Re: [PATCH 3/3] riscv: dts: spacemit: Add Milk-V Jupiter board (M1)

[Aurelien Jarno <aurelien@aurel32.net>]

On 2026-07-28 21:28, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] Missing unique board-specific `compatible` string for the M1 Jupiter board variant, which inherits the identical compatible string as the K1 variant and breaks DTB matching.
> --
> 
> --- Patch [3]: [PATCH 3/3] riscv: dts: spacemit: Add Milk-V Jupiter board (M1) ---
> commit 2d54397a3a4baaa8a2eee569d6fe3f912a3c7661
> Author: Aurelien Jarno <aurelien@aurel32.net>
> 
> riscv: dts: spacemit: Add Milk-V Jupiter board (M1)
>     
> This commit adds support for the Milk-V Jupiter board with the M1 CPU.
> The M1 is similar to the K1 but allows running at up to 1.8 GHz, adding
> a new device tree and an OPP table for this higher frequency.
> 
> > diff --git a/arch/riscv/boot/dts/spacemit/m1-milkv-jupiter.dts b/arch/riscv/boot/dts/spacemit/m1-milkv-jupiter.dts
> > new file mode 100644
> > index 0000000000000..32e1468385ca0
> > --- /dev/null
> > +++ b/arch/riscv/boot/dts/spacemit/m1-milkv-jupiter.dts
> > @@ -0,0 +1,11 @@
> > +// SPDX-License-Identifier: GPL-2.0 OR MIT
> > +/*
> > + * Copyright (C) 2026 Aurelien Jarno <aurelien@aurel32.net>
> > + */
> > +
> > +#include "k1-milkv-jupiter-common.dtsi"
> > +#include "m1-opp.dtsi"
> > +
> > +/ {
> > +	model = "Milk-V Jupiter (M1)";
> 
> [Severity: Medium]
> Since m1-milkv-jupiter.dts includes k1-milkv-jupiter-common.dtsi, does this
> board inherit the identical root compatible string as the K1 variant?
> 
> If a bootloader like U-Boot uses a FIT image containing both DTBs, it relies
> on the root compatible string to select the correct one. If they are
> identical, could the bootloader mistakenly apply this M1 DTB to a K1 board?
> 
> If that happens, would the K1 board be allowed to run at the 1.8 GHz
> operating point defined here, potentially causing instability or hardware
> damage since it is only rated for 1.6 GHz? 
> 
> Should a unique board-specific compatible string be added to this root node
> to uniquely identify the M1 board hardware?
> 
> > +};

That's a valid point, I'll fix that in the next version. I believe we 
want to keep the spacemit,k1 compatible even with the m1, as the 
difference is rather a property of the board and not the SoC.

What could be the compatible for the board? "milkv,jupiter-m1"? Maybe 
the existing board should have been called "milkv,jupiter-k1"...

-- 
Aurelien Jarno                          GPG: 4096R/1DDD8C9B
aurelien@aurel32.net                     http://aurel32.net