[PATCHv3 0/2] Driver for TI tlc59116 16 Channel i2c LED driver

[R, Vignesh]

On 1/16/2015 9:25 PM, Andrew Lunn wrote:
> On Fri, Jan 16, 2015 at 04:52:12PM +0200, Tomi Valkeinen wrote:
>> Hi,
>>
>> On 15/01/15 01:15, Andrew Lunn wrote:
>>> This patchset is a driver for the TI tlc59116 16 Channel i2c LED
>>> driver. This driver is used on the Belkin WRT1900AC access point and
>>> the C code is derived from code Belkin contributed to OpenWRT.
>>> However it has been extensively re-written, and a device tree binding
>>> added to replace platform data.
>>
>> We have a TLC59108 on one of our boards, and with a quick glance it
>> looks about the same as TLC59116, except the amount of outputs. Vignesh
>> wrote a driver for it and was about to send it for review.
>>
>> However, Vignesh implemented it as a PWM driver. We use it for LCD
>> backlight (via pwm-backlight).
>>
>> I'm not very familiar with LED and PWM drivers, but doesn't implementing
>> this as a LED driver prevent us from using it as a backlight? Whereas it
>> looks like PWM can be used as a led via pwm-leds (I think).
>>
> Hi Tomi
> 
> I've no idea about backlighting....
> 
> But the driver does fully implement brightness. So on my board:
> 
> echo 128 > /sys/class/leds/wrt1900ac\:white\:wan/brightness
> 
> will set that LED to 128/256 brightness. You have 255 steps of
> brightness.
> 
> The reason i did not model it as a PWM, is that it cannot fulfil the
> PWM interface. The configure interface is:
> 
> int pwm_config(struct pwm_device *pwm, int duty_ns, int period_ns);
> 
> However with this device, you have no control over the period. It is
> fixed, from a 97-kHz clock. All you can change is the duty as a ratio
> between 0 and 256. There is the group PWM, where you do have control
> over the period. But as the name implies, this PWM is shared for all
> outputs, which is not what the PWM interface expects, it wants to be
> able to control them individually. Also, it only has a range of 24 Hz
> to 10.73 s. Again, i have no idea about backlights, but i suspect if
> it is driven at 24Hz, i'm going to get a headache. 
> 
> The LED interface can be fully implemented. So that is what i have
> done.

I understand that period of TLC chip cannot be changed and hence cannot
fully implement PWM interface. But, suppose I want to control brightness
of an LCD screen, with your current design, my LCD driver can never can do
something like:
	pwm_get(chip);
	pwm_update_brightness();
	pwm_remove();
I will always have to rely on sysfs entries to control brightness.
 
If the driver is implemented as pwm-backlight, then brightness can
be controlled both via sysfs entries and pwm-core callbacks(in kernel).

> 
> So i think modeling this as an LED driver is correct, and maybe you
> need to implemented an led_bl.c driver?
> 

But, as far as I understand, leds-pwm.c does almost similar function.  

Regards
Vignesh R