keyboard/touchpad not working properly as wakeup sources

[marcelo]

The keyboard/touchpad are not working properly as wakeup sources.

[wmb@…]

This bug needs a lot more detail.

[cjb]

The EC code hasn't been written to have keyboard/touchpad events generate wakeup events yet.

[David.Lin]

I will add this function on next version.

[kimquirk]

Would like to fix this on the B4 boards.

[rsmith]

Replying to kimquirk:

Would like to fix this on the B4 boards.

We tested some pre-release EC code from Quanta and wakeup from keyboard worked but wakeup from touchpad did not.

Also on wakeup from keyboard the character that was pressed for the wakeup was not delivered to the kernel.

[rsmith]

I have test code from quanta where this works.

A test .rom is at

http://dev.laptop.org/~rsmith/q2cd71.rom

Touchpad events wake up correctly.

The EC properly sends up a key on wake from keypress. You can test this with openFirmware by going into suspend with 's' and then pressing a key. That keypress will show up on the command line when OFW resumes.

The kernel however still does not register the keypress. Re-assigning as a kernel bug.

[wmb@…]

One idea: Perhaps the problem is that the keyboard driver is not waking up because there is no PS2 interrupt in the wakeup case. On a keyboard wakeup, the wakeup happens via the SCI interrupt, and there is no hardware interrupt on the PS2 IRQ (and there cannot be one because the interrupt controller is powered off when the event happens).

[dilinger]

This is now working correctly in the kernel. Note that one of the issues we have is that we don't wake up quickly enough for touchpad interaction to be smooth, but we'll have to tweak that later.

Should we wait for a firmware release that includes support before closing this?

[cjb]

Replying to dilinger:

This is now working correctly in the kernel.

I think #2401 is a dealbreaker for using this.

[jg]

Fixed in C20 firmware and current builds