Re: [PATCH 4/4] Input: leds - force the LED status after .probe()

From: Benjamin Tissoires
Date: Fri Jan 27 2017 - 13:29:02 EST

On Jan 27 2017 or thereabouts, Samuel Thibault wrote:
> Benjamin Tissoires, on Fri 27 Jan 2017 18:13:18 +0100, wrote:
> > Once the host has set the initial setting of the LEDs after a new
> > USB keyboard gets connected, they appear to have a SET_IDLE command
> > sent which turns of the LEDs.
> > This means that the initial step of setting the LEDs on a keyboard is
> > just reset once .probe() finishes.
> > To be sure we set the LEDs, start a delayed task to try to resend
> > them after 1 sec.
> If there is such issue, I'd say fix the issue (what is sending
> SET_IDLE?) instead of working around it.

That is the question. I am not even sure who sends the SET_IDLE, and if
this is something that can be controlled without disturbing too many
devices. I'll dig further then.