Latest 30 Comments

It looks promising, but currently its more like a tech demo (..), will see when it gets more developed

I was about to say Xpander is already warming up their steering wheel 😄

View this comment - View article - View full comments - Report this comment

No LLVM makes life so much easier when compiling for Android and/or Waydroid. If I don't disable LLVM, clang++ can't find certain headers and fails very quickly lol

View this comment - View article - View full comments - Report this comment

I have multiple friends and family who previously had no interest in the Deck and certainly no interest in Linux who are now planning to buy the Machine and some also interested in the Frame.

When I asked why, the Machine has really resonated with them. They've seen my Deck and think it's good but wasn't for them as they just didn't want a handheld and it is now a bit underpowered for their tastes. Most are gamers and either already have much older machines they were planning to relace and so this is still a upgrade for them, or they have a decent PC already but wanted something small and quiet in an HTPC style for their lounge or bedroom. They really like the small form factor, the decent but not power/thermal hungry specs, the resulting low noise factor, and the console ease of use of the device, while still allowing them flexibility as a PC for game modding and custom changes if they decide to.

A number of them were not exactly anti-Linux but they were definitely Windows first and generally shied away from Linux in any form. Seeing SteamOS over the last few years, even on a handheld they didn't want, has changed their minds.

View this comment - View article - View full comments - Report this comment

Can someone please make a GabeCube startup video...

It's not exactly what you want, but it's close.

https://steamdeckrepo.com/post/YozvX/gabecube_startup_video

View this comment - View article - View full comments - Report this comment

Can someone please make a GabeCube startup video where all the cubes rolled look like companion cube colors edges when the cube jumps into place in the middle it becomes a heart.

That would be legit.

View this comment - View article - View full comments - Report this comment

yeah, meson is confusing with false vs disabled. disabled is probably fine for it.

View this comment - View article - View full comments - Report this comment

https://www.gamingonlinux.com/2025/11/wine-10-19-released-as-we-head-towards-wine-11/?comment_id=285780

thanks now appear this:

meson setup build64 --prefix=/usr --sysconfdir=/etc --localstatedir=/var --libdir=lib/x86_64-linux-gnu -Ddri-drivers-path=/usr/lib/x86_64-linux-gnu/dri -Dplatforms=x11,wayland -Dvulkan-drivers=amd -Dvulkan-layers=anti-lag,device-select,overlay -Dtools=drm-shim -Dglx-direct=true -Dgbm=enabled -Dgallium-extra-hud=true -Dlmsensors=enabled -Dllvm=disabled -Damd-use-llvm=false -Dgallium-va=enabled -Dvideo-codecs="all" -Dgallium-drivers=radeonsi,zink -Dbuildtype=release


mesa 26.0.0-devel

Directories
prefix : /usr
libdir : lib/x86_64-linux-gnu
includedir : include

Common C and C++ arguments
c_cpp_args : -mtls-dialect=gnu2

OpenGL
OpenGL : YES
ES1 : YES
ES2 : YES
GLVND : YES

DRI
Platform : drm
Driver dir : /usr/lib/x86_64-linux-gnu/dri

GLX
Enabled : YES
Provider : dri

EGL
Enabled : YES
Drivers : builtin:egl_dri2 builtin:egl_dri3
Platforms : x11 wayland surfaceless drm xcb

GBM
Enabled : YES
External libgbm : NO
Backends path : /usr/lib/x86_64-linux-gnu/gbm

Vulkan
Drivers : amd
Platforms : x11 wayland surfaceless drm xcb
ICD dir : share/vulkan/icd.d
Layers : anti-lag device-select overlay
Intel Ray tracing: NO

Video
Codecs : vc1dec h264dec h264enc h265dec h265enc av1dec av1enc vp9dec
APIs : va vulkan

LLVM
Required : disabled

Gallium
Enabled : YES
Drivers : radeonsi zink
Platforms : x11 wayland surfaceless drm xcb
Frontends : mesa va
HUD lm-sensors : YES

Perfetto
Enabled : NO

Teflon (TensorFlow Lite delegate)
Enabled : NO

User defined options
amd-use-llvm : false
buildtype : release
dri-drivers-path : /usr/lib/x86_64-linux-gnu/dri
gallium-drivers : radeonsi,zink
gallium-extra-hud: true
gallium-va : enabled
gbm : enabled
glx-direct : true
libdir : lib/x86_64-linux-gnu
llvm : disabled
lmsensors : enabled
localstatedir : /var
platforms : x11,wayland
prefix : /usr
sysconfdir : /etc
tools : drm-shim
video-codecs : all
vulkan-drivers : amd
vulkan-layers : anti-lag,device-select,overlay

Found ninja-1.12.1 at /usr/bin/ninja

Only need change -Dllvm=false to -Dllvm=disabled as suggest meson a minor thing

View this comment - View article - View full comments - Report this comment

Oh! And I just tought about the fact that the Steam Frame, with FEX, could become the development basis for an Arm based Steam Deck 2. Who knows?

View this comment - View article - View full comments - Report this comment

One thing is particularly funny: they let you play from any shady Windows machine. But they block an console-like system.

View this comment - View article - View full comments - Report this comment

Crazy, I tried Steam Link yesterday and today and still getting the Steam VR - Error 499 (Video Encoder) issue on Arch (EOS) and Fedora 43 - two separate AMD systems.

ALVR continues to work without issue.

Its funny because a couple beta versions ago it was working perfectly with Steam Link :(.

View this comment - View article - View full comments - Report this comment

@Mohandevir Many anti-cheat vendors offer Linux compatibility nowadays, but there're often reasons not to use this product/feature.
VAC is the most obvious solution.
EAC and BattleEye work too.
Denuvo can be considered an anti-cheat provider that works on Linux.
You also have that anti-cheat solution whose name I always forget that only works on Steam Deck.
Fairplay is an interesting story. It's [certainly not an internally developed solution.](https://www.fairplay.ac/) It's one of the more extreme interpretations of server side anti-cheat. By doing most to all of the gaming at the server side it's easy to check for cheating.

Interesting thing I found:

Spoiler, click me

The FACEIT league actually has its own anti-cheat system, but if that one worked on Linux I would be pretty surprised.

View this comment - View article - View full comments - Report this comment

@CIAPA Do the xinput_calibrator and or xinput count.
xinput_calibrator is an x.org frontend for xinput, but straightly assumes you want your mouse to point up and do nothing fancy.
xinput is a pretty intuitive cli tool for managing this.

View this comment - View article - View full comments - Report this comment

May be I'll open a Mesa bug to suggest the above.

View this comment - View article - View full comments - Report this comment

I don't understand why people recommend balena etcher, when you have something as efficient, lightweight and easy to use as those 2 small tools.

Because the writer tool only seems to work well with Linux Mint installers. It corrupted the installers for some of the other distros I've tested. Tuxedo OS comes to mind...

View this comment - View article - View full comments - Report this comment

Rumour has it that this uses Denuvo. That, the launcher and the microtransactions are enough for me to say "no thanks."

@soulsource

On Switch, at least, you could work around the login-requirement by disconnecting from the internet. I don't know if that's an option on PC.

I can confirm that, at least in Windows, you can set the launcher to offline mode. It's a bit of a hassle, however.

View this comment - View article - View full comments - Report this comment

Strange. If it's an official stable version, odd that it's not listed as an available option on the Compatibility tab.

As Liam said, however, it can be found by searching the Steam Library manually.

I've tried Proton 10 and GE 10 on a couple of games over the past month or two and had zero success even getting them to run. Had to fall back to 9 which has been flawless. I'll probably wait a while to try 10 again.

I've experienced this too. Steam is set to default to Experimental for me. A number of my games wouldn't launch when Experimental moved to the 10.0 base. I had to switch them to a stable release based on 9.0 to fix the issue.

On the other hand, I just picked up Stray (52% off at Fanatical). It actually runs with better frame rates on the 10.0-based Experimental (and GE-Proton 10.25). It was appreciably slower on Proton 9.0-4.

View this comment - View article - View full comments - Report this comment

Interesting issue i trying compile without llvm but in my case dont work show llvm is required for radeonsi

It works fine without llvm now. Radeonsi recently switched to aco compiler by default. You need to configure the build with meson to turn off build dependency on llvm like this:

...
-D llvm=false \
-D amd-use-llvm=false \
...

Not sure why they didn't make the second switch be automatically set to false when the first switch is false, may be they'll fix it, but for now the above works.

If build fails, I recommend reading meson.build to figure out why. The above was clear just by reading the file.

See: https://gitlab.freedesktop.org/mesa/mesa/-/blob/main/meson.build?ref_type=heads#L226

View this comment - View article - View full comments - Report this comment

IMO the trends are showing that competitive shooters is a bubble. Not only are many of them struggling to get the sales they want lately, but given what's going on with Arc Raiders it's clear there's a narrative shift (and consequently why so many games "journalists" are talking negatively about it).

I'm not saying competitive/online shooters will go away, but that the level of competition, and the amount of it, people want isn't actually what everyone thinks. A lot of this is being driven by streaming; whether that be fortnite (which in it's own regard will sustain itself), DayZ, rust, or Tarkov, so much of this is propelled by streamers.
Ironically I think the streaming world also is causing a drive for cheats/hacks, since there's a large enough percentage of each game's community that play that game for 8+ hours a day which forces you to be way better in order to just have fun at the end of a long day at work.

Couple all that with the move to external cheating solutions and Kernel A/C is going to collapse sooner or later.

Time will tell, and Arc Raiders will be one of the best indicators on where things will be going, but I can't imagine that invasive A/C actually has a future.

View this comment - View article - View full comments - Report this comment

All this got me wondering... Totally aware I'm probably off the mark, because things have probably changed, since then, but some years ago, I was an active World of Tanks player. I was using PlayOnLinux and cheating didn't seem much of an issue. I may be wrong about that, though... This said, what Anti-cheat Wargaming is/were they using? Fairplay, if I understand correctly? It's not kernel level AC? Is it an homemade solution?

Hearing the overall discourse about AC, It feels like EAC and Battleye are the only viable solutions, for Linux support... Why such a focus around these two systems?

View this comment - View article - View full comments - Report this comment

View this comment - View article - View full comments - Report this comment

https://www.gamingonlinux.com/2025/11/wine-10-19-released-as-we-head-towards-wine-11/?comment_id=285774

Interesting issue i trying compile without llvm but in my case dont work show llvm is required for radeonsi

for more information can see this:

https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38070

View this comment - View article - View full comments - Report this comment

If a game deliberately doesn't support Proton or Linux, they're not serious about getting my money.

View this comment - View article - View full comments - Report this comment

A few cool things happened recently:

* Wine finally works without dependencies on 32-bit.
* AMD drivers (radv and more importantly radeonsi) now work without any dependencies on llvm, so you can build AMD related Mesa without installing llvm either.

View this comment - View article - View full comments - Report this comment

insane this guy wants to be taken seriously when he says

When we stopped supporting Linux, users made up less than .01% of the total player base, even if that number has doubled, or tripled, it's not worth it.

they really are just making shit up to defend their bone-ass decision making

I’m very comfortable saying that if a game supports Proton or Linux, they’re not serious about anti-cheat

and yet consistently games like Gunfire Reborn or ARC Raiders have less cheaters than your slop

Developers won’t be saying this kind of thing repeatedly if it was a flat out lie

We pretty consistently see no change in cheating when these things are rolled out. It's pretty safe to say it's a flat-out lie. Mandatory I'd say. Dishonest to say otherwise dare I say.

View this comment - View article - View full comments - Report this comment

Client-side anti-cheat is nothing more than a vector for deployment of back-doors and root-kits.
Furthermore, client-side anti-cheat is nothing more than a pointless placebo that can never, ever, prevent cheating.

View this comment - View article - View full comments - Report this comment

How sim-heavy are these games ? I'm highly interested in a good Rally game, as long as it's fun with a controller and doesn't require the steering wheel etc.

Open to other Rally game recommendations as well

Assetto Corsa Rally (ACR) is far heavier on the sim side than the arcade side. If you try to drive it like it's DRIVE Rally or WRC (both more arcade-y, more controller-friendly games), you're gonna have issues. I've found WRC 9 to work pretty well on the Steam Deck (just had to bind the back paddles to shift up/down).

View this comment - View article - View full comments - Report this comment

I know I'm late, but very happy to hear this news.

I am hoping they are also planning another vr release. HL Alyx has been fantastic but it is time for a new release to go with the new hardware.

Long live Valve's love for linux. :)

View this comment - View article - View full comments - Report this comment

"Microsoft is moving towards limiting Windows kernel-level access for security reasons, a shift that will impact kernel anti-cheat systems over time. Instead of granting direct kernel access, Microsoft is developing its own security sensors in the kernel that user-mode applications, including anti-cheat software, can utilize for checks."

I think that the real reason to not enable it on linux, is that we are just not enough users that they care enough about us to find a solution. The good news is that the linux gamers are constantly increasing, so that at some moment they will find a solution. Probably in the form of a linux module that one would have to install. Let us just hope that this will not take another 10 years.

View this comment - View article - View full comments - Report this comment

@phil995511 I couldn't find commercial cheating software, so you appear to be right about that part.
[I did find open source cheating software.](https://alternativeto.net/software/cheat-engine/?platform=linux)
[Here an opensource cheat for old versions of half-life.](https://github.com/UnkwUsr/hlhax) and I know that several Windows cracks are also Linux compatible.
@BlackBloodRum might have more insight in the commercial side though.

View this comment - View article - View full comments - Report this comment

@kmturley Linux does something [similair/better](https://en.wikipedia.org/wiki/Linux_namespaces).
The problem isn't and has never been Linux features.
Android and TeslaOS are Linux distros with a high focus on security and deeply trusted by games.
RHEL is deeply trusted by production software.
Most to all their features are integrated directly into the Linux kernel you use.

The problem is trust: the games don't trust it, because you could've modified the kernel to lie to them and the users don't trust it, because they've no easy way of checking this is truly what is happening.

So, either one of the parties has to gain trust or the source of trust for the vendors has to be moved outside the kernel, since I'm a programmer and not a social sciences student I propose the second.

Edit:
Places where this trust can be moved to: the development environment(homeomorphic encryption), external servers(game streaming) or hardware modules(TPM)

View this comment - View article - View full comments - Report this comment