EA / Respawn now block Apex Legends from running on Linux and Steam Deck

Genuinely asking here, what is the path forwards for this situation? I do not see how it can be solved. I love linux and the fact it is completely open, but that of course means it's also undoubtedly going to be easier to be malicious in. I will never switch to windows as I haven't for the few decades I've been alive (and we never had many of these games working on linux without a headache back then, aside from ID software <3). I am wondering however if my playable library is going to keep dwindling when it comes to multiplayer games. I love PVP shooters; I have many hours in Titanfall 2, COD WWII, Urban Terror & Wolfenstein, Q3A, even a chunk in Apex and Battlefield 4/V/1 which of course is no longer possible to play.

Unless something like PlaysafeID actually takes off - and let's be real, how many of us really want to submit our ID to play games? I don't see the pathway to take. Server side anticheat is unfortunately a dream (Poor Counter Strike), and cheats only keep getting more sophisticated. Is this problem solvable?

Quoting: BeamboomWhat's the server-side difference between a player with great game sense, and one with wallhacks?

You don't send the position of other players to someone when they're not in view.

if the ammount of linux players is THAT SMALL as they claim, they should have no problem refunding then for all their in game purchases right? RIGHT?

Quoting: Liam Dawe

Quoting: BeemerThey should put up stats. I want to see:

• Total Linux user count pre-block


• Total of Windows user count


• Pre-block count of users allegedly cheating


• Post-block count of users allegedly cheating

The statements of "we've identified Linux OS as being a path for a variety of impactful exploits and cheats" and "this will impact a small number of Apex players," seemed a bit at odds. If it's such a small number, how are the hacks "impactful".

They install a cheat into your system to prevent cheating. It honestly should be illegal to do that.

They have all the metrics they need on their end to kick users or ban them. This is exactly what A.I. expert systems are for.

This type of comment *always* seems to appear, and I'm seeing a lot of it across social media in reply to this news. I'm really surprised people don't understand, so here's it in simple terms.

There's a difference between counted players, and cheaters when these situations appear.

The number of Linux players will be (in percentage terms) low. The statements aren't at odds at all. Even a small number of cheaters can cause huge problems for the whole of the player-base, of which is mostly not-Linux.

You could have 100 Linux players, 1000,000,000 Windows players and 2 people on Linux doing cheats that affect everyone. That's what they mean. We've seen this info repeated by various developers, that Linux enables the cheats because it's harder for devs to block.

Hope people get that now.

Sure. But that's going by the assumption that people on Windows don't cheat or that they can effectively block 100 percent of cheaters on Windows. Which they can't.

The argument of "People on Linux cheat so fuck them!" is completely bonkers, as that would mean they should also block people using Windows from playing, because they cheat.

Who cares if 2 of the cheaters use Linux? If they are that concerned they should release an Apex Machine and only allow playing the game with that.

It's just treating Linux users as second class citizens because they are a minority. Nothing more.

Quoting: TurkeysteaksGenuinely asking here, what is the path forwards for this situation? I do not see how it can be solved. I love linux and the fact it is completely open, but that of course means it's also undoubtedly going to be easier to be malicious in. I will never switch to windows as I haven't for the few decades I've been alive (and we never had many of these games working on linux without a headache back then, aside from ID software <3). I am wondering however if my playable library is going to keep dwindling when it comes to multiplayer games. I love PVP shooters; I have many hours in Titanfall 2, COD WWII, Urban Terror & Wolfenstein, Q3A, even a chunk in Apex and Battlefield 4/V/1 which of course is no longer possible to play.

Unless something like PlaysafeID actually takes off - and let's be real, how many of us really want to submit our ID to play games? I don't see the pathway to take. Server side anticheat is unfortunately a dream (Poor Counter Strike), and cheats only keep getting more sophisticated. Is this problem solvable?

I have several conflicting answers.

Theoretical and practical.

Theoretical:
In theory kernel side checking doesn't actually add much.
It can always be edited out in the binary and mathematically users with physical access have root.
In that sense Linux openness doesn't do much damage.

Techinical:
If you ask the anti-cheat makers like easy anti-cheat and battle eye. They will tell you. "No problem."

This hides a lot of complexity here's a small slice of the technical state of the matter:

There're ways to do anti-cheat in a system you don't have superior control over, but each of these ways work on Windows too.
Server side code.
Obfuscation.
Captcha's.
sgx
etc.
Some of these ways are currently in use and others aren't.

Yet Windows has a significant amount of cheaters, so apparently it doesn't work.
I currently suspect only one way of doing this that might be Linux specific and nobody has tried it yet in that sense.
Ring 0- resource hogging(My hypothesis is that you might be able to hash check entire resource states in ways that would if subverted on the same system result in detectable longer load times. You're allowed to ask further I'm pretty proud of the trick.).

On Windows you can't do much you can't do on Linux as an anti-cheat writer, but you can feel safe knowing that you're one of the few with access to the kernel, while in Linux that number is potentially infinite.


By the way:
Kernel checks still work on Linux they're just slightly less effective.



Market-political:
Personally I suspect this is currently what we're encountering.
There're several players in the industry to which Linux gaming is not a profitable business anyway(not enough market share), but an interesting chess piece you can freely move.
EA has chosen to move against Valve who has incorporated the piece in their strategy, so breaking Linux is a good business move.

A comparable much bigger problem in this case is hardware attestation Android keyboxes:
Basically all android phones have hardware backed keys deep in the firmware(TEE).
These keys are the root of trust used for drm, custom Rom blocking, anti-cheat etc..
Each android phone Vendor has their own keys.
In the past Google(Seller of Pixel and Winedivine drm) has revoked several vendor keys(keyboxes), because the devices proved vulnerable to attack and the keys could be extracted out of the device, breaking all the functions reliant on these keys.
Not long ago someone pulled this on the beta version of the Pixel phone. At first Google revoked these key too, but that significantly angered their customers who want to be able to use a phone as such they reinstated the key and now there is a consistent crack for all these functions I mentioned before.
The same has happened before with all Qualcomm devices until 2013.
Google could revoke all these keys, but they would do too much collateral damage and as such they accept the damage.

For this problem I see two fixes(I'm a programmer not a marketer)
Positive reinforcement: grow the Linux userbase and general Linux gaming profit margin.
Negative reinforcement: Launch a fork of Proton/wine that specifically fixes detection attempts in blocked games.


My conclusion: the current primary obstacle for anti-cheat on Linux is market-political and the currently employed method against it is growing the Linux gamer userbase.


Edit:
Also a true and tested way nobody uses to do anti-cheat would be to utilize ASLR(Address space layout randomization).
This works on both Linux and Windows, but with the right features enabled in the kernel Linux does it much better.
It's easy to test for.
This is also something a motivated programmer/Valve could do to make Linux stronger in general and more attractive to anti-cheat.
Implement distribution time ASLR in to their choice a game engine or the Linux kernel.
Currently ASLR is implemented at boot or compile time.
Neither case would be enough for anti-cheat.
Compile time makes people capable of cracking it for entire distros and a bootkit could crack boot ASLR.
What you would need is a distribution script that implements a new seed every download and if you implement it for the kernel also a way to check that we're dealing with a kernel with the feature enabled.
The best way I can think off is a distro signed cryptographic secure hash of the seed used.

Implementing it in the kernel has as "benefit" that it doesn't benefit Windows.
Implementing it in a game engine is more freedom preserving(you can make code modifications of specifically the game hard instead of the entire kernel)



Boot ASLR can be kind of for checked for by checking the bootloader.
The real problem is detecting compile time seed generation and seeds that are distributed with the kernel.
Compile time seed generation can be pretty well detected with a blacklist(the only issue was that it could be too predictable), Seeds distributed with the kernel can be pretty well detected with signing an arbitrary part of the ASLR'ed kernel.

A way this could be improved would be to create assymetric ASLR, so not only verified distros can launch anti-cheat compatible kernels, but that is a dream. I'm not yet convinced that's even possible.


ASLR was designed against return-to-libc attacks. I don't actually think this is the best way(I would advise for VAC(voluntary access control) if programs themselves clearly state what they don't need they can prevent themselves from being exploited with tricks they don't use themselves.


Last edited by LoudTechie on 5 November 2024 at 12:23 pm UTC

Quoting: TurkeysteaksGenuinely asking here, what is the path forwards for this situation? I do not see how it can be solved.

The only way you can prevent cheating is by controlling (restricting) people’s actions. The only way you can catch cheaters in the act is by scrutinizing every action they take. This applies to everything outside simply video games. Sports competitions have blood tests to detect drug use. Exams have invigilators to spot concealed cheat sheets and other tricks. Card games have people watching for a literal card up your sleeve. Casinos only trust their own dice. Some methods are more invasive than others, but in each example they have the considerable advantage of holding the competition in person to visually keep everyone in check.

Online play lacks that last lever and is built on eminently fallible platforms with multiple points of vulnerability. User input must be sanitized to account for what’s humanly possible, but in real-time 3D games with freedom of movement, that’s a boggling amount of parameters to account for and since latency kills those games, you have very little time to make a decision when it comes to proactive processing (after-the-fact analysis and sanctions are another story of course). I imagine it’s a lot easier to validate a player’s actions in a turn-based game with a small set of well-defined possible moves.

So you have to restrict the player’s actions, and watch them closely in a way that you can trust what you are seeing. That’ll never happen if you don’t control the machine they use to interface with the game. And even if you do, there’s always things like having accomplices watch a competitor to gain illegitimate strategic knowledge in near-real-time. Remote competitions are a flawed concept altogether, as far as I see it. Rootkits are both necessary and still insufficient: you can’t spy on the player without them, but they still won’t catch everything either. So make your choice: accept that there will be cheaters and take reactive action to punish and disincentivize them, or don’t compete via the Internet. And since cheaters have no trouble evading bans in free-to-play games, the former is also a lost cause as long as you’re not playing in person.

Yeah I don't EA will curb cheating by blocking the Steam Deck/Linux gaming. Just saying.

They don't care about gamers at all. Firstly they blocked support for EA Sports WRC on steam deck adding kernel level anti cheat for a single player game. Now they are blocking game with anti cheat compatible with Linux. I understand that they want to fight with cheaters more effectively but I don't think blocking Linux users (and steam deck) is a good way to do that. They written that stupid things only to tell us they don't know what to do and they think they do well. They care only about money, not gamers. Hoped that MS will block that sh*t but I see they won't. Pity.


Last edited by marcin1509 on 1 November 2024 at 3:33 pm UTC

My (admittedly debased) two cents on an already well-beaten dead horse... all this will do is further drive the cheaters to improve the cheating tools on Linux.

Quoting: jensWhy don’t you just apply for the role to fix this?

I don't like corporations enough to want to work for one.