[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20211001143601.5f57eb1a@thinkpad>
Date: Fri, 1 Oct 2021 14:36:01 +0200
From: Marek BehĂșn <kabel@...nel.org>
To: Pavel Machek <pavel@....cz>,
Jacek Anaszewski <jacek.anaszewski@...il.com>,
Rob Herring <robh+dt@...nel.org>
Cc: Andrew Lunn <andrew@...n.ch>,
"linux-leds@...r.kernel.org" <linux-leds@...r.kernel.org>,
netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
devicetree@...r.kernel.org
Subject: lets settle the LED `function` property regarding the netdev
trigger
Hello Pavel, Jacek, Rob and others,
I'd like to settle DT binding for the LED function property regarding
the netdev LED trigger.
Currently we have, in include/dt-bindings/leds/common.h, the following
functions defined that could be interpreted as request to enable netdev
trigger on given LEDs:
activity
lan
rx tx
wan
wlan
The "activity" function was originally meant to imply the CPU
activity trigger, while "rx" and "tx" are AFAIK meant as UART indicators
(tty LED trigger), see
https://lore.kernel.org/linux-leds/20190609190803.14815-27-jacek.anaszewski@gmail.com/
The netdev trigger supports different settings:
- indicate link
- blink on rx, blink on tx, blink on both
The current scheme does not allow for implying these.
I therefore propose that when a LED has a network device handle in the
trigger-sources property, the "rx", "tx" and "activity" functions
should also imply netdev trigger (with the corresponding setting).
A "link" function should be added, also implying netdev trigger.
What about if a LED is meant by the device vendor to indicate both link
(on) and activity (blink)?
The function property is currently a string. This could be changed to
array of strings, and then we can have
function = "link", "activity";
Since the function property is also used for composing LED classdev
names, I think only the first member should be used for that.
This would allow for ethernet LEDs with names
ethphy-0:green:link
ethphy-0:yellow:activity
to be controlled by netdev trigger in a specific setting without the
need to set the trigger in /sys/class/leds.
Opinions?
Marek
Powered by blists - more mailing lists