refactor(split): Refactor split code for extension
Extract central/peripheral code to allow for plugging in alternate
transports, instead of tying all split logic to BT.
feat(split): Add full-duplex wired split support
* Depends on full-duplex hardware UART for communication.
* Supports all existing central commands/peripheral events, including
sensors/inputs from peripherals.
* Only one wired split peripheral supported (for now)
* Relies on chosen `zmk,split-uart` referencing the UART device.
docs: Add wired split config docs.
Migrate split to its own dedicated config file, and add details
on wired split config.
Co-authored-by: Nicolas Munnich <98408764+Nick-Munnich@users.noreply.github.com>
fix: Properly override stack size on RP2040
Move the system work queue stack size override on RP2040 ouf of
a `ZMK_BLE` conditional so it is properly applied generally for that
SoC.
---------
Co-authored-by: Nicolas Munnich <98408764+Nick-Munnich@users.noreply.github.com>
* Studio will use stable layer IDs to refer to layers, so that
layer reordering doesn't affect behavior bindings, so update to match.
* Fix a few layer metadata entries that missed being refactored.
* Avoid doing duplicate calls to setings_load_subtree, which iterates
NVS fully each time under the hood, and instead use on settings_load
later in the lifecycle.
* Add a new feature for tracking a given behavior by a new concept
of a "behavior local ID" which is a stable 16-bit identifier for
a given behavior, that is resilient to new behaviors being added
and requires no additional work on the part of the behavior
authors.
* Add implementations for either settings lookup table, or CRC16
hashing of behavior device names for generating behavior local
IDs.
* For upcoming ZMK studio work, make a set of rich metadata available
to provide a friendly name for a behavior, and allow super flexible
descriptions of the parameters the behaviors take.
* Add ability to validate a zmk_behavior_binding against
the behavior metadata available.
Added BEHAVIOR_DT_DEFINE() and BEHAVIOR_DT_INST_DEFINE(), which work
exactly like the DEVICE_*_DEFINE() macros, except they also register the
device as a behavior by adding a pointer to it to a memory section.
Added zmk_behavior_get_binding(), which works like device_get_binding()
except that it only searches the devices that have been registered as
behaviors. This ensures that behaviors cannot have name collisions with
other devices defined by the SoC, which will be important when we remove
the label property from behaviors so they are given their node names.
As an added benefit, this is faster since it searches a smaller list.
Some basic benchmark code I wrote indicates it takes 30-70% as long,
depending on where the behavior is in the list and whether the name
string is an exact pointer match.
From now on, behaviors should use BEHAVIOR_*_DEFINe() instead of
DEVICE_*_DEFINE(), and any code that looks up a behavior by name should
use zmk_behavior_get_binding() instead of device_get_binding().
Pete Johanson
Cem Aksoylar
Joel Spadin