Re: [RFC PATCH 1/5] lib: Implement find_{first,next,nth}_notandnot_bit, find_first_andnot_bit

[Yury Norov]

On Tue, Aug 20, 2024 at 10:45:17PM +0200, Mathieu Desnoyers wrote:
> On 2024-08-20 19:19, Mathieu Desnoyers wrote:
> > On 2024-08-19 21:19, Yury Norov wrote:
> [...]
> > > > +/**
> > > > + * find_next_notandnot_bit - find the next bit cleared in both
> > > > *addr1 and *addr2
> > > > + * @addr1: The first address to base the search on
> > > > + * @addr2: The second address to base the search on
> > > > + * @size: The bitmap size in bits
> > > > + * @offset: The bitnumber to start searching at
> > > > + *
> > > > + * Returns the bit number for the next bit cleared in both
> > > > *addr1 and *addr2.
> > > > + * If no such bits are found, returns @size.
> > > > + */
> > > > +static inline
> > > > +unsigned long find_next_notandnot_bit(const unsigned long *addr1,
> > > > +        const unsigned long *addr2, unsigned long size,
> > > > +        unsigned long offset)
> > > > +{
> > > > +    if (small_const_nbits(size)) {
> > > > +        unsigned long val;
> > > > +
> > > > +        if (unlikely(offset >= size))
> > > > +            return size;
> > > > +
> > > > +        val = (~*addr1) & (~*addr2) & GENMASK(size - 1, offset);
> > > > +        return val ? __ffs(val) : size;
> > > > +    }
> > > > +
> > > > +    return _find_next_notandnot_bit(addr1, addr2, size, offset);
> > > > +}
> > > > +#endif
> > > > +
> > > 
> > > It's not said explicitly, but some naming conventions exist around bitmap
> > > searching.
> > > 
> > > If you're looking for a clear (unset) bit in a mask, you'd use a 'zero'
> > > modifier. We have only 2 such functions now: find_{first,next}_zero_bit,
> > > both taking one bitmap. I think it's time to extend this rule for
> > > many bitmaps and write down the naming rules.
> > > 
> > > With the following, the find_next_notandnot_bit() should be named
> > > like; find_next_zero_and_bit(). It's not perfect, but still sounds
> > > better to me than 'notandnot' thing.
> 
> Actually, now that I come to think of it in terms of logic gates:
> 
> ~A & ~B == ~(A | B)
> 
> So this "notandnot" is simply a "NOR" gate.
> 
> I therefore intend to name it "find_next_nor_bit" if that's OK with
> you.

Yes, I'm OK.

To me, if you can put definition of a logical operation inside
FIND_NEXT_BIT() macro directly, you can name it correspondingly.
So in this case, find_next_nor_bit would be a:

  FIND_NEXT_BIT(~(addr1[idx] | addr2[idx]), /* nop */, size)

Correspondingly, instead of 'zero_or' we should use a 'nand' notation,
if it will be needed. I'll notice that in the naming manual.

Good catch.

Thanks,
Yury