On Wed, Mar 4, 2015 at 5:51 PM, Conselvan De Oliveira, Ander <[email protected]> wrote: >> The tricky bit here is that we must have all the connectors added to the >> drm_atomic_sate for the given crtc. Otherwise there might be no connector >> at all and we'd return a bogus answer. drm_atomic_add_affected_connectors >> is the function which does this for you, and I think we need to call it >> somewhere in the top-level compute_config code. And I haven't spotted that >> call anywhere in your series. > > I did add it in patch 12, but now I realize it won't be called for the > disable case, since the call is in intel_modeset_pipe_config(). I'll > move that to intel_modeset_compute_config().
Indeed. I've used my mail client to search for it, no idea why that didn't show up ... For safety maybe we should have a WARN_ON if there are no connectors for a given crtc? Normally there's no cloning going on hence should catch all such bugs. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx
