Re: [PATCH 2/2] platform/x86: add CMOS battery monitoring for simatic IPCs

[Hans de Goede]

Hi,

On 7/13/23 10:28, Henning Schild wrote:
> Am Wed, 12 Jul 2023 20:05:54 +0200
> schrieb Hans de Goede <hdegoede@xxxxxxxxxx>:
> 
>> <resend with Lee's new email address>
>>
>> Hi,
>>
>> On 7/6/23 17:48, Henning Schild wrote:
>>> Siemens Simatic Industrial PCs can monitor the voltage of the CMOS
>>> battery with two bits that indicate low or empty state. This can be
>>> GPIO or PortIO based.
>>> Here we model that as a hwmon voltage. The core driver does the
>>> PortIO and provides boilerplate for the GPIO versions. Which are
>>> split out to model runtime dependencies while allowing fine-grained
>>> kernel configuration.
>>>
>>> Signed-off-by: Henning Schild <henning.schild@xxxxxxxxxxx>  
>>
>> So I tried to merge this, but it does not apply because:
>>
>> "[PATCH 1/1] leds: simatic-ipc-leds-gpio: add new model BX-21A"
>> https://lore.kernel.org/platform-driver-x86/20230531155457.31632-2-henning.schild@xxxxxxxxxxx/
>>
>> has not been merged yet.
>>
>> I think it would be best to split the:
>>
>> drivers/platform/x86/simatic-ipc.c
>> include/linux/platform_data/x86/simatic-ipc-base.h
>> include/linux/platform_data/x86/simatic-ipc.h
>>
>> bits of
>> https://lore.kernel.org/platform-driver-x86/20230531155457.31632-2-henning.schild@xxxxxxxxxxx/
>>
>> out into its own prep patch named e.g. :
>> "platform/x86: simatic-ipc: add BX-21A model"
>>
>> And then post a new v2 series for
>> "leds: simatic-ipc-leds-gpio: add new model BX-21A"
>> consisting of the prep patch + the actual new LED driver.
>>
>> Then I can do an immutable branch based on 6.5-rc1 with
>> the prep patch in there and send a pull-req to Lee Jones
>> for that, so that he can apply the LED driver patch on
>> top of the immutable branch.
>>
>> This way we can continue with merging all the pending
>> simatic IPC work without being dependent on Lee having
>> time to review the LED driver.
> 
> Sounds like a plan, i will send another version of "leds:
> simatic-ipc-leds-gpio: add new model BX-21A" with the pdx86 stuff as
> separate patch.

Great.

> But just to note one thing, we would be registering a platform device
> before there is a driver for it. ("simatic_ipc_leds_gpio_elkhartlake")
> I think that is fine>

Yes that really is not an issue, e.g. the ACPI subsystem registers
a platform device for every device in the ACPI tables and not nearly
all of them has a driver. Having a platform_device without a driver
around is not an issue.

Regards,

Hans