Received-SPF: pass (google.com: domain of linux-kernel+bounces-36610-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Message-ID: <f40f9aa9-e722-0264-2d4d-bc9062472f12@linux.dev>
Date: Wed, 24 Jan 2024 16:08:05 +0800
Precedence: bulk
MIME-Version: 1.0
Subject: Re: [PATCH v2 1/2] sched/fair: Return NULL when entity isn't a task
 in task_of()
Content-Language: en-US
To: Steven Rostedt <rostedt@goodmis.org>
Cc: mingo@redhat.com, peterz@infradead.org, juri.lelli@redhat.com,
 vincent.guittot@linaro.org, dietmar.eggemann@arm.com, bsegall@google.com,
 mgorman@suse.de, bristot@redhat.com, vschneid@redhat.com,
 linux-kernel@vger.kernel.org
References: <20231206063400.3206346-1-yajun.deng@linux.dev>
 <20231206063400.3206346-2-yajun.deng@linux.dev>
 <20240122224852.6ab9b498@gandalf.local.home>
From: Yajun Deng <yajun.deng@linux.dev>
In-Reply-To: <20240122224852.6ab9b498@gandalf.local.home>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit


On 2024/1/23 11:48, Steven Rostedt wrote:
> On Wed,  6 Dec 2023 14:33:59 +0800
> Yajun Deng <yajun.deng@linux.dev> wrote:
>
>> Before calling task_of(), we need to make sure that the entity is a task.
>> There is also a warning in task_of() if the entity isn't a task. That
>> means we need to check the entity twice. If the entity isn't a task,
> Does it really check it twice? Have you disassembled it to see if the code
> is any better?
>
> #define entity_is_task(se)	(!se->my_q)
> static inline struct task_struct *task_of(struct sched_entity *se)
> {
> 	SCHED_WARN_ON(!entity_is_task(se));
> 	return container_of(se, struct task_struct, se);
> }
>
> The above is a macro and a static inline, which means that the compiler
> should optimized out that second check.


Yes, the second check should be optimized.

>
>> return the task struct is meaningless.
>>
>> Return NULL when entity isn't a task in task_of(), and call task_of()
>> instead of entity_is_task() when we need a task_struct.
> I'm not against the change, as it could be considered a clean up. But it is
> up to the sched maintainers to decide if it's worth the churn.


Return NULL in task_of() makes the code cleaner.

>
> -- Steve
>
>
>> Signed-off-by: Yajun Deng <yajun.deng@linux.dev>