Changes

mce-io: initialise error to NULL to prevent segfaults

mce.io: close fd in mce_unregister_io_monitor

mce apparently never closes the backing fd of a GIOChannel, this was
evident in /proc/<mce-pid>/fd, after hotplugging and removing input
devices a few times. (Note the fds that are still open, pointing to
/dev/input/event* files that are no longer there - 'deleted')

    # ls -lsh /proc/1397/fd
    total 0
    0 lrwx------ 1 root root 64 Feb 11 13:50 0 -> /dev/pts/4
    0 lrwx------ 1 root root 64 Feb 11 13:50 1 -> /dev/pts/4
    0 lr-x------ 1 root root 64 Feb 11 13:50 10 -> anon_inode:inotify
    0 lrwx------ 1 root root 64 Feb 11 13:50 11 -> socket:[232872]
    0 lr-x------ 1 root root 64 Feb 11 13:50 12 -> /dev/input/event2 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 13 -> /dev/input/event4 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 14 -> /dev/input/event3 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 15 -> /dev/input/event2 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 16 -> /dev/input/event2 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 17 -> /dev/input/event6
    0 lr-x------ 1 root root 64 Feb 11 13:50 18 -> /dev/input/event1
    0 lr-x------ 1 root root 64 Feb 11 13:50 19 -> /dev/input/event0
    0 lrwx------ 1 root root 64 Feb 11 13:50 2 -> /dev/pts/4
    0 lrwx------ 1 root root 64 Feb 11 13:50 20 -> /dev/fb0
    0 lr-x------ 1 root root 64 Feb 11 13:50 21 -> /dev/input/event2 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 22 -> /dev/input/event3 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 23 -> /dev/input/event4 (deleted)
    0 lr-x------ 1 root root 64 Feb 11 13:50 24 -> /dev/input/event2
    0 lr-x------ 1 root root 64 Feb 11 13:50 25 -> /dev/input/event2
    0 lr-x------ 1 root root 64 Feb 11 13:50 26 -> /dev/input/event3
    0 lr-x------ 1 root root 64 Feb 11 13:50 27 -> /dev/input/event4
    0 lr-x------ 1 root root 64 Feb 11 13:50 28 -> /dev/input/event5
    0 lrwx------ 1 root root 64 Feb 11 13:50 3 -> anon_inode:[eventfd]
    0 lrwx------ 1 root root 64 Feb 11 13:50 4 -> socket:[232111]
    0 lrwx------ 1 root root 64 Feb 11 13:50 5 -> socket:[234536]
    0 lrwx------ 1 root root 64 Feb 11 13:50 6 -> socket:[234537]
    0 lrwx------ 1 root root 64 Feb 11 13:50 7 -> socket:[234540]
    0 lrwx------ 1 root root 64 Feb 11 13:50 8 -> anon_inode:[eventfd]
    0 lrwx------ 1 root root 64 Feb 11 13:50 9 -> anon_inode:[eventfd]

After this patch, the result looks like this:

    # ls -lsh /proc/3558/fd
    total 0
    0 lrwx------ 1 root root 64 Feb 11 13:54 0 -> /dev/pts/4
    0 lrwx------ 1 root root 64 Feb 11 13:54 1 -> /dev/pts/4
    0 lr-x------ 1 root root 64 Feb 11 13:54 10 -> anon_inode:inotify
    0 lrwx------ 1 root root 64 Feb 11 13:54 11 -> /dev/fb0
    0 lr-x------ 1 root root 64 Feb 11 13:55 12 -> /dev/input/event2
    0 lr-x------ 1 root root 64 Feb 11 13:55 13 -> /dev/input/event2
    0 lr-x------ 1 root root 64 Feb 11 13:55 14 -> /dev/input/event3
    0 lr-x------ 1 root root 64 Feb 11 13:55 15 -> /dev/input/event4
    0 lr-x------ 1 root root 64 Feb 11 13:55 16 -> /dev/input/event5
    0 lr-x------ 1 root root 64 Feb 11 13:55 17 -> /dev/input/event6
    0 lr-x------ 1 root root 64 Feb 11 13:55 18 -> /dev/input/event1
    0 lr-x------ 1 root root 64 Feb 11 13:55 19 -> /dev/input/event0
    0 lrwx------ 1 root root 64 Feb 11 13:54 2 -> /dev/pts/4
    0 lrwx------ 1 root root 64 Feb 11 13:55 20 -> socket:[236198]
    0 lrwx------ 1 root root 64 Feb 11 13:54 3 -> anon_inode:[eventfd]
    0 lrwx------ 1 root root 64 Feb 11 13:54 4 -> socket:[237579]
    0 lrwx------ 1 root root 64 Feb 11 13:54 5 -> socket:[236179]
    0 lrwx------ 1 root root 64 Feb 11 13:54 6 -> socket:[237580]
    0 lrwx------ 1 root root 64 Feb 11 13:54 7 -> socket:[235421]
    0 lrwx------ 1 root root 64 Feb 11 13:54 8 -> anon_inode:[eventfd]
    0 lrwx------ 1 root root 64 Feb 11 13:54 9 -> anon_inode:[eventfd]

debian/changelog: bump for 1.8.128.13

Minor fixes

Fix warning: cannot optimize possibly infinite loops

Fix warning: variable 'module_info' set but not used

Fix warning: warning: implicit declaration of function 'close'

debian/control: separate deps and conflicts with newlines

mce-io: add glob writing functions

This allows for easier generic matching.

modules: keypad: add generic keyboard backlight support

Works at least on the Motorola Droid 4.

Add systemd support

Update debian/changelog (1.8.128.14)

modules: display: also use dpms when (un)blanking

When using DRM for display management, as opposed to just fbdev, just
blanking /dev/fb0 is not enough to make devices actually enter deeper
idle states, DPMS does the trick by telling X11 to send the right DRM
ioctls.

tklock: use X11 to enable and disable the touchscreen

This will make sure that:

1. We do not get touchscreen input events when the device is locked (and
   not accidentally start/stop programs/apps)
2. X will close the fd X has to the device on disable, so that the
   device can idle (power management wise), once mce also starts to release
   the handle to this device.

Add cross-compiling support

Update debian/changelog (1.8.128.15)

filter-brightness-als: support droid4 ALS

Supporting dimming the display backlight brightness, the LEDs and the
keyboard backlight based upon the Ambient Light Sensor in the Droid 4.

The values in the profiles are work in progress, I have yet to test
very dark rooms.

Close touchscreens when locked

This allows touchscreen to enter idle mode, this saves power.
Add a new datapipe to explicitly suspend touchscreens, as opposed to
something relying on tk_lock_pipe and duplicating logic.

debian/changelog: bump for 1.8.128.16

debian/control: add libx11-dev, libxi-dev deps