Op 25-09-2019 om 06:17 schreef Matt Roper: > On Fri, Sep 20, 2019 at 01:42:19PM +0200, Maarten Lankhorst wrote: >> This can all be done from the intel_update_crtc function. Split out the >> pipe update into a separate function, just like is done for the planes. >> >> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@xxxxxxxxxxxxxxx> > The code here all looks logically correct, but I think our various > pipe functions are still somewhat confusing. E.g., > intel_update_pipe_config() and commit_pipe_config() have names that > make it sound like they're doing the same thing. At the moment > intel_update_pipe_config() is basically the fastset-specific stuff, but > is there any reason most of those operations can't just be done on all > commits? Re-writing the pfit/scaler registers or pipe chicken seems > like a pretty small number of registers to try to avoid. > > If we want to keep them separate, I'd suggest renaming > intel_update_pipe_config() (and probably the 'update_pipe' flag too) to > make it more clear that it's meant for the fastset special case. > I'm sending a new version with intel_update_pipe_config inlined. It makes sense to avoid, especially the chicken bits because that function performs a read as well. _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
