Re: [PATCH 04/14] sched/deadline: Return EBUSY if dl_bw_cpus is zero

[Andrea Righi <arighi@nvidia.com>]

On Mon, Oct 20, 2025 at 03:38:12PM +0200, Andrea Righi wrote:
> On Mon, Oct 20, 2025 at 11:49:51AM +0200, Juri Lelli wrote:
> > Hi!
> > 
> > On 17/10/25 11:25, Andrea Righi wrote:
> > > From: Joel Fernandes <joelagnelf@nvidia.com>
> > > 
> > > Hotplugged CPUs coming online do an enqueue but are not a part of any
> > > root domain containing cpu_active() CPUs. So in this case, don't mess
> > > with accounting and we can retry later. Without this patch, we see
> > > crashes with sched_ext selftest's hotplug test due to divide by zero.
> > > 
> > > Signed-off-by: Joel Fernandes <joelagnelf@nvidia.com>
> > > ---
> > >  kernel/sched/deadline.c | 7 ++++++-
> > >  1 file changed, 6 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
> > > index 4aefb34a1d38b..f2f5b1aea8e2b 100644
> > > --- a/kernel/sched/deadline.c
> > > +++ b/kernel/sched/deadline.c
> > > @@ -1665,7 +1665,12 @@ int dl_server_apply_params(struct sched_dl_entity *dl_se, u64 runtime, u64 perio
> > >  	cpus = dl_bw_cpus(cpu);
> > >  	cap = dl_bw_capacity(cpu);
> > >  
> > > -	if (__dl_overflow(dl_b, cap, old_bw, new_bw))
> > > +	/*
> > > +	 * Hotplugged CPUs coming online do an enqueue but are not a part of any
> > > +	 * root domain containing cpu_active() CPUs. So in this case, don't mess
> > > +	 * with accounting and we can retry later.
> > 
> > Later when? It seems a little vague. :)
> 
> Yeah, this comment is actually incorrect, we're not "retrying later"
> anymore (we used to do that in a previous version), now the params are
> applied via:
> 
>   ext.c:handle_hotplug() -> dl_server_on() -> dl_server_apply_params()
> 
> Or via scx_enable() when an scx scheduler is loaded. So, I'm wondering if
> this condition is still needed. Will do some tests.

Looks like I can't reproduce the error with the hotplug kselftest anymore
(and it was happening pretty quickly).

Then I guess we can drop this patch or maybe add a WARN_ON_ONCE(!cpus) just
to safe?

Thanks,
-Andrea