[PATCH] riscv: Fix unaligned access info messages

[Alexandre Ghiti]

Hi Drew,

On 09/04/2025 17:36, Andrew Jones wrote:
> Ensure we only print messages about command line parameters when
> the parameters are actually in use. Also complain about the use
> of the vector parameter when vector support isn't available.
>
> Fixes: aecb09e091dc ("riscv: Add parameter for skipping access speed tests")
> Reported-by: Geert Uytterhoeven <geert at linux-m68k.org>
> Closes: https://lore.kernel.org/all/CAMuHMdVEp2_ho51gkpLLJG2HimqZ1gZ0fa=JA4uNNZjFFqaPMg@mail.gmail.com/
> Closes: https://lore.kernel.org/all/CAMuHMdWVMP0MYCLFq+b7H_uz-2omdFiDDUZq0t_gw0L9rrJtkQ@mail.gmail.com/
> Signed-off-by: Andrew Jones <ajones at ventanamicro.com>
> ---
>   arch/riscv/kernel/unaligned_access_speed.c | 35 +++++++++++++---------
>   1 file changed, 21 insertions(+), 14 deletions(-)
>
> diff --git a/arch/riscv/kernel/unaligned_access_speed.c b/arch/riscv/kernel/unaligned_access_speed.c
> index 585d2dcf2dab..b8ba13819d05 100644
> --- a/arch/riscv/kernel/unaligned_access_speed.c
> +++ b/arch/riscv/kernel/unaligned_access_speed.c
> @@ -439,29 +439,36 @@ static int __init check_unaligned_access_all_cpus(void)
>   {
>   	int cpu;
>   
> -	if (unaligned_scalar_speed_param == RISCV_HWPROBE_MISALIGNED_SCALAR_UNKNOWN &&
> -	    !check_unaligned_access_emulated_all_cpus()) {
> -		check_unaligned_access_speed_all_cpus();
> -	} else {
> -		pr_info("scalar unaligned access speed set to '%s' by command line\n",
> -			speed_str[unaligned_scalar_speed_param]);
> +	if (unaligned_scalar_speed_param != RISCV_HWPROBE_MISALIGNED_SCALAR_UNKNOWN) {
> +		pr_info("scalar unaligned access speed set to '%s' (%lu) by command line\n",
> +			speed_str[unaligned_scalar_speed_param], unaligned_scalar_speed_param);
>   		for_each_online_cpu(cpu)
>   			per_cpu(misaligned_access_speed, cpu) = unaligned_scalar_speed_param;
> +	} else if (!check_unaligned_access_emulated_all_cpus()) {
> +		check_unaligned_access_speed_all_cpus();
> +	}
> +
> +	if (unaligned_vector_speed_param != RISCV_HWPROBE_MISALIGNED_VECTOR_UNKNOWN) {
> +		if (!has_vector() &&
> +		    unaligned_vector_speed_param != RISCV_HWPROBE_MISALIGNED_VECTOR_UNSUPPORTED) {
> +			pr_warn("vector support is not available, ignoring unaligned_vector_speed=%s\n",
> +				speed_str[unaligned_vector_speed_param]);
> +		} else {
> +			pr_info("vector unaligned access speed set to '%s' (%lu) by command line\n",
> +				speed_str[unaligned_vector_speed_param], unaligned_vector_speed_param);
> +		}
>   	}
>   
>   	if (!has_vector())
>   		unaligned_vector_speed_param = RISCV_HWPROBE_MISALIGNED_VECTOR_UNSUPPORTED;
>   
> -	if (unaligned_vector_speed_param == RISCV_HWPROBE_MISALIGNED_VECTOR_UNKNOWN &&
> -	    !check_vector_unaligned_access_emulated_all_cpus() &&
> -	    IS_ENABLED(CONFIG_RISCV_PROBE_VECTOR_UNALIGNED_ACCESS)) {
> -		kthread_run(vec_check_unaligned_access_speed_all_cpus,
> -			    NULL, "vec_check_unaligned_access_speed_all_cpus");
> -	} else {
> -		pr_info("vector unaligned access speed set to '%s' by command line\n",
> -			speed_str[unaligned_vector_speed_param]);
> +	if (unaligned_vector_speed_param != RISCV_HWPROBE_MISALIGNED_VECTOR_UNKNOWN) {
>   		for_each_online_cpu(cpu)
>   			per_cpu(vector_misaligned_access, cpu) = unaligned_vector_speed_param;
> +	} else if (!check_vector_unaligned_access_emulated_all_cpus() &&
> +		   IS_ENABLED(CONFIG_RISCV_PROBE_VECTOR_UNALIGNED_ACCESS)) {
> +		kthread_run(vec_check_unaligned_access_speed_all_cpus,
> +			    NULL, "vec_check_unaligned_access_speed_all_cpus");
>   	}
>   
>   	/*


Tested-by: Alexandre Ghiti <alexghiti at rivosinc.com>

Reviewed-by: Alexandre Ghiti <alexghiti at rivosinc.com>

I'll wait for Geert feedback before picking it.

Thanks,

alex