Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 2 Oct 2020 17:14:16 +0200
From:   Oleg Nesterov <oleg@redhat.com>
To:     Jens Axboe <axboe@kernel.dk>
Cc:     linux-kernel@vger.kernel.org, io-uring@vger.kernel.org,
        peterz@infradead.org, tglx@linutronix.de
Subject: Re: [PATCH 3/3] task_work: use TIF_TASKWORK if available
Message-ID: <20201002151415.GA29066@redhat.com>
References: <20201001194208.1153522-1-axboe@kernel.dk>
 <20201001194208.1153522-4-axboe@kernel.dk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20201001194208.1153522-4-axboe@kernel.dk>
User-Agent: Mutt/1.5.24 (2015-08-30)
Precedence: bulk

Heh. To be honest I don't really like 1-2 ;)

Unfortunately, I do not see a better approach right now. Let me think
until Monday, it is not that I think I will find a better solution, but
I'd like to try anyway.

Let me comment 3/3 for now.

On 10/01, Jens Axboe wrote:
>
> +static void task_work_signal(struct task_struct *task)
> +{
> +#ifndef TIF_TASKWORK
> +	unsigned long flags;
> +
> +	/*
> +	 * Only grab the sighand lock if we don't already have some
> +	 * task_work pending. This pairs with the smp_store_mb()
> +	 * in get_signal(), see comment there.
> +	 */
> +	if (!(READ_ONCE(task->jobctl) & JOBCTL_TASK_WORK) &&
> +	    lock_task_sighand(task, &flags)) {
> +		task->jobctl |= JOBCTL_TASK_WORK;
> +		signal_wake_up(task, 0);
> +		unlock_task_sighand(task, &flags);
> +	}
> +#else
> +	set_tsk_thread_flag(task, TIF_TASKWORK);
> +	set_notify_resume(task);
> +#endif

Again, I can't understand. task_work_signal(task) should set TIF_TASKWORK
to make signal_pending() = T _and_ wake/kick the target up, just like
signal_wake_up() does. Why do we set TIF_NOTIFY_RESUME ?

So I think that if we are going to add TIF_TASKWORK we should generalize
this logic and turn it into TIF_NOTIFY_SIGNAL. Similar to TIF_NOTIFY_RESUME
but implies signal_pending().

IOW, something like

	void set_notify_signal(task)
	{
		if (!test_and_set_tsk_thread_flag(task, TIF_NOTIFY_SIGNAL)) {
			if (!wake_up_state(task, TASK_INTERRUPTIBLE))
				kick_process(t);
		}
	}

	// called by exit_to_user_mode_loop() if ti_work & _TIF_NOTIFY_SIGNAL
	void tracehook_notify_signal(regs)
	{
		clear_thread_flag(TIF_NOTIFY_SIGNAL);
		smp_mb__after_atomic();
		if (unlikely(current->task_works))
			task_work_run();
	}

This way task_work_run() doesn't need to clear TIF_NOTIFY_SIGNAL and it can
have more users.

What do you think?

Oleg.