[PATCH] soc: imx: Kconfig: force using OF when COMPILE_TEST

[Peng Fan]

> Subject: Re: [PATCH] soc: imx: Kconfig: force using OF when COMPILE_TEST
> 
> On Wed, Jun 17, 2020 at 02:05:33AM +0000, Peng Fan wrote:
> > > Subject: Re: [PATCH] soc: imx: Kconfig: force using OF when
> > > COMPILE_TEST
> > >
> > > On 2020-06-16 16:59, Uwe Kleine-König wrote:
> > > > On Mon, Jun 15, 2020 at 06:21:25PM +0800, peng.fan at nxp.com wrote:
> > > >> From: Peng Fan <peng.fan at nxp.com>
> > > >>
> > > >> Fix the build warning with x86_64-randconfig
> > > >>>> drivers/soc/imx/soc-imx8m.c:150:34: warning: unused variable
> > > >>>> 'imx8_soc_match' [-Wunused-const-variable]
> > > >> static const struct of_device_id imx8_soc_match[] = { ^
> > > >>
> > > >> Fixes: fc40200ebf82 ("soc: imx: increase build coverage for imx8m
> > > >> soc
> > > >> driver")
> > > >> Reported-by: kernel test robot <lkp at intel.com>
> > > >> Signed-off-by: Peng Fan <peng.fan at nxp.com>
> > > >> ---
> > > >>   drivers/soc/imx/Kconfig | 2 +-
> > > >>   1 file changed, 1 insertion(+), 1 deletion(-)
> > > >>
> > > >> diff --git a/drivers/soc/imx/Kconfig b/drivers/soc/imx/Kconfig
> > > >> index
> > > >> d515d2cc20ed..aadf13c9d396 100644
> > > >> --- a/drivers/soc/imx/Kconfig
> > > >> +++ b/drivers/soc/imx/Kconfig
> > > >> @@ -19,7 +19,7 @@ config IMX_SCU_SOC
> > > >>
> > > >>   config SOC_IMX8M
> > > >>   	bool "i.MX8M SoC family support"
> > > >> -	depends on ARCH_MXC || COMPILE_TEST
> > > >> +	depends on ARCH_MXC || (COMPILE_TEST && OF)
> > > >
> > > > A bit prettier (IMHO) would be:
> > > >
> > > > config SOC_IMX8M
> > > >   	bool "i.MX8M SoC family support"
> > > >   	depends on ARCH_MXC || COMPILE_TEST
> > > > +	depends on OF
> > > >   	default ARCH_MXC && ARM64
> > > >   	select SOC_BUS
> > > >   	help
> > >
> > > That's not just prettier, it's logically correct. If some code needs
> > > OF to build correctly, then it depends on OF, regardless of *why*
> > > it's being built at any given time.
> > >
> > > That said, if the only issue in this particular case is the warning
> > > above, then it's hardly a real dependency; simply marking the
> > > of_device_id data as __maybe_unused (like various other drivers do)
> > > should suffice. Limiting COMPILE_TEST coverage instead of actually
> > > fixing simple issues that it exposes seems a bit backwards...
> >
> > ok, I'll use __maybe_unused in v2, but it actually not make sense to
> > compile this file for x86_64.
> 
> This is wrong in a subtle way. It makes sense to compile the driver on
> x86_64 (and other archs) for QA reasons. It doesn't make sense to use a
> kernel on an x86 machine with this driver enabled. That's why we have
> COMPILE_TEST which you disable if you want to create a kernel to actually
> run it.

Shawn,

What's your idea? I am not very clear on which path to go.

Thanks,
Peng.

> 
> Best regards
> Uwe
> 
> 
> --
> Pengutronix e.K.                           | Uwe Kleine-König
> |
> Industrial Linux Solutions                 | https://www.pengutronix.de/
> |