linux-kernel - Re: [PATCH v1 4/7] thermal: introduce the Power Allocator governor

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20150226175253.GA9883@e104805>
Date:	Thu, 26 Feb 2015 17:52:53 +0000
From:	Javi Merino <javi.merino@....com>
To:	Eduardo Valentin <edubezval@...il.com>
Cc:	"linux-pm@...r.kernel.org" <linux-pm@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Punit Agrawal <Punit.Agrawal@....com>,
	"broonie@...nel.org" <broonie@...nel.org>,
	Zhang Rui <rui.zhang@...el.com>
Subject: Re: [PATCH v1 4/7] thermal: introduce the Power Allocator governor

On Wed, Feb 25, 2015 at 07:00:38PM +0000, Eduardo Valentin wrote:
> On Wed, Feb 25, 2015 at 02:48:49PM +0000, Javi Merino wrote:
> > Hi Eduardo,
> > 
> > On Tue, Feb 24, 2015 at 06:21:26PM +0000, Eduardo Valentin wrote:
> > > On Wed, Jan 28, 2015 at 05:00:35PM +0000, Javi Merino wrote:
> > > > +
> > > > +k_d
> > > > +---
> > > > +
> > > > +`k_d` configures the PID loop's derivative term constant.  It's
> > > > +recommended to leave it as the default: 0.
> > > > +
> > > 
> > > I know we are considering K_d = 0. However, ...
> > > 
> > > <yet another big cut>
> > > 
> > > > +	/*
> > > > +	 * Calculate the derivative term
> > > > +	 *
> > > > +	 * We do err - prev_err, so with a positive k_d, a decreasing
> > > > +	 * error (i.e. driving closer to the line) results in less
> > > > +	 * power being applied, slowing down the controller)
> > > > +	 */
> > > > +	d = mul_frac(tz->tzp->k_d, err - params->prev_err);
> > > 
> > > 
> > > ... Shouldn't the above d component consider the rate of changes over time of the error?
> > > 
> > > I would expect you should do:
> > > d = k_d * (dE / dt)
> > > 
> > > or
> > > 
> > > d = K_d * ((err - params->prev_err) / sampling_period)
> > > 
> > > in plain C:
> > > 
> > > +	d = mul_frac(tz->tzp->k_d, err - params->prev_err);
> > > +	d /= tz->passive_polling; /* might require fixed point division */
> > 
> > Could do.  To be honest, both k_d and passive_polling are constants so
> 
> Yes, I agree that they are constants. But if you deploy the thermal zone
> with different sampling period on different devices, then the behavior
> will change.
> 
> > I don't think you get anything by doing this other than the added
> > complexity of the fixed point division.  As you said, the default k_d
> > is 0, so I'm not strongly against it.
> 
> OK. Then I would prefer to add the division, as it makes the code aligned
> to the concept.

Fair enough, I've added it to the code.  We will send a v2 later today
if all regressions pass.

Cheers,
Javi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/