On Wed, Oct 31, 2018 at 07:00:33PM +0100, Borislav Petkov wrote:
> +#ifdef CONFIG_ACPI_ADXL
> const char * const *adxl_get_component_names(void);
> int adxl_decode(u64 addr, u64 component_values[]);
> +#else
> +static inline const char * const *adxl_get_component_names(void) { return NULL; }
> +static inline int adxl_decode(u64 addr, u64 component_values[]) { return false; }
> +#endif
>
> #endif /* _LINUX_ADXL_H */
It's not exactly ideal to build the skx_edac.c driver with these
stubs. Sure, it make the kernel link without errors. But now you
silently end up with a driver that doesn't really do all you want
it to do.
Perhaps this isn't a huge issue. Only "randconfig" would try to
build a kernel without ACPI. The user will find other stuff is
broken in an ACPI=n kernel long before they notice the lack of
EDAC error reporting.
So:
Acked-by: Tony Luck <tony.luck@xxxxxxxxx>
-Tony
