Date: Thu, 8 Nov 2012 11:33:53 +0100
From: Alban Bedel <alban.bedel@avionic-design.de>
To: Roland Stigge <stigge@antcom.de>
Cc: Thierry Reding <thierry.reding@avionic-design.de>,
        LKML <linux-kernel@vger.kernel.org>,
        Alexandre Pereira da Silva <aletes.xgr@gmail.com>
Subject: Re: [PATCH] pwm: lpc32xx - Fix the PWM polarity
Message-ID: <20121108113353.4fdbb25c@avionic-0020.adnet.avionic-design.de>
In-Reply-To: <509B80A7.1060700@antcom.de>
References: <20121106181939.61c50233@avionic-0020.adnet.avionic-design.de>
	<1352301955-15960-1-git-send-email-alban.bedel@avionic-design.de>
	<509B80A7.1060700@antcom.de>
Organization: Avionic Design
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1744
Lines: 65

On Thu, 08 Nov 2012 10:51:35 +0100
Roland Stigge <stigge@antcom.de> wrote:

> On 07/11/12 16:25, Alban Bedel wrote:
> > Signed-off-by: Alban Bedel <alban.bedel@avionic-design.de>
> > ---
> >  drivers/pwm/pwm-lpc32xx.c |    6 +++++-
> >  1 files changed, 5 insertions(+), 1 deletions(-)
> > 
> > diff --git a/drivers/pwm/pwm-lpc32xx.c b/drivers/pwm/pwm-lpc32xx.c
> > index adb87f0..0dc278d 100644
> > --- a/drivers/pwm/pwm-lpc32xx.c
> > +++ b/drivers/pwm/pwm-lpc32xx.c
> > @@ -51,7 +51,11 @@ static int lpc32xx_pwm_config(struct pwm_chip *chip, struct pwm_device *pwm,
> >  
> >  	c = 256 * duty_ns;
> >  	do_div(c, period_ns);
> > -	duty_cycles = c;
> > +	if (c == 0)
> > +		c = 256;
> > +	if (c > 255)
> > +		c = 255;
> > +	duty_cycles = 256 - c;
> 
> Except for the range check (for the original c > 255), this results in:
> 
> 	duty_cycles = 256 - c
> 
> except for (c == 0) where
> 
> 	duty_cycles = 1

No it lead to duty_cycles = 0

> which actually is
> 
> 	duty_cycles = (256 - c) - 255
> 
> (think with the original c)
> 
> i.e. nearly a polarity inversion in the case of (c == 0).
> 
> Why is the case (c == 0) so special here? Maybe you can document this,
> if it is really intended?

It is intended, the formular for duty value in the register is:

duty = (256 - 256*duty_ns/period_ns) % 256

But the code avoid the modulo by clamping '256*duty_ns/period_ns' to 1-256.

Perhaps something like:

if (c > 255)
	c = 255;
duty_cycles = (256 - c) % 256;

would be easier to understand.

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