On Wed, Jan 29, 2014 at 3:56 PM, Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> wrote: > I think once we get to working on D0i2, we'll need to move the PSR > wakeup to happen from a workqueue since it essentially requires a > full modeset. Even now in your code it's somewhat questionable > since you're doing stuff like aux transfers while holding > struct_mutex. Imo we can fix this up later on by pushing it into a workqueue. It's not great that we block while hodling struct_mutex, but imo there's too many loose pieces around psr so I prefer we start by merging something pessimistic, but known to work properly. Then we can tune and improve with the knowledge that the crc testcase will catch regressions. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
