xf86-video-armsoc, Mali and exynos4

[Luc Verhaegen]

On Tue, Nov 12, 2013 at 04:08:27PM -0600, Daniel Drake wrote:
> 
> Yes, it does work reasonably well. Unfortunately this does not support
> xrandr resolution switching though, which is now made available "for
> free" by the underlying exynos_drm display driver.
> 
> Maybe we could add KMS support to fbturbo, I was just interested in
> exploring things from the armsoc side, as the original announcement
> talks about the combination of X/Mali integrated with exynos, DRM and
> KMS. http://lists.x.org/archives/xorg-devel/2012-May/031250.html
> Maybe that level of integration has not yet been reached in public
> code though :(

I haven't tried recent kernels on exynos, or on sunxi. It's generally 
hard enough to get anything halfway working in the ARM world, and i am 
having to deal with too many distractions from lima already, even though 
some are quite necessary.

Having had to deal with the armsoc driver on exynos5, my general 
impression of armsoc is not too good though. But perhaps some of that is 
due to the kernel side of things.

As for the binaries you are using, I guess it all depend on which 
binaries you are using and what they were built for. The binaries for 
sunxi are either for fbdev (involving nothing in the likes of ump, drm, 
or dmabuf) or X (which uses ump and drm to get the buffers sorted). I 
would expect that these binaries are not too keen on working with 
drm/kms if dmabuf is used, or if ump is not properly used. It all 
depends on the egl code in the binaries.

Perhaps things can gradually get solved when i finally have full 
glmark2-es2 working, and i push out my lima mesa code.

Luc Verhaegen.