on purpose? To help identify the sshd in question to the attacker which nodes are compromised? What reason(s) could there be?
He’s talking about Debian’s slowness in getting new versions to stable, and how the meme ignores security backports.
Ohh that makes way more sense, thanks. I haven’t used Debian in like 10 years but it was obviously the same back then too.
If the above decides to continue, the code appears to be parsing the symbol tables in memory. This is the quite slow step that made me look into the issue.
That is from the original find. Not sure the relevance of it and this being proof for it being “on purpose”. But that is the origin of the slowness.
I doubt that was intentional, they would likely want to hide that latency but the CPU time required to scan everything just is what it is.
https://bsky.app/profile/filippo.abyssdomain.expert/post/3kowjkx2njy2b
The hooked RSA_public_decrypt verifies a signature on the server’s host key by a fixed Ed448 key, and then passes a payload to system().
It’s RCE, not auth bypass, and gated/unreplayable.
Maybe Manjaro should delay update even longer to make it extra secure /s
Ngl manjaro beeing slow as saved my ass more than once
I recently ran into a bug in the latest version of cmake that breaks it completely in my system, can’t compile shit and it just does a coredump.
What is worse is that I can’t even report the bug because I can’t get the registration email from the cmake gitlab. I checked the manjaro repos and their cmake version is 2 versions older than the one that has the bug that left me thinking for a while lol.
Tbh Manjaro has made my system unbootable before with a system update, base Arch has never done that for me. I think Manjaro is just poorly constructed, or maybe it’s bc of all the packages that come pre installed with it causing problems. Minimal installs ftw
i had bootfails on both systems, shit happens just boot from usb then changeroot into your system to repair
In this case the slower updates payed off. There are many things wrong with Manjaro but slower updates is not one of them.
deleted by creator
Well honestly holding packages is one of the only good things I hear
It’s not though, because the malicious release happened more than two weeks ago and manjaro had to fast track the patched xs from arch git repo. This is why manjaro should extend their delayed update policy to catch this kind of issue in the future (maybe 2 months instead of 2 weeks) /s
They honestly should as it probably would fix a lot of issues. Then again, Manjaro is so broken that it probably doesn’t matter
deleted by creator
How about you just use any stable system. For instance, stable Fedora wasn’t affected
This was the slow process for my hungover brain today.
