Re: [PATCH v2 0/3] leds: blink resolution improvements

[Jacek Anaszewski]

On 04/29/2015 01:26 PM, Stas Sergeev wrote:
> 28.04.2015 15:58, Jacek Anaszewski ÐÐÑÐÑ:
> > On 04/28/2015 12:12 PM, Stas Sergeev wrote:
> > > 28.04.2015 11:57, Jacek Anaszewski ÐÐÑÐÑ:
> > > > Hi Stas,
> > > >
> > > > Have you tested it?
> > > Of course I did.
> > > Works with gpio driver and provides up to 10usec precision on
> > > armada-xp board.
> > > This is 1000 times better than without my patch - the precision
> > > was 10ms (jiffy).
> >
> > Please take into account that this could work reliably only for gpio
> > LEDs. For the LEDs driven though a bus (e.g. I2C) delays below 1ms
> > might be hard to achieve. The minimum available delay would vary from
> > driver to driver.
> >
> > We could think of adding the hr_timer mode to the led-class.
> > The mode could be turned on with use of a new led_set_high_res_timer
> > API. The API would be called by drivers/leds/leds-gpio.c driver when
> > a dedicated sysfs attribute was set adequately.
> > The other drivers could also set this mode if they controlled device
> > with a suitable LED switching rate. The minimum delay value could
> > be made configurable by the driver and readable through sysfs
> > when in hr_timer mode.
> I've found the SET_BRIGHTNESS_ASYNC and SET_BRIGHTNESS_SYNC flags.
> Sounds interesting for my patch, but the only "documentation" I was
> able to find, is this:
> ---
>          /* Setting a torch brightness needs to have immediate effect */
>          led_cdev->flags &= ~SET_BRIGHTNESS_ASYNC;
>          led_cdev->flags |= SET_BRIGHTNESS_SYNC;
> ---
> Aren't these flags mutually exclusive, and so just one could have
> been used?

Yes they are.

> Anyway, from that comment I can try to guess that if the driver
> supports ASYNC mode, it should be fast enough and without sleeps.
> The drivers that do i2c transfers with sleeps, should be using SYNC
> mode. Or was the intention for these flags entirely different?

Sync mode was designed for flash LED devices, for which turning
torch immediately has a big importance because they are often
synchronized with camera sensor.

Triggers use led_set_brightness_async API anyway.

> My intention is to use either a work-queue or a direct hrtimer
> callback, depending on whether the driver supports sync or async
> mode. This is instead of the driver being able to set the minimum
> delay - much simpler to implement. Makes sense?
> Can I use the above flags for that purpose, or will I need
> a new one?

These flags are intended for internal LED core use.


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