Re: [PATCH 1/3] firmware: imx: Add stub functions for SCMI MISC API

["Arnd Bergmann" <arnd@arndb.de>]

On Fri, Aug 22, 2025, at 04:35, Peng Fan wrote:
>> Subject: Re: [PATCH 1/3] firmware: imx: Add stub functions for SCMI

>> The logic is the same here in the end, but the link failure is easier to
>> trigger and repair if someone gets it wrong.
>> 
>> Also, for drivers that actually need the exported interface, the
>> dependency becomes the simpler 'depends on IMX_SCMI_MISC_DRV'.
>
> Yeah, but since consumer drivers supports multiple platforms,
> if platform A not requires the real API in IMX_SCMI_MISC_DRV,
> no need to link the real API.
>
>> 
>> Which driver using this symbol are you actually looking at? I see you
>> have three similar patches for a couple of interfaces, and want to make
>> sure the added complexity is really needed here. I do a lot of
>> randconfig build tests, so quite often I end up being the one that runs
>> into the subtle link failures from these.
>
> The patch is here
> https://lore.kernel.org/linux-remoteproc/20250821-imx95-rproc-1-v5-0-e93191dfac51@nxp.com/T/#macc660f61873742de447ac8c20d34f2d494ff712
>
> Please help give a look.

Thanks for the pointer. Yes, I agree this makes sense then,
and seems harmless since there are only a few drivers that
use the interfaces. I still think that the use conditional
stubs for optional features is problematic when the definition
is in a loadable module, and want to make sure this
is not just done out of habit, but because of a specific
documented requirement.

Please clarify the changelog then, changing the 'Fixes' tags
to a description that explains why you choose to revert my
earlier patches.

      Arnd