On Thu 2020-07-23 20:13:18, Marek Behún wrote:

Hi,

this is v2 of my RFC adding support for LEDs connected to Marvell PHYs.

The LED subsystem patches are not contained:
- the patch adding support for LED private triggers is already accepted
  in Pavel Machek's for-next tree.
  If you want to try this patch on top of net-next, please also apply
  https://git.kernel.org/pub/scm/linux/kernel/git/pavel/linux-leds.git/commit/?h=for-next&id=93690cdf3060c61dfce813121d0bfc055e7fa30d
- the other led-trigger patch is not needed in this version of the RFC

The main difference from v1 is that only one trigger, named
"hw-control", is added to the /sys/class/leds/<LED>/trigger file.

When this trigger is activated, another file called "hw_control" is
created in the /sys/class/leds/<LED>/ directory. This file lists
available HW control modes for this LED in the same way the trigger
file does for triggers.

Example:

  # cd /sys/class/leds/eth0\:green\:link/
  # cat trigger
  [none] hw-control timer oneshot heartbeat ...
  # echo hw-control >trigger

Make this "hw-phy-mode" or something. hw-control is a bit too generic.

  # cat trigger
  none [hw-control] timer oneshot heartbeat ...
  # cat hw_control
  link/nolink link/act/nolink 1000/100/10/nolink act/noact
  blink-act/noact transmit/notransmit copperlink/else [1000/else]
  force-hi-z force-blink
  # echo 1000/100/10/nolink >hw_control
  # cat hw_control
  link/nolink link/act/nolink [1000/100/10/nolink] act/noact
  blink-act/noact transmit/notransmit copperlink/else 1000/else
  force-hi-z force-blink

The benefit here is that only one trigger is registered via LED API.
I guess there are other PHY drivers which too support HW controlled
blinking modes. So of this way of controlling PHY LED HW controlled
modes is accepted, the code creating the hw-control trigger and
hw_control file should be made into library code so that it can be
reused.

What do you think?

So.. you have 10 of them right now. I guess both hw_control and making
it into the trigger directly is acceptable here.

In future, would you expect having software "1000/100/10/nolink"
triggers I could activate on my scrollock LED (or on GPIO controlled
LEDs) to indicate network activity?

Best regards,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html