[PATCH 2/5] arm64: dts: exynos: Move the clock-frequency property

[Krzysztof Kozlowski]

On Wed, Feb 8, 2017 at 12:09 PM, Hoegeun Kwon <hoegeun.kwon at samsung.com> wrote:
> On 02/08/2017 04:32 PM, Krzysztof Kozlowski wrote:
>>
>> On Wed, Feb 8, 2017 at 9:24 AM, Hoegeun Kwon <hoegeun.kwon at samsung.com>
>> wrote:
>>>>
>>>> "Remove the ports node abd move burst and esc clock frequency properties
>>>> to the parent (DSI node)."
>>>>
>>>> The information which is missing is the answer for WHY? Why are you
>>>> doing this?
>>>
>>>
>>> The current mipi-dsi must have at least one OF graph.
>>> However, for example, dsi and panel are parent-child relationships,
>>> so OF graph is not needed, and fimd and dsi are not connected to the OF
>>> graph.
>>> In this case, an error occurred in dsi_parse in the code before patch
>>> (1/5).
>>
>> The error occurred in case of DSI + FIMD? I do not get it whether you
>> are removing the something which is not needed or fixing something.
>>
>>> So I modified dsi_parse_dt.
>>>
>>>> Does the patch depends on 1/5?
>>>
>>>
>>> Yes, it is.
>>> The 2/5 to 5/5 patches depend on the 1/5 patch.
>>
>> So that's a break of DT ABI for no clear (yet) reasons. Please mention
>> this in 1/5 patch and explain what is really fixed.
>
>
> I would like to post the s6e63j0x03 panel driver for exynos3250.
> However, as Rob Herring noted, dsi + panel is a parental relationship,
> so OF grpah is not needed.
> Therefore, the current dsi_parse_dt function will throw an error,
> because there is no linked OF graph for case such as fimd + dsi + panel.
> I think that the OF graph of dsi should be deleted even if it is not
> the purpose of the s6e63j0x03 panel driver.
>
> So the 1/5 patch parse the Pll, burst and esc clock frequency properties
> in dsi_parse_dt and modified to create a bridge_node only if there is
> an OF graph associated with dsi.
>

Thanks, now it makes sense. Such clear explanation should be part of
commit 1/5 as a proof that ABI breakage is needed.

BR,
Krzysztof