Re: [PATCH bpf-next v2 03/13] bpf: Add spis_*() helpers for 4-byte stack slot bitmasks

[Eduard Zingerman <eddyz87@gmail.com>]

On Fri, 2026-04-10 at 01:50 +0000, bot+bpf-ci@kernel.org wrote:
> > diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h
> > index 410939d7b325..124828c9a916 100644
> > --- a/include/linux/bpf_verifier.h
> > +++ b/include/linux/bpf_verifier.h
> 
> [ ... ]
> 
> > +static inline bool spis_test_bit(spis_t s, u32 slot)
> > +{
> > +	return s.v[slot / 64] & BIT_ULL(slot % 64);
> > +}
> > +
> > +static inline void spis_or_range(spis_t *mask, u32 lo, u32 hi)
> > +{
> > +	u32 w;
> > +
> > +	for (w = lo; w <= hi && w < STACK_SLOTS; w++)
> > +		mask->v[w / 64] |= BIT_ULL(w % 64);
> > +}
> > +
> > +static inline spis_t spis_one_bit(u32 slot)
> > +{
> > +	if (slot < 64)
> > +		return (spis_t){{ BIT(slot), 0 }};
> > +	else
> > +		return (spis_t){{ 0, BIT(slot - 64) }};
> > +}
> 
> Should BIT() here be BIT_ULL() to match spis_test_bit() and
> spis_or_range() above?
> 
> The bitmask is stored in u64 v[2], and BIT() expands to
> (1UL << (nr)) which is 32-bit wide on 32-bit architectures.
> For slot values >= 32, this would be undefined behavior.
> 
> spis_single_slot() passes values up to 127 through here via
> spis_one_bit(spi * 2 + 1), so on a 32-bit build both branches
> would hit shifts >= 32.
> 
> Note: this function is removed later in the series by commit
> 6314c1b64c52 ("bpf: simplify liveness to use (callsite, depth)
> keyed func_instances"), but it is used in the intermediate commit
> 3199040a2cb9 ("bpf: make liveness.c track stack with 4-byte
> granularity") via spis_single_slot().

Yeap, BIT_ULL is needed

> 
> ---
> AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
> See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
> 
> CI run summary: https://github.com/kernel-patches/bpf/actions/runs/24221779050