From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1751834AbdG0Tzl (ORCPT <rfc822;w@1wt.eu>);
        Thu, 27 Jul 2017 15:55:41 -0400
Received: from bombadil.infradead.org ([65.50.211.133]:34491 "EHLO
        bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751769AbdG0Tzj (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Thu, 27 Jul 2017 15:55:39 -0400
Date: Thu, 27 Jul 2017 21:55:31 +0200
From: Peter Zijlstra <peterz@infradead.org>
To: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: linux-kernel@vger.kernel.org,
        "Paul E . McKenney" <paulmck@linux.vnet.ibm.com>,
        Boqun Feng <boqun.feng@gmail.com>
Subject: Re: [RFC PATCH] membarrier: expedited private command
Message-ID: <20170727195531.GE28975@worktop>
References: <20170727185943.11570-1-mathieu.desnoyers@efficios.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170727185943.11570-1-mathieu.desnoyers@efficios.com>
User-Agent: Mutt/1.5.22.1 (2013-10-16)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Thu, Jul 27, 2017 at 02:59:43PM -0400, Mathieu Desnoyers wrote:
> diff --git a/kernel/membarrier.c b/kernel/membarrier.c
> index 9f9284f37f8d..8c6c0f96f617 100644
> --- a/kernel/membarrier.c
> +++ b/kernel/membarrier.c
> @@ -19,10 +19,81 @@
>  #include <linux/tick.h>
>  
>  /*
> + * XXX For cpu_rq(). Should we rather move
> + * membarrier_private_expedited() to sched/core.c or create
> + * sched/membarrier.c ?

The later perhaps.

> +static void membarrier_private_expedited(void)
> +{
> +	int cpu, this_cpu;
> +	cpumask_var_t tmpmask;
> +
> +	if (num_online_cpus() == 1)
> +		return;
> +
> +	/*
> +	 * Matches memory barriers around rq->curr modification in
> +	 * scheduler.
> +	 */
> +	smp_mb();	/* system call entry is not a mb. */
> +
> +	if (!alloc_cpumask_var(&tmpmask, GFP_NOWAIT)) {

Why GFP_NOWAIT ? and falback. There seems to be a desire to make this a
nonblocking syscall. Should we document this somewhere?

> +		/* Fallback for OOM. */
> +		membarrier_private_expedited_ipi_each();
> +		goto end;
> +	}
> +
> +	this_cpu = raw_smp_processor_id();

This is a tad dodgy, you might want to put in a comment on how migrating
this thread is ok.

> +	for_each_online_cpu(cpu) {

One would also need cpus_read_lock() if you rely on the online mask.

> +		struct task_struct *p;
> +
> +		if (cpu == this_cpu)
> +			continue;
> +		rcu_read_lock();
> +		p = task_rcu_dereference(&cpu_rq(cpu)->curr);
> +		if (p && p->mm == current->mm)
> +			__cpumask_set_cpu(cpu, tmpmask);
> +		rcu_read_unlock();
> +	}
> +	smp_call_function_many(tmpmask, ipi_mb, NULL, 1);
> +	free_cpumask_var(tmpmask);
> +end:
> +	/*
> +	* Memory barrier on the caller thread _after_ we finished
> +	* waiting for the last IPI. Matches memory barriers around
> +	* rq->curr modification in scheduler.
> +	*/
> +	smp_mb();	/* exit from system call is not a mb */
> +}

> @@ -2737,6 +2757,7 @@ context_switch(struct rq *rq, struct task_struct *prev,
>  
>  	mm = next->mm;
>  	oldmm = prev->active_mm;
> +	membarrier_expedited_mb_after_set_current(mm, oldmm);
>  	/*
>  	 * For paravirt, this is coupled with an exit in switch_to to
>  	 * combine the page table reload and the switch backend into

As said on IRC, we have finish_task_switch()->if (mm)
mmdrop(mm)->atomic_dec_and_test() providing a smp_mb(). We just need to
deal with the !mm case.