[PATCH v2 4/6] RISC-V: Detect unaligned vector accesses supported.

[Jesse Taube]

On 6/14/24 04:40, Conor Dooley wrote:
> On Fri, Jun 14, 2024 at 09:36:55AM +0100, Conor Dooley wrote:
>> On Thu, Jun 13, 2024 at 03:16:13PM -0400, Jesse Taube wrote:
>>> --- a/arch/riscv/kernel/unaligned_access_speed.c
>>> +++ b/arch/riscv/kernel/unaligned_access_speed.c
>>> @@ -19,7 +19,8 @@
>>>   #define MISALIGNED_BUFFER_ORDER get_order(MISALIGNED_BUFFER_SIZE)
>>>   #define MISALIGNED_COPY_SIZE ((MISALIGNED_BUFFER_SIZE / 2) - 0x80)
>>>   
>>> -DEFINE_PER_CPU(long, misaligned_access_speed);
>>> +DEFINE_PER_CPU(long, misaligned_access_speed) = RISCV_HWPROBE_MISALIGNED_UNKNOWN;
>>> +DEFINE_PER_CPU(long, vector_misaligned_access) = RISCV_HWPROBE_VEC_MISALIGNED_UNSUPPORTED;
>>>   
>>>   #ifdef CONFIG_RISCV_PROBE_UNALIGNED_ACCESS
>>>   static cpumask_t fast_misaligned_access;
>>> @@ -268,12 +269,18 @@ static int check_unaligned_access_all_cpus(void)
>>>   
>>>   	if (riscv_has_extension_unlikely(RISCV_ISA_EXT_ZICCLSM)) {
>>>   		for_each_online_cpu(cpu) {
>>> +#ifdef CONFIG_RISCV_VECTOR_MISALIGNED
>>> +			per_cpu(vector_misaligned_access, cpu) = RISCV_HWPROBE_VEC_MISALIGNED_FAST;
>>> +#endif
>>> +#ifdef CONFIG_RISCV_MISALIGNED
>>>   			per_cpu(misaligned_access_speed, cpu) = RISCV_HWPROBE_MISALIGNED_FAST;
>>> +#endif
>>
>> Can you IS_ENABLED()-ify these two as well please?
> 
> Ah, you can't cos the variable doesn't exist in the other case.

Yeah kinda just dealing with how it was originally written ideally we 
would use IS_ENABLED. I don't really want to have a 500+ diff patch
IS_ENABLED()-ifying the original code as well. I can do that if 
necessary though.

Thank,
Jesse Taube