FOSDEM 2013: The future of X.org on non-Linux systems

[fonz]

It's obvious that Xorg is a mess and needs replacement

I don't remember exactly when, but at some point FreeBSD switched from XFree86 to X.org. I'd like to think that that decision wasn't taken lightly.

Hypothetically, what would be necessary to get such a project started.

Sheer manpower goes a long way. But do keep in mind that a complete X11 suite is very large and complex, consisting of many components. We're not talking about a CS BSc or even MSc project here.

 

[kpa]

I don't remember exactly when, but at some point FreeBSD switched from XFree86 to X.org. I'd like to think that that decision wasn't taken lightly.

Sheer manpower goes a long way. But do keep in mind that a complete X11 suite is very large and complex, consisting of many components. We're not talking about a CS BSc or even MSc project here.

I would say that X11 as a whole is a mess and needs a replacement that offers the good stuff from X (network transparency etc.) but also offers a solid and platform for graphics, be it the standard desktop stuff or 3D graphics for games.

 

[fonz]

I would say that X11 as a whole is a mess and needs a replacement

Now that's a statement that has been true for many years.

 

[throAU]

But do keep in mind that a complete X11 suite is very large and complex, consisting of many components. We're not talking about a CS BSc or even MSc project here.

Don't need a complete X11 suite.

There are plenty of X11 servers which can sit on top of whatever graphics subsystem you have (e.g., XQuartz - open source, could hack it up to have it render direct to SDL or whatever low level rendering engine on FreeBSD instead of Quartz on OS X).


edit:

SDL supports Linux, Windows, Windows CE, BeOS, MacOS, Mac OS X, FreeBSD, NetBSD, OpenBSD, BSD/OS, Solaris, IRIX, and QNX. The code contains support for AmigaOS, Dreamcast, Atari, AIX, OSF/Tru64, RISC OS, SymbianOS, and OS/2, but these are not officially supported.

Porting an X11 display server to SDL probably makes a lot of sense anyhow, given the above.

In fact, I'd say its a no brainer.

Treat SDL as a first-class library on FreeBSD, enable GPU drivers to plug into the base system to support SDL properly and we're well on the way...

So in terms of layering from bottom to top (something like)
kernel -> video driver -> openGL <- not sure? -> libSDL -> X11


edit:
Side effect of using SDL: you could run X11 on top of aalib. Terminal friendly!

 

[fonz]

Don't need a complete X11 suite.

There are plenty of X11 servers which can sit on top of whatever graphics subsystem you have

I was referring to a complete rewrite from scratch but sure, if you're able to use existing components that might make the job considerably easier.

 

[Crivens]

Hello incompatible server Nr n+1.
Seems we are not alone with the worries, but let's see what is going to happen to them.

 

[arp242]

Hello incompatible server Nr n+1.
Seems we are not alone with the worries, but let's see what is going to happen to them.

Why is it incompatible?

Quoted from the linked page:

we want to provide a graphics stack that works across different platforms and driver models by limiting our assumptions to a bare minimum

with our protocol- and platform-agnostic approach, we can make sure that we reach our goal of a consistent and beautiful user experience across platforms and device form factors

The system should easily be adaptable and portable to different platforms and use-cases

keeps cross-platform use-cases in mind.

In fact, the word `Linux' in mentioned only twice in the article, both in the introduction.

This, ofcourse, doesn't mean it's compatible, but they do seem to *care* about it, which is an excellent first step and the most you can reasonably expect.

 

[Crivens]

There is some heat now between Mir and Wayland, which I can understand from what I have read so far. What I think is going to happen is that now there is another display server, and it will be different in some places.

For application development, this will mean that it is going to be harder to have some "write-once-run-anywhere" apps. Given the current direction of Canonical, which seems to be the tablet and phone. Canonical seems to have the same worries w.r.t. the Xorg direction, and should they succeed - great. If it is portable, or more portable than other solutions, even better. Here, the new focus of tablet/phone may be good, we may be able to participate in this project much easier than getting KMS/GEM/... for all different GPUs FreeBSD can use.
And maybe this was the reason for them not to participate too much in Wayland, which IIRC is also using some (for now) Linux-only APIs.

As I said, let's see what is going to happen. I have seen it just now, still reading up on it, but I'll hope for the best and prepare for the worst.