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 S1756266AbdEDSmT (ORCPT <rfc822;w@1wt.eu>);
        Thu, 4 May 2017 14:42:19 -0400
Received: from bombadil.infradead.org ([65.50.211.133]:39021 "EHLO
        bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1752118AbdEDSmL (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Thu, 4 May 2017 14:42:11 -0400
Date: Thu, 4 May 2017 20:42:00 +0200
From: Peter Zijlstra <peterz@infradead.org>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@kernel.org>,
        Clark Williams <williams@redhat.com>,
        Daniel Bristot de Oliveira <bristot@redhat.com>,
        John Kacur <jkacur@redhat.com>, Scott Wood <swood@redhat.com>
Subject: Re: [PATCH tip/sched/core v2] sched/rt: Simplify the IPI rt
 balancing logic
Message-ID: <20170504184200.3gjfciwqulgwpeln@hirez.programming.kicks-ass.net>
References: <20170424114732.1aac6dc4@gandalf.local.home>
 <20170504153256.fbmglqe2zjo6ika2@hirez.programming.kicks-ass.net>
 <20170504132538.157f8ca7@gandalf.local.home>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170504132538.157f8ca7@gandalf.local.home>
User-Agent: NeoMutt/20170113 (1.7.2)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Thu, May 04, 2017 at 01:25:38PM -0400, Steven Rostedt wrote:
> > I think you want to write that as:
> > 
> > 	struct root_domain *rd = rq->rd;
> > 	int cpu, next;
> > 
> > 	/* comment */
> > 	for (;;) { 
> > 		if (rd->rto_cpu >= nr_cpu_ids) {
> 
> If we go with your change, then this needs to be:
> 
> 		if (rd->rto_cpu < 0) {
> 
> > 			cpu = cpumask_first(rd->rto_mask);
> > 			rd->rto_cpu = cpu;
> > 			return cpu;
> > 		}

No you can leave it out entirely.

> > 
> > 		cpu = cpumask_next(rd->rto_mask);
> 
> cpumask_next() requires two parameters.

Indeed it does:

		cpu = cpumask_next(rd->rto_cpu, rd->rto_mask);

will be cpumask_first() when rto_cpu == -1, see for example
for_each_cpu().

> > > +static inline bool rto_start_trylock(atomic_t *v)
> > > +{
> > > +	return !atomic_cmpxchg(v, 0, 1);  
> > 
> > Arguably this could be: !atomic_cmpxchg_acquire(v, 0, 1);
> 
> Yes agreed. But if you remember, at the time I was basing this off of
> tip/sched/core, which didn't have atomic_cmpxchg_acquire() available.

No that's the try_cmpxchg stuff, the _acquire stuff is long in.

> Thanks for the review. I'll spin up a new patch. Unfortunately, I no
> longer have access to the behemoth machine. I'll only be testing this
> on 4 cores now, or 8 with HT.

I have something with 144 CPUs in or thereabout, if you have the
testcase handy I can give it a spin.