On Mon, Oct 29, 2018 at 12:03:46PM +0000, Emil Velikov wrote: > On Fri, 26 Oct 2018 at 14:40, Daniel Vetter <daniel@xxxxxxxx> wrote: > > > > On Fri, Oct 26, 2018 at 01:06:47PM +0100, Emil Velikov wrote: > > > From: Emil Velikov <emil.velikov@xxxxxxxxxxxxx> > > > > > > VGEM doesn't do anything modeset specific, so in a way exposing a > > > primary node is 'wrong'. At the same time, we extensively use if for > > > creating dumb buffers, fences, prime fd <> handle imports/exports. > > > > > > To the point that we explicitly annotate the vgem fence ioctls as > > > DRM_RENDER_ALLOW and have an IGT test which opens the render node. > > > > > > close(drm_open_driver_render(DRIVER_VGEM)) > > > > Huh, I guess that test doesn't pass? > > > It does since the IGT code seems bonkers... Namely it silently falls > back to the primary/card node. > It seems to be for historical reasons... back in the days where render > nodes were still experimental and had to be explicitly enabled via > kernel arg. > > I'll try and follow-up to address that. I guess enforcing this with igt is a bit harder, since not much with render nodes is generic across drivers. Hence a lot of the tests are very specific. But anything we can increase the odds for people using render nodes is good I guess, and if it's just using vgem as a driver copypaste template. -Daniel > > > > Better late than never, let's flip the switch. > > > > > > Cc: David Airlie <airlied@xxxxxxxx> > > > Cc: Daniel Vetter <daniel.vetter@xxxxxxxx> > > > Signed-off-by: Emil Velikov <emil.velikov@xxxxxxxxxxxxx> > > > > Acked-by: Daniel Vetter <daniel.vetter@xxxxxxxx> > > Thanks. I'll give it a few more days for others to review/comment. > > -Emil -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
