Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
References: <20200311181601.18314-1-valentin.schneider@arm.com> <20200311181601.18314-9-valentin.schneider@arm.com> <c0a1f683-f5c5-cd95-a586-28cc1da9fc3d@arm.com>
User-agent: mu4e 0.9.17; emacs 26.3
From:   Valentin Schneider <valentin.schneider@arm.com>
To:     Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc:     linux-kernel@vger.kernel.org, mingo@kernel.org,
        peterz@infradead.org, vincent.guittot@linaro.org
Subject: Re: [PATCH v2 8/9] sched/fair: Split select_task_rq_fair want_affine logic
In-reply-to: <c0a1f683-f5c5-cd95-a586-28cc1da9fc3d@arm.com>
Date:   Thu, 19 Mar 2020 12:06:48 +0000
Message-ID: <jhjsgi4o8fb.mognet@arm.com>
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


On Thu, Mar 19 2020, Dietmar Eggemann wrote:
>> +	 */
>>      for_each_domain(cpu, tmp) {
>> -		/*
>> -		 * If both 'cpu' and 'prev_cpu' are part of this domain,
>> -		 * cpu is a valid SD_WAKE_AFFINE target.
>> -		 */
>> -		if (want_affine && (tmp->flags & SD_WAKE_AFFINE) &&
>> +		if ((tmp->flags & SD_WAKE_AFFINE) &&
>>                  cpumask_test_cpu(prev_cpu, sched_domain_span(tmp))) {
>>                      if (cpu != prev_cpu)
>>                              new_cpu = wake_affine(tmp, p, cpu, prev_cpu, sync);
>>
>> -			sd = NULL; /* Prefer wake_affine over balance flags */
>> +			/* Prefer wake_affine over SD lookup */
>
> I assume that 'balance flags' stands for (wakeup) load balance, i.e.
> find_idlest_xxx() path. So why change it?
>
>

You mean the comment part, right? I was hoping to clarify it a bit - if
we go through the want_affine condition, we'll override whatever SD we
picked with the highest_flag_domain() lookup (and the cached version in
9/9). Hence me referring to the SD lookup there.

> [...]