linux-kernel - Re: [PATCH v2 2/4] pwm-backlight: add support for pwm-delay-us property

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]

Date:   Thu, 6 Jul 2017 10:17:18 +0100
From:   Daniel Thompson <daniel.thompson@...aro.org>
To:     Pavel Machek <pavel@....cz>,
        Thierry Reding <thierry.reding@...il.com>
Cc:     Enric Balletbo i Serra <enric.balletbo@...labora.com>,
        Lee Jones <lee.jones@...aro.org>,
        Jingoo Han <jingoohan1@...il.com>,
        Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
        Rob Herring <robh+dt@...nel.org>,
        Richard Purdie <rpurdie@...ys.net>,
        Jacek Anaszewski <jacek.anaszewski@...il.com>,
        Heiko Stuebner <heiko@...ech.de>, linux-pwm@...r.kernel.org,
        linux-fbdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        groeck@...omium.org, linux-rockchip@...ts.infradead.org,
        huang lin <hl@...k-chips.com>
Subject: Re: [PATCH v2 2/4] pwm-backlight: add support for pwm-delay-us
 property

On 06/07/17 10:12, Pavel Machek wrote:
> On Thu 2017-07-06 10:01:32, Thierry Reding wrote:
>> On Fri, Jun 30, 2017 at 01:21:07PM +0200, Enric Balletbo i Serra wrote:
>>> From: huang lin <hl@...k-chips.com>
>>>
>>> Some panels (i.e. N116BGE-L41), in their power sequence specifications,
>>> request a delay between set the PWM signal and enable the backlight and
>>> between clear the PWM signal and disable the backlight. Add support for
>>> the new pwm-delay-us property to meet the timing.
>>>
>>> Note that this patch inverts current sequence. Before this patch the
>>> enable signal was set before the PWM signal and vice-versa on power off.
>>>
>>> I assumed that this sequence was wrong, at least it is on different panel
>>> datasheets that I checked, so I inverted the sequence to follow:
>>>
>>>    On power on, set the PWM signal, wait, and set the LED_EN signal.
>>>    On power off, clear the LED_EN signal, wait, and stop the PWM signal.
>>
>> I think this should be two separate patches to make it easier to revert
>> the inverted sequence should it prove to regress on other panels.
> 
> Don't make this overly complex. This is trivial. No need to split it
> into more patches.

Agree. IMHO getting the code that reads the (optional) new parameter 
correct is the best way to manage risk of regression since in most cases 
the delay will be skipped anyway.


Daniel.


> 
> 
>> Two more comments below.
>>
>>> Signed-off-by: huang lin <hl@...k-chips.com>
>>> Signed-off-by: Enric Balletbo i Serra <enric.balletbo@...labora.com>
>>> ---
>>> Changes since v1:
>>>   - As suggested by Daniel Thompson
>>>     - Do not assume power-on delay and power-off delay will be the same
>>>   - Move the check of dt property to the parse dt function.
>>>
>>> v1: https://lkml.org/lkml/2017/6/28/219
>>>
>>>   drivers/video/backlight/pwm_bl.c | 24 ++++++++++++++++++++----
>>>   include/linux/pwm_backlight.h    |  1 +
>>>   2 files changed, 21 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/video/backlight/pwm_bl.c b/drivers/video/backlight/pwm_bl.c
>>> index 002f1ce..0f5470e 100644
>>> --- a/drivers/video/backlight/pwm_bl.c
>>> +++ b/drivers/video/backlight/pwm_bl.c
>>> @@ -10,6 +10,7 @@
>>>    * published by the Free Software Foundation.
>>>    */
>>>   
>>> +#include <linux/delay.h>
>>>   #include <linux/gpio/consumer.h>
>>>   #include <linux/gpio.h>
>>>   #include <linux/module.h>
>>> @@ -35,6 +36,7 @@ struct pwm_bl_data {
>>>   	struct gpio_desc	*enable_gpio;
>>>   	unsigned int		scale;
>>>   	bool			legacy;
>>> +	unsigned int		pwm_delay[2];
>>>   	int			(*notify)(struct device *,
>>>   					  int brightness);
>>>   	void			(*notify_after)(struct device *,
>>> @@ -54,10 +56,14 @@ static void pwm_backlight_power_on(struct pwm_bl_data *pb, int brightness)
>>>   	if (err < 0)
>>>   		dev_err(pb->dev, "failed to enable power supply\n");
>>>   
>>> +	pwm_enable(pb->pwm);
>>> +
>>> +	if (pb->pwm_delay[0])
>>> +		usleep_range(pb->pwm_delay[0], pb->pwm_delay[0] + 2000);
>>
>> 2000 us is kind of arbitrary. What if pwm_delay[0] is on the order of 20
>> us? Making the delay 2 ms longer (in the worst case) seems somewhat
>> excessive. Why not something like:
>>
>> 	usleep_range(pb->pwm_delay[0], pb->pwm_delay[0] * 2);
>>
>> ?
>>
>>> +
>>>   	if (pb->enable_gpio)
>>>   		gpiod_set_value_cansleep(pb->enable_gpio, 1);
>>>   
>>> -	pwm_enable(pb->pwm);
>>>   	pb->enabled = true;
>>>   }
>>>   
>>> @@ -66,12 +72,15 @@ static void pwm_backlight_power_off(struct pwm_bl_data *pb)
>>>   	if (!pb->enabled)
>>>   		return;
>>>   
>>> -	pwm_config(pb->pwm, 0, pb->period);
>>> -	pwm_disable(pb->pwm);
>>> -
>>>   	if (pb->enable_gpio)
>>>   		gpiod_set_value_cansleep(pb->enable_gpio, 0);
>>>   
>>> +	if (pb->pwm_delay[1])
>>> +		usleep_range(pb->pwm_delay[1], pb->pwm_delay[1] + 2000);
>>> +
>>> +	pwm_config(pb->pwm, 0, pb->period);
>>> +	pwm_disable(pb->pwm);
>>> +
>>>   	regulator_disable(pb->power_supply);
>>>   	pb->enabled = false;
>>>   }
>>> @@ -174,6 +183,12 @@ static int pwm_backlight_parse_dt(struct device *dev,
>>>   		data->max_brightness--;
>>>   	}
>>>   
>>> +	/* read pwm to enable pre/post delays from DT property */
>>
>> This comment is confusing. This isn't reading anything from the PWM.
>>
>> Thierry
> 
> 
>