Re: [PATCH] arm64/efi: fix variable 'si' set but not used

[Dave Martin]

On Tue, Jul 30, 2019 at 05:23:48PM -0400, Qian Cai wrote:
> GCC throws out this warning on arm64.
> 
> drivers/firmware/efi/libstub/arm-stub.c: In function 'efi_entry':
> drivers/firmware/efi/libstub/arm-stub.c:132:22: warning: variable 'si'
> set but not used [-Wunused-but-set-variable]
> 
> Fix it by making free_screen_info() a static inline function.
> 
> Signed-off-by: Qian Cai <cai@xxxxxx>
> ---
>  arch/arm64/include/asm/efi.h | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/include/asm/efi.h b/arch/arm64/include/asm/efi.h
> index 8e79ce9c3f5c..76a144702586 100644
> --- a/arch/arm64/include/asm/efi.h
> +++ b/arch/arm64/include/asm/efi.h
> @@ -105,7 +105,11 @@ static inline unsigned long efi_get_max_initrd_addr(unsigned long dram_base,
>  	((protocol##_t *)instance)->f(instance, ##__VA_ARGS__)
>  
>  #define alloc_screen_info(x...)		&screen_info
> -#define free_screen_info(x...)
> +
> +static inline void free_screen_info(efi_system_table_t *sys_table_arg,
> +				    struct screen_info *si)
> +{
> +}

Is this issue caused by the EFI stub being built with non-default
CFLAGS?

The toplevel Makefile specifies -Wno-unused-but-set-variable, which
would silence this warning.

It's debatable whether calling an empty inline function "uses" the
arguments, so I think your patch only silences the warning by accident:
different GCC versions, or clang, might still warn.


I wonder if we're missing any other options that would make sense for
the EFI stub.

[...]

Cheers
---Dave