| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
|
|
|
| |
If the button bounces between the triggering interrupt and the button
state check, we could lose the final state change of the button and
send an incorrect event to the application. In the worst case, the button
would end up in exactly the wrong state, toggling in the wrong direction.
Use the fast timer to poll all buttons instead so that there is only
one check of each button at each poll interval (instead of the
interrupt and the state check). This makes buttons reliably debounced.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
This was used with the 8051 bit-addressing mode to provide
single-instruction access to GPIO pins.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
| |
TeleLCO v2.0 has more than 4
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
| |
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
This returns the current state of the button; useful to check what the
value is at startup.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
| |
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of waiting for a while after the transition to decide if it
has stuck, signal the event right away and then ignore other
transitions for the debounce interval.
This seems to work just as reliably, but has the benefit of
eliminating button latency at press time.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
This "can't" happen, but it's not unreasonable to have a default
return value.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
This splits the fast-timer portion out of the debounce helper code and
shares that with the quadrature driver which now uses it directly.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
| |
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
These were neglected in the commit which was supposed to include them
Signed-off-by: Keith Packard <keithp@keithp.com>
|