Hi Geert, On Thursday 10 March 2016 09:17:40 Geert Uytterhoeven wrote: > On Thu, Mar 10, 2016 at 3:32 AM, Laurent Pinchart wrote: > > On Wednesday 09 March 2016 20:18:42 Geert Uytterhoeven wrote: > >> Signed-off-by: Geert Uytterhoeven <geert+renesas@xxxxxxxxx> > >> > >> --- /dev/null > >> +++ b/include/dt-bindings/power/r8a7795-sysc.h > >> @@ -0,0 +1,42 @@ > >> > >> +/* Always-on power area */ > >> +#define R8A7795_PD_ALWAYS_ON 32 > > > > Shouldn't we also define the always-on power domain for the other SoCs > > (patches 2/7 to 6/7 in this series) ? I know they're already covered by > > the cpg power domain, but going forward I believe that standardizing on > > the SYSC power domains would be beneficial. We of course have to keep > > backward compatibility in the implementation. > > Yes, that's the plan. I didn't want to make that change now, as someone > may object against the always-on power domain. On r8a7795 it feels more > natural, as it also has I/O devices in SYSC power areas, unlike R-Car Gen2 > and H1. That's fine with me. > Note that it also complicates the rcar-sysc core driver: it has to choose > between the cpg_mssr_*() or cpg_mstp_*() callbacks, depending on SoC, and > also depending on DTS if we migrate the older SoCs to CPG/MSSR and want to > maintain backwards-compatibility. > > Time to finish the "renesas,apmu" enable-method work, and celebrate one big > flag day for all the new DT evolutions? I'll bring champagne ;-) -- Regards, Laurent Pinchart
