X.Org Server 21.1 Will Aim To Release In The Coming Weeks

[mdedetrich]

I get it. Those people also missed that its the host file descriptors/handle that is the real problem.

https://lwn.net/Articles/517375/ sorry I got the year wrong it was not 2010 it was 2012. This change from using GEM Buffer or some other GPU vendor buffer solution to host file descriptors/handle is to allow working with the Linux/BSD OS host security. Yes this bring mandatory need for DMA BUF support and GBM features also need to use file handles. So Nvidia has been stubborn for over 9 years on this point at least they are finally coming into line.

Thanks for illustrating this, I was also unaware the Wayland protocol changed to use file descriptors. In hindsight there may be an argument made about whether this was the right decision, I mean it was obviously solving issues but as you stated before it shoehorned NVidia into a position that caused issues for everyone involved.

Security also appears to be a topic where Wayland put the cart before the horse and I am getting the honest impression that Wayland is solving what is today a largely non existent problem (and to save the problem well you really do need a permissions/capabilities based system like what Android has and what Google is creating with Fuschia as a core capability in the Kernel.

I think its also important to realize that the unix philosophy of "treat everything as a file" has its own host of problems, it is the unix way of doing things but its not the silver bullet to solve all problems.

[Teggs]

Thank you Povilas Kanapickas and the others who have contributed to X and to those who now take on the task of actually bringing their work to end users.

[arokh]

I've blacklisted you 'cause I'm tired of your aggression, inability to argue and Trump-like statements (stating the desired instead of facts). As to what I've done for Linux, look no further - it's not about the article, scroll it down and find the list of contributions. To cool you down a bit you may need to know that I've been using Linux since the late 90s, probably 10 times more than you have. Cheers!

Facts are obviously not important to you, that mind-boggling website is proof of that. That must be the largest collection of pure FUD that I have ever come across. It's basically a long rant about stuff that doesn't work the way you expect/want them to. The "solutions" you present are even more laughable than the FUD that comes before it. That whole page is so disconnected from reality that it's hard to believe anyone at all bothered to read the entirety of it.

Your "contributions" are hilarious, basically all of them are "bug reports". I've looked through a few of them and they were dismissed by the developers as non-bugs, basically you didn't understand how the software works. You're not a contributor, you're a guy who complains a lot. In fact, it seems you've been complaining since 1999, it seems you are the loudest complainer in the history of open source. For someone complaining that long, it would be odd if some of the complaining did not yield some results. Some of your bug reports do indeed turn out to be real bugs that are resolved by someone else. Well, good on you. I don't know if you should be so proud of your "contributions" though, considering lots of them were not real bugs and actually wasted the time of the developers who had to spend time reading through your misconceptions and trying to explain them to you. Time that could have been spent on actual bugs.

If you actually wanted to contribute to open source, first thing you should do is remove that huge collection of crap from the internet. If you spent as much time on learning as you did complaining, you would have a list of git commits to show for instead of that horror show of a website. Who knows, we might even have had the year of the Linux desktop already, if not for your complaining!

[oiaohm]

So a "trivial" Xorg update triggers this whole mess? Worrying!
I've read exaggerations from those who are outraged by this Xorg update and the usual Wayland jokes.
I want to remind you all, that to date the only DE that uses wayland by default is Gnome.
KDE is starting to get good support, I use it myself on a daily basis, but it is still under strong development.
Xfce does not have a wayland session yet, Mate does not have a wayland session yet, I will not go further in the list as it is long.
It is not a question of will, it is a question of means ... for how wayland was created, the small DEs are struggling to switch to wayland, because it takes a lot of work.
Kde has contributed to the creation of APIs that didn't exist (they existed to work only in Gnome) and that will help even minor DEs to move to wayland, but it takes time.
So what's the deal if Xorg gets an update? What should the Xfce user who has no wayland support do? A Xorg, there is no alternative.
If they wanted a quick switch to Wayland, they would have to create something other than a protocol and that's it.

Its more complicated than this. Reality here we have had more DE on Linux than we have had personal to support properly. The wayland change has brought this lack of resources to head.

https://www.forbes.com/sites/jasonevangelho/2019/10/23/bold-prediction-kde-will-steal-the-lightweight-linux-desktop-crown-in-2020/

Yes the lack of resources for items like XFCE are leading to some horrible things. Like XFCE in fact needing more memory than KDE. Why KDE has been able to fix up their code base and move to more modern interfaces that result in using less memory running the desktop even under X11.

There are quite a few new wayland compositors that have no legacy baggage that have appeared. The smaller ones also have the question will they have enough developers.

Also this is not exactly a trivial update. As Xwayland will go forwards as it own tree under Redhat/Suse development and Bare metal X.org Server will go forwards as it own tree. This now means Xwayland and bare metal x.org can have different X11 protocol support. Worst case incompatible.

[Mez']

This is the exact wrong way to look at this.

Nope. This is the exact correct way too look at it.
If you don't like spinach, mushrooms or chicory, force it down your throat is the exact way to worsen your dislike of it.
If you need to force something, it means its advantages/strengths are not convincing enough for people to embrace it. Hence it will irritate everyone to be forced to it, and it can be considered as a failure. The electric car is a good example of that, as it is being forced in some European countries, without success.

[tildearrow]

I've blacklisted you 'cause I'm tired of your aggression, inability to argue and Trump-like statements (stating the desired instead of facts). As to what I've done for Linux, look no further - it's not about the article, scroll it down and find the list of contributions. To cool you down a bit you may need to know that I've been using Linux since the late 90s, probably 10 times more than you have. Cheers!

birdie on every AMD or X thread:

1. Turn on the lighter, post something with "fanboy" on it
2. Wait for response
3. Start flamewar
4. "I've blacklisted you"
5. Post that link all over again
6. ???
7. Profit?

Why can't you stop burning the forum? This ain't Ventonix.
I already debunked your facts. (scroll to the end for that part)

[oiaohm]

Thanks for illustrating this, I was also unaware the Wayland protocol changed to use file descriptors. In hindsight there may be an argument made about whether this was the right decision, I mean it was obviously solving issues but as you stated before it shoehorned NVidia into a position that caused issues for everyone involved.

The reality here is before changing to file handles they tried 20 different other options that always lead back to failure. Nvidia did put up solution because they were not interested in taking part in development at that time. This is always bad move to let someone plan you future.

Security also appears to be a topic where Wayland put the cart before the horse and I am getting the honest impression that Wayland is solving what is today a largely non existent problem (and to save the problem well you really do need a permissions/capabilities based system like what Android has and what Google is creating with Fuschia as a core capability in the Kernel.

This is not a solid arguement. Fuschia is a everything is a object.

I think its also important to realize that the unix philosophy of "treat everything as a file" has its own host of problems, it is the unix way of doing things but its not the silver bullet to solve all problems.

This is common argue that turns out to be total garbage. https://en.wikipedia.org/wiki/Plan_9_from_Bell_Labs plan 9 proved that a OS with everything as a file in fact works. Turns out there are not a stack of major problems.

How to move to everything is a file without breaking all the legacy applications that use features that are not everything is a file that is tricky bit. Unix is not everything is a file. PID the process ID number this was done in early unix because like you people said everything as a file would put too much overhead on process handling. This lead to stacks of race conditions that was finally fixed by the introduction of pidfd in 2019 in Linux.

mdedetrich there is a catch here. When design OS you really do need to pick 'everything is a [X]' answer. If you don't you security framework end up coming more and more complex and buggy. Unix/posix early design documents did state the "everything as a file" so their X was file but then when you checking the selected X file was not case for everything.

If you have a OS that is 90% everything is a file and 10% is not results in: 10% that has bugs that the other 90% does not have. This is about being able to share developer time making your security and anti-race condition solutions so getting enough peer review that you get these things right.

Remember Windows is everything is a object. Yes this means windows security system is designed around process the object/handles. Race condition handling is around objects and handles....

Linux and BSD and Unix have had a split personality problem where people state that unix is everything is a file but in reality it was everything is a file for 90% of cases with the 10% of cases not file but using domain particular things that turned out not enough peer review so be buggy and not to integrate with the host OS security system properly.

So it just one of those things. If Linux is going to be a everything is a file os might as well be properly that and reduce the problems. Of course fixing up this problem has put Nvidia straight in the cross hairs.

[Myownfriend]

Nope. This is the exact correct way too look at it.
If you don't like spinach, mushrooms or chicory, force it down your throat is the exact way to worsen your dislike of it.
If you need to force something, it means its advantages/strengths are not convincing enough for people to embrace it. Hence it will irritate everyone to be forced to it, and it can be considered as a failure. The electric car is a good example of that, as it is being forced in some European countries, without success.

The reason it was the wrong way to look at it is because, like I said, ever-updating software is never finished. That sentiment doesn't even remotely translate to your food analogy. In regards to forcing Wayland on people, that's not happening, not even remotely. OS's making Wayland the default session is no more or less forcible than defaulting to X11.

Even that electric car example you're using is a blatant misunderstanding of what the word "forced" means.

[tildearrow]

Show me a single person who's been hacked by Spectre/Meltdown? As if you somehow would know you had been hacked.... X.org being a swiss cheese of security holes is a known fact. Your disbelief in reality has nothing to do with Wayland, you need to see a shrink.

As usual, any X.org/Wayland thread on here is filled with rants from incompetent, paranoid and delusional people. Instead of actually contributing to open source, you lot seem to be content repeating the same misconceptions over and over again.

Problem is, Wayland devs think that the only way to fix X's security nightmare is by removing everything. Wayland has no data query protocols (e.g. retrieving key or mouse events/position) and instead leaves this up to the compositor, which leads to fragmentation and increased developer work (read the post about Synergy on the second page for an example of a real use case of this feature!).
The right way to fix X's lack of security is by adding a permission system. macOS did this, and it works very well. But Wayland devs will never realize that.........

	Allows keyloggers	Allows benign data query (e.g. Synergy)
X11	Yes	Yes
Wayland	No	No
macOS	No	Yes

[smitty3268]

Why can't you stop burning the forum? This ain't Ventonix.

birdie is what he is, that's never going to change.

I blame you (or Michael) or whoever has the capability to permaban these people from Phoronix but chooses not to.

Phoronix has cultivated this kind of community for itself.