Wayland on FreeBSD

[Tomasz CEDRO]

On Mon, Apr 20, 2020 at 11:14 AM Niclas Zeising wrote:
> On 2020-04-20 03:38, Tomasz CEDRO wrote:
> > On Fri, Apr 17, 2020 at 10:34 PM Tomasz CEDRO wrote:
> >> I would like to use Enlightenment (0.23.1) + Wayland.
> >
> > I did manage to run Wayland + Hikari so Wayland works on FreeBSD :-)
> > Ports version is 1.18 and lags behind only one release (1.20 is 1.19
> > with some readme added).
>
> Are you referring to the wayland versions here?  Looking at the source
> for both wayland and wayland-protocols, FreeBSD ports are at the latest
> versions, wayland 1.18, and wayland-protocols 1.20.

Exactly! I mean we are really up to date but there seems no use of it
yet..? People (including me) are confused with Xorg drivers while
there may be totally better solution right next to it :-)

I am new to Wayland, I saw some really minimalistic WM for it, but I
know Enlightenment, it is known to work on Wayland, and this is the
beautiful fully featured WM that we all know from Xorg. This could
provide nice transition for people that want to try out the Wayland
and maybe work on it everyday. This could also gather community and
improve both Wayland and Enlightenment on FreeBSD as a strong base,
then focus on more subtle works.

I can understand the need for something smaller and faster than Xorg.
I remember playing with MESA back then in 1998 with first 3D
acceleration cards. I remember DirectFB to skip Xorg for graphics on
embedded systems. I even remember when Drag-n-Drop in Xorg had several
implementations. Embedded Systems and IoT will enforce using Wayland
sooner or later. In its core concept its just drawing over the frame
buffer memory so what's the big deal?


It seems the only problem to run Enlightenment on Wayland is hardcoded
Linux DMA code in EFL:

https://github.com/Enlightenment/efl/blob/master/src/lib/ecore_wl2/ecore_wl2_buffer.c

Do you know Niclas if current DRM drivers provide such DMA
capabilities? Maybe this is also part of KMS standard? Maybe I should
try with generic bus_dma(9)  with no efficiency loss?

Any hints welcome :-)
Tomek

-- 
CeDeROM, SQ7MHZ, http://www.tomek.cedro.info