Re: [PATCH 06/14] sched_ext: Add a DL server for sched_ext tasks

[Andrea Righi <arighi@nvidia.com>]

On Fri, Oct 17, 2025 at 09:04:23AM -1000, Tejun Heo wrote:
> On Fri, Oct 17, 2025 at 08:58:35PM +0200, Andrea Righi wrote:
> > On Fri, Oct 17, 2025 at 05:47:45AM -1000, Tejun Heo wrote:
> > > On Fri, Oct 17, 2025 at 11:25:53AM +0200, Andrea Righi wrote:
> > > > +static struct task_struct *
> > > > +ext_server_pick_task(struct sched_dl_entity *dl_se, struct rq_flags *rf)
> > > > +{
> > > > +	return pick_task_scx(dl_se->rq, rf);
> > > > +}
> > > 
> > > I wonder whether we should tell pick_task_scx() to suppress the
> > > rq_modified_above() test in this case as a fair or RT task being enqueued
> > > has no reason to restart the picking process. While it will behave fine on
> > > retry, it's probably useful to be explicit here.
> > 
> > Yeah, that's a valid point. Maybe we can add a new flag to rq->scx.flags?
> > Something like SCX_RQ_DL_SERVER_PICK?
> 
> We can factor out the internals of pick_task_scx() into a separate function
> and add a flag there?

Much better, I like that. Ok, I'll incorporate this change and send a new
version.

Thanks,
-Andrea