RE: [PATCH v32 6/7] arm64: dts: imx8ulp: add secure enclave node

> arm64: dts: imx8ulp: add secure enclave node
> 
> This patch adds the EdgeLock Enclave (ELE) secure-enclave node to the
> i.MX8ULP SoC dtsi. The node is kept disabled by default so it does not impose
> a reserved-memory requirement on every board.
> 
> Closes: https://sashiko.dev/#/patchset/20260724-imx-se-if-v30-0-
> [email protected]?part=6
> 
> > diff --git a/arch/arm64/boot/dts/freescale/imx8ulp.dtsi
> > b/arch/arm64/boot/dts/freescale/imx8ulp.dtsi
> > index c6d1bb9edf388..38233dd74ee33 100644
> > --- a/arch/arm64/boot/dts/freescale/imx8ulp.dtsi
> > +++ b/arch/arm64/boot/dts/freescale/imx8ulp.dtsi
> 
> [ ... ]
> 
> > @@ -185,6 +185,14 @@ scmi_sensor: protocol@15 {
> >  				#thermal-sensor-cells = <1>;
> >  			};
> >  		};
> > +
> > +		hsm0: secure-enclave {
> > +			compatible = "fsl,imx8ulp-se-ele-hsm";
> > +			mbox-names = "tx", "rx";
> > +			mboxes = <&s4muap 0 0>, <&s4muap 1 0>;
> > +			sram = <&sram0>;
> > +			status = "disabled";
> > +		};
> 
> [Severity: Low]
> Will this cause a devicetree schema validation warning during dtbs_check?
> 
> Because the dt-binding schema requires the 'memory-region' property for the
> "fsl,imx8ulp-se-ele-hsm" compatible string, omitting it here means any board
> that includes imx8ulp.dtsi without enabling the enclave will fail validation,
> even though the node is disabled.
> 
In this series it does not fire, though: the only boards that build
imx8ulp.dtsi (imx8ulp-evk.dts, and 9x9-evk which includes it) also
#include imx8ulp-firmware.dtsi, which overrides &hsm0 with
'memory-region = <&ele_reserved>' and status = "okay". dtbs_check runs
on the final assembled .dtb, so those DTBs carry a complete node and
pass cleanly. I have run `make dtbs_check` for imx8ulp and there is no
new warning.


> Could we avoid validation warnings by defining a dummy memory-region, or
> by restructuring how this node is included for boards that do not use the
> enclave?
> 
No, adding a dummy memory-region - pointing a real device at a
fake/empty reserved-memory is misleading and can mask a genuine missing
pool.

Request you to consider this as: not an issue.
> >  	};
> >
> >  	cm33: remoteproc-cm33 {
> 
> --
> Sashiko AI review · https://sashiko.dev/#/patchset/20260730-imx-se-if-v32-0-
> [email protected]?part=6