On Tue 2016-02-23 21:55:14, Johannes Berg wrote:

On Tue, 2016-02-23 at 21:45 +0100, Pavel Machek wrote:

quoted

So... you add LED trigger to display the state of the airplane
mode. Ok, why not.

Yes. However, consider that "the airplane mode" isn't a well-defined
concept; some systems may want to light up that LED even when wifi is
still enabled, since you're nowadays quite likely to be allowed to use
wifi (but not enable e.g. LTE) while in-flight.

Well "the airplane mode" is well defined. It means no intentional
transmitting at radio frequencies.

The fact that you are allowed to use WIFI on certain flights does not
change anything.

quoted

But now you add an way to override it? Why? If someone wants to
change
the led state, he can just change trigger to none, and then control
the LED manually...

Yes, but now you've forced every application that wants to deal with
this to know about every single LED that might be used with this
trigger... that won't work for some generic userland tool that might
want to implement an "airplane-mode policy".

I see that "airplane mode" trigger might be a tiny bit
useful... dunno, for a LED near the airplane mode switch, when vendor
replaced hardware toggle with a key. But policy should have nothing to
do with that. If you argue additional "policy daemon" is needed for
that... forget it, that's overdesigned.

(Besides, finding all LEDs with given trigger is trivial
operation. Besides... there should never be more than one).

quoted

BTW what happens when the device contains both radios controlled by
kernel (wifi, bluetooth) and radios controlled by userspace (GSM
modem)?

You'd better have the userspace to control the LED :)

Yes, so lets forget that and no additional triggers? Good ;-).
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html