Not trying to clickbait but is this the end for kernel-level anti-cheat?
Not trying to clickbait but is this the end for kernel-level anti-cheat?
From the article
Microsoft has officially announced its intent to move security measures out of the kernel, following the Crowdstrike disaster a few short months ago. The removal of kernel access for security solutions would likely revolutionise running Windows games on the Steam Deck and other Linux systems.
I'm okay with that. I don't want some program I don't control having access to the kernel of my system.
game developers and publishers are hesitant to enable Linux compatibility,
And I am hesitant to spend money on their games.
Also more than half of games with AC do in fact run on linux right now, and the world hasn't ended.
I would reply with something like "Arrrr matey" but new games aren't even worth pirating anymore.
And nothing was lost.
Could very well be possible. Apple did the same thing with macOS Catalina in 2019. Since then, there are no kernel extensions, meaning no third-party code running at kernel level. This greatly improves the security of macOS, and other desktop operating systems should do the same.
They have a hardcore micro kernel though, windows and Linux are a long way off from that Valhalla, gnu Mach notwithstanding.
It's probably going to move to hardware attestation similar to what Android and iOS are doing. This may or may not be a good thing.
Yeah, idk why everyone seems to legitimately think devs are going to just quietly revert back to usermode anticheat. I could see Riot patching an actual root kit before that happens.
But yeah, more likely MSFT will lobby for hw that is more annoying than secure boot or TPM to get working with linux, every windows app after that point will rely on it "because turnkey security!", and if you ever manage to disable it none of those apps will work on your machine in any OS (if they even worked through proton at all).
Can anyone give an ELI5 on how hardware attestation works or is implemented on current OSes?
If it works on Linux, I'm fine with it, since I'm not cheating. Just like hardware attestation works on GrapheneOS, because it doesn't decrease the security of Android, in fact, it greatly improves it.
Believe it [or] not battlefield still won't work on Linux because fuck you peasants
If it doesn't run on Linux it doesn't deserve my money at this point.
As a long time cs gamer I approve of this change but I warn ye regardless that there is no alternative or viable solution to actually stop cheaters right now.
And if you've only heard stories and don't really experience cs (vac kind of does nothing)
Ive kept track of players for months/years who have not been banned. I find it strange that they eventually do get banned several months after cheating. It took one account nearly 2 years to get banned.
I hope that a clever solution comes out, a man can dream right ?
well... you see back in my day we had cool bros in "clans" running their servers mostly paying for it themselves with some donations. admins would boot bad faith actors as needed.
then something happened to that model... and here we are now... FPS genre has no been the same IMHO
The competitive scene happened. Can't have meaningful competitive matchmaking against the same 100 players. People don't just want to frag noobs, they want to grind the ladder to be able to say "I'm GE and you're Gold, therefore I know for a fact I'm better than you".
This is a global phenomenon. Even goddamn chess has this, first thing players ask each other nowadays is "what's your chess.com ELO".
I'm not a competitive player myself but I get why people rush after ELO progression. And it's not much of a stretch to say CS, Valo, and especially chess wouldn't have seen such widespread success without competitive ELO-based matchmaking.
Self moderation has been way more effective at controlling cheaters than automated systems. Counterstrike did some good with overwatch and phone verification but I've always enjoyed manual server moderation if it's maintained.
Csgo and 2 have a "trust" system to keep track of player behavior and put you in games with others of similar trust value. So if you get reported often or have a history of bad behavior, you're more likely to be put in games with other bad actors, and vice versa. Idk how effective it is though.
Honestly there isn't a great solution, which is kind of why I avoid competitive multiplayer games. Even kernel level anticheats can be circumvented.
The nice thing about vac is that theres pretty much no false positives. And valve will occasionally update it, catching a ton of cheaters off guard and getting them banned.
There's always A.I. powered anticheat, and server side anticheat. Both work with anticheat client side that's not kernel level.
There was news that battlefield one would stop working because they were implementing fairfight(?) but it's still working and someone in game chat told me it wasn't kernel level in battlefield one version of the anti cheat. Any facts to this?
I wouldn't get my hopes up. Them announcing something like this looks good PR-wise, so they'll do it, even if they don't actually expect this effort to lead to anything.
But even if they do implement such an API, companies won't start adopting this API until its capabilities are roughly comparable to the kernel-level solution AND it's available on most Windows systems in the wild. So, we're likely talking more than a decade before this sees sufficient adoption...
I just found this thread as well could not be a good thing for us Linux users.
A locked down Windows "gaming OS" is probably what Xbox wants to go towards in some respects. It gives Microsoft the walled garden that they want, can lock out Valve as much as they fancy, and will likely be paired with some new APIs to set back Proton/WINE a few years. Hell, they could even still release XBOX hardware for that niche.
I can definitely see Microsoft releasing a "gaming os" that will lock down your PC and basically make it a console in all but name. I would be surprised if it's not in the works already for a handheld device.
I think I need more info. It seems like userspace is very hackable, so thus kernel level anti-cheat was born to control stuff like synthetic inputs and manipulation of memory / frame analysis. This anti-cheat would be held together by the fact that the kernel/drivers are proprietary and not very easy to edit. Obviously still possible because it's on your own computer, but challenging and invasive. Do I have that right?
In which case I don't see how going back to userspace would help. What is the solution? There probably isn't one outside of hardware (buying a hacking chip and soldering it in is annoying for most)
When I was doing game dev we focussed on AI-style analytics of user behavior. Of course a good enough bot could always look human. A real cat and mouse game wasting lots of time
Does there need to be a solution?
Do E-Sports competitions on identical certified hardware and otherwise ban people caught cheating.
Root kits aren't necessary for having fun in a game.lol, anti-cheat isn’t just about esports, or high level play. It’s not even just about cheating.
It’s broadly about harassment and griefing and just shitty behavior mediated by hacking in online games.
What is the solution?
My guess is that Microsoft wants provide some kind of kernel level anti-cheat, possibly directly integrated with directx, and it will use cryptography which will make it impossible to emulate with Wine/Proton.
The same kernel software cryptography could certainly be marketed for single player games and proprietary applications as a solution to piracy.
Don't like kernel anti cheat in your multiplayer games? here's kernel anticheat for your single player games!
Removing 3rd party kernel access will probably also make cheating harder. Kernel anticheat is necessary largely in part due to cheat software using exploits in the 3rd party extension system to get kernel privileges itself and evade user mode anticheat.
Am I misremembering to think Genshin Impact was a cause of one of these major security disasters?
It wasn’t even people who installed Genshin that were victims - it was like, Microsoft signed a driver made by Mihoyo to scan for cheat apps. But mihoyo, being a game company with a rapid release cycle and imperfect security, had a vulnerability in the driver. So, malware authors could include that driver in their packages to elevate access on Windows installs even when no one had any idea what a Genshin is.
Not quite the same thing as Crowdstrike I guess though.