AMD Radeon Linux Gaming Performance At Parity Between KDE Plasma 6.0 X11 vs. Wayland

[anda_skoa]

The introduction of Posix ACL allows non system administrators to create there own allowed users and group lists on any file they control.

Very good point, I hadn't thought about using ACLs to have even better and finer grained control of access to system resources such as sockets.

[mSparks]

Lot of people think on a per user base.

because for accounting, HR and project management applications the primary security concern is keep the application and its data safe from the user. lest they f' with the audit log... or worse.

This is mostly a consequence of not having to carry all the cruft a full X11 server implementation would need to have.
There are X11 server implementations that ditch all (or most) of the cruft and thus become quite lightweight as well.

"the cruft" being all the security architecture and distributed compute functionality.....
At least such lightweight implementations of X11 still have applications to use on them - the elephant in the room here is whether the wayland protocol was actually superior than X11. Practical examples are universally showing it is not.

That difference does not even exist.

Neither protocol carries information about the owning user of a client process.
Their server side implementations handle each client connection equally, regardless of whether one client has created multiple connections, or multiple clients have created

Protocol is one thing, implementation something else.
Wayland threw out both under what is now clearly the flawed assumption the protocol was over complicated and the implementation poorly implemented.

Scrambling to get xwayland up and working is a valiant attempt to recover from that monumental fuck up, what xwayland can not do is fix waylands poor implmentation of a subpar protocol, and without that xwayland has no reason to exist.

[oiaohm]

because for accounting, HR and project management applications the primary security concern is keep the application and its data safe from the user. lest they f' with the audit log... or worse.

Altering the XACE audit log is one of faults of XACE implementation in X11 protocol.

How are you really keeping the data safe when every user who connects to a X11 server can alter every other users data on that X11 server and every application that those users run can alter every other application resources on the X11 server..

Keeping the application data safe equals you need per application granularity so one application cannot cause another application data to be corrupted. The lack application granularity is how you are able to mess with the XACE audit logs.

Taking from from the user point of view is where you are now stuffed. You have missed applications are threats to other applications.

Accounting/HR/project management bring the security design of what you use with filing cabinets and files into the computer space. The problem is the computer space with applications is more like dealing with animals than files.

Remember X11 protocol does not even implement correct security between users either to keep one users applications safe from different users applications. Like every user on the system being able to connect to the X11 socket so able to perform attack on the X11 server to access what ever application is running on that X11 server and extract data is only starting point of o my god bad. x.org X11 server is very much depending on not having a hostile user.

Please note I am not saying this is not fixable but there is lots and lots of work required to fix all this stuff up and along the way there are going to be application breakages to fix.

[anda_skoa]

"the cruft" being all the security architecture and distributed compute functionality.....

True, one could consider the bolted on security as part of the cruft, especially in such settings which would have custom X11 server implementations.
However I was more thinking along lines of features and sub system that have not been used for decades.
Like the font server, server side rendering, TCP/IP capability, server side printing, etc.

At least such lightweight implementations of X11 still have applications to use on them

If any still exist.
As I said most system that had those have moved to single application interfaces or Wayland if they need to support multiple UI processes.

the elephant in the room here is whether the wayland protocol was actually superior than X11

Again not really an important difference given how similar the two system are on this level.
Obviously the Wayland protocol was created with all the knowledge on what went wrong during the X11 design and extension phases.
But that doesn't necessarily make it "superior", just an application of "lessons learned"

Protocol is one thing, implementation something else.

True but they are also very similar in implementation.

There used to be a difference when X11's implementation was still mostly manually written code while Wayland's code got generated.

However one of the main overhauls of X11 in the more recent past was the transfer of the protocol description into a machine understandable format and thus gaining the ability to generate code for it as well.

XCB, for example, would have been a major effort without that. It really only became viable with the given development resources due to code generation.

Again not really a fundamental or conceptual difference, the similarities far outweigh those.

[mSparks]

True but they are also very similar in implementation.

There used to be a difference when X11's implementation was still mostly manually written code while Wayland's code got generated.

However one of the main overhauls of X11 in the more recent past was the transfer of the protocol description into a machine understandable format and thus gaining the ability to generate code for it as well.

XCB, for example, would have been a major effort without that. It really only became viable with the given development resources due to code generation.

Again not really a fundamental or conceptual difference, the similarities far outweigh those.

So your position is wayland and X11 are the same, except X11 has universal adoption and hundreds of billions of dollars in application revenues each year, and wayland doesn't.

Still sounds like you are saying wayland is dead to me.

slight tangent, still interesting/related reading.

https://bbs.archlinux.org/viewtopic.php?id=87281

[oiaohm]

So your position is wayland and X11 are the same,

That the problem what you were point as differences in reality are that different. The differences are so minor that not important.

The granularity of the security designed into the protocol is very different between X11 and Wayland. Wayland protocol does have a finer grained security than X11 protocol.

except X11 has universal adoption and hundreds of billions of dollars in application revenues each year, and wayland doesn't.

https://www.x.org/wiki/XorgFoundation/Reports/treasurer-2017/

All that so called money and the income of the x.org foundation is less than 10000USD a year in income. Yes those reports are meant to be put out every year guess what 2017 is the most recent.

The reality is the x.org project yearly income is less than what KDE and Gnome foundations individually pay in tax every year. KDE and Gnome both can spend individually 1 million dollars a year. Reality x.org basically has zero budget to spend on anything once you cover paying operation costs.

x.org X11 xserver implementation some of the reason it so bad is lack money to pay people to fix problems that no one else wants to touch. KDE and Gnome can make board agreement say X problem need fixing we will employ a developer to fix that because no one wants to fix if of their own free will.

Cosmo from popos that the rust based Wayland compositor also has funding where they can employ developers on need to fix their problem cases.

The lack of money at x.org is in fact one of the problems why x.org X11 server is in such a bad state.

[mSparks]

The lack of money at x.org is in fact one of the problems why x.org X11 server is in such a bad state.

well, all this revenue is off the back of linux/X11

https://www.reuters.com/technology/nvidia-forecasts-first-quarter-revenue-above-estimates-2024-02-21/

It's not their fault x org devs were unwilling to follow their lead. It is wayland devs fault for not providing what people were willing to pay for.

My 2c.
Maybe if x.org stop wasting effort on a dead display servers like wayland/xwayland and actually invest in improving things people actually want and use they wouldnt be so short of money.

Are you really telling me you are surprised that xwayland nearly reaching parity with plain xorg-server is not something people are willing to get their cheque books out for?

[oiaohm]

My 2c.
Maybe if x.org stop wasting effort on a dead display servers like wayland/xwayland and actually invest in improving things people actually want and use they wouldnt be so short of money.

My 2c when it comes to this money. "well, all this revenue is off the back of linux/X11" Very little of that revenue has gone to X11 close to none. The reality is most of x.org/xfree86 development was not paid for by HPC/Desktop... The reality it majority of xFree86 and x.org was paid for by embedded users.

The problem has a name. Enshittification https://en.wikipedia.org/wiki/Enshittification and it way broader than most presume This is effect of major companies being in the mix is going to causes as they attempt to increase their own income..

At first lot of focus of putting features into X11 to get the market share. Then over time they keep on working out is what is the lowest amount of money we can invest in x11 development and users will still buy our Distribution and pocket most of the money ourselves.

This pocketing money themselves applies to KDE and Gnome as well they depended on x.org X11 to exist but then were collecting all the money and giving nothing to x.org..

mSparks here a shocker and a half. Has Nvidia even once done a money donation x.org or xfree86 before that. The answer is no. Has Nvidia ever had a member of the board at x.org this is again no. Nvidia pocketed the money alright but has not invested in x.org while they were pocketing the money and they are not alone with this behavior..

Are you really telling me you are surprised that xwayland nearly reaching parity with plain xorg-server is not something people are willing to get their cheque books out for?

Its funny. Nvidia has currently going o crud our solution using X11 bare metal is going to stop working we are going to fund development to make Wayland work even start supporting open source drivers. Yet Nvidia will not open their cheque books and give x.org some cash or put a person on x.org board or put forward .x.org bare metal maintainer(yes this seat is still technically open to be taken). Alan from Oracle is not the main maintainer he is the bug fix maintainer this is why we are not seeing x.org bare metal new releases there is no active main maintainer to make them.

For a item that you make out as such a important thing mSparks when you follow the money you find Enshitification that causes x.org not to get the money required to have developed x.org security and the like correctly. Not getting this money means having to beg people employed by different companies to-do things.

Allowing parties to make their own Wayland compositors. Reality is Gnome and KDE are better funded to make a Wayland compositor or X11 server than what x.org project is.

So its not just about implementing features people want. The reality is about implement features for people who will pay for them this is how we get Wayland. What wayland is happened to be the features people will pay to have developed and the features have appeared in order that people are willing to pay for the features. Once you wake up to this this gives you a really clear idea where to look in x.org X11 xserver to find highly under funded and poorly maintained and written code.

Embedded users were first to pay.

Think about it when a company pays redhat and so on mSparks how many make sure that money is in fact going into the projects they need.

mSparks here the horrible part is x.org yearly revenue went up once they were supporting Wayland not down. 2000-2008 how does staying alive on a 5 thousand dollar and less budget sound to you that what x.org project was living on before Wayland.

The money side is like the X11 security side. A broken mess that is basically unworkable to keep x.org project alive long term and that was with and without wayland. Without wayland x.org would have gone bankrupt all ready and maybe that would have been for the best.. Best so people woke up how bad the money location is.

mSparks my problem is I cannot see who is going fund x.org X11 bare metal development. At every chance everyone currently involved in X11 bare metal development is going to-do Enshittification where they are going to min amount work into it so it runs nothing more. To fix X11 security issues there need to be a serous investment of cash. Maybe we get some because of Xwayland maybe not. Some of the reasons we need to admit X11 protocol is in a hell of a mess.

[WileEPyote]

Why are you putting 'words into my mouth?' Don't invent things I didn't say. I am just saying the implementation and progress of Wayland so terribly slow - is anyone here saying it's 'dead?' I doubt it. It's just a very slow progress and I don't think it's anything to get excited about. It seems that there's still issues and serious ones - and lots of ppl who are negative towards Wayland say, 'it's broken by design.' Why do ppl say that? 'Care to explain that one?

I just perceive it similar to the 'AMD hysteria' here and other places - that amd gpus are so great in Linux when they're actually not - problems all over the place and functionality is limited with so many problems yet there's cheerleaders as if they were part of the developer team or somehow related to the ppl or something. I just find it very peculiar and odd.

Yes, somebody is saying it's dead. You quoted my post, that was a response to Msparks specifically saying that wayland is dead. I assumed you knew the context of the post. My response was based on that. If that's not what you were trying to say, then I apologize.

I'm actually of the same opinion as you I think. I don't understand the people that latch onto something so completely that they discount oother possibilities.

I'm tired of the ridiculous fanboy ranting on these threads. Both protocols have flaws and both have strengths. I actually still have both on my system. I find Wayland better as a daily usage display server because of it's superior scaling and font fine tuning abilities, and x11 is still better for gaming and similar workloads. I'll continue to use both until one does everything I want.

[anda_skoa]

So your position is wayland and X11 are the same

Not at all.
I am saying that they are very similar in a lot of aspects, both in terms of concept and implementation.
They are different in some areas, sometimes in concepts, sometimes in implementation and sometimes in both.

Therefore I find it puzzling when people try to find differences in areas of similarity instead of areas of actual difference.

It is somewhat understandable as technology is complicated and can be confusing for people without the relevant technical background.

However it is also strange that people who lack the fundamental knowledge and understanding would engage in such difference finding missions which are doomed to fail.