On Mon, Mar 06, 2023 at 05:44:41PM +0100, Jason A. Donenfeld wrote:
> On Mon, Mar 6, 2023 at 5:39 PM Yury Norov <yury.norov@xxxxxxxxx> wrote:
> >
> > On Tue, Mar 07, 2023 at 12:06:51AM +0800, Vernon Yang wrote:
> > > After commit 596ff4a09b89 ("cpumask: re-introduce constant-sized cpumask
> > > optimizations"), the cpumask size is divided into three different case,
> > > so fix comment of cpumask_xxx correctly.
> > >
> > > Signed-off-by: Vernon Yang <vernon2gm@xxxxxxxxx>
> > > ---
> > > include/linux/cpumask.h | 46 ++++++++++++++++++++---------------------
> > > 1 file changed, 23 insertions(+), 23 deletions(-)
> > >
> > > diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
> > > index 8fbe76607965..248bdb1c50dc 100644
> > > --- a/include/linux/cpumask.h
> > > +++ b/include/linux/cpumask.h
> > > @@ -155,7 +155,7 @@ static __always_inline unsigned int cpumask_check(unsigned int cpu)
> > > * cpumask_first - get the first cpu in a cpumask
> > > * @srcp: the cpumask pointer
> > > *
> > > - * Returns >= nr_cpu_ids if no cpus set.
> > > + * Returns >= small_cpumask_bits if no cpus set.
> >
> > There's no such thing like small_cpumask_bits. Here and everywhere,
> > nr_cpu_ids must be used.
> >
> > Actually, before 596ff4a09b89 nr_cpumask_bits was deprecated, and it
> > must be like that for all users even now.
> >
> > nr_cpumask_bits must be considered as internal cpumask parameter and
> > never referenced outside of cpumask code.
>
> What's the right thing I should do, then, for wireguard's usage and
> for random.c's usage? It sounds like you object to this patchset, but
> if the problem is real, it sounds like I should at least fix the two
> cases I maintain. What's the right check?
Everywhere outside of cpumasks internals use (cpu < nr_cpu_ids) to
check if the cpu is in a valid range, like:
cpu = cpumask_first(cpus);
if (cpu >= nr_cpu_ids)
pr_err("There's no cpus");
