It is a bit complicated. Any kernel level program that crashes will cause the entire operating system to crash. But it won’t cause the system to continuously blue screen because it isn’t a required program in the way that crowdstrike was.
Crowdstrike is basically an antivirus program so it has to run when the operating system starts up and if it isn’t running then the operating system should not boot for safety reasons. The problem is that if it must be loaded, and it has a crash, then it loads and kills the system. So you get an infinite loop you cannot get out of.
Vanguard only has to run when you’re playing online though, so it’s not loaded when the system runs, or at least it doesn’t have to be. So it won’t cause a recurring boot loop. It might fail to load and you wouldn’t be able to play online games that require it until they fix it, but it isn’t going to prevent the computer from running.
Wait, wasn’t Vanguard coming in form of a driver? I don’t use Windows and don’t play games with intrusive software requirements, but I believe I saw someone installing it and showing how it works on YouTube, and if I don’t misremember it, it was in fact a virtual device driver, not just a fully privileged process.
Yup, I seem to remember the same thing and I also seem to remember it being loaded at boot. Iirc you can stop it when the system has started and you’ve logged in but then you have to reboot in order for it to load again so you can plain. But same, I also don’t use windows and haven’t played leaugh for many many years.
I could be wrong but I think it runs on boot. So it would loop into bsod.
It has comparable access, yes, but assuming no malicious intentions, it’s extremely unlikely that they achieve something as catastrophic.
If they fucked up in a similar fashion, that would cause your PC to bluescreen, too, but since League does not start up during boot, you could still use your PC, just not League.
Vanguard doesn’t care if LoL or valorant or any other game is running. Vanguard is in your kernel and will be starting regardless.
This is correct, as in windows a driver is the most straightforward method to runlevel0 access. It absolutely could at any time do exactly what crowdstrike did. But also so could Nvidia/amd with GPU drivers, your motherboard manufacturer with chipset and RGB drivers, etc. it’s not quite the smoking gun people make it out to be, as there are a lot of legitimate reasons to have this kind of system access.
The egregious part was that crowdstrike users agreed to allow a vendor to bypass canary channels and deploy straight to their endpoints.
And that’s the problem, like CrowdStrike Vanguard will update itself in the background unlike your GPU driver which you need to go through an update process explicitly, so if the same thing happens where they pushed a bad update, the same outcome of causing failed boots without prompt could happen.
Does Vanguard not seek testing and validation by Microsoft before pushing updates?
I saw the recent video from the Task Manager designer Dave’s Garage on YouTube, lack of thorough official validation seemed to be an important part of the CrowdStrike problem.
What non algorith streaming sites are there? The reason these people appear to be in the majority is because most people who agree wont post " i agree with the protests" under those yea but windows amd etc get through the microsoft driver signing wich is the process where microsoft checks if the drivernis broken or not. The crowdstrike driver got its updates via microcode. Think off the driver as a engine that runs code from a file. The druver was signed but the code it exevuted was broken. I dont know how vanguard handels updates but i guess they take a similar approach as crowdstroke did and only got their “engine” signed but kot the actual code that the driver executes. Else they need to resign their driver every time they donupdates and that wouöd be costry and slow.
Of course it’s not a smoking gun. That’s the wrong metaphor. It’s an extra stick of dynamite that isn’t needed, just waiting to explode at the flip of a coin. That there are other sticks of dynamite doesn’t negate the risk posed by this one.
Huh, seems like you’re right:
Riot Vanguard is an on-boot application. That means if you do choose to disable it and later decide you’d like to play VALORANT, you will have to restart your computer.
https://support-valorant.riotgames.com/hc/en-us/articles/360046160933-What-is-Vanguard
I guess, it’s only user-space drivers which Windows can load at runtime then?
At least, I’m hoping that’s a technical limitation of Windows. Otherwise, this is fucking stupid.Well, it always is fucking stupid, but it would be even more so.
Yes, the key difference being that nobody’s playing Valorant on airport displays. Just yesterday I installed a new early access game for two accounts at home and discovered that it just wouldn’t work with the non-admin account because of anti-cheat. All of this is making me consider going back to running games under flatpak.
Yes, works on the same layer.
In theory, yes. Vanguard uses ring 0 access; and Failures/crashes on the code that are running on that level will lead to BSOD.
In practice, Riot very likely tests Vanguard on various hardware as parts of their tests before shipping updates on it, as it’s used by all players that play Lol and Valorant; and a fuckup like that would mess the trust they’ve built between the players. Players are trusting them to run ring 0 code on their computer, so they can have a cheatless experience after all.
In practice, CrowdStrike very likely tests Falcon on various hardware as parts of their tests before shipping updates on it, as it’s used by a huge amount of enterprises; and a fuckup like that would mess the trust they’ve built with those enterprises. Enterprises are trusting them to run ring 0 code on their computer, so they can have a malware-less experience after all.
Welp, they’re a good example of what happens if they don’t do proper testing.
I’m far from an expert, but Vanguard is a kernel-level program. If a kernel-level program crashed, the whole system crashes. So yes, any kernel-level program could do the same thing CrowdStrike did, intentionally or not.
Kernel-level programs can do whatever the hell they want.
Yes, and I’ve seen it happening. Usually it doesn’t instantly brick every PC, but it can sometimes brick certain PCs with specific configurations. Then it will be silently patched without acknowledgement for the bug.
I’ve seen it mess with (and crash) graphics and network drivers, rendering PCs useless until forced reboot. It can also mess up other games, processes, and even updates.
People have been warning gamers about kernel level anticheats since they were introduced, because no userland code should run with that level of privileges, period. However, people still installed those games not really understanding the threat, and that’s why we have so many games with a kernel anticheat.
Because without the software, we can’t play online. Full stop. Valve has tried to find another way without it and VAC2, but they keep winning and Valve gave up (seriously, play CS2, they’re everywhere)
You don’t give your house keys to your home security system provider. Giving kernel access to anything, even if it’s for your own good, is dumb. People don’t understand the risks that come with it. People just think what the companies tell them to think. As a matter of fact, there are still cheaters in valorant. Vanguard isn’t perfect, it can still be bypassed. VAC works fine for what it is, and it could still be refined. It bans more people monthly than Vanguard.
The biggest reason for kernel level anticheats is your sweet sweet data and more control of your computer. You don’t need them. We have been playing online games since the 90s, and none used kernel anticheats. It was never necessary to sell your computer to Tencent in order to play a game which, again, still has cheaters.
Valve must be doing something right behind the scenes. I haven’t encountered a hacker since May and play regularly. To be fair, back then the game was unplayable.
Helldivers 2 fucked my PC up after one of their updates in May. Game literally became unplayable and corrupted my Steam database twice (causing me to have to reinstall Steam both times).
In PVP games, I can sort of understand the players’ desire to have a cheat free experience, but in purely PvE coop games, it really feels so pointless and is such overkill. Regardless, there are better ways to accomplish anticheat that don’t involve gaining kernel level access. The risk isn’t worth it.
It’s also potentially a infiltration vector for malicious activity.
Genshin impacts anti-cheat has been used to enable ransomware taking over windows computers, and you don’t even need to have Genshin installed.
It was a danger to all windows users just by existing, because the ransomware just came with genshin anti-cheat, which it would install on its own. Because it was a “verified” piece of software windows would just go “oh ok seems cool, go right ahead” and the ransomware would gain complete control of the system through the anti-cheat.
I’m confused. If you don’t install the game, how do you end up with the game specific anti cheat software on your system?
The anti cheat software’s code is in a separate file from the game files, which malware can ship themselves.
Yup.
Yeah really not much else needed to be said here. What happened with Crowdstrike is exactly the sort of exploit Kernel Level Anti-Cheat in general has been critized for enabling on consumer hardware.
And why most Linux users would rather not play these games than allow that garbage on our PCs.
Preface: I’m not an expert in this yet but I’m pretty interested in learning about systems-level topics so if I’m wrong please correct me!
Yes, the thing about anticheats and anti viruses is that they are only useful when they have access to the underlying resources that a virus or cheat engine might try to modify. In other words, if cheating software is going to use kernel-level access to modify the game, then an anticheat would also need kernel-level access to find that software. It very quickly became an arms race to the lowest level of your computer. It’s the same with anti viruses.
IMO the better strategy would be to do verification on a server level, but that probably wouldn’t be able to catch a lot of cheats like wall hacks or player outlines. At some point you just have to accept that some cheaters are going to get through and you’ll have to rely on a user-reporting system to get cheaters because there will always be a way to get past the anticheats and installing a separate rootkit for each game isn’t exactly a great idea.
Could they not hash the contents of the game’s folders and send that back to the server to confirm it’s not been tampered with?
So save files exist. Also custom user content. So the hash will change accordingly. Plus some cheats don’t require a modification of game files anyway, they use memory analysis to get, say, the location of other player objects, then they manipulate local information to give the player an advantage. This is how aim hacks and wall hacks work.
Cheats are hard to prevent for the sole reason of you don’t own the computer they could be running on. You can’t trust the user or the machine, and have to design accordingly. This leads many to the “solution” that is kernel level anticheat, it gives total access to the system.
Games probably do this in some way already with something like a checksum but the problem is you could have some separate program reading from game state/display at runtime to get around this. That’s part of why a lot of cheats are installed at a kernel-level.
What stops you from tampering with the game folder, and changing the function that sends the hash, to send a pre-calculated and valid value, instead of calculating it from real files you are running?
They do do a lot of verification on the server side. Since unreal introduced their server-side-lagged-approval networking model, all local movement and most shooting can be retracted by the server.
But what a ring 0 level driver is looking for is other software, like aimbots, modified assets (transparent walls, custom shaders etc) etc. To be able to detect all that it needs to be level 0.
What I would trust more is if Microsoft acquired one of these companies and worked across the industry to root cheating out. Giving some random company ring 0 access feels completely off to me.
Couldn’t aimbots be picked up as odd movement and be detectable on a server though? Kind of similar to how those “not a robot” checks can tell if a human is clicking on the box just by looking at the movements of the cursor.
In addition, things like textures and game-modifications could be picked up in part by things like checksum verification to make sure the client is unmodified (assuming the files are modified on the disk and not in memory)
I feel like most client-side changes like see-through walls or player highlighting make themselves pretty obvious when aggregated over multiple games. A good user-reporting system could probably catch most of these.
I definitely agree though, allowing multiple random companies to install ring 0 rootkits should not be the norm. Honestly, even a Windows-level anticheat would be problematic because it would only worsen the monopoly Microsoft has on competitive games as a platform. A new solution would need to be cross-platform or else it would only be marginally better than what already exists.
Aimbots dont need to do a lot to provide advantage at the highest level. Moving “perfect aim” from 1x1 pixel to 3x3 pixels, but with 33% probability would provide a huge advantage and be undetectable.
Modified assets cannot be verified unless you lock the system down, like an Xbox. On a PC? No way. You can combat it by sitting in ring 0 (which is what anti cheat software does) but you couldn’t just check some checksums.
In terms of aggregating data and spotting something like see-through walls, there isn’t the statistical method to discern between great intution built over years of playing the same map and having see through assets.
I used to work in AAA game development, across most of low level (graphics, networking, memory, assets etc) so unfortunately I know this problem is nigh on impossible to solve unless you have a locked platform.
One Minecraft server I played on installed a program for blocking x-ray hackers (a type of hack that lets you see valuable ores through walls so you know exactly where to mine).
The anti-xray mod worked by reporting to the user that the blocks behind a wall are a jumble of completely random blocks, preventing X-ray from revealing anything meaningful.
This mod resulted in massive lag, because when you are mining, every time you break a block, the server now needs to report that the blocks behind it are now something different. It basically made the game unplayable.
The server removed the mod and switched to having moderators use a different type of x-ray mod to look at the paths people mine in the ground. Those using x-ray hacks would have very suspicious looking mines, digging directly from one vein to another, resulting in erratic caves. Normal mining results in more regular patterns, like long straight lines or grids, where the strat is to reveal all blocks in an area while breaking as few as possible.
Once moderators started banning people with suspicious mining patterns, hacking basically stopped.
It’s possible to still hack and avoid the mods in this kind of system by making your mines deliberately look like legitimate patterns, but then the hacker is at best only slightly more efficient than a non-hacker would be.
That’s kind of my point with hacks like player highlighting, I feel like a good user-reporting system would get us a lot of the way there. E.g. If someone is using see through wall hacks in an FPS I feel like it would be pretty obvious for other players to tell in a lot of cases. Other times things like erratic movements from aimbots could probably be detected by the server.
Theoretically it should only be running during gameplay, and that’s probably true as I’m sure security researchers would’ve pointed it out if games installed a persistently running rootkit. So it’s different than Crowdstrike which was running immediately from boot.
So there is that, if it caused your PC to crash it should be fine after reboot. The driver has God power though as far as your PC goes so if it was the point of entry for a malicious attack you could be really screwed.
It runs all the time and launches during boot. A ring0 anticheat that only runs while the game is running would be even more pointless
Vanguard is always running at all times.
Honestly no idea why it isn’t considered malware.
For the same reason spyware and adware is not anymore malware categories scanned for by antivirus software: our fucked up economy, and because average people accept everything no questions asked.
Besides that it could, we don’t really know what it does during regular use.
It could transmit sensitive data it collected from your computer and send it to tencent and we wouldn’t be the wiser.
Don’t worry , it runs all the time
Helldivers 2 does the same thing. If this continues it will be extremely advisable to move any non-gaming use-cases to a different computer as you have no idea what the “anti-cheat” is doing with that level of authority over your computer.
I play HD2 under proton. Even if there is a rootkit, it’s sandboxed.
Proton is not actually sandboxed the way an actual container is.
A) if the program running in proton was given root access in some way, say by tricking people into entering their root password for a claimed update, it would have complete normal control of your entire system just like normal.
B)apps running in proton still have access to the regular file system.
Wine isn’t an emulator or a vm.
Or just dont buy those games.
That works until all* games come with root level anti cheat. It was the same with micro transactions which people still defend despite being utter shit.
- Realistically this will never be 100% but it will be enough of the mass market AAA games like CoD etc to mean that if you functionally want to play a game made in the last X number of years you will need to accept this or stop playing games altogether. I think most people will continue to play games. Most people will continue to install root level anti cheat, knowingly or otherwise, and all of them will get fucked by an exploit of that software. They may never even know about it.
That’s why piracy must continue, pirates cracking can just extend to pulling those pesky anticheats out of the single player games.
This is pretty much what I’m afraid of. Even my knowledgeable friends dismissed the implications of a root kit so they could play Helldivers. Like 2 months later and they’re sick of it, and the damage is done. That game was a massive success despite what should have stopped everyone in their tracks.
I think that’s the point that I start just eating the performance hit and running those games in sandboxed VMs where possible
Pro tip: don’t install rootkits.
Things we didn’t think we’d have to tell people in the future.
That unfortunately means, you can’t play a lot of games. And for most people it’s practically unknowable what the installer is doing, they don’t expect a game to nuke their computer.
There needs to be accountability and a certain level of trust. Microsoft shouldn’t allow kernel drivers for crap like anti cheat.
Yet another reason to use Linux. You don’t have to know weather the installer comes with a root kit, the installer will just fail 😎
That’s… not remotely true? Linux can absolutely install kernel drivers. If you mean running windows games under wine then sure, but then we’re no longer talking apples:apples. You could do the same thing on windows by running a game in a VM.
Yup, kernel level “anti-cheat” is a rootkit spyware that “pinky swears” it’s only spying for a good reason.
I don’t understand why a game like Genshin would need a fucking anticheat other than monetization protections
You just answered it for yourself.
Also to absolutely not provide a backdoor for a certain hostile government.
