[kvm-unit-tests PATCH v2 01/11] riscv: sbi: Drop fwft upper bits test

[Clément Léger]

On 27/02/2025 15:19, Andrew Jones wrote:
> The test should be checking for SBI_SUCCESS and that the behavior of
> feature = BIT(32) is the same as feature = 0 (MISALIGNED_EXC_DELEG).
> However, enabling MISALIGNED_EXC_DELEG doesn't always lead to traps
> in S-mode (the platform may support misaligned accesses and not
> trap at all). Drop this upper bits of feature ID test for now.
> We'll add it back with another feature that has a deterministic
> behavior (such as with the ADUE feature).
> 
> Signed-off-by: Andrew Jones <andrew.jones at linux.dev>
> ---
>  riscv/sbi-fwft.c | 15 ---------------
>  1 file changed, 15 deletions(-)
> 
> diff --git a/riscv/sbi-fwft.c b/riscv/sbi-fwft.c
> index f73ae52e7397..f3408d8de081 100644
> --- a/riscv/sbi-fwft.c
> +++ b/riscv/sbi-fwft.c
> @@ -57,21 +57,6 @@ static void fwft_check_base(void)
>  	fwft_check_reserved(SBI_FWFT_GLOBAL_RESERVED_START);
>  	fwft_check_reserved(SBI_FWFT_GLOBAL_RESERVED_END);
>  
> -#if __riscv_xlen > 32
> -	/* Check id > 32 bits */
> -	{
> -		struct sbiret ret;
> -
> -		ret = fwft_get_raw(BIT(32));
> -		sbiret_report_error(&ret, SBI_ERR_INVALID_PARAM,
> -				    "get feature with bit 32 set error");
> -
> -		ret = fwft_set_raw(BIT(32), 0, 0);
> -		sbiret_report_error(&ret, SBI_ERR_INVALID_PARAM,
> -				    "set feature with bit 32 set error");
> -	}
> -#endif
> -
>  	report_prefix_pop();
>  }
>  

Hi Andrew,

Looks good to me.

Reviewed-by: Clément Léger <cleger at rivosinc.com>

Thanks,

Clément