Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753091AbaFDM46 (ORCPT ); Wed, 4 Jun 2014 08:56:58 -0400 Received: from mail-vc0-f179.google.com ([209.85.220.179]:63300 "EHLO mail-vc0-f179.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752746AbaFDM4y convert rfc822-to-8bit (ORCPT ); Wed, 4 Jun 2014 08:56:54 -0400 MIME-Version: 1.0 From: =?UTF-8?Q?Rapha=C3=ABl_Poggi?= Date: Wed, 4 Jun 2014 14:56:13 +0200 X-Google-Sender-Auth: y9CWeIUnUWytvvM5Bg81j4f3bvM Message-ID: Subject: pwm: atmel: problem when disable pwm To: linux-kernel@vger.kernel.org Cc: thierry.reding@gmail.com Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hello, I'm using the pwm-atmel driver to control led, but i'm facing a problem. When writing "echo 0 > /sys/class/leds/d1/brightness" (to switch off the led) sometimes the led stayed active. After investigation I think this problem is due to this function: static void atmel_pwm_disable(struct pwm_chip *chip, struct pwm_device *pwm) { struct atmel_pwm_chip *atmel_pwm = to_atmel_pwm_chip(chip); atmel_pwm_writel(atmel_pwm, PWM_DIS, 1 << pwm->hwpwm); clk_disable(atmel_pwm->clk); } By disable the pwm channel, I think the duty cycle "0" has not been update, and the channel is disable and stayed in active state (high or low depending of polarity). I have fixing the problem with this patch: pwm: atmel: improve disable of pwm Signed-off-by: Raphaël Poggi --- drivers/pwm/pwm-atmel.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/pwm/pwm-atmel.c b/drivers/pwm/pwm-atmel.c index 0adc952..a227c89 100644 --- a/drivers/pwm/pwm-atmel.c +++ b/drivers/pwm/pwm-atmel.c @@ -236,16 +236,17 @@ static int atmel_pwm_enable(struct pwm_chip *chip, struct pwm_device *pwm) return ret; } - atmel_pwm_writel(atmel_pwm, PWM_ENA, 1 << pwm->hwpwm); - return 0; } static void atmel_pwm_disable(struct pwm_chip *chip, struct pwm_device *pwm) { struct atmel_pwm_chip *atmel_pwm = to_atmel_pwm_chip(chip); + unsigned long clk_rate = clk_get_rate(atmel_pwm->clk); + unsigned long clk_period = 1000000000/clk_rate; + unsigned int pwm_period = atmel_pwm_ch_readl(atmel_pwm, pwm->hwpwm, PWMV2_CPRD); - atmel_pwm_writel(atmel_pwm, PWM_DIS, 1 << pwm->hwpwm); + ndelay(pwm_period * clk_period); clk_disable(atmel_pwm->clk); } With this patch instead of disable / enable pwm channel, we wait nsec depending of MCK rate and pwm period and then disable the clock. By doing this, we are sure that the duty cycle has been update and disable in good state. Is this seems right for you or my patch/fix is stupid ? Maybe my problem is not due to this function, if so do you have any idea of where it can come from ? Thanks, Raphaël -- 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/