Re: [PATCH] riscv: cpufeature: use bitmap_equal() instead of memcmp()

[Clément Léger]

On 11/02/2025 08:41, Alexandre Ghiti wrote:
> Hi Clément,
> 
> On 10/02/2025 16:56, Clément Léger wrote:
>> Comparison of bitmaps should be done using bitmap_equal(), not memcmp(),
>> use the former one to compare isa bitmaps.
>>
>> Signed-off-by: Clément Léger <cleger@xxxxxxxxxxxx>
>> ---
>>   arch/riscv/kernel/cpufeature.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/riscv/kernel/cpufeature.c b/arch/riscv/kernel/
>> cpufeature.c
>> index c6ba750536c3..40ac72e407b6 100644
>> --- a/arch/riscv/kernel/cpufeature.c
>> +++ b/arch/riscv/kernel/cpufeature.c
>> @@ -479,7 +479,7 @@ static void __init riscv_resolve_isa(unsigned long
>> *source_isa,
>>               if (bit < RISCV_ISA_EXT_BASE)
>>                   *this_hwcap |= isa2hwcap[bit];
>>           }
>> -    } while (loop && memcmp(prev_resolved_isa, resolved_isa,
>> sizeof(prev_resolved_isa)));
>> +    } while (loop && !bitmap_equal(prev_resolved_isa, resolved_isa,
>> RISCV_ISA_EXT_MAX));
>>   }
>>     static void __init match_isa_ext(const char *name, const char
>> *name_end, unsigned long *bitmap)
> 
> 
> Since bitmap does not guarantee the value of the 'extra' bits (https://
> elixir.bootlin.com/linux/v6.13.1/source/lib/bitmap.c#L24), this is a fix
> so I would add a Fixes tag:
> 
> Fixes: 625034abd52a8c ("riscv: add ISA extensions validation callback")

Yeah, I wasn't sure since it does not fix anything per se. We aligned
the ISA size to a multiple of 64 (128) so we actually do not have any
extra undefined bits. But I agree that it's better to track it correctly
as a fix.

That makes me think that we could actually size RISCV_ISA_EXT_MAX to be
exactly the number of ISA extensions we support rather than a larger
arbitrary value.

> 
> You can also add:
> 
> Reviewed-by: Alexandre Ghiti <alexghiti@xxxxxxxxxxxx>

Thanks !

Clément

> 
> Thanks,
> 
> Alex
>