Re: [PATCH v3 01/16] bitops: Change parity8() return type to bool

[H. Peter Anvin]

On March 7, 2025 4:13:26 AM PST, Ingo Molnar <mingo@xxxxxxxxxx> wrote:
>
>* Jiri Slaby <jirislaby@xxxxxxxxxx> wrote:
>
>> On 07. 03. 25, 12:38, Ingo Molnar wrote:
>> > 
>> > * Jiri Slaby <jirislaby@xxxxxxxxxx> wrote:
>> > 
>> > > On 06. 03. 25, 17:25, Kuan-Wei Chiu wrote:
>> > > > Change return type to bool for better clarity. Update the kernel doc
>> > > > comment accordingly, including fixing "@value" to "@val" and adjusting
>> > > > examples. Also mark the function with __attribute_const__ to allow
>> > > > potential compiler optimizations.
>> > > > 
>> > > > Co-developed-by: Yu-Chun Lin <eleanor15x@xxxxxxxxx>
>> > > > Signed-off-by: Yu-Chun Lin <eleanor15x@xxxxxxxxx>
>> > > > Signed-off-by: Kuan-Wei Chiu <visitorckw@xxxxxxxxx>
>> > > > ---
>> > > >    include/linux/bitops.h | 10 +++++-----
>> > > >    1 file changed, 5 insertions(+), 5 deletions(-)
>> > > > 
>> > > > diff --git a/include/linux/bitops.h b/include/linux/bitops.h
>> > > > index c1cb53cf2f0f..44e5765b8bec 100644
>> > > > --- a/include/linux/bitops.h
>> > > > +++ b/include/linux/bitops.h
>> > > > @@ -231,26 +231,26 @@ static inline int get_count_order_long(unsigned long l)
>> > > >    /**
>> > > >     * parity8 - get the parity of an u8 value
>> > > > - * @value: the value to be examined
>> > > > + * @val: the value to be examined
>> > > >     *
>> > > >     * Determine the parity of the u8 argument.
>> > > >     *
>> > > >     * Returns:
>> > > > - * 0 for even parity, 1 for odd parity
>> > > > + * false for even parity, true for odd parity
>> > > 
>> > > This occurs somehow inverted to me. When something is in parity means that
>> > > it has equal number of 1s and 0s. I.e. return true for even distribution.
>> > > Dunno what others think? Or perhaps this should be dubbed odd_parity() when
>> > > bool is returned? Then you'd return true for odd.
>> > 
>> > OTOH:
>> > 
>> >   - '0' is an even number and is returned for even parity,
>> >   - '1' is an odd  number and is returned for odd  parity.
>> 
>> Yes, that used to make sense for me. For bool/true/false, it no longer does.
>> But as I wrote, it might be only me...
>
>No strong opinion on this from me either, I'd guess existing practice 
>with other parity functions should probably control. (If a coherent 
>praxis exists.).
>
>Thanks,
>
>	Ingo

Instead of "bool" think of it as "bit" and it makes more sense