Re: [RFC PATCH 0/1] Add generic inverted led triggers

[Tobias Schramm]

Hi Pavel,

thanks for the feedback.

> Not a big fan (sorry).
> 
> We have already _way_ too many triggers, we don't want to have twice
> that much.

True. Doubling the amount of triggers is probably not a good idea.

> 
> Better implementation might be to have a trigger attribute doing the
> inverting.

I agree. Especially since Jacek pointed out that some triggers do that
already.

> 
> Inverting really does not work with all the triggers; numlock-inverted
> will not get too many
> users. always-on-inverted... blink-inverted.... I guess it does make
> sense for disk activity (but be warned disk can be continuously active
> for quite a while).
> 
> What triggers do you think make sense inverted?

I think all kinds of activity indicators (disk-activity, mmc, usb, ide,
nand, cpu, network, etc.) make sense. Guess I'll add a flags field to
the led_trigger struct and have an invertible flag that specifies
whether a trigger should be invertible or not.


Thanks again,

Tobias