Re: [PATCH v3] selftests/cgroup: Adjust cpu test duration based on HZ

[Joe Simmons-Talbott]

On Thu, Jun 25, 2026 at 10:23:22AM +0200, Michal Koutný wrote:
> Hi.
> 
> On Wed, Jun 24, 2026 at 12:03:57PM -0400, Joe Simmons-Talbott 
> <jo...@redhat.com> wrote:
> > +/*
> > + * Best effort attempt to get the kernel's HZ value from the config.
> > + * Return the HZ value if found otherwise return -1 to indicate failure.
> > + */
> > +static long
> > +_get_config_hz(void)
> 
> drop underscore from the static function
> 
> > +{
> > +   long hz = -1;
> 
> use the default 1000 here to simplify the callers
> 
> > +   FILE *f;
> > +   char cmd[256] = "zcat /proc/config.gz 2>/dev/null | grep '^CONFIG_HZ='";
> > +
> > +   f = popen(cmd, "r");
> > +
> > +   if (!f)
> > +           return hz;
> > +
> > +   if (fscanf(f, "CONFIG_HZ=%ld", &hz) == EOF)
> > +           goto out;
> > +
> > +out:
> > +   pclose(f);
> > +   return hz;
> > +}
> > +
> >  /*
> >   * This test creates a cgroup with some maximum value within a period, and
> >   * verifies that a process in the cgroup is not overscheduled.
> > @@ -646,15 +670,21 @@ test_cpucg_nested_weight_underprovisioned(const char 
> > *root)
> >  static int test_cpucg_max(const char *root)
> >  {
> >     int ret = KSFT_FAIL;
> > +   long hz = _get_config_hz();
> >     long quota_usec = 1000;
> >     long default_period_usec = 100000; /* cpu.max's default period */
> > -   long duration_seconds = 1;
> > +   long duration_seconds;
> >  
> > -   long duration_usec = duration_seconds * USEC_PER_SEC;
> > +   long duration_usec;
> >     long usage_usec, n_periods, remainder_usec, expected_usage_usec;
> >     char *cpucg;
> >     char quota_buf[32];
> >  
> > +   if (hz == -1)
> > +           hz = 1000;
> > +   duration_seconds = 1000 / hz;
> > +   duration_usec = duration_seconds * USEC_PER_SEC;
> 
> I'd do the calculation in usecs
> 
>       duration_usec = duration_seconds * USEC_PER_SEC * 1000 / hz;
> 
> so that actual duration is more precise (for hz=300 which is the only
> that doesn't divide 1000)
> 
> All in all, make the adjustments for HZ with less code (since I expect
> this will need adjustments for SMPs in future).

Hi Michal,

Thanks for your feedback.  I'll make the changes you have suggested in v4.

Thanks,
Joe