Re: [PATCH] sched/core: expand sched_getaffinity(2) to return number of CPUs

From: Peter Zijlstra
Date: Fri Apr 05 2019 - 07:49:49 EST


On Fri, Apr 05, 2019 at 01:08:58PM +0200, Florian Weimer wrote:
> * Peter Zijlstra:
>
> > On Fri, Apr 05, 2019 at 12:16:39PM +0200, Florian Weimer wrote:
> >
> >> > True; but I suppose glibc already does lots of that anyway, right? It
> >> > does contain the right information.
> >>
> >> If I recall correctly my last investigation,
> >> /sys/devices/system/cpu/possible does not reflect the size of the
> >> affinity mask, either.
> >
> > Strictly speaking correct; the bitmap can be longer than the highest
> > possible cpu number, however the remainder would be 0-padding and could
> > thus be stripped without issue.
>
> Doesn't the kernel still enforce the larget bitmap in sched_getaffinity,
> even if the bits are always zero?

Oh crap, you're right. That's unfortunate I suppose.