2015-11-02 22:07:14

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH RESEND v2 net-next] net: hisilicon: updates HNS config and documents

On Saturday 31 October 2015 02:18:19 Salil Mehta wrote:
> On 10/31/2015 1:40 AM, huangdaode wrote:
> > On 2015/10/30 22:20, Arnd Bergmann wrote:
> >> On Tuesday 27 October 2015 19:16:34 huangdaode wrote:
> >>> mdio@803c0000 {
> >>> #address-cells = <1>;
> >>> #size-cells = <0>;
> >>> - compatible = "hisilicon,mdio","hisilicon,hns-mdio";
> >>> + compatible = "hisilicon,hns-mdio","hisilicon,mdio";
> >>> reg = <0x0 0x803c0000 0x0 0x10000>;
> >>>
> >> Does "hisilicon,mdio" actually have a specific meaning? Is that just
> >> there
> >> for legacy reasons?
> >>
> >> Arnd
> >>
> >> .
> >>
> > hi Arnd,
> > "hisilicon,mdio" is the gernerical configuation compatible for the
> > default hisilicon chip,
> > We use generic hisilicon since our MDIO/PHY is same across flavour of
> > SoCs.
> >
> Hi Arnd,
> Yes, "hisilicon,mdio" exists for the legacy reasons to support older SoC
> Hip04.

I see. In that case, that compatible string should probably be mentioned
only as 'optional' in the binding, and not used for devices other than
hip04.

Arnd


2015-11-12 02:11:45

by huangdaode

[permalink] [raw]
Subject: Re: [PATCH RESEND v2 net-next] net: hisilicon: updates HNS config and documents

On 2015/11/3 6:05, Arnd Bergmann wrote:
> On Saturday 31 October 2015 02:18:19 Salil Mehta wrote:
>> On 10/31/2015 1:40 AM, huangdaode wrote:
>>> On 2015/10/30 22:20, Arnd Bergmann wrote:
>>>> On Tuesday 27 October 2015 19:16:34 huangdaode wrote:
>>>>> mdio@803c0000 {
>>>>> #address-cells = <1>;
>>>>> #size-cells = <0>;
>>>>> - compatible = "hisilicon,mdio","hisilicon,hns-mdio";
>>>>> + compatible = "hisilicon,hns-mdio","hisilicon,mdio";
>>>>> reg = <0x0 0x803c0000 0x0 0x10000>;
>>>>>
>>>> Does "hisilicon,mdio" actually have a specific meaning? Is that just
>>>> there
>>>> for legacy reasons?
>>>>
>>>> Arnd
>>>>
>>>> .
>>>>
>>> hi Arnd,
>>> "hisilicon,mdio" is the gernerical configuation compatible for the
>>> default hisilicon chip,
>>> We use generic hisilicon since our MDIO/PHY is same across flavour of
>>> SoCs.
>>>
>> Hi Arnd,
>> Yes, "hisilicon,mdio" exists for the legacy reasons to support older SoC
>> Hip04.
> I see. In that case, that compatible string should probably be mentioned
> only as 'optional' in the binding, and not used for devices other than
> hip04.
>
> Arnd
>
> .
>
Hi, Arnd
ok, ill send a patch to specify it, please Ack then.
thanks
Daode.

--
Best Regards
Daode Huang