[PATCH 2/8] ARM: dts: OMAP36xx: move CPU OPP tables to device tree

Jon Hunter jon-hunter at ti.com
Sat Mar 16 01:58:43 EST 2013


On 03/15/2013 09:38 AM, Nishanth Menon wrote:
> On 09:26-20130315, Jon Hunter wrote:
>>
>> On 03/15/2013 08:56 AM, Nishanth Menon wrote:
>>> On 16:44-20130314, Jon Hunter wrote:
>>>>
>>>> On 03/14/2013 03:58 PM, Nishanth Menon wrote:
>>>>> Add DT OPP table for OMAP36xx family of devices. This data is
>>>>> decoded by OF with of_init_opp_table() helper function. This
>>>>> overrides the default OMAP34xx CPU OPP table definition.
>>>>
>>>> Not sure I following the last sentence. The tables are in a different
>>>> dtsi file and only the relevant file should be included, right?
>>> omap3630.dsi includes omap3.dtsi (which is meant for OMAP34xx).
>>> The opp tables introduced by this patch in omap36xx.dtsi will override
>>> the ones defined on omap3.dtsi. Will the following rephrase help clarify
>>> this?
>>>
>>> Original:
>>> This overrides the default OMAP34xx CPU OPP table definition.
>>> Suggested;
>>> This overrides the default OMAP34xx CPU OPP table definition in
>>> omap3.dtsi.
>>
>> Sorry, I just missed that the omap3430 opps were in omap3.dtsi and not
>> omap34xx.dtsi. I guess I am not familiar with how the DTC overrides
>> nodes, however, at least from a readability standpoint it would seem
>> nice to have the omap3430 opps in a omap3430 specific dtsi and not
>> omap3.dtsi. However, thats just my opinion.
> most of omap3630 is based off omap3430. I know from an readability point
> of view, it might have been good to split that to omap3-common.dtsi,
> omap34xx.dtsi, omap36xx.dtsi etc.. But there is no real need at this
> point in time to do that. Unless, ofcourse, we'd like to set that up as
> an standard for all OMAP SoCs...

How would omap3-common.dtsi be any different from omap3.dtsi? I don't
wish us to go nuts with creating dtsi files either, but having an
omap3430.dtsi does not seem unreasonable to me, but that is just my opinion.

Jon



More information about the devicetree-discuss mailing list