Re: [PATCH v1 22/22] riscv: dts: starfive: jhb100: Add clocks and resets nodes

[Conor Dooley]

On Fri, Apr 03, 2026 at 01:07:48AM +0000, Changhuang Liang wrote:
> Hi, Conor
> 
> > On Thu, Apr 02, 2026 at 03:55:23AM -0700, Changhuang Liang wrote:
> > > Add clocks and resets nodes for JHB100 RISC-V BMC SoC. They contain
> > > sys0crg/sys1crg/sys2crg/per0crg/per1crg/per2crg/per3crg.
> > >
> > > Signed-off-by: Changhuang Liang <changhuang.liang@xxxxxxxxxxxxxxxx>
> > > ---
> > >  arch/riscv/boot/dts/starfive/jhb100.dtsi | 198
> > > ++++++++++++++++++++++-
> > >  1 file changed, 195 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/arch/riscv/boot/dts/starfive/jhb100.dtsi
> > > b/arch/riscv/boot/dts/starfive/jhb100.dtsi
> > > index 4d03470f78ab..700d00f800bc 100644
> > > --- a/arch/riscv/boot/dts/starfive/jhb100.dtsi
> > > +++ b/arch/riscv/boot/dts/starfive/jhb100.dtsi
> > > @@ -4,6 +4,8 @@
> > >   */
> > >
> > >  /dts-v1/;
> > > +#include <dt-bindings/clock/starfive,jhb100-crg.h>
> > > +#include <dt-bindings/reset/starfive,jhb100-crg.h>
> > >
> > >  / {
> > >  	compatible = "starfive,jhb100";
> > > @@ -268,12 +270,96 @@ pmu {
> > >  			<0x00 0x22 0xFFFFFFFF 0xFFFFFF22 0x00007FF8>;	/* Event
> > ID 34 */
> > >  	};
> > >
> > > -	clk_uart: clk-uart {
> > > -		compatible = "fixed-clock"; /* Initial clock handler for UART */
> > > +	osc: osc {
> > > +		compatible = "fixed-clock";
> > >  		#clock-cells = <0>;
> > >  		clock-frequency = <25000000>;
> > >  	};
> > 
> > Is this really on the SoC?
> 
> This is not on the SoC.
> 
> > 
> > >
> > > +	pll0: pll0 {
> > > +		compatible = "fixed-clock";
> > > +		#clock-cells = <0>;
> > > +		clock-frequency = <2400000000>;
> > > +	};
> > 
> > What's providing all of these PLLs? Are they all fixed-frequency on-chip PLLs
> > without an off-chip reference? I find that somewhat unlikely.
> > 
> > Since devicetrees are now being imported into U-Boot, it's important to make
> > sure that I'm not merging fixed-clocks that later get replaced by dedicated
> > drivers that U-Boot won't have.
> > 
> > To that end, I won't apply the series this depends on without this patch being
> > applied at the same time.
> 
> I am preparing a PLL driver series, but PLL0 and PLL1 will still retain fixed frequencies. 
> The reference clock for each PLL comes from the osc. Perhaps I can use "fixed-factor-clock" 
> to indicate the relationship of the reference clock.

I'll reserve judgement until I see that series so, but it wasn't as if
any of this was going into 7.1 anyway (or maybe even 7.2) so not a
problem.

Attachment: signature.asc
Description: PGP signature