simonvanderveldt / openAV-Ctlra

Thanks @foolswood for noticing!

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

After testing, the sensetivity of the pads needed some updates.
Future commits will amend and improve these values more.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Increased threshold, seems to trigger exactly as desired,
need to work on linearity of velocity yet.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Using standard qsort() method for now. This provides a
large improvement in single-pad handling. Multiple pads
still seeing issues.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This commit enables the blit to screen function to push a single
channel bitmap of binary data from the application to the screen
on the maschine. The maschine driver performs the required
multiplexing of and xfer header mushing to make it happen.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This commit adds basic screen infrastructure.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Lights verified as working now

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Need to rework filters for new arrival rate of data,
which means that in general the "feel" for the pads
will need to be re-evaluated.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This open the libUSB backend to maschine, removing the requirement
on the /dev/hidraw interface. Currently the bytes are only read,
and they are not decoded properly, due to an offset difference between
the HidRaw buffer (including endpoint as first byte) and libUSB data
(without said first byte).

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Using the HIDRAW API, time read and write durations.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

During debugging, a usleep() was added to allow USB writes
progress to the device. This has since been fixed in the USB
sub-system, to wait for a specific amount of time or until any
outstanding writes have completed (whichever happens first),
see commit 859ea9a for details.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Relates to feature request openAVproductions#18

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This commit fixes a possible segfault if a controller
is virtualized, and the device returns NULL instead of
a valid device_t from its connect() function.

Previously a NULL pointer + offset would be passed to the
application accept_dev_func(), which would often result
in a segfault. Returning early before the app dev accept
call handles this.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This gives the USB subsystem / kernel / device a chance to
turn off some LEDs before we terminate the USB connection,
(and in doing so cancel any outstanding events..).

Use the inflight write count to skip out of waiting if it
isn't required.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Also usleep to allow the usb transactions to retire.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Note that stable means that it is consistently performing the writes
in a performant way, without drops of event messages.

The colour table index still needs to be accessed to fix the colour-id
lookups.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Still hacky code and an issue with writing both grid and touchstrips
during the same function.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This commit removes #ifdef-ed code out of the driver, relying on
libUSB to perform all communication. Testing shows that with the
refactored USB code no events are dropped, so there's no reason
to not switch to it.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Also nuke some dead code

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This removes "garbage" stuff that was being drawn by eg:
device_test as it was reading off the end of an array and
representing it as AVTKA widgets..

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Previously, the value would always jump to 0 as the touch-depart
event message was interpreted as a value message.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Fix Screen/LED Strip define conflict, general improvements

Previously these two bitmasks had the same value, causing
confusion between screens and LED strips when drawn.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Makes no sense to test a generic MIDI implementation, as the
implementation is generic. Ignoring it allows the next driver
to create the AVTKA UI - so that this example still serves its
purpose.

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

Generic MIDI backend merge, thanks @foolswood and @simonvanderveldt!

Thanks for suggestion @foolswood! (relates to pr openAVproductions#57)

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>

This commit removes device-specific defines from the generic
MIDI backend. These values are ignored anyway, so this should
not impact the operation in any way.

Re PR openAVproductions#57, thanks @simonvanderveldt for the review!

Signed-off-by: Harry van Haaren <harryhaaren@gmail.com>