Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 4 Mar 2019 18:40:28 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     Quentin Perret <quentin.perret@arm.com>
Cc:     =?utf-8?B?V2FuZywgVmluY2VudCAo546L5LqJKQ==?= 
        <Vincent.Wang@unisoc.com>,
        =?utf-8?B?WmhhbmcsIENodW55YW4gKOW8oOaYpeiJsyk=?= 
        <Chunyan.Zhang@unisoc.com>, Ingo Molnar <mingo@redhat.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Chunyan Zhang <zhang.lyra@gmail.com>,
        "Rafael J. Wysocki" <rjw@rjwysocki.net>
Subject: Re: =?utf-8?B?562U5aSNOiBbUEFUQ0ggVjRdIHNj?=
 =?utf-8?Q?hed=2Fcpufreq=3A_initializ?= =?utf-8?Q?e?= iowait_boost_max and
 iowait_boost with cpu capacity
Message-ID: <20190304174028.GO32494@hirez.programming.kicks-ass.net>
References: <1550831866-32749-1-git-send-email-chunyan.zhang@unisoc.com>
 <20190222105957.wxhlcmoag5f3i4fi@queper01-lin>
 <9099990618e242e1bab77ce3f9d9b1e3@BJMBX02.spreadtrum.com>
 <20190304135810.rq2ojnbn5vezrab3@queper01-lin>
 <20190304152616.GM32494@hirez.programming.kicks-ass.net>
 <20190304164816.4fnxxesjwzdoqria@queper01-lin>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20190304164816.4fnxxesjwzdoqria@queper01-lin>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Mar 04, 2019 at 04:48:16PM +0000, Quentin Perret wrote:

> > @@ -837,7 +818,9 @@ static int sugov_start(struct cpufreq_policy *policy)
> >  		memset(sg_cpu, 0, sizeof(*sg_cpu));
> >  		sg_cpu->cpu			= cpu;
> >  		sg_cpu->sg_policy		= sg_policy;
> > -		sg_cpu->iowait_boost_max	= policy->cpuinfo.max_freq;
> > +		sg_cpu->min			=
> > +			(SCHED_CAPACITY_SCALE * policy->cpuinfo.min_freq) /
> > +			policy->cpuinfo.max_freq;
> 
> The 'funny' thing is that on big little this 'min' can end up being
> larger than 'max' ...
> 
> On juno r0 for example, min_freq and max_freq for little CPUs are
> respectively 450MHz and 850MHz. So you get sg_cpu->min=542, but
> sg_cpu->max=446 ... So you'll max out after the first IO wakeup.
> And since iowait_boost is reset whenever it is smaller than sg_cpu->min,
> you end up with something that can either force max freq or apply no
> boost at all ...
> 
> Perhaps you could keep the 'util' and 'max' pointers in
> sugov_iowait_apply() and overwrite them like before, but in the
> SCHED_CAPACITY_SCALE scale as you suggest ?

Urgh; but then we're back to having that boostrap problem.

Now; at this time; @max is in fact scale_cpu_capacity, so can't we
change this:

-       /*
-        * Apply the current boost value: a CPU is boosted only if its current
-        * utilization is smaller then the current IO boost level.
-        */
-       boost_util = sg_cpu->iowait_boost;
-       boost_max = sg_cpu->iowait_boost_max;
-       if (*util * boost_max < *max * boost_util) {
-               *util = boost_util;
-               *max = boost_max;
-       }
+       sg_cpu->iowait_boost_pending = false;
+
+       return min(max(util, sg_cpu->iowait_boost), max);
}

to something like:

	/*
	 * @util is already in capacity scale, convert iowait_boost
	 * into the same scale so we can compare.
	 */
	boost = (sg_cpu->iowait_boost * max) >> SCHED_CAPACITY_SHIFT;
	util = max(boost, util);
	return min(util, max);