[2] [8.476104] [DBUS] power button pressed (blank=0)
[2] [10.190857] event 4 (screen press)
[2] [10.367950] event 5 (screen release)
[2] [11.911529] event 4
[2] [12.115478] event 5 (another screen tap)
[2] [13.247558] event 4
[2] [13.326446] event 5 (and a third time to unmap ASUI)
[2] [13.326446] window_unmap()
[2] [30.369629] [DBUS] power button pressed (blank=0)
[2] [30.376190] window_map() - can't get keyboard, unmap
[2] [31.483551] [DBUS] signal: com.nokia.mce.signal.tklock_mode_ind (lock)
[2] [34.750763] displayed turns off briefly, until power button is pressed
[2] [36.074096] [DBUS] power button pressed (blank=0)
[2] [36.265930] [DBUS] signal: com.nokia.mce.signal.tklock_mode_ind (unlock)

You pressed power button three times, a menu or something prevented it from mapping ASUI on the second press but it did lock the screen&keys and the third press unlocked the screen&keys. It didn't receive any key events so you should've only seen the "press dpad" notification when pressing power button, was it also appearing when pressing the keys? Or was it continuously reappearing when not pressing keys?

When power button is pressed the screen and keys are unlocked but MCE remains in a locked state so there would be no sounds. ASUI must then tell MCE to unlock if no secondary key or manually disables the screen and makes sure keys are not disabled so you can press the secondary key. It could be that my functions to control keypad and screen aren't working correctly on n800.

Can you confirm via SSH with device unlocked and working correctly that the following locks and then unlocks the hardware keys (n800)?
Does /sys/devices/platform/i2c_omap.2/i2c-0/0-0045/disable_kp (n810) exist and does it lock/unlock keys when writing 1 and 0 to it?

Does writing 1 and 0 to /sys/devices/platform/omap2_mcspi.1/spi1.0/disable_ts lock (n800 and n810) and unlock the screen?

I have been rewriting all of the power button handler code in an attempt to only have it activate on short presses. It failed to work since MCE doesn't send the short press signal when device is locked, screen is locked or when alarm is active. It did however result in much cleaner handling code and eliminated several issues but I need to test it more before breaking anyones tablets and I doubt it would resolve your current problem.


The dpad arrows should only change orientation when the screen is rotated and never when locking the screen. I'll look at the rotate code and make sure it always adjusts the arrows.


Let me know if this is still a problem after I release the new code, and if it is try to let me know what conditions cause it.

Kroll mentioned some abnormal battery drain a couple releases back and I optimized a few of the hardware polls but none of the changes shouldv'e made any noticable changes in battery life. The loop goes to sleep when screen is off or ASUI is unmapped and only dbus signals will wake it briefly. You can test this by installing the debug binary and watching its log while the screen is off or ASUI is unmapped.

For a couple months now I've had a 0.4%/minute drain, with no load and screen off, after turning off wifi or entering flight mode. Even happened after I reformated and hadn't yet installed ASUI. I had to reboot every time I used wifi. Recently found out that if I stop ssh and all wifi services the problem goes away. So your 0.8%/hour is nothing!