Wayland. It comes up a lot: “Bug X fixed in the Plasma Wayland session.” “The Plasma Wayland session has now gained support for feature Y.” And it’s in the news quite a bit lately with the announcement that Fedora KDE is proposing to drop the Plasma X11 session for version 40 and only ship the Plasma Wayland session. I’ve read a lot of nervousness and fear about it lately. So today, let’s talk about it!
I replaced my Nvidia with an AMD graphics card last year. Ever since ive been using Wayland on KDE Plasma without any issue. I have 2 VRR Monitors connected with different refresh rates, which felt clunky on X11 and now feels fluid and just brilliant to use. I don’t use X11 Sessions at all anymore and only have XWayland for stuff that requires it
Wayland+Plasma 5.27 feel pretty close to usable, so I’m hopeful that with Plasma 6 I can finally just pick Wayland and stick with it and not have to back to X11.
If Nate thinks that wayland only on Plasma 6 is the way to go then I feel like we as a user should trust him.
Currently I use AMD hardware and for me KDE + Wayland is fantastic experience. Some crashes here and there, with which I’m fine with.
My brother is running Fedora Kinoite + Wayland with Nvidia and he is total noob when it comes to Linux. He is happy and never complained about his system.
So, I would only assume for normal use case wayland is already in good shape currently. It is only going to get better with plasma 6.
If Nate thinks that wayland only on Plasma 6 is the way to go then I feel like we as a user should trust him.
Well… that is not what he is saying. X11 will be supported for a long time still. In fact KDE has not even set a deadline for ending support for X11^*^. It will definitely not end with the release of Plasma 6.
The point is that not adapting software to Wayland is a mistake. It may be a pain, but X11 is virtually abandonware. The developers have moved on (to Wayland) and there are no new versions coming out – unless someone forks it, of course, but that would probably be another mistake, as the codebase is an unsustainable mess.
This implies that, yes, when most software projects have got their applications working on Wayland, X11 will be phased out as a platform Plasma works on, but there is no date for that yet.
– ^*^ Other projects are less coy. Fedora is considering removing support X11 from their very soon, maybe in their next releases. This is what sparked the discussion. Not KDE.
Are you using the Jerboa client? I think they recently introduced an option to open links in a private tab which is on by default for some reason. It confused me too until I found the setting.
One problem that has long plagued X11 is that any app can snoop on any other app, including things like keystrokes and displayed information, even from within containers like Flatpak. (This is understandable, since it was designed at a time when spyware was rare, so there was no need for isolation more fine-grained than the user level.)
IIRC, Wayland didn’t address that problem in its early days, but in these modern times of surveillance capitalism, I suspect it has been getting more attention. It would be nice to see it solved.
The risk existed but did it plague X11? I never heard about any app logging keystrokes and sending theme somewhere. Where there any attacks using this? I don’t think normal uses had to worry about it.
Yes, and it still does. Practically every X11 installation is vulnerable.
(If you’re nitpicking my use of the word plagued, though, note that I am talking about the vulnerability, not the exploit.)
I never heard about any app logging keystrokes and sending theme somewhere.
That’s because of a variety of external factors, including:
X11 desktops aren’t common enough to be priority malware targets, yet.
People who run only open-source software typically get it from trustworthy channels, like their OS distro’s package repository.
Devices likely to attract malware, such as game consoles and mobile phones, have avoided X11. (Android phones and Steam Deck are examples.) This is no accident; lack of app isolation was a factor in that decision.
I don’t think normal uses had to worry about it.
We’ve been lucky so far, in that our circumstances have kept us mostly safe. However: Linux malware is on the rise. Commercial games, both on their own and through anti-cheat systems, are making opaque software more common on our desktops. Flathub is working on paid apps, which could likewise create malware opportunities that weren’t there before. The Epic Game Store has already been caught collecting data from other apps, so the intent is clearly present already.
It’s generally just a matter of time before exploitable systems become exploited systems. We would do well to close the door on unauthorized key logging, clipboard snooping, screen scraping, and input injection.
And all the arguments are like this. “It’s good to use it”, “it has features”, “it’s better code”. But it’s never “it has essential features that people need”. Because it doesn’t. If it did people would use it.
You have misunderstood me. I don’t use or promote Wayland, mainly for the very reasons you just listed. But I do recognize that it has the potential to solve real problems that are deeply embedded in X11. If/when it gets there, and fixes various deficiencies that it has today, I expect I will have a good reason to switch.
But it’s never “it has essential features that people need”. Because it doesn’t. If it did people would use it.
Actually, I believe it does have such features for people with certain hardware setups. I just don’t happen to have such a setup.
It does for me. For some reason my touchpad has really high scroll sensitivity with libinput. It’s borderline unusable. The only desktop environment that exposes the ability to change this sensitivity is plasma Wayland. AFAIK there’s technical reasons it can’t be done on xorg without hacky workarounds. This is the killer feature for me.
In addition both plasma and gnome only have 1:1 touchpad gestures on their Wayland sessions. Obviously I could use third party tools for trackpad gestures under x11 but those aren’t 1:1.
Also while I’m aware that fractional scaling on Wayland is a mess and hacky but I still find the fractional scaling implementation on KDE Wayland to be the best, followed by KDE on xorg. I need fractional scaling for things to be appropriate sizes on my laptop screen.
For my desktop I still use x11 because of nvidia but I would definitely benefit from the multi monitor improvements under Wayland since I have two monitors of differing refresh rates and it causes issues.
Nope, I also use it for many of these things. They’re not alone by a long shot.
If you want to continue to use X11, you are free to simply not update your machine any further. It’s unlikely you value security, so this shouldn’t be an issue for you.
Maybe a while ago that was true. But there’s so much cool stuff that KDE devs are spearheading with Wayland.
For example, You’ll be able to reboot the window server without ending your session and losing your app state! They’ve demonstrated being able to swap between GNOME, KDE, Sway, and other WMs without logging out or crashing apps. This could also be used for swapping active GPU configurations without relogging, which would make Gaming laptops way less shitty to use.
Wayland can also store window state to disk, which isn’t possible on X11. Another useful feature that could allow for more fluid hibernate and reboot behaviour.
Touchpad gestures! You need a lot of dev effort to get them working on X11 but on Wayland it’s very fluid.
Wayland is also partly why there’s been new effort to standardise the desktop experience on Linux with stuff like XDG.
For the end user, X11 is fine. You don’t need to particularly care how your windows are drawn. As an app or desktop dev you’ll be way more empowered to build a next generation desktop experience with Wayland, in a way that X11 just wasn’t able to support because of its underlying design.
But that’s also why the change can’t come from begging end users to migrate: we have to rely on distros dropping support for X and making Wayland the default.
mixed DPI scaling (can drag a window from a 1x screen to a 1.75x screen and have it look correct on both at the same time, even when halfway across each)
Mixed refresh rate (my center monitor is higher refresh rate than my side monitors, X11 just baselines all monitors to the lowest common denominator).
Mixed variable refresh rate (center monitor is VRR capable, side monitors are not).
HDR support soon (already exists in GameScope).
Mixed HDR/SDR output across monitors
Performance:
Lower resources
Smoother operation (can be felt in mouse cursor movements, window drags, composited animations, etc)
Better VR headset isolation compared to X11 (allows the headset to run separately and not interrupt regular monitor layout, and also lets it run freely at the correct refresh rate)
Other:
Better security between apps (yes I actually use this and count it as a feature)
App video isolation leads to pipe wire functionality, which is a bonus and makes OBS work better overall
I know for a fact I’m forgetting something because this list was longer the last time I wrote it out, but I think you get the point.
Wayland works fine on Nvidia. And as long as NVidia is coddled with Xorg the longer they’ll not bother with Wayland. About time they got their act together
Again, you’re not the target of my comments. I’m talking about people who continue to buy nvidia after switching to Linux, and then bitching that it doesn’t work, especially with Wayland.
The reason I’m speaking up is that I am sick and tired of people buying nvidia and then bitching that it doesn’t work. Not people who already had nvidia hardware or received it secondhand. People who keep buying nvidia laptops and cards and bitching that it doesn’t work all the time, especially with the transition to Wayland.
I stated the reason that this is the case, confirmed by the leader of the kernel, and you’re turning it into “I don’t care what Linus thinks.” It’s not elitism. The fact is that nvidia doesn’t care about Linux as much as Intel and AMD do. That’s just facts. And there’s no hope of this ever changing unless Linux users start boycotting nvidia.
Why do you care so much about those people? Why not just let them live with the consequences of their own actions?
As a Linux user, you will never be able to boycott Nvidia. Linux users make up about 3% of computer users. It won’t matter to Nvidia if 3% of anything boycott their products.
I’ve used Linux long enough to know that refusing to be complacent can lead to positive change. I’ve seen it firsthand.
We didn’t always have such good hardware support on Linux. People refused to accept crappy binary blobs and ndiswrapper for other things, and won. Having the attitude that you don’t want to listen to Linus because you love nvidia so much doesn’t help.
Okay. I’d like to know how a boycott will lead to positive change.
According to Steam’s latest hardware survey in August, Linux systems make up 1.82% of all hardware. I’d prefer to use this over the StatCounter statistic, since Steam’s survey data more accurately represents Nvidia’s target demographic for their graphics cards.
Let’s say all Linux users start boycotting Nvidia. I will assume that 60% of them are already on AMD cards since as you’ve described, they have better Linux support.
So the remaining 40% of those 1.82% (approx. 0.73%) can now start boycotting Nvidia and lose them additional sales.
So of the 98.9% of the gamer demographic (100% - AMD Linux users) that Nvidia could market to, they would lose 0.73 / 98.9 = approx. 0.7% of their sales.
How can a boycott that lowers their sales by up to 0.7% at best improve hardware support? I get that each individual person will be improving their own situation by switching to a card with better support, but I don’t understand how it will incentivize Nvidia to improve their Linux support.
I’ve been using Linux since 2004. Back then, it didn’t even have nearly the marketshare it does today, and Android didn’t exist, but boycotts and protests have worked anyway. Many times. Even nvidia themselves changed their tune with their motherboard chipset drivers.
By your logic, all these hardware manufacturers should just give up and refuse to support Linux at all. It sounds like that’s what you are advocating for.
You said that Linux users should boycott Nvidia. I’m asking you how that will incentivize Nvidia to improve their Linux support. Can you answer that question or can you not?
I’ve said numerous times that this has worked multiple times in the last 20 years I’ve dabbled with Linux. You refuse to listen and throw numbers and “I don’t care what Linus thinks” at me. I’m done here.
System76 is a Linux-first hardware OEM, but not open source first. Nvidia’s GPUs using proprietary drivers function almost as well as AMD’s open source drivers and have the added functionality with NVENC and Cuda. It really depends on your use case.
The problem is that those drivers are awful if you plan to keep your computer for more than a year or two. Most Linux-first OEMs are shipping Nvidia, not just System76. I’ve had two computers I got secondhand with Nvidia GPUs, and that damn GPU was the bane of my existence, and from what I’m seeing, that situation hasn’t changed for the better at all.
Ideally, I would love to see things change, but it definitely seems like the majority of Linux users and OEMs are still using Nvidia GPUs, so Nvidia has no incentive to change.
If you can avoid buying Nvidia I’m in favor of it. AMD’s all around a more supportive company when it comes to Linux and Open Source. But some people are stuck relying on Nvidia for their hardware.
Add comment