Thread (16 messages) 16 messages, 3 authors, 2021-11-12

Re: [RFC PATCH v4 1/8] leds: add support for hardware driven LEDs

From: Randy Dunlap <hidden>
Date: 2021-11-11 02:24:11
Also in: linux-doc, linux-leds, lkml, netdev

On 11/10/21 5:34 PM, Ansuel Smith wrote:
quoted hunk ↗ jump to hunk
diff --git a/Documentation/leds/leds-class.rst b/Documentation/leds/leds-class.rst
index cd155ead8703..0175954717a3 100644
--- a/Documentation/leds/leds-class.rst
+++ b/Documentation/leds/leds-class.rst
@@ -169,6 +169,38 @@ Setting the brightness to zero with brightness_set() callback function
  should completely turn off the LED and cancel the previously programmed
  hardware blinking function, if any.
  
+Hardware driven LEDs
+===================================
+
+Some LEDs can be driven by hardware (for example a LED connected to
                                                     an LED
+an ethernet PHY or an ethernet switch can be configured to blink on activity on
+the network, which in software is done by the netdev trigger).
+
+To do such offloading, LED driver must support this and a supported trigger must
+be used.
+
+LED driver should declare the correct control mode supported and should set
+the LED_SOFTWARE_CONTROLLED or LED_HARDWARE_CONTROLLED bit in the flags
+parameter.
+The trigger will check these bit and fail to activate if the control mode
                                 bits
+is not supported. By default if a LED driver doesn't declare a control mode,
+bit LED_SOFTWARE_CONTROLLED is assumed and set by default.
drop the second:                                  by default
+
+The LED must implement 3 main API:
                                  APIs:
+- hw_control_status(): This asks the LED driver if hardware mode is enabled
+    or not.
+- hw_control_start(): This will simply enable the hardware mode for the LED
+    and the LED driver should reset any active blink_mode.
+- hw_control_stop(): This will simply disable the hardware mode for the LED.
+    It's advised to the driver to put the LED in the old state but this is not
+    enforcerd and putting the LED off is also accepted.
        enforced
+
+If LED_HARDWARE_CONTROLLED bit is the only contro mode set (LED_SOFTWARE_CONTROLLED
                                               control
+not set) set hw_control_status/start/stop is optional as the LED supports only
+hardware mode and any software only trigger will reject activation.
+
+On init a LED driver that support a hardware mode should reset every blink mode
            an LED
quoted hunk ↗ jump to hunk
+set by default.
  
  Known Issues
  ============
diff --git a/drivers/leds/Kconfig b/drivers/leds/Kconfig
index ed800f5da7d8..bd2b19cc77ec 100644
--- a/drivers/leds/Kconfig
+++ b/drivers/leds/Kconfig
@@ -49,6 +49,17 @@ config LEDS_BRIGHTNESS_HW_CHANGED
  
  	  See Documentation/ABI/testing/sysfs-class-led for details.
  
+config LEDS_HARDWARE_CONTROL
+	bool "LED Hardware Control support"
+	help
+	  This option enabled Hardware control support used by leds that
	              enables                                  LEDs
+	  can be driven in hardware by using supported triggers.
+
+	  Hardware blink modes will be exposed by sysfs class in
+	  /sys/class/leds based on the trigger currently active.
+
+	  If unsure, say Y.

-- 
~Randy
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help