On Fri 20 Aug 07:09 PDT 2021, Sibi Sankar wrote:
> On 2021-08-19 10:07, Vinod Koul wrote:
> > Hi Sibi,
> >
> > On 19-08-21, 09:06, Sibi Sankar wrote:
> >
> > > diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi
> > > b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> > > index 5e4f4f3b738a..894106efadfe 100644
> > > --- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
> > > +++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> > > @@ -48,6 +48,16 @@
> > > #size-cells = <2>;
> > > ranges;
> > >
> > > + hyp_mem: memory@80000000 {
> > > + reg = <0x0 0x80000000 0x0 0x600000>;
> > > + no-map;
> >
> > This should conflict with the memory defined in this file:
> >
> > memory@80000000 {
> > device_type = "memory";
> > /* We expect the bootloader to fill in the size */
> > reg = <0 0x80000000 0 0>;
> > };
> >
> > I think this should be updated?
>
> Vinod,
>
> I prefer we leave ^^ node untouched. For platforms using hyp_mem, the
> regions defined in the memory map are valid and for the other
> platforms not using hyp_mem we would just delete them in the board
> files anyway.
Logically this node describes where there is RAM, the reserved-memory
then subtracts blocks of memory out of that. So I think it's perfectly
legal for a region at the base to be marked as no-map.
That said, isn't the address in the memory node just a placeholder?
Regards,
Bjorn
