Re: [PATCH 2/5] arm64: dts: ls1028a: describe the Lynx 10G SerDes

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] The `reg` property for the `serdes` node incorrectly sets the region size to `0xffff` instead of `0x10000`, causing an off-by-one error in the resource length.
--

commit a39aeb25ec050b4deb66ed84479cc1400f6ce47b
Author: Vladimir Oltean <vladimir.oltean@nxp.com>

arm64: dts: ls1028a: describe the Lynx 10G SerDes

This commit describes the Lynx 10G SerDes block and its 4 SerDes lanes found
on the LS1028A SoC. The node is left disabled at the SoC level so board
device trees can enable it once consumer nodes use it.

> diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi b/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
> --- a/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
> +++ b/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
[ ... ]
> @@ -250,6 +250,35 @@ ls1028a_uid: unique-id@1c {
>  			};
>  		};
>  
> +		serdes: phy@1ea0000 {
> +			compatible = "fsl,ls1028a-serdes";
> +			reg = <0x00 0x1ea0000 0x0 0xffff>;

[Severity: Low]
Does this reg property specify a length that is off by one byte?

The reg property size field represents the length of the region in bytes. If
this is a standard 64KB hardware block, should the size be 0x10000 instead
of 0xffff?

Setting the length to 0xffff tracks a memory region that is 1 byte smaller
than the actual block. While ioremap aligns to the page boundary and avoids
runtime side effects, this tracks an incorrect resource length and leaves the
very last byte unmapped by the resource manager.

> +			#address-cells = <1>;
> +			#size-cells = <0>;
> +			#phy-cells = <1>;
[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260630110459.516364-1-ioana.ciornei@nxp.com?part=2