Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <e72d9b4a-a986-3b73-d4e5-63df40a0bee1@bytedance.com>
Date:   Tue, 28 Jun 2022 11:39:38 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:91.0)
 Gecko/20100101 Thunderbird/91.10.0
Subject: Re: [PATCH v4 1/7] sched/fair: default to false in test_idle_cores
Content-Language: en-US
To:     Josh Don <joshdon@google.com>
Cc:     Peter Zijlstra <peterz@infradead.org>,
        Mel Gorman <mgorman@suse.de>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Chen Yu <yu.c.chen@intel.com>,
        Tim Chen <tim.c.chen@linux.intel.com>,
        K Prateek Nayak <kprateek.nayak@amd.com>,
        "Gautham R . Shenoy" <gautham.shenoy@amd.com>,
        linux-kernel <linux-kernel@vger.kernel.org>
References: <20220619120451.95251-1-wuyun.abel@bytedance.com>
 <20220619120451.95251-2-wuyun.abel@bytedance.com>
 <CABk29NtfVjXQZORGB1owmBS6C9LZJn8ci15gv4nQm+2DNLEPBA@mail.gmail.com>
From:   Abel Wu <wuyun.abel@bytedance.com>
In-Reply-To: <CABk29NtfVjXQZORGB1owmBS6C9LZJn8ci15gv4nQm+2DNLEPBA@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi Josh, thanks for your comments!

On 6/28/22 6:53 AM, Josh Don Wrote:
> On Sun, Jun 19, 2022 at 5:05 AM Abel Wu <wuyun.abel@bytedance.com> wrote:
>>
>> It's uncertain whether idle cores exist or not if shared sched-domains
>> are not ready, so returning "no idle cores" usually makes sense.
>>
>> While __update_idle_core() is an exception, it checks status of this
>> core and set to shared sched-domain if necessary. So the whole logic
>> depends on the existence of shared domain, and can bail out early if
>> it isn't available. Modern compilers seems capable of handling such
>> cases, so remove the tricky self-defined default return value.
> 
> I don't think the compiler will be able to bail out of the smt
> iteration early, since it'll have to do another rcu dereference for
> the sd_llc in set(). But I also don't think this case needs
> optimization, since it should be transient while the domain isn't
> ready.
> 
> Reviewed-by: Josh Don <joshdon@google.com>

Obviously I failed to comprehend the difference between the changed
assembly code, my bad..

Thanks,
Abel