Re: [PATCH v3 2/2] arm64: dts: imx8mq: Deduplicate PCIe clock-names property

[Marek Vasut <marex@denx.de>]

On 1/17/23 05:00, Hongxing Zhu wrote:

Hi,

[...]

>> diff --git a/arch/arm64/boot/dts/freescale/imx8mq.dtsi
>> b/arch/arm64/boot/dts/freescale/imx8mq.dtsi
>> index faed28e3ffa17..98fbba4c99a99 100644
>> --- a/arch/arm64/boot/dts/freescale/imx8mq.dtsi
>> +++ b/arch/arm64/boot/dts/freescale/imx8mq.dtsi
>> @@ -1542,6 +1542,11 @@ pcie0: pcie@33800000 {
>>   			                <0 0 0 4 &gic GIC_SPI 122
>> IRQ_TYPE_LEVEL_HIGH>;
>>   			fsl,max-link-speed = <2>;
>>   			linux,pci-domain = <0>;
>> +			clocks = <&clk IMX8MQ_CLK_PCIE1_ROOT>,
>> +				 <&clk IMX8MQ_CLK_PCIE1_PHY>,
>> +				 <&clk IMX8MQ_CLK_PCIE1_PHY>,
> Why there are two PHY clocks?
> Do you want to define IMX8MQ_CLK_PCIE1_PHY clock as "pcie_bus" clock here,
>   and then change "pcie_bus" clock to "pcie0_refclk" in the evk board dts later?

Yes

> How about to set the "pcie_bus" clock as " pcie#_refclk " directly in i.MX8MQ
>   dtsi file?
> Thus, it can avoid more duplicated codes in boards dts files further.

I can do that, but is that really the right approach ?

Consider a board which supplies 'pcie_bus' from some I2C PCIe clock 
generator, such board would now have a useless /pcie#-clock {} node in 
DT , right ?

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel