Re: [PATCH] interconnect: qcom: Add the missing MODULE_LICENSE

[Georgi Djakov <djakov@kernel.org>]

Hi Yiwei,

On 13.09.22 13:06, Leo Yan wrote:
> Changed to Bjorn's updated email.
> 
> On Tue, Sep 13, 2022 at 04:30:50PM +0800, Huang Yiwei wrote:
>> Hi Leo,
>>
>> On 9/13/2022 Tue 15:08, Leo Yan wrote:
>>> Hi Yiwei,
>>>
>>> On Tue, Sep 13, 2022 at 02:27:21PM +0800, Huang Yiwei wrote:
>>>> Since icc-common.c can be compiled as module, add the missing
>>>> MODULE_LICENSE to avoid compile errors.
>>>
>>> Just curious how you can enable config INTERCONNECT_QCOM as "m"
>>> (module)?
>>>
>>> I checked the config INTERCONNECT_QCOM which is dependent on
>>> INTERCONNECT, and INTERCONNECT has below dependency:
>>>
>>>     - DRM_TEGRA [=n] && HAS_IOMEM [=y] && (ARCH_TEGRA [=y] || ARM && COMPILE_TEST [=n]) && COMMON_CLK [=y] && DRM [=m] && OF [=y]
>>>     - TEGRA_MC [=n] && MEMORY [=y] && (ARCH_TEGRA [=y] || COMPILE_TEST [=n] && COMMON_CLK [=y])
>>>
>>> Seems to me, it's impossible to enable INTERCONNECT as "m" because the
>>> dependency ARCH_TEGRA/OF/COMMON_CLK cannot be configured as "m".
>>>
>>> Thanks,
>>> Leo
>>>
>> Found the problem, INTERCONNECT_QCOM is modified to tristate internally
>> so we can compile it as a module, and the new added icc-common.c uses
>> INTERCONNECT_QCOM config to compile, then cause the problem.
> 
> Thanks for confirmation, Yiwei.
> 
>> So shall we change INTERCONNECT_QCOM config to tristate?
> 
> I personally think modularization is the right thing to do; to avoid
> misleading you, I would leave your question to interconnect
> maintainers.

Thanks for clarifying! It sounds like the right approach. So this should
be a single patch that allows icc-common.c to be compiled as a module and
also adds MODULE_LICENSE where needed to avoid any compile errors.

Thanks,
Georgi