amdgpu/sw was just a (not so good) example. The only user that I can think of is fbdev, although with st/egl gone we're rather far from it.
-Emil On 28 April 2015 at 09:37, Marek Olšák <mar...@gmail.com> wrote: > I was asking about winsys/amdgpu/sw. I had no problem with moving > radeon_winsys.h. > > Marek > > On Tue, Apr 28, 2015 at 9:08 AM, Emil Velikov <emil.l.veli...@gmail.com> > wrote: >> On 28 April 2015 at 07:51, Chia-I Wu <olva...@gmail.com> wrote: >>> On Tue, Apr 28, 2015 at 2:19 PM, Emil Velikov <emil.l.veli...@gmail.com> >>> wrote: >>>> On 28 April 2015 at 03:30, Michel Dänzer <mic...@daenzer.net> wrote: >>>>> On 28.04.2015 03:57, Marek Olšák wrote: >>>>>> Can you elaborate on what amdgpu/sw would be good for? Frankly I don't >>>>>> see a point, therefore we don't need the "drm" subdirectory. >>>>> >>>>> I think Emil's main point (which I agree with) is that radeon_winsys.h >>>>> should live somewhere under src/gallium/drivers/, probably >>>>> src/gallium/drivers/radeon/, because the winsys interface is defined by >>>>> the drivers, not the winsys. All other drivers have their *_winsys.h >>>>> headers there. >>>>> >>>>> Maybe amdgpu/sw wasn't a good example, but at some point there might be >>>>> e.g. amdgpu/haiku. ;) >>>>> >>>> Well said Michel - this is exactly what I meant. >>> ilo moved its intel_winsys.h to src/gallium/drivers/ilo because Emil >>> wouldn't let me put it in src/gallium/winsys/intel, which is exactly >>> what this patch tries to do for radeon :) >>> >> Here we (will) have two different winsys implementations (radeon, >> amdgpu), providing the same interface for the same radeon driver(s). I >> hope that my suggestion for consistency did not come as pushy. >> >> -Emil >> _______________________________________________ >> mesa-dev mailing list >> mesa-dev@lists.freedesktop.org >> http://lists.freedesktop.org/mailman/listinfo/mesa-dev _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
