Hi Geert, On Tuesday, 26 September 2017 15:46:59 EEST Geert Uytterhoeven wrote: > On Tue, Sep 26, 2017 at 2:39 PM, Laurent Pinchart wrote: > > On Tuesday, 26 September 2017 14:47:35 EEST Geert Uytterhoeven wrote: > >> On Tue, Sep 26, 2017 at 12:05 PM, Laurent Pinchart wrote: > >> > The FCP instances associated with a VSPI must be treated differently > >> > than the ones associated with another type of VSP. Add a new compatible > >> > string to allow telling them apart. > >> > > >> > Signed-off-by: Laurent Pinchart > >> > <laurent.pinchart+renesas@xxxxxxxxxxxxxxxx> > >> > --- > >> > > >> > Documentation/devicetree/bindings/media/renesas,fcp.txt | 3 ++- > >> > 1 file changed, 2 insertions(+), 1 deletion(-) > >> > > >> > diff --git a/Documentation/devicetree/bindings/media/renesas,fcp.txt > >> > b/Documentation/devicetree/bindings/media/renesas,fcp.txt index > >> > 3ec91803ba58..c1f28736e2d6 100644 > >> > --- a/Documentation/devicetree/bindings/media/renesas,fcp.txt > >> > +++ b/Documentation/devicetree/bindings/media/renesas,fcp.txt > >> > @@ -11,7 +11,8 @@ are paired with. These DT bindings currently support > >> > the FCPV and FCPF. > >> > > >> > - compatible: Must be one or more of the following > >> > - - "renesas,fcpv" for generic compatible 'FCP for VSP' > >> > + - "renesas,fcpvi" for generic compatible 'FCP for VSPI' > >> > + - "renesas,fcpv" for generic compatible 'FCP for VSP' (all other > >> > VSPs) > >> > - "renesas,fcpf" for generic compatible 'FCP for FDP' > >> > >> You may want to update the paragraph before that, referring to three > >> types of FCP. > > > > As far as I know the FCP-VSPI and FCP-VSP[^I] are the same type of device, > > with different SoC integration. > > OK. So they should use the same compatible value? > > >> Just wondering: as FCPVI vs. FCPV is programmable through the FCPVSEL > >> bit, can you deduce the type using the renesas,fcp link in the > >> corresponding VSP node in DT? > > > > You could possibly, if you had a link from the FCP to the VSP in DT. As > > there's none, the type can't be currently inferred from DT at probe time. > > We could pass the type from the VSP driver to the FCP driver, but I'm not > > sure I want to go that way. If the FCP-VSPI and VSP[^I] are identicaly > > and need different software configuration due to being used by different > > types of VSPs it would make sense, but if there are hardware differences > > between the FCPs then I think they should be described in DT. > > Given the FCPVSEL bit, it looks like there's no difference, and software is > supposed to specify configuration. That's hard to tell for sure, it could be that there are hardware differences that require a different software configuration. I'll keep your comment in mind for the next version though, and will try to get more information from Renesas. -- Regards, Laurent Pinchart
