Re: [PATCH v3 3/4] lib/find_bit: optimize find_next_bit() functions

[Yury Norov <yury.norov@gmail.com>]

On Fri, Sep 09, 2022 at 02:24:31PM +0200, Sven Schnelle wrote:
> Hi Yury,
> 
> Yury Norov <yury.norov@gmail.com> writes:
> 
> > On Wed, Sep 07, 2022 at 05:27:08PM +0100, Valentin Schneider wrote:
> >> On 27/08/22 10:58, Yury Norov wrote:
> >> > +#define FIND_NEXT_BIT(FETCH, MUNGE, size, start)				\
> >> > +({										\
> >> > +	unsigned long mask, idx, tmp, sz = (size), __start = (start);		\
> >> > +										\
> >> > +	if (unlikely(__start >= sz))						\
> >> > +		goto out;							\
> >> > +										\
> >> > +	mask = MUNGE(BITMAP_FIRST_WORD_MASK(__start));				\
> >> > +	idx = __start / BITS_PER_LONG;						\
> >> > +										\
> >> > +	for (tmp = (FETCH) & mask; !tmp; tmp = (FETCH)) {			\
> >> > +		if (idx > sz / BITS_PER_LONG)					\
> >> 
> >> Does that want to be
> >
> > Yes, I already fixed this.
> >  
> >>                 if (idx + 1 >= sz / BITS_PER_LONG)
> >> 
> >> ?
> 
> Did you push that already? We're still seeing crashes in CI, and the
> 'idx + 1' doesnt seem to be in next-20220908. Adding it makes the
> out-of-bound access go away, but the kernel will crash later in the
> block mq code:

Hi Swen,

I removed the whole series and will resend it with an appropriate fixes
at the weekend. Hopefully it will disappear in next-20220909 or 10.

Thanks,
Yury