Richard Purdie | 75c1d31 | 2006-03-31 02:31:03 -0800 | [diff] [blame] | 1 | LED handling under Linux |
| 2 | ======================== |
| 3 | |
| 4 | If you're reading this and thinking about keyboard leds, these are |
| 5 | handled by the input subsystem and the led class is *not* needed. |
| 6 | |
| 7 | In its simplest form, the LED class just allows control of LEDs from |
| 8 | userspace. LEDs appear in /sys/class/leds/. The brightness file will |
| 9 | set the brightness of the LED (taking a value 0-255). Most LEDs don't |
| 10 | have hardware brightness support so will just be turned on for non-zero |
| 11 | brightness settings. |
| 12 | |
| 13 | The class also introduces the optional concept of an LED trigger. A trigger |
| 14 | is a kernel based source of led events. Triggers can either be simple or |
| 15 | complex. A simple trigger isn't configurable and is designed to slot into |
| 16 | existing subsystems with minimal additional code. Examples are the ide-disk, |
| 17 | nand-disk and sharpsl-charge triggers. With led triggers disabled, the code |
| 18 | optimises away. |
| 19 | |
| 20 | Complex triggers whilst available to all LEDs have LED specific |
| 21 | parameters and work on a per LED basis. The timer trigger is an example. |
| 22 | |
| 23 | You can change triggers in a similar manner to the way an IO scheduler |
| 24 | is chosen (via /sys/class/leds/<device>/trigger). Trigger specific |
| 25 | parameters can appear in /sys/class/leds/<device> once a given trigger is |
| 26 | selected. |
| 27 | |
| 28 | |
| 29 | Design Philosophy |
| 30 | ================= |
| 31 | |
| 32 | The underlying design philosophy is simplicity. LEDs are simple devices |
| 33 | and the aim is to keep a small amount of code giving as much functionality |
| 34 | as possible. Please keep this in mind when suggesting enhancements. |
| 35 | |
| 36 | |
| 37 | LED Device Naming |
| 38 | ================= |
| 39 | |
| 40 | Is currently of the form: |
| 41 | |
Richard Purdie | 6c152be | 2007-10-31 15:00:07 +0100 | [diff] [blame] | 42 | "devicename:colour:function" |
Richard Purdie | 75c1d31 | 2006-03-31 02:31:03 -0800 | [diff] [blame] | 43 | |
| 44 | There have been calls for LED properties such as colour to be exported as |
| 45 | individual led class attributes. As a solution which doesn't incur as much |
| 46 | overhead, I suggest these become part of the device name. The naming scheme |
Richard Purdie | 6c152be | 2007-10-31 15:00:07 +0100 | [diff] [blame] | 47 | above leaves scope for further attributes should they be needed. If sections |
| 48 | of the name don't apply, just leave that section blank. |
Richard Purdie | 75c1d31 | 2006-03-31 02:31:03 -0800 | [diff] [blame] | 49 | |
| 50 | |
Márton Németh | 4c79141 | 2007-10-31 15:07:12 +0100 | [diff] [blame^] | 51 | Hardware accelerated blink of LEDs |
| 52 | ================================== |
| 53 | |
| 54 | Some LEDs can be programmed to blink without any CPU interaction. To |
| 55 | support this feature, a LED driver can optionally implement the |
| 56 | blink_set() function (see <linux/leds.h>). If implemeted, triggers can |
| 57 | attempt to use it before falling back to software timers. The blink_set() |
| 58 | function should return 0 if the blink setting is supported, or -EINVAL |
| 59 | otherwise, which means that LED blinking will be handled by software. |
| 60 | |
| 61 | The blink_set() function should choose a user friendly blinking |
| 62 | value if it is called with *delay_on==0 && *delay_off==0 parameters. In |
| 63 | this case the driver should give back the chosen value through delay_on |
| 64 | and delay_off parameters to the leds subsystem. |
| 65 | |
| 66 | Any call to the brightness_set() callback function should cancel the |
| 67 | previously programmed hardware blinking function so setting the brightness |
| 68 | to 0 can also cancel the blinking of the LED. |
| 69 | |
| 70 | |
Richard Purdie | 75c1d31 | 2006-03-31 02:31:03 -0800 | [diff] [blame] | 71 | Known Issues |
| 72 | ============ |
| 73 | |
| 74 | The LED Trigger core cannot be a module as the simple trigger functions |
| 75 | would cause nightmare dependency issues. I see this as a minor issue |
| 76 | compared to the benefits the simple trigger functionality brings. The |
| 77 | rest of the LED subsystem can be modular. |
| 78 | |
Richard Purdie | 75c1d31 | 2006-03-31 02:31:03 -0800 | [diff] [blame] | 79 | |
| 80 | Future Development |
| 81 | ================== |
| 82 | |
| 83 | At the moment, a trigger can't be created specifically for a single LED. |
| 84 | There are a number of cases where a trigger might only be mappable to a |
| 85 | particular LED (ACPI?). The addition of triggers provided by the LED driver |
| 86 | should cover this option and be possible to add without breaking the |
| 87 | current interface. |
| 88 | |