* Extensible split data sync capability through transactions.
- Split common transport has been split up between the transport layer
and data layer.
- Split "transactions" model used, with convergence between I2C and
serial data definitions.
- Slave matrix "generation count" is used to determine if the full slave
matrix needs to be retrieved.
- Encoders get the same "generation count" treatment.
- All other blocks of data are synchronised when a change is detected.
- All transmissions have a globally-configurable deadline before a
transmission is forced (`FORCED_SYNC_THROTTLE_MS`, default 100ms).
- Added atomicity for all core-synced data, preventing partial updates
- Added retries to AVR i2c_master's i2c_start, to minimise the number of
failed transactions when interrupts are disabled on the slave due to
atomicity checks.
- Some keyboards have had slight modifications made in order to ensure
that they still build due to firmware size restrictions.
* Fixup LED_MATRIX compile.
* Parameterise ERROR_DISCONNECT_COUNT.
* Intended usage is data validation in split transport code.
* Default space efficient algorithm.
* Opt-in fast table based algorithmn with #define CRC8_USE_TABLE switch.
* Define switches for size and speed optimized versions, the default is size
optimized by using uint_least8_t as datatype for calculations.
* #define CRC8_OPTIMIZE_SPEED uses uint_fast8_t as datatype for
calculations, this only affects 32-bit Archs like ARM and RISC-V.
* Placeholder crc_init() function for hardware backed crc calculation,
not implemented yet.
This commit makes atmel-dfu and chibios-dfu bootloaders retry to detect the bootloader
every 0,5 seconds (now configurable via the BOOTLOADER_RETRY_TIME makefile variable),
and a period is printed after every try. This is a much more pleasant behaviour than
the 5s retry timeout.
* Add fast_timer_t that is 16-bit or 32-bit based on architecture
A 16-bit timer will overflow sooner but be faster to compare on AVR.
* Avoid 8-bit timer overflows in debounce algorithms
Count down remaining elapsed time instead of trying to do 8-bit timer
comparisons.
Add a "none" implementation that is automatically used if DEBOUNCE is
0 otherwise it will break the _pk/_pr count down.
* Avoid unnecessary polling of the entire matrix in sym_eager_pk
The matrix only needs to be updated when a debounce timer expires.
* Avoid unnecessary polling of the entire matrix in sym_eager_pr
The matrix only needs to be updated when a debounce timer expires.
The use of the "needed_update" variable is trying to do what
"matrix_need_update" was added to fix but didn't work because it only
applied when all keys finished debouncing.
* Fix sym_defer_g timing inconsistency compared to other debounce algorithms
DEBOUNCE=5 should process the key after 5ms, not 6ms
* Add debounce tests
* add readPort() and some API to 'tmk_core/common/*/gpio.h'
The following macros have been added to gpio.h.
* readPort(port)
* setPortBitInput(port, bit)
* setPortBitInputHigh(port, bit)
* setPortBitOutput(port, bit)
* writePortBitLow(port, bit)
* writePortBitHigh(port, bit)
* add data type 'port_data_t' into gpio.h
* rename qmk_pin to pin
Debian bullseye (testing at the moment, but seems close to release) has
avr-libc 1:2.0.0+Atmel3.6.2-1.1 with some changes taken from the
Atmel-distributed toolchain. In particular, the <avr/io.h> header for
ATmega32A (avr/iom32a.h) now defines the FLASHEND constant as `0x7FFFU`,
and that `U` suffix breaks the firmware size check code, because the
shell arithmetic expansion that is used to calculate `MAX_SIZE` does not
support those C-specific suffixes.
As a workaround, add `-D__ASSEMBLER__` to the C preprocessor invocation
that is used to expand those macros; in this case avr/iom32a.h defines
`FLASHEND` without the `U` suffix, and everything works as it did before
with older avr-libc versions.
The exact same code is present in two places; they are both changed,
even though the code in `tmk_core/avr.mk` is actually never used for
ATmega32A (and the header for ATmega32U4 does not add that `U` suffix to
`FLASHEND` for some reason).
* Add Per Key functionality for AutoShift (#11536)
* LED Matrix: Reactive effect buffers & advanced indicators (#12588)
* [Keyboard] kint36: switch to sym_eager_pk debouncing (#12626)
* [Keyboard] kint2pp: reduce input latency by ≈10ms (#12625)
* LED Matrix: Split (#12633)
* [CI] Format code according to conventions (#12650)
* feat: infinite timeout for leader key (#6580)
* feat: implement leader_no_timeout logic
* docs(leader_key): infinite leader timeout docs
* Format code according to conventions (#12680)
* Update ADC driver for STM32F1xx, STM32F3xx, STM32F4xx (#12403)
* Fix default ADC_RESOLUTION for ADCv3 (and ADCv4)
Recent ChibiOS update removed ADC_CFGR1_RES_10BIT from the ADCv3 headers
(that macro should not have been there, because ADCv3 has CFGR instead of
CFGR1). Fix the default value for ADC_RESOLUTION to use ADC_CFGR_RES_10BITS
if it is defined (that name is used for ADCv3 and ADCv4).
* Update ADC docs to match the actually used resolution
ADC driver for ChibiOS actually uses the 10-bit resolution by default
(probably to match AVR); fix the documentation accordingly. Also add
both ADC_CFGR_RES_10BITS and ADC_CFGR1_RES_10BIT constants (these names
differ according to the ADC implementation in the particular MCU).
* Fix pinToMux() for B12 and B13 on STM32F3xx
Testing on STM32F303CCT6 revealed that the ADC mux values for B12 and
B13 pins were wrong.
* Add support for all possible analog pins on STM32F1xx
Added ADC mux values for pins A0...A7, B0, B1, C0...C5 on STM32F1xx
(they are the same at least for STM32F103x8 and larger F103 devices, and
also F102, F105, F107 families). Actually tested on STM32F103C8T6
(therefore pins C0...C5 were not tested).
Pins F6...F10, which are present on STM32F103x[C-G] in 144-pin packages,
cannot be supported at the moment, because those pins are connected only
to ADC3, but the ChibiOS ADC driver for STM32F1xx supports only ADC1.
* Add support for all possible analog pins on STM32F4xx
Added ADC mux values for pins A0...A7, B0, B1, C0...C5 and optionally
F3...F10 (if STM32_ADC_USE_ADC3 is enabled). These mux values are
apparently the same for all F4xx devices, except some smaller devices may
not have ADC3.
Actually tested on STM32F401CCU6, STM32F401CEU6, STM32F411CEU6 (using
various WeAct “Blackpill” boards); only pins A0...A7, B0, B1 were tested.
Pins F3...F10 are inside `#if STM32_ADC_USE_ADC3` because some devices
which don't have ADC3 also don't have the GPIOF port, therefore the code
which refers to Fx pins does not compile.
* Fix STM32F3xx ADC mux table in documentation
The ADC driver documentation had some errors in the mux table for STM32F3xx.
Fix this table to match the datasheet and the actual code (mux settings for
B12 and B13 were also tested on a real STM32F303CCT6 chip).
* Add STM32F1xx ADC pins to the documentation
* Add STM32F4xx ADC pins to the documentation
* Add initial support for tinyuf2 bootloader (when hosted on F411 blackpill) (#12600)
* Add support for jumping to tinyuf2 bootloader. Adds blackpill UF2 example.
* Update flashing.md
* Update chconf.h
* Update config.h
* Update halconf.h
* Update mcuconf.h
* eeprom driver: Refactor where eeprom driver initialisation (and EEPROM emulation initialisation) occurs to make it non-target-specific. (#12671)
* Add support for MCU = STM32F446 (#12619)
* Add support for MCU = STM32F446
* Update platforms/chibios/GENERIC_STM32_F446XE/configs/config.h
* Restore mcuconf.h to the one used by RT-STM32F446RE-NUCLEO64
* stm32f446: update mcuconf.h and board.h for 16MHz operation, with USB enabled, and other peripherals disabled.
* Format code according to conventions (#12682)
* Format code according to conventions (#12687)
* Add STM32L433 and L443 support (#12063)
* initial L433 commit
* change to XC
* fix L433
* disable all peripherals
* update system and peripheral clocks
* 433 change
* use its own board files
* revert its own board files
* l433 specific change
* fix stm32l432xx define
* remove duplicate #define
* fix bootloader jump
* move to L443xx and add i2c2, spi2, usart3 to mcuconf.h
* move to L443
* move to L443
* fix sdmmc in mcuconf.h
* include STM32L443
* add L443
* Include L443 in compatible microcontrollers
* Include L443 in compatible microcontrollers
* Update config bootloader jump description
* Update ChibiOS define reasoning
* Update quantum/mcu_selection.mk
* fix git conflict
* Updated Function96 with V2 files and removed chconf.h and halconf.h (#12613)
* Fix bad PR merge for #6580. (#12721)
* Change RGB/LED Matrix to use a simple define for USB suspend (#12697)
* [CI] Format code according to conventions (#12731)
* Fixing transport's led/rgb matrix suspend state logic (#12770)
* [CI] Format code according to conventions (#12772)
* Fix comment parsing (#12750)
* Added OLED fade out support (#12086)
* fix some references to bin/qmk that slipped in (#12832)
* Resolve a number of warnings in `qmk generate-api` (#12833)
* New command: qmk console (#12828)
* stash poc
* stash
* tidy up implementation
* Tidy up slightly for review
* Tidy up slightly for review
* Bodge environment to make tests pass
* Refactor away from asyncio due to windows issues
* Filter devices
* align vid/pid printing
* Add hidapi to the installers
* start preparing for multiple hid_listeners
* udev rules for hid_listen
* refactor to move closer to end state
* very basic implementation of the threaded model
* refactor how vid/pid/index are supplied and parsed
* windows improvements
* read the report directly when usage page isn't available
* add per-device colors, the choice to show names or numbers, and refactor
* add timestamps
* Add support for showing bootloaders
* tweak the color for bootloaders
* Align bootloader disconnect with connect color
* add support for showing all bootloaders
* fix the pyusb check
* tweaks
* fix exception
* hide a stack trace behind -v
* add --no-bootloaders option
* add documentation for qmk console
* Apply suggestions from code review
* pyformat
* clean up and flesh out KNOWN_BOOTLOADERS
* Remove pointless SERIAL_LINK_ENABLE rules (#12846)
* Make Swap Hands use PROGMEM (#12284)
This converts the array that the Swap Hands feature uses to use PROGMEM,
and to read from that array, as such. Since this array never changes at
runtime, there is no reason to keep it in memory. Especially for AVR
boards, as memory is a precious resource.
* Fix another bin/qmk reference (#12856)
* [Keymap] Turn OLED off on suspend in soundmonster keymap (#10419)
* Fixup build errors on `develop` branch. (#12723)
* LED Matrix: Effects! (#12651)
* Fix syntax error when compiling for ARM (#12866)
* Remove KEYMAP and LAYOUT_kc (#12160)
* alias KEYMAP to LAYOUT
* remove KEYMAP and LAYOUT_kc
* Add setup, clone, and env to the list of commands we allow even with broken modules (#12868)
* Rename `point_t` -> `led_point_t` (#12864)
* [Keyboard] updated a vendor name / fixed minor keymap issues (#12881)
* Add missing LED Matrix suspend code to suspend.c (#12878)
* LED Matrix: Documentation (#12685)
* Deprecate `send_unicode_hex_string()` (#12602)
* Fix spelling mistake regarding LED Matrix in split_common. (#12888)
* [Keymap] Fix QWERTY/DVORAK status output for kzar keymap (#12895)
* Use milc.subcommand.config instead of qmk.cli.config (#12915)
* Use milc.subcommand.config instead
* pyformat
* remove the config test
* Add function to allow repeated blinking of one layer (#12237)
* Implement function rgblight_blink_layer_repeat to allow repeated blinking of one layer at a time
* Update doc
* Rework rgblight blinking according to requested change
* optimize storage
* Fixup housekeeping from being invoked twice per loop. (#12933)
* matrix: wait for row signal to go HIGH for every row (#12945)
I noticed this discrepancy (last row of the matrix treated differently than the
others) when optimizing the input latency of my keyboard controller, see also
https://michael.stapelberg.ch/posts/2021-05-08-keyboard-input-latency-qmk-kinesis/
Before this commit, when tuning the delays I noticed ghost key presses when
pressing the F2 key, which is on the last row of the keyboard matrix: the
dead_grave key, which is on the first row of the keyboard matrix, would be
incorrectly detected as pressed.
After this commit, all keyboard matrix rows are interpreted correctly.
I suspect that my setup is more susceptible to this nuance than others because I
use GPIO_INPUT_PIN_DELAY=0 and hence don’t have another delay that might mask
the problem.
* ensure we do not conflict with existing keymap aliases (#12976)
* Add support for up to 4 IS31FL3733 drivers (#12342)
* Convert Encoder callbacks to be boolean functions (#12805)
* [Keyboard] Fix Terrazzo build failure (#12977)
* Do not hard set config in CPTC files (#11864)
* [Keyboard] Corne - Remove legacy revision support (#12226)
* [Keymap] Update to Drashna keymap and user code (based on develop) (#12936)
* Add Full-duplex serial driver for ARM boards (#9842)
* Document LED_MATRIX_FRAMEBUFFER_EFFECTS (#12987)
* Backlight: add defines for default level and breathing state (#12560)
* Add dire message about LUFA mass storage bootloader (#13014)
* [Keyboard] Remove redundant legacy and common headers for crkbd (#13023)
Was causing compiler errors on some systems.
* Fix keyboards/keymaps for boolean encoder callback changes (#12985)
* `backlight.c`: include `eeprom.h` (#13024)
* Add changelog for 2021-05-29 Breaking Changes merge (#12939)
* Add ChangeLog for 2021-05-29 Breaking Changes Merge: initial version
* Add recent develop changes
* Sort recent develop changes
* Remove sections for ChibiOS changes per tzarc
No ChibiOS changes this round.
* Add and sort recent develop changes
* add notes about keyboard moves/deletions
* import changelog for PR 12172
Documents the change to BOOTMAGIC_ENABLE.
* update section headings
* re-sort changelog
* add additional note regarding Bootmagic changes
* remove changelog timestamp
* update dates in main Breaking Changes docs
* fix broken section anchors in previous changelogs
* add link to backlight/eeprom patch to changelog
* highlight some more changes
* link PRs from section headers
* Restore standard readme
* run: qmk cformat --core-only
This converts the array that the Swap Hands feature uses to use PROGMEM,
and to read from that array, as such. Since this array never changes at
runtime, there is no reason to keep it in memory. Especially for AVR
boards, as memory is a precious resource.
When the USB device is connected, FreeBSD creates not one, but three
device nodes in /dev, e.g.: /dev/ttyU0, /dev/ttyU0.init, and
/dev/ttyU0.lock.
As a result, this leads to the USB variable containing 3 paths
(and therefore, whitespace) and messages like this one:
Device /dev/ttyU0
/dev/ttyU0.init
/dev/ttyU0.lock has appeared; assuming it is the controller.
This changes fixes the use of the -z flag of "[" (see test(1)). Also, it
removes undesired paths from the USB variable, leaving only
one path there (i.e., "/dev/ttyU0").
* Fix how USB queue overflow is handled in chibios.
This commit reverts PR 12472 (commit c823fe2d3f),
and it implements the original intent of the commit in a better way.
The original intent of the above mentioned commit was to not deadlock the
keyboard when console is enabled, and hid_listen is not started.
The above mentioned commit had a few drawbacks:
1) When a lot of data was printed to the console, the queue would get full,
and drop data, even if hid_listen was running. (For example having matrix debug
enabled just didn't work right at all)
2) I believe the function in which this was implemented is used by all other
USB endpoints, so with the above change, overflow, and data loss could
happen in other important functions of QMK as well.
This commit implements deadlock prevention in a slightly similar way to how
it's done on AVR. There is an additional static local variable, that memorizes
whether the console has timeouted before. If we are in the timeouted=false
state, then we send the character normally with a 5ms timeout. If it does
time out, then hid_listen is likely not running, and future characters should
not be sent with a timeout, but those characters should still be sent if there
is space in the queue. The difference between the AVR implementation and this
one is that the AVR implementation checks the queue state directly, but this
implementation instead attempts to write the character with a zero timeout.
If it fails, then we remain in the timeouted=true state, if it succeeds, then
hid_listen started removing data from the queue, so we can go out of the
timeouted=true state.
* Added comment explaining the timeouted logic to console flow control.
* Console flow control: refactor chibios flowcontrol code to make it more readable, and rename the timeouted variable to timed_out on both chibios and lufa. Changed comments to says timed_out is an approximation of listener_disconnected, to make it clear that it's not the same thing
* fix typo
Before this commit, attaching an ARM-based (i.e. ChibiOS-based) keyboard that
uses CONSOLE_ENABLE = yes and produces debug messages would deadlock the
keyboard unless one was running hid_listen.
With this commit, dead-locking writes to the queue are detected and prevented.
fixes#5631
