Hi Henrique,
On 6/16/07, Henrique de Moraes Holschuh <[EMAIL PROTECTED]> wrote:
Dmitry, input gurus,
When a ThinkPad does an sleep cycle, the firmware explicitly kills all
pending hotkey state. That reminds me that I also need to do the same in
thinkpad-acpi.
How does one tell the input layer that it should clear its idea of the
current state of all events in a input device? I couldn't find any
input_reset() function...
Or is such a reset not needed? The recent issue with ACPI and EV_SW SW_LID
seems to indicate that in some cases these resets are indeed needed. For
EV_SW, it appears that I'd have to interate over all my switches and reset
them. But what about EV_KEY and the others?
We do need to refresh switch state upon resume because use may change
it while box is sleeping. With keys we shoudl just assume that they
are all released at the time box is waking up.
Now that I have input core converted to "struct device" I will add
suspend/resume support to input devices to mark all keys as "released'
unpon suspending and restore led state and repeat rate for keyboards
upon resuming.
--
Dmitry
