[Venn1]

This week we closed the doors on our Linux gaming podcast, which has run continuously for the past 13 years. No fuss, no drama. With the announcement of Steam Machine II (we also covered the original launch), it just seemed like the right time. Proton has evolved to the point where most things work out of the box. Few people are bothering with native support, and it’s become difficult to find new things to cover.

It really feels like everything is lined up for the year of Linux in the living room, and it’s great to see.

[J_McQuade]

I never listened to the podcast, but I see where you're coming from and thanks for doing it anyway.

Twenty years ago I was in university and had a Debian install on a cheap-ass Acer laptop and I managed to get exactly two and a half games working under Wine: the first two Fallouts and about three hours of Civ IV before crash. Getting games to run at all was A THING so a podcast for that makes a lot of sense.

Today I have a full-time job and deleted the Windows partition from my expensive PC about three years ago... pretty much every game I've ever wanted to play since then has just WORKED. Better than on modern Windows, even. Not a lot to talk about there, I guess.

One thing I wish is that Valve could publish a 'Proton spec' that people could build against to ensure compatibility, but I imagine that that this would be an IP nightmare.

[johanvts]

Anticheat is a big issue that nobody seems to mention. I had to go back to windows for online games and it’s my understanding that there are deep technical reasons why anticheat on linux can’t be done the same way as on windows.

[sph]

Not sure what you mean, in every thread there's someone that mentions anticheat as if to stress why Proton is never gonna be good enough.

You can be a true gamer™ even if you don't play the latest $90 AAA multiplayer FPS. To me not having a proprietary rootkit is a feature, and Windows is always there for those that are OK with being spied upon.

[piva00]

Anti-cheat is the only reason why I had to build a Win11 machine for games, and only games, some months ago.

Hadn't touched Windows in more than 10 years, and it's as bad as I remember it, everything is clunky, badly designed, no polish whatsoever.

The moment developers find a way to get their anti-cheat working in Linux I have absolutely no reason to ever boot a Windows machine again...

[AnthonyMouse]

> The moment developers find a way to get their anti-cheat working in Linux I have absolutely no reason to ever boot a Windows machine again...

The trouble is that kernel-level anti-cheat sounds like something useful but it doesn't actually get you anything because the cheat developers are going to analyze and modify the anti-cheat code the same as they do the game code. And then having it running in kernel mode on the non-cheater's PC doesn't buy you anything when the anti-cheat code you wrote isn't actually running unmodified on the cheater's PC.

The cheat developers do have to put in the effort to analyze what it's doing in order for that to work, but the same is true of user-mode anti-cheat. Being in kernel mode doesn't solve or improve anything, it just creates a hazard because then bugs or malware in the anti-cheat code can compromise the entire system and are effectively granting themselves access to things you didn't approve, e.g. a game running as the kid's user account can't normally access the parent's tax returns, but in kernel mode it can. So what you want is for them to stop doing that.

Meanwhile the Windows kernel and the Linux kernel are completely different, so you're not going to be able to take Windows kernel anti-cheat code and run it in the Linux kernel even if you're not attempting to cheat. You'd have to have them to make a Linux-specific one, but you don't want them to, because they shouldn't be doing it at all.

[CuriouslyC]

This is entirely possible using TDX/SEV-SNP, running in a vm alongside a host OS. It's just a big engineering lift. They're almost certainly already working on it.

[wiseowise]

Anticheat is a big non-issue that multiple people mention whenever Linux gaming is brought up. 5 popular anti-cheat games do not outweigh the whole ecosystem.

[snowram]

It is when those anticheats gatekeep the most popular PC games. For most gamers, they can't compromise on what they play and there is still a very large amount of potential games that would forbid a switch to another OS. See : https://areweanticheatyet.com/

[wiseowise]

Most of those are addiction machines and f2p shitholes, that they isolate themselves from the system is a win in my book.

[2OEH8eoCRo0]

sour grapes

[pluralmonad]

Well, sometimes the grapes actually are sour...

[nottorp]

> For most gamers, they can't compromise on what they play

I'm sorry, but 98% of video games are not competitive multiplayer IAP fests.

[Gareth321]

Because it's intractable on Linux and advocates don't want to admit that. The entire security model on Linux is resistant to deeper levels of access and control for applications, which is required for kernel level anti-cheat. While these forms of anti-cheats can't stop cheating, they are clearly more effective than user-space anti-cheats. For 99% of users, we gladly accept these more "invasive" anti-cheats because it means less cheating in the games we enjoy. Linux developers will never allow this kind of access because it is antithetical to their ideological beliefs around security. They gladly exclude any kernel level cheats to maintain the security model. It is a permanent impasse. One which I believe will never be solved with user-space or server-side detection. This is why the most common retort is: "just play different games."

To be frank, the argument that kernel level anti-cheats are invasive has never been all that accurate or compelling. Any user-space application already has numerous privileges which could ruin your day. You trust a developer and application every time you run it, irrespective of its access level. Valve has an opportunity now with SteamOS to impose technologies like SecureBoot and "safe" deeper layer anti-cheats which actually work. Yes, Linux enthusiasts would be up in arms, but it would mean that the most popular online FPS games would be supported on Linux, and I think that's far more important.

[hhmc]

> For 99% of users, we gladly accept these more "invasive" anti-cheats because it means less cheating in the games we enjoy.

The modal user likely doesn’t even know anti-cheat exists, and if they did, wouldn’t care at all. They just want to play the game.

[Gareth321]

They want to play the game without cheaters. That's why studios use anti-cheats.

[Mindwipe]

Well, it's not intractable if it's pushed to the underlying hardware and signed drivers.

Valve could build something into their chipset and start signing the Steam Deck drivers, create secure boot etc and essentially create an Apple SIP equivalent. Wouldn't work for the rest of the Linux ecosystem or other devices, and people would absolutely howl about it, but they could do it.

[keyringlight]

The other side is linux totally permits you to do whatever you like to your system, and then it's similar discussion to DRM (digital rights management, not direct rendering manager). When you're trying to the user from doing things they're not allowed to and the same user can fiddle with the system, there's no starting point for trust.

[sph]

I run Steam in flatpak, so my games are sandboxed and do not have access to my home directory. I don’t have to trust anyone.

[Gareth321]

That's an added layer of protection but it's hardly foolproof. A malicious game/app can still:

* Exfiltrate personal data from allowed Flatpak directories

* Steal data you intentionally open via portals (e.g., documents, password files, wallet backups)

* Store malware or persistence files inside the Flatpak sandbox

* Use network access to phone home data or join botnets

* Abuse CPU/GPU for crypto mining

* Delete or modify files in your home directory if granted --filesystem=home

* Read browser cookies, auth tokens, SSH keys, cloud credentials if home is exposed

* Install persistence via ~/.config/systemd/user/ services

* Global keystroke logging on X11

* Screenshot entire desktop on X11

* Inject fake input events to the system (mouse/keyboard) on X11

* Record screen via portals if user once granted permission

* Gain full FS access if granted --filesystem=host

* Abuse DBus to change system settings or trigger polkit actions

* Install software outside the sandbox (e.g., ~/.local/bin or autostart scripts)

* Interact with hardware via /dev if granted --device=all

* Trigger kernel or driver privilege-escalation vulnerabilities

* Load or execute unsafe third-party mods, DLLs, or anti-cheat binaries

* Malicious patchers or mod loaders downloading external payloads

* Replace shell history or alter aliases to hide malicious activity

* Encrypt local or network-mounted files (ransomware)

* Spread laterally via stolen SSH keys to other machines

* Manipulate GPU/driver calls for rootkit-like persistence

* Abuse Wine/Proton compatibility layers to escape sandbox using native loaders

* Modify dotfiles (.bashrc, .profile) for stealth persistence

* Abuse LAN trust to attack other devices on the network

* Disrupt system performance via thermal abuse (extreme sustained loads)

* Exfiltrate browser sessions or wallet seeds stored in plaintext

* Execute background processes whenever game is launched without user awareness

[likeclockwork]

Same.

This is The Way.

Bonus: No game files junking up my home directory.

[abnercoimbre]

Wait, what was the reason for winding down the podcast?

> Few people are bothering with native support

Was the podcast an attempt to increase porting efforts to Linux? But Proton (and now Steam Machine II) took the wind out of your sails?

[pabs3]

Can we have a link to the podcast?

[watermelon0]

I'm quite sure it's this one: https://linuxgamecast.com/podcasts/

[pabs3]

Hmm, nothing about shutting down on that site?

[tropicalfruit]

> where most things work out of the box

i really doubt this very much. i hope i am wrong.