On Wed, Mar 19, 2014 at 12:28:43PM +0100, Daniel Vetter wrote: > On Tue, Mar 18, 2014 at 05:22:48PM -0700, Matt Roper wrote: ... > > + > > + /* > > + * set_config() adjusts crtc->primary->fb; however the DRM setplane > > + * code that called us expects to handle the framebuffer update and > > + * reference counting; save and restore the current fb before > > + * calling it. > > + */ > > + tmpfb = plane->fb; > > + ret = crtc->funcs->set_config(&set); > > I wonder whether we should have an oppportunistic path using the page_flip > interface here. Otoh we must have a fallback to ->set_config anyway since > the drivers currently only allow pageflips on identical buffers. E.g. i915 > rejects tiling changes and stride changes and position changes. So I think > having a pageflip fastpath isn't worth the trouble. > > Also I think you need to use set_config_internal here to get the > refcounting right.
Hmm. I specifically didn't use set_config_internal here because I thought drm_mode_setplane() (and presumably the future atomic ioctl which may also call into this) will already handle the refcounting for us. Am I overlooking something? Matt -- Matt Roper Graphics Software Engineer ISG Platform Enabling & Development Intel Corporation (916) 356-2795