It's Becoming Very Easy To Run Wayland

[werfu]

If Wayland takes off, wonder if nVidia and AMD/ATI will consider providing support for Wayland in their proprietary display drivers?

If I understood well, all Wayland needs is KMS and GEM (and I guess DRI2).
So, once these are implemented in NVIDIA or AMD driver, Wayland should run without problem.

[Panix]

I don't if this is a valid question and/or if it's a dumb question but if so, I will bow out gracefully and appreciate any corrections.

I was wondering if it needs KMS, what does that mean for situations in which kernel modesetting is disabled. That wouldn't be an (optimal) option anymore or?

[werfu]

I don't if this is a valid question and/or if it's a dumb question but if so, I will bow out gracefully and appreciate any corrections.

I was wondering if it needs KMS, what does that mean for situations in which kernel modesetting is disabled. That wouldn't be an (optimal) option anymore or?

KMS is somewhat the holy grail of graphic mode setting. It promise flicker free graphic context changes. In that sense, every body will want to use it. It will become the norm.

[bridgman]

Panix; good question.

Today the older UMS code paths provide a good "fail safe" alternative when the newer KMS code has problems. Over time the expectation is that the KMS issues will get resolved (I think it's mostly a matter of remembering/reinventing old hardware-specific hacks and putting something similar for each into KMS), hopefully before the UMS code bitrots and stops being a useful fallback. This all applies mainly to older (ie already supported) hardware.

For new hardware support, nearly all of it is being implemented KMS-only anyways.

[89c51]

why i remember that i read somewhere that KMS is not needed and the protocol can be implemented with different ways.


i might be wrong though