[PATCH v2 00/11] Add Milk-V Pioneer RISC-V board support

Conor Dooley conor at kernel.org
Fri Sep 22 03:50:29 PDT 2023


On Fri, Sep 22, 2023 at 06:24:25PM +0800, Chen Wang wrote:
> Regards,
> 
> unicornx
> 
> Conor Dooley <conor.dooley at microchip.com> 于2023年9月20日周三 18:01写道:
> >
> > On Wed, Sep 20, 2023 at 02:33:48PM +0800, Chen Wang wrote:
> > > Milk-V Pioneer [1] is a developer motherboard based on SOPHON SG2042 [2]
> > > in a standard mATX form factor. Add minimal device
> > > tree files for the SG2042 SOC and the Milk-V Pioneer board.
> >
> > Please run dtbs_check with W=1 set & fix the below issues:
> >
> > sg2042-cpus.dtsi:1600.25-1607.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 0: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1609.25-1616.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 1: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1618.25-1625.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 2: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1627.25-1634.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 3: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1636.25-1643.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 4: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1645.25-1652.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 5: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1654.25-1661.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 6: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1663.25-1670.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 7: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1672.25-1679.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 8: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1681.25-1688.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 9: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1690.27-1697.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 10: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1699.27-1706.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 11: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1708.27-1715.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 12: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1717.27-1724.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 13: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1726.27-1733.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 14: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:1735.27-1742.5: Warning (unit_address_vs_reg): /cpus/l2-cache at 15: node has a unit name, but no reg or ranges property
> > sg2042-cpus.dtsi:256.15-275.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 0: duplicate unit-address (also used in node /cpus/l2-cache at 0)
> > sg2042-cpus.dtsi:277.15-296.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 1: duplicate unit-address (also used in node /cpus/l2-cache at 1)
> > sg2042-cpus.dtsi:298.15-317.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 2: duplicate unit-address (also used in node /cpus/l2-cache at 2)
> > sg2042-cpus.dtsi:319.15-338.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 3: duplicate unit-address (also used in node /cpus/l2-cache at 3)
> > sg2042-cpus.dtsi:340.15-359.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 4: duplicate unit-address (also used in node /cpus/l2-cache at 4)
> > sg2042-cpus.dtsi:361.15-380.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 5: duplicate unit-address (also used in node /cpus/l2-cache at 5)
> > sg2042-cpus.dtsi:382.15-401.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 6: duplicate unit-address (also used in node /cpus/l2-cache at 6)
> > sg2042-cpus.dtsi:403.15-422.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 7: duplicate unit-address (also used in node /cpus/l2-cache at 7)
> > sg2042-cpus.dtsi:424.15-443.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 8: duplicate unit-address (also used in node /cpus/l2-cache at 8)
> > sg2042-cpus.dtsi:445.15-464.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 9: duplicate unit-address (also used in node /cpus/l2-cache at 9)
> > sg2042-cpus.dtsi:466.17-485.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 10: duplicate unit-address (also used in node /cpus/l2-cache at 10)
> > sg2042-cpus.dtsi:487.17-506.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 11: duplicate unit-address (also used in node /cpus/l2-cache at 11)
> > sg2042-cpus.dtsi:508.17-527.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 12: duplicate unit-address (also used in node /cpus/l2-cache at 12)
> > sg2042-cpus.dtsi:529.17-548.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 13: duplicate unit-address (also used in node /cpus/l2-cache at 13)
> > sg2042-cpus.dtsi:550.17-569.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 14: duplicate unit-address (also used in node /cpus/l2-cache at 14)
> > sg2042-cpus.dtsi:571.17-590.5: Warning (unique_unit_address_if_enabled): /cpus/cpu at 15: duplicate unit-address (also used in node /cpus/l2-cache at 15)
> >
> > Thanks,
> > Conor.
> 
> I find the rootcause is due to following code in DTS:
> ```
>                 l2_cache0: l2-cache at 0 {
>                         compatible = "cache";
>                         cache-block-size = <64>;
>                         cache-level = <2>;
>                         cache-size = <1048576>;
>                         cache-sets = <1024>;
>                         cache-unified;
>                 };
> ```
> To differ 16 l2-cache controller, we add <unit-address>, but due to
> the l2-cache controller on sg2042 has no address, no reg property is
> added here. That's why dtbs_check warns "node has a unit name, but no
> reg or ranges property". I just double-confirmed with sophgo
> engineers, they  told me there is really no address for the cache
> controller for sg2042.
> 
> One solution I use here is to provide unique name for the l2-cache
> node. I learn this from  "arch/arm64/boot/dts/hisilicon/hi3660.dtsi"
> and seems this work and pass dtbs_check with W=1. For example:
> ```
>                 l2_cache0: l2-cache0 {
>                         compatible = "cache";
>                         cache-block-size = <64>;
>                         cache-level = <2>;
>                         cache-size = <1048576>;
>                         cache-sets = <1024>;
>                         cache-unified;
>                 };
> 
>                 l2_cache1: l2-cache1 {
>                         compatible = "cache";
>                         cache-block-size = <64>;
>                         cache-level = <2>;
>                         cache-size = <1048576>;
>                         cache-sets = <1024>;
>                         cache-unified;
>                 };
>                 ......
> ```
> But I remember as mentioned in
> https://devicetree-specification.readthedocs.io/en/latest/chapter2-devicetree-basics.html#generic-names-recommendation,
> node names should be generic. So I have two questions here:
> - Should I use "cache-controller" instead of "l2-cache", though I see
> "l2-cache" is used in many places but not listed in
> devicetree-specification.
> - Even if I replace "l2-cache" with "cache-controller", I think
> "cache-controller0", "cache-controller1" ... are not generic name, but
> due to sg2042 does not have address for cache controller, how to
> handle this problem?

I would go for "cache-controller-0" "cache-controller-1". Close as
possible to the generic node name while having the required differences
between nodes. There's already some examples in-tree (eg renesas) that
do this.

Thanks,
Conor.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-riscv/attachments/20230922/34135035/attachment.sig>


More information about the linux-riscv mailing list