On 2022-05-02 12:41:37, Dmitry Baryshkov wrote: > On 02/05/2022 11:43, Marijn Suijten wrote: > > On 2022-05-02 01:44:20, Dmitry Baryshkov wrote: > >> [sni[ > >>> In any case, given that you've already sent this patch and another three > >>> patches [2] fixing/cleaning up the series tells me it's far from ready. > >>> Most of this should just be handled - or have been handled - in review > >>> and amended? > >> > >> During the review time we agreed that [2] would come as a separate > >> change It is an API change that would make using panel-bridge easier, > >> but isn't otherwise required. > >> > >> I have been working towards more logical drm_bridge/drm_bridge_connector > >> chains employing panel-bridge and display-connector where required, [2] > >> is a part of that effort (as well as few other patches that hit > >> dri-devel in the last few days). > > > > I understand what is going on now. Since the DSC patches have already > > been queued up in the 5.19 pull I won't hurry to review them; rather > > will go over them when time allows me to play with the many phones here > > that require DSC for the screen to work. I've been told the series > > didn't result in positive screen output way back in its infancy, but > > I'll re-evaluate and send fixes or improvements if/when necessary. > > Sure, thank you! > > They work on Pixel3 (sdm845, non-active CTLs, no ping-pong binding to > intf). I still didn't have time to test them on P4 (sm8150, active CTLs, > PPs bound to the intf in runtime). The devices mentioned above were all recent SoCs with active CTLs. My ping-pong binding to intf patch recently "fixed" sm6125 (non-DSC) but I have been told it didn't make a difference on the more powerful SoCs (sm8[123]50) with DSC panels. There might indeed be a problem with either active CTLs and CMDmode in general (we still have patches in the works that move PP features to the INTF block) or DSC + actice CTL, or both. To be continued... - Marijn
