Whoops, There's A Big Problem For Wayland GTK+

[Nille]

Some people in this thread really believe that EGL and KMS never can be implemented in proprietary drivers.

Maybe but this is not opinion. My post is only related to you tegra comment The tegra driver is different from the "nvidia" driver so you can't compare both drivers.

[RussianNeuroMancer]

The tegra driver is different from the "nvidia" driver so you can't compare both drivers.

I doesn't compare it, if you didn't notice.

[Serafean]

Some people in this thread really believe that EGL and KMS never can be implemented in proprietary drivers. AMD embedded drivers and nVidia Tegra drivers is good living example that proof one simple fact - this is possible. So when Wayland will be ready - nVidia and AMD driver developers have no choice but implement needed functional in proprietary drivers for desktop and laptops GPU's.

I for one am not saying it can't be done, I'm saying I believe said companies won't be bothered to do so. As I said before, X isn't going anywhere for a while, binary blob primary targets being workstations, what is their incentive? Someone mentioned using wayland on a worksation is possible, I agree, but AFAIK most workstations use applications written for X, some of them proprietary. I'm not saying it won't happen, I just don't believe it will; but would however love to be proven wrong.
As for the not having any choice, I'm afraid they do have a choice : they can simply flip us off.

[asdx]

WOW, I've just tried Weston/Wayland and damn... it feels incredibly smooth.

X in comparison feels clunky, slow and with very high latency.

I can't wait for the day we can use full DEs like Gnome or KDE on Wayland.

Amazing work, hats off to the developers.

[RussianNeuroMancer]

As for the not having any choice, I'm afraid they do have a choice : they can simply flip us off.

They doesn't have reasons to flip us off because all needed stuff is already written, they just need to port it from embedded drivers to desktop drivers.

[mrugiero]

They doesn't have reasons to flip us off because all needed stuff is already written, they just need to port it from embedded drivers to desktop drivers.

They do have reasons, the main one would be they must pay for that porting, and it's not focused to their main target.

[RussianNeuroMancer]

They do have reasons, the main one would be they must pay for that porting, and it's not focused to their main target.

Laptops with hybrid graphics is not main target for AMD, but they support it in fglrx, HTPC is not main target for nVidia, but they support HDMI audio bitstreaming. They already pay for support this and other features, even less important. Since Wayland is more important they will support it anyway.

[Nille]

they just need to port it from embedded drivers to desktop drivers.

In case of AMD the have an egl Implementation in the fglrx driver.

[RussianNeuroMancer]

In case of AMD the have an egl Implementation in the fglrx driver.

Really? Bridgman say some time ago there is packaging problems with EGL in desktop driver. You mean desktop driver or embedded driver? Embedded driver have EGL, yes.

[Ansla]

5 pages of comments and nobody read/understood the article, the usual for Phoronix. Anyway, if anybody is interested in the real problem here is a summary:

1. When building GTK+ with Wayland support memory usage will increase when running on regular X11, so it has nothing to do with nVidia supporting Wayland.

2. This is because nVidia builds their drivers with position dependant code, which is slightly faster but can not be shared between different programs (so it gets duplicated for each program running that uses GTK+)

3. It only happens for 32 bit applications, because the AMD64 architecture doesn't support shared libraries built with position dependant code so nVidia was forced to fix their 64 bit driver.