[Tizen General] Wayland in future?

Carsten Haitzler (The Rasterman) tizen at rasterman.com
Sat Jan 14 01:08:03 GMT 2012


On Sat, 14 Jan 2012 00:31:20 +0200 Kiril Vladimirov <kiril at vladimiroff.org>
said:

> Hello, guys.
> I've just read this article on Phoronix:
> http://www.phoronix.com/scan.php?page=news_item&px=MTA0Mjc
> And I'm wondering, if this ever goes stable, will be possible to run Tizen
> env via Wayland, instead of X11?
> 
> In other words, what are the obstacles to run wayland on Tizen?

1. drivers. (egl/gles needs modifications for wayland to run and all drivers
for the gpu's found in arm soc's and even medfield etc. are closed and do not
have this support).
2. toolkit ports to wayland are all pretty immature. at this moment i believe
efl has the most "advanced" port, but its definitely not complete, well tested
etc.
3. we will need to replace the current wm/compositor - ie write a whole new
wayland compositor with the same features.
4. we will still need to extend core wayland with more protocol at this stage
than it currently has
5. ports are still working around some bugs in wayland's demo composiotr
6. video support (yuv formats and video layer access)
7. testing, testing, testing, time, trust etc.

wayland needs time and work. the benefit of wayland is actually just a simpler
wm/compositor and no xorg server (and libraries). so really we save a bit of
memory (maybe 2mb or so - that ballpark), and a little disk space. (again - a
few mb too). performance-wise we actually stay still - well compared to the
drivers we have currently that properly optimize swap paths to actually swap
and not copy on the soc's we have here.

we currently already have tear-free composited display in x11 with the "assumed
performance benefits" of wayland because what we have is implemented pretty
decently in an x environment, so the delta in gain going to wayland is much
smaller.

-- 
Carsten Haitzler (The Rasterman) <tizen at rasterman.com>


More information about the General mailing list