Hi Lucas, Pardon for jumping it like this.
On 19 October 2015 at 10:25, Lucas Stach <l.st...@pengutronix.de> wrote: [snip] > This series implements a way to fuse multiple DRM devices into a single > EGL device, solely on the base that current EGL users expect that a > single device is able to render and scanout at the same time. This is a > flawed premise on itself and this series actively tries to keep the > illusion that this is true, even for SoC and prime systems. While doing > so it is putting a lot device specific knowledge into the implementation > of the midlayer/helpers. > I'm not sure if one must enforce a higher level (EGL) decision onto a lower level component such as the DRI module. A list of reasons, in no particular order: - DRI does not want/need to know anything about the winsys specifics. It's up-to the relevant loader to sort that out. - The proprietary drivers from nvidia/amd did not support EGL streams (last time I've looked). - The wayland/weston people (amongst others) do not seem too keen on EGL streams - Neither of the suggested extensions exists in mesa - The EGL device patches (and equivalent libdrm helpers) have been around for a while with little to no review :( - GLX ? I agree that when (if?) EGL streams come into play, one reshuffle things as you've explained. Although until then, having the original idea of wrapping the scanout dri module around the render one, looks like a very good compromise. Esp since in most of the cases the two devices are tightly coupled. -Emil P.S. The half-way argument always holds true, but considering the limited activity on the etnaviv front, half-way is better than nothing. _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
