Re: [RFC] [PATCH] cpuset: fix wrong calculation of relax domain level

[Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>]

Paul Jackson wrote:
> Question:
> 
>     In the case that the top_cpuset's sched_load_balance is -not- set,
>     is there code already present that sets the sched_relax_domain_level
>     in overlapping cpusets to the largest value in any of the overlapping
>     cpusets?
> 
>     If so, where is that code?
> 

My humble answer:

static void rebuild_sched_domains(void)
{
 :
	while (__kfifo_get(q, (void *)&cp, sizeof(cp))) {
	//  pick up cpusets with sched_load_balance = 1
	}
 :
restart:
	/* Find the best partition (set of sched domains) */
	for (i = 0; i < csn; i++) {
	// check overlap and set proper partition number
	}
 :
	for (nslot = 0, i = 0; i < csn; i++) {
 :
				if (apn == b->pn) {
					// make map and attr from all cpusets
					// having same partition number
					cpus_or(*dp, *dp, b->cpus_allowed);
					b->pn = -1;
					if (dattr)
						update_domain_attr(dattr
								   + nslot, b);
                                }
 :
        }
 :
rebuild:
 :
done:
 :
}

So the codes you searching is near by 'update_domain_attr' above, I guess.

Thanks,
H.Seto