Easy Anti Cheat (Epic Games) is asking Linux distros for OS lockdown -- how should Ubuntu respond?

Earlier today, Epic made a blog post with a list of demands from Linux distros in order for them to bring Fortnite to Linux. The post hints they have already contacted maintainers, so I think it would be helpful for users to also illustrate their feeling about Easy Anti Cheat (EAC).

I am sure there will be fair criticism from Linux users who prefer to compile their own kernel, have written custom modules, use a different boot manager, have run-time kernel patching enabled, or don’t feel comfortable with EAC reading and reporting their loaded processes. It’s been some years since I compiled my own kernel, so I will defer to others to better champion these concerns than I can. On my post below, I would like to raise specifically the user consent issues since that is an area EAC is quite bad at and there is a potential of history repeating itself.

EAC has a deeply problematic history with user consent

Some years ago back when I used Windows, I had a negative experience of their Easy Anti Cheat software being installed into the ring0 of my kernel with no user consent involved. I simply launched a game from steam, and foolishly accepted the admin privilege request thinking it was just a DirectX redistributable. I was stunned when I saw the Easy Anti Cheat banner briefly appear and disappear and realized it had installed itself without my consent. (At the time, EAC use was not required to be disclosed on Steam’s store, and Steam does not warn you what dependencies it will be installing when it installs dependencies.) I verified in wireshark that it was sending data to its servers without any agreement from myself. Even worse, uninstalling the game did not uninstall Easy Anti Cheat – and EAC itself did not publish any user facing instructions on how to completely remove it.

These days, I am much more careful about root installs and run Steam as its own user and via flatpak. But I want to make sure I don’t accidentally install EAC, especially if it comes to the official repositories. One of Ubuntu’s strengths is the ease of use it has for end users, and that’s been great for installing Nvidia drivers and such. But I worry about a scenario where I enable some package to install, not realizing it’s EAC because it’s named something that obscures it, like systemd-authenticity or something like that. Or perhaps it really is named ‘systemd-eac’ but I don’t immediately realize that’s Easy Anti Cheat and will result in my process list being reported to Epic Games.

Additional EAC controversies

EAC, for many reasons, is a particularly controversial piece of software. Beyond the reporting of user data, it also has a history of ignoring user consent and it (via Epic) has a minority (40%) ownership stake from a state-run entertainment conglomerate whose state has a heavy surveillance history. (Though in fairness it is worth referencing Wikipedia here that Epic has stated that Tencent does not have access to private data nor provides this to the Chinese government.)

More broadly, there have been controversies with other kernel level anti-cheat software on Windows. A notable one is Genshin Impact’s anticheat being used by ransomware actors. (To my knowledge, EAC itself hasn’t been breached in such a way (yet).)

How should Ubuntu respond?

So this post is a long way of saying I would like to advocate to the Ubuntu team to use maximal caution when dealing with any potential EAC integration or EAC packages.

Likely there will be some end users who want to install EAC and are not concerned about the safety or privacy risks. I’m not arguing to stand in the way of these users or make it unnecessarily difficult for them to make their own choices about their computing. It is their choice how to use their computer.

The key being, users should be reasonably informed before taking such a risk. I think any such EAC package should not be in the main repos, but instead in a third party repository to ensure users really want to install it and aren’t accidentally tricked into it by some package sneaking in a long line of other dependent packages. Given EAC’s past consent-hostile tactics, I think this is a reasonable concern and worth raising.

Thanks for reading!

4 Likes

That is a really interesting topic! My understanding is that a sandbox limits what an app can access, but not necessarily what it does with that data, which I believe is mostly the developer’s responsibility. In essence, are you suggesting that, in the future, when installing a Snap from the App Center or the command line, the list of required permissions should be clearly displayed before the installation is confirmed?

2 Likes

I agree with the concerns raised in this thread around consent, privacy and the level of access anticheat software requires, and appreciate the history provided that I honestly wasn’t even aware of.

Separate to the privacy concerns, I think there’s an implementation question worth considering: how could Ubuntu support users who choose to meet Epic’s requirements for competitive gaming without imposing those restrictions on their everyday environment?

My proposal would be to support two separate boot configurations:

  • Normal Ubuntu: Everyday desktop environment, retaining the flexibility to use custom kernels, development tools and other software.

  • Competitive gaming: A dedicated boot entry configured to meet Epic’s requirements, including Secure Boot, measured boot, kernel lockdown in integrity mode, IOMMU support and appropriately signed kernel modules.

Ideally both entries would be selectable from GRUB with the gaming configuration clearly labelled. If necessary it could use a separate kernel and initramfs so that the gaming specific restrictions don’t affect my everyday environment.

This is how I’d plan to approach it personally. I’d opt into the more restrictive configuration when I want to play a game that requires EAC v2, rather than making those restrictions a permanent part of my desktop setup.

However, this doesn’t resolve the consent and privacy questions raised here. Those definitely deserve consideration in their own right. It’s simply a potential implementation approach for accommodating users who want to use the anticheat while retaining a more flexible everyday Ubuntu environment.

5 Likes

Using the officially signed and provided packages, usually doesn’t bother most of the users.

If i run the out of the box settings and packages like bootloader and kernel, what apart from compiling my own kernel is no longer possible under these requirements?

Being protected by a measured and signed boot chain and key material stored away safely in an existing tpm , if one is available, combined with disk encryption, sounds good on paper. Isn’t protecting the users data as well as possible a large cornerstone of the OS?.

We could even have a third “privacy mode” on top for users looking for advanced control? Something like this? https://www.whonix.org/wiki/Qubes

I remember when Ubuntu Core Desktop was presented in 2023, it was mentioned that users could easily switch kernels depending on their use case (specifically citing gaming as an example, in terms of performance). It is possible that the answer to these concerns, as well as being able to play Fortnite and other multiplayer titles, might lie in leveraging all the technology currently being developed in Snapd. Ubuntu Core Desktop - Deep dive

1 Like

It’s an interesting topic and I am vaguely aware that EAC is controversial (not a gamer myself) but I think most of the features they are asking for are enabled in the Ubuntu kernel already, for different reasons…

1 Like

The Windows anti-cheat is way more invasive, though. But for Linux, as you said, right now they’re only asking for basic stuff that totally makes sense for 2026. The real question is if they’ll stop there…

3 Likes

I’m not a gamer, so this doesn’t affect me at all. But, I fully appreciate what you are saying!

Many packages are already in third-party repositories. Examples include Tailscale, TeamViewer, Google Chrome, GIMP.

We have to take those companies at their word that there is no problematic behaviour.

By putting EAC into the official repositories, Canonical could ensure that the package has the appropriate warnings. By forcing EAC into its own repository, Canonical has no control whatsoever.

Therefore, I strongly disagree with your conclusion. The opposite applies: The EAC package should be placed in the official repositories, so that Canonical can have oversight.


An alternative is to have EAC packaged in snap or flatpak (the former is a bit more secure, while the latter would be accessible to many more distributions). Then the sandbox comes into play. EAC might even benefit from using the sandbox, because they could control what affects the running of their product.