Re: [PATCH v16 4/7] sched: Fix runtime accounting w/ split exec & sched contexts

[Juri Lelli <juri.lelli@redhat.com>]

On 16/04/25 16:30, John Stultz wrote:
> On Mon, Apr 14, 2025 at 2:28 AM Juri Lelli <juri.lelli@redhat.com> wrote:
> > On 11/04/25 23:02, John Stultz wrote:
> > > +static s64 update_se_times(struct rq *rq, struct sched_entity *se)
> > >  {
> > >       u64 now = rq_clock_task(rq);
> > >       s64 delta_exec;
> > >
> > > -     delta_exec = now - curr->exec_start;
> > > +     delta_exec = now - se->exec_start;
> > >       if (unlikely(delta_exec <= 0))
> > >               return delta_exec;
> > >
> > > -     curr->exec_start = now;
> > > -     curr->sum_exec_runtime += delta_exec;
> > > +     se->exec_start = now;
> > > +     if (entity_is_task(se)) {
> > > +             struct task_struct *running = rq->curr;
> > > +             /*
> > > +              * If se is a task, we account the time against the running
> > > +              * task, as w/ proxy-exec they may not be the same.
> > > +              */
> > > +             running->se.exec_start = now;
> > > +             running->se.sum_exec_runtime += delta_exec;
> > > +     } else {
> > > +             /* If not task, account the time against se */
> > > +             se->sum_exec_runtime += delta_exec;
> > > +     }
> >
> > ...
> >
> > > @@ -1213,7 +1224,7 @@ s64 update_curr_common(struct rq *rq)
> > >       struct task_struct *donor = rq->donor;
> > >       s64 delta_exec;
> > >
> > > -     delta_exec = update_curr_se(rq, &donor->se);
> > > +     delta_exec = update_se_times(rq, &donor->se);
> > >       if (likely(delta_exec > 0))
> > >               update_curr_task(donor, delta_exec);
> >
> > Considering that we calculate delta_exec in updated_se_times using
> > exec_start of the sched_entity passed as argument, is it correct to use
> > donor in the above?
> 
> Sorry, I'm not sure I quite understand your concern here.  Why are you
> thinking using donor might be problematic here?  We're passing the
> donor->se in to calculate the delta_exec.
> 
> I'll grant that "update_curr_common" maybe isn't the best name for the
> calling function anymore, as we're really only working on the donor
> here. Is that what your concern stems from?

Ah, I think I was just confused similarly to Peter. A comment explaining
this point might help (I believe you were going to add that based on
the discussion with Peter).

Thanks,
Juri