Replying to @⁨schipelblorp@sh.itjust.works⁩

Their fingerprinting is flawed if this story is true. If I take someone else’s CPU and drop it into what is otherwise different hardware then the fingerprint should absolutely change. If it doesn’t then their fingerprinting is broken and this is going to happen whether any particular CPU has been banned or not.

It seems far more plausible that someone got themselves banned and they’re lying to try to get out of it.

Replying to @⁨echo@lemmy.today⁩

Fingerprinting that is that hyperspecific is the one that’s actually flawed. If you could circumvent a hardware ban by rearranging your RAM sticks, it wouldn’t be much good.

As it is, the CPU is the one most stable, expensive, and difficult piece to replace, so it makes sense to ban on that alone.

But, I do agree that cheaters are really slimy fuckers, so what you’re saying is certainly possible. I remember a kid that had his grandmother bring his computer into the shop to see we could evade his hw ban, talked his grandmother into agrreing to buy him a new GPU right in front of us before we turned him down flat.

Replying to @⁨schipelblorp@sh.itjust.works⁩

A CPU serial number or other ID is more unique than a MAC address, as those can be spoofed, and are not guaranteed to be unique, either. I heard one case of someone tearing their hair out over a mystery network issue, eventually discovering two NICs with the same MAC. And also the case of someone buying a whole pallet of NICs, which turned out to all have the same MAC address.

But I don’t think that math checks out. A player driven away might return, but a banned player can’t.

Replying to @⁨frongt@lemmy.zip⁩

That pallet of NIC’s: oof! I thought there were complicated methods in place to assure that didn’t happen.

As for the math, cheating is a problem because it drives players away. Whatever the math is, there is a number of innocent players that could get swept up in a ban that would be acceptable if enough cheaters were also kicked out. I’m not talking morally, just the viability of an online game.

Replying to @⁨yesman@lemmy.world⁩

I would be very interested to know how they are able to identify a specific CPU. The article speculates but does not answer this question.

The fact that this is even possible is a security nightmare.

OK, I've done a bit more digging and pieced it together. Acknowledgement to @dual_sport_dork for prompting me to try (again) to find a command that might do it among other things.

IMO, this sort of information shouldn't be in processors in the first place, but apparently it is there and needs admin/root access to, well, access it.

... and the game in question requires kernel-level anti-cheat, so of course, it has the necessary access.

The command for Linux is sudo dmidecode -t processor, which dumps a lot of info including an ID field, which is supposedly the CPU's specific identifier. It can be grepped for or otherwise filtered out.

My own Ryzen gives 8 hexadecimal pairs, so it's a 64-bit value.

I would not know that that was unique to my specific individual processor if I had not been told, but I'll assume that at least part of it is not unique between Ryzens of the same calibre. (I won't be posting any of it for comparison, for obvious reasons.)

So the upshot is, if you don't want to be identified right down to your CPU, don't run programs as root if there's any chance of them phoning home.

Replying to @⁨palordrolap@fedia.io⁩

You can extract the chip’s serial number in software trivially. In Windows, from a commandline:

wmic cpu get pricessorid

I’m positive you can do the same with some Linux command, and failing that certainly with basically every vaguely modern benchmarking tool. Same deal with hard drives and probably also video cards. My BIOS/UEFI even enumerates all my connected hard drive serial numbers on screen if I interrupt its graphical boot screen.

Given that the anti-cheat schemes these competitive games are literal rootkits, I don’t think gathering the user’s hardware serial numbers in order to attempt identify them is much of a stretch.

Replying to @⁨inari@piefed.zip⁩

This is not running it as root (i.e. privileged userland, still ring 3), this is running it as ring 0 (i.e. the level that the kernel runs at, with no restrictions on access to memory and hardware at all). A tiny bug in ring 0 code can take down the system or open security vulnerabilities, where the kernel can provide more hardening to protect applications to an extent.

You could say this is the inner circle of hell.

Replying to @⁨palordrolap@fedia.io⁩

Question is, what happens if you don’t have dmidecode installed? Even more interesting, what happens if you’ve replaced it with a fake that spews back values of your choosing? If they’re bundling it, given that it appears to be GPL2, they’re opening themselves to a lawsuit unless they provide source upon request.

(Somehow, I don’t think they’re actually using dmidecode.)

Replying to @⁨nyan@lemmy.cafe⁩

What I've read just now is what I'd deliberately not gone looking for up to this point in case I found this out.

Linux and Windows happen to require that the executables that access that information be run with root privileges, but it looks like that's merely an affectation.

It seems that any old piece of software, without root or other privileges, can independently run the CPUID instruction that obtains a processor's serial number.

I do not like this one bit.

Replying to @⁨frongt@lemmy.zip⁩

Personally I tend to avoid running untrustworthy programs outside of a sandbox or VM.

what does that mean? are you not running any programs at all? you cannot know just by looking at the name if a program reads such sensitive info. and its not unlikely that some developer figured it would be a good idea to include it in automatic crash reports

yes, this is sensitive information. its like your fingerprint, you cannot change it. I think its quite obvious. web browsers firefox is fighting to hide more and more that even indirectly could lead to identification. websites can’t read this ID, this is just a comparison.

Replying to @⁨TheGreenWizard@lemmy.zip⁩

To be honest I’ve never found the mechanics in Counter-Strike games enjoyable. Not trying to hate, I’m sure it’s great for those who like it, I just don’t think it’s for me.

I think the two pvp fps games I’ve found to have the closest to ideal mechanics is Doom 2016, and Red Eclipse 2. In Doom I liked that it had some of the more Call of Duty-esque features like loadouts, and a multiplayer progression system. And I especially liked how the actual combat preserved that classic arena shooter feel in kind of an brilliant way. They made it so right click had that iron sights feature, but which also applied a secondary fire ability - but crucially it was entirely optional for basically every weapon. Like in classic arena shooters you could be in constant motion and shoot from the hip without any penalties to your aiming.

Red Eclipse does this too (plus has crazy parkour features), but last I checked lacks any kind of progression system for better and worse.

Titanfall 2 is pretty fun as well, but I am really bad at that game in multiplayer.