On Tue, May 03, 2016 at 12:03:08PM +0200, Boris Brezillon wrote: > On Tue, 3 May 2016 11:53:18 +0200 > Daniel Vetter <daniel@xxxxxxxx> wrote: > > > On Tue, May 3, 2016 at 11:39 AM, Boris Brezillon > > <boris.brezillon@xxxxxxxxxxxxxxxxxx> wrote: > > >> > + .detect = sii902x_connector_detect, > > >> > + .fill_modes = drm_helper_probe_single_connector_modes, > > >> > + .destroy = sii902x_connector_destroy, > > >> > +}; > > >> > > >> I'm guessing this is to support both atomic and !atomic drivers. I > > >> thought it was working automatically? > > > > > > Hm, the dw-hdmi driver is providing both, but maybe it's not needed. > > > Daniel, any comments? > > > > You only need this in case you actually have atomic and !atomic users > > of your bridge. We discussed whether we should have a dpms helper that > > dtrt automatically, but with just 1 driver that doesn't seem worth it. > > You wouldn't need to have 2 entier vfunc tables, you could just switch > > on DRIVER_ATOMIC. > > Ok, I'll drop the !atomic helpers. Should I still test > drm_core_check_feature(drm, DRIVER_ATOMIC) in the ->attach() > implementation and return an error if the DRM device does not support > atomic operations? If you feel like. If someone tries to use this bridge driver with a non-atomic driver then - they'll have a big fireworks show either way - a really good reason to fix up their driver to be atomic, like it should be ;-) Cheers, Daniel -- 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
