I agree with your review. I’ve been using Linux since Slack in the mid 90’s and I switched over most of my machines to Elementary. An Alienware with 3090, Airbus laptop with 1080, and a Lenovo with an AMD 550.
Except for NVidia proprietary drivers:
Fastest OS install. I want to play games, not wait for an OS to install and give me 50 pedantic options to step through.
Boots very fast, shuts down just as fast.
Fast Sleep and wake up every time on desktop and laptop. WiFi works, video normal
Clean, stable, consistent GUI that doesn’t do weird things
Bluetooth and audio worked great with no fuss.
As you mentioned, Flatseal is a must. However, I use AppImages as much as possible. I get the security benefit of flatpaks, but all this sandboxing and containerizing creates too many problems with apps that need to communicate with one another, and accessing my files was a serious PITA because of permission issues that needed to be corrected. There are no permission issues with AppImage, but security benefits aren’t there either. However, both work wonderfully with Elementary.
Use AppImageLauncher to automatically create your Application menu items
Heroic Games Launcher was written by wonderful humans!
Cyberpunk won’t work, need to dualboot to Windows. But many windows games work well.
Now, about NVidia: The proprietary driver takes all the horrible fiddling Linux has a reputation for, but reality, is that NVidia drivers are closed source and AMD works with the community. OOTB experience with AMD is flawless.
3090 came up and everything was green, a problem with the Nouveau driver.
1080 everything looked ok
Ran the install, installed the kernel headers, the dev/build packages, mucked around a bit and it works great. However, every time there is a new kernel, the new linux headers and Nvidia module aren’t automatically installed and compiled so it boots to the command line. I know how to manually install them and get back and running, but I haven’t figured out what the problem is yet. Never ran into this on Ubuntu, Fedora or RHEL before.
Kernel & Wayland versions matter. Elementary ships older versions than SteamOS and Nobara. I don’t know how far behind if at all is EOS in relation to Debian 12.
You could try systemd-inhibit as a prefix to any game launch command, much in the same way as gamemoderun is used.
Sidebar: this could be considered an example of the XY Problem. You should consider editing the title since the problem you’re really hoping to solve is your laptop going to sleep while you play a game.
While the XY problem sort of applies to this post, I think labeling it as such might be a little too dismissive, since this is one of those rare cases where OP has correctly identified the root cause of the observed issue.
(Easier workarounds exist, but the ideal solution would indeed be fixing the hardware manager’s input idle detection such that it no longer ignores game controllers.)
Hey folks, just curious, is it a bad idea to switch between different versions of proton-ge? How does it apply fixes related to the prefix creation on a game that already had a prefix?
I don’t know about your second question, but with almost every proton-ge update I’ve gone into Steam and switched over to the new version with no issues so far.
Maybe that’s a no-no for some games but, again, no issues so far.
EDIT: Same goes for wine-ge whenever I have games I bought through GOG and play through Lutris.
Thinking about it, using a new version of proton/wine-ge on lutris does show a setup dialog similar to the one that shows when you first run it so I would assume wine refreshes things on a new version.
It’s not clickbait. The Linux version of Steam consisted of a binary, as well as a script called steam.sh which could be used to launch Steam. There actually was a line in that script with a comment above it saying # Scary!. The script was not supposed to rm -rf the entire hard drive, but if something went wrong, exactly that happened.
Yeah I don’t want to insult the KDE folks but I miss Gnome as well.
It is embarrassing as hell that the Gnome folks haven’t merged that commit after this many years. They also don’t have any concrete steps laid out to the contributors to get it merged. It feels like they just don’t give a shit about section of their community and it’s pretty disrespectful to the original contributor to give them no path forward. End rant.
sched-ext is the proof of concept. It has a demo video if you follow the link to the mailing list in the article showing it improving performance in games that are normally CPU bottlenecked when the CPU is given a heavy background task.
It’s scheduler improvements though, so it’s not really a measurable improvement (or at least not easily measurable) to performance as a whole but it can result is a more responsive system under load.
I have been using battlenet for a couple of years now without account issues. Blizzard has been pretty decent towards Linux users in the past, even fixing issues that broke functionality in wine (there is an example of this on the blizz forums somewhere). Can’t speak for the rest of the launchers since I do not use them.
While I take issue with oversimplification on this, the ship has sailed: Generally speaking, many/most PC anti-cheat solutions are “rootkits” in that they require elevated, often kernel level, permissions to detect running processes and the like.
The linux implementation, while it varies, generally has drastically reduced permissions and is more or less restricted to the wine/whatever container as far as the higher level permissions. But many still have user level permissions and that is pretty much sufficient. At a high level: It can basically run ps -elf and all the fun stuff like that.
So yes, a linux version of an anti-cheat can potentially create more cheaters. Imagine that I run my aimbot/maphack process as a privileged user (because random ass hackers are totally who I want to give root on my machine…) and it either finds a way to overlay on my screen (ha, good luck with the mess that is x11 and wayland) or outputs to a device over the network so I can glance at a laptop or tablet. Without sufficiently elevated permissions, this can’t be detected.
But… I am not aware of any indications of this ever happening. And, for funsies, me and a few buddies used a webcam, off the (github) shelf computer vision libraries, and a bit of off the (github) shelf ML to set up predictive aiming that more or less replicates the idea of “pre-firing” and learning how to watch corners with a near real-time overlay (when we can be bothered to run it on a GPU box). Which does absolutely nothing on the player computer so…
Which is why we are seeing an increasing push for server side anti-cheat solutions. It increases the cost of the server (which can be amortized by server farms) but basically lets “anomalies” be detected. So it doesn’t matter if they can detect you are running a Tarkov map hack. All they need to do is detect that you are tracking enemies you should not be able to see. And we already see suspicions of similar tools being used for things like MOBAs.
So to answer the question: Yes. But there is no actionable data saying it has ever occurred. Although, grain of salt, that also wouldn’t be widely shared.
Okay. All the fanboys please stop reading. I am just going to talk about… uhm… Eternalism and Warframe and the implication on taxation in the Origin System. Yeah, you don’t want to hear that. So please leave.
Seriously. Please leave.
Okay, are they gone? Good.
Yeah, it actually very much can. Not because more linux players are cheaters. But because it increases the support burden. I am not aware of the resources behind The Finals, but adding linux support means you add a LOT more SKUs to have to deal with. And even if your policy is “Linux, go fuck yourself” you still have to deal with the negative PR of “I tried to play this on my nvidia wayland machine using open source drivers and the framerate was low. FUCKING LAZY DEVS!! YOU LIVE AT <insert address here> AND I AM GOING TO HUNT YOU DOWN AND <insert horrifying threats here>” and so forth. You know, Gamer shit.
Which means there is now less time and resources to go around to preemptively test and fix potential vulnerabilities. Because if EAC or Battleye or whatever spots a massive vulnerability, it is only a matter of time until a sufficiently different process exploits the same vulnerability.
Which is the story of everything. Linux adds a support burden (and “improved bug reports” only go so far and are also incredibly subjective) and that means a development burden. Wine/proton goes a long way toward not making that a big one. But if you are trying to push The Next Big Thing in multiplayer gaming, even those first few hours after launch matter a lot for your survival. The moment your game is “dead”, it is pretty much guaranteed to die.
And that also ignores stuff like (maybe?) FUD being spread. The best example is probably that Chinese battle royale game that everyone and their mother accused of being a crypto miner or whatever it was. Not sure if anyone ever conclusively decided, but the game basically was radioactive during the mad rush to find the next ApeLegs.
So if a studio decides they don’t think the improved sales to us linux gamers is worth the potential headaches at a vital time? I don’t like it, but I am not going to blame them.
I’ve been trying to help my roommate get games working on pop_os via proton / lutris because his windows install broke, and we just cannot make any game work. Games that say they should work on protondb just won’t start no matter what we try. I wish I knew what we were doing wrong. Would love to play Baldurs Gate 3 with him again.
linux_gaming
Top
This magazine is from a federated server and may be incomplete. Browse more on the original instance.