On 19 December 2016 at 12:53, StDenis, Tom <Tom.StDenis at amd.com> wrote: > > Hi Emil, > Hey there, Tom. Seems my "ignoring the heat around DC/DAL ..." wasn't enough and things have gone there :-( Cutting down to the important piece... > > It's a bit of a catch. The public (re: Linux) wants open, stable, efficient, and featureful drivers but then they want to dictate to the vendor exactly how that's going to be executed. That doesn't really work in practice or theory. The kernel team has standards that AMD has to meet but the process to get there won't always make everyone happy. > ... here. Intel has managed to get there and it would be amazing if the whole AMD team paves in that direction. It's not quick or easy route, yet you can use any of the materials (presentations, etc.) from Daniel Vetter's blog. I'm stupid^Wnice enough to repeat one unwritten rule(?) in a hope that people who do not know it will take note. And yes, rules are _not_ mean to be all fun all/most of the time :-\ To reiterate in a different light: upstream kernel should not (and does not most of the time) cater for non-upstream kernel code. Thanks Emil P.S. Please try to avoid html emails and/or top-posting.
