Re: [PATCH v2 RESEND] ARM: dts: hpe: BUG: Correct GXP register ranges

[Arnd Bergmann]

On Fri, Nov 10, 2023, at 20:48, charles.kearney@xxxxxxx wrote:
> -		ahb@c0000000 {
> +		ahb@80000000 {
>  			compatible = "simple-bus";
>  			#address-cells = <1>;
>  			#size-cells = <1>;
> -			ranges = <0x0 0xc0000000 0x30000000>;
>  			dma-ranges;
> +			ranges = <0x00000000 0x80000000 0x20000000>,
> +					 <0x40000000 0xc0000000 0x3fff0000>;
> 
> -			vic0: interrupt-controller@eff0000 {
> +			vic0: interrupt-controller@4eff0000 {
>  				compatible = "arm,pl192-vic";
> -				reg = <0xeff0000 0x1000>;
> +				reg = <0x4eff0000 0x1000>;
>  				interrupt-controller;
>  				#interrupt-cells = <1>;
>  			};
> 
> -			vic1: interrupt-controller@80f00000 {
> +			vic1: interrupt-controller@f00000 {
>  				compatible = "arm,pl192-vic";
> -				reg = <0x80f00000 0x1000>;
> +				reg = <0xf00000 0x1000>;
>  				interrupt-controller;
>  				#interrupt-cells = <1>;
>  			};
> 
> -			uarta: serial@e0 {
> +			uarta: serial@400000e0 {

Are these the actual addresses in the datasheet or
did you just make up a mapping?

If you have two separate address ranges, that would
typically indicate that there are two distinct ahb
instances that should be represented as separate
bus nodes. With the new mapping, it looks like all the
devices using vic0 are part of the same address range
as vic0, which also indicates that they might
be on one bus.

If this is a single ahb instance with multiple distinct
address ranges, I would expect that you need to adapt
the dma-ranges as well.

     Arnd