* fix __force_inline for different GCC versions in cpp mode (the affected versions are different from c mode), and build kitchsink.c as a cpp file too for testing
* silly bug in host platform.h
* Platform updates
- Add PICO_RP2040=1 to rp2040 builds
- Add new PICO_RP2040_B0/1/2_SUPPORTED macros and retailer chip specific code to use
- Add doxygen to platform.h
- Make pico.h includable from assembly (because header order is important and tricky) - split out platform_asm.h
- Switch to using PICO_RP2040_B0_SUPPORTED in board headers
mutex_t and mutex_ are reverted to non recursive versions (pre SDK1.2.0) and new recursive_mutex_t and recursive_mutex_ functions have been added
PICO_MUTEX_ENABLE_SDK120_COMPATIBILITY flag has been added to allow old SDK1.2.0 compatibility (i.e. mutex_t can be used recursively or not) but this is slower (and is will be removed in a future version)
add watchdog_enable_caused_reboot. Note there is a small backwards incompatibility in so far as watchdog_enable sets scratch[4] to a magic value not 0 now, however this behavior was not documented before.
* add stdio_usb_connected() method
* add PICO_STDIO_USB_CONNECT_WAIT_TIMEOUT_MS to allow waiting for CDC connection during init(*
* add puts_raw and putchar_raw to skip any CR/LF translation
Due to the well known electronic market situation, we were forced to mount an alternative part number in a batch of Arduino Nano RP2040 Connect.
These flash chips, from ISSI, need yet another way to configure the QE sticky bit :|
At the moment, the safest way to handle the dual sourcing is to fallback using W25Q080 loader, and requiring that the sticky bit has already been programmed during production.
Fixes build when pico_stdlib isn't included in the target libraries:
pico-sdk/src/rp2_common/pico_divider/divider.S:8:10: fatal error: hardware/divider_helper.S: No such file or directory
Also fixes the same error in pico_double, though I'm not sure how/why:
pico-sdk/src/rp2_common/pico_double/double_aeabi.S:9:10: fatal error: hardware/divider_helper.S: No such file or directory
* use PICO_DISABLE_SHARED_IRQ_HANDLERS exclusively as config for no shared handler support (rather than also PICO_MAX_SHARED_IRQ_HANDLERS == 0)
additionally make irq_add_shared_irq_handler() call irq_set_exclusive_handler() so that single usage of an IRQ still works
* Comment typo
Co-authored-by: Luke Wren <wren6991@gmail.com>
* Add missing DREQ_s
* store actual clock frequency in clock_configure (fixes#368)
* use dma DREQ values defined in dreqs/dma.h
* Fix hw_is_claimed, and add xxx_is_claimed APIs
* Add some PIO irq helper methods
* Add DMA channel IRQ status getter and clear methods
* Implement the correct PIO IRQ status/clear methods (good to have methods here as the h/w interrupt registers are super confusing)
* fix pico_multicore dependencies
* add missing wrapper func __aeabi_f2d
* Further DMA/PIO IRQ API cleanup (and review fixes)
* add PICO_INT64_OPS_IN_RAM flag
Most build related items have moved into <tinyusb>/hw/bsp/rp2040/family.cmake which is now the source of truth
force merging as same code save submodule was already reviewed
* - fix interp_claim_lane (in case of interp1 and lane==1 bit was 0b11 instead of 0b1000)
- added missing function interp_unclaim_lane_mask
* - interp_hw_save_t are not I/O registers
* Fix warnings about some unused parameters in pico_stdio_usb
* Use `__unused` for the unused parameter in tud_descriptor_configuration_cb
* Remove redundant inclusions of `pico/platform.h`
This header is included by the other library headers (via `pico.h`)
Fixes the following warning when building for host
```
[...]/pico-sdk/src/common/pico_time/time.c: In function 'alarm_pool_dump_key':
[...]/pico-sdk/src/common/pico_time/time.c:282:15: warning: format '%ld' expects argument of type 'long int', but argument 2 has type 'uint64_t' {aka 'long long unsigned int'} [-Wformat=]
printf("%ld", to_us_since_boot(get_entry(pool, id)->target));
~~^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
%I64d
```
I'm using this in my program to get a better idea of how many entries I
need to allocate to avoid blocking or losing data, and to debug performance
issues.
Fixes the following warning:
```
[...]/pico-sdk/src/host/hardware_divider/include/hardware/divider.h:81:26: warning: operand of ?: changes signedness from 'int' to 'uint32_t' {aka 'unsigned int'} due to unsignedness of other operand [-Wsign-compare]
return b ? (a / b) : -1;
^~
```
- The divider state needs to be saved for __aeabi_ddiv, __aeabi_fdiv, __aeabi_dtan and __aeabi_ftan or they won't work in interrupts *(probably not used much youd hope), or on an RTOS context switch
- Refactored code out for the integer and floating point cases
- Improved the floating point 'tests' in passing to check more return values against GCC implementations
- Added floating point usage to the IRQ nesting test case
Adds a new cmsis_core INTERFACE library, which provides the basic CMSIS macros/definitions, and RP2040.h and system_RP2040.h
By default the isr_ methods in the vector table are renamed to their CMSIS versions, but this can be overridden by defining PICO_CMSIS_RENAME_EXCEPTIONS=0
- Add recursive_mutex
- Make all locking primitives and sleep use common overridable wait/notify support to allow RTOS
implementations to replace WFE/SEV with something more appropriate
- Add busy_wait_ms
* Change some cmake output to DEBUG level
Make SDK build more consistent with other libraries (use an INTERFACE marker library for inclusion tests)
Add PICO_SDK_PRE_LIST_FILES, PICO_SDK_POST_LIST_FILES build vars
* fix typo
* remove leftover debugging message
* SPI Definitions for SparkFun boards
Added the #defines for the default SPI pins on SparkFun boards
* SPI Definitions for SparkFun boards
Added the #defines for the default SPI pins on SparkFun boards
Co-authored-by: Kirk <blowme@cow.com>
This prevents a comparison between a signed and an unsigned number which
will create a warning tripping -Werror.
Also added a check for the alignment of the spin lock structure
* i2c: set hold time of SDA during transmit to 2 for TCS34725 color sensor
* i2c: fix issues in i2c_write_blocking_internal
* i2c: rename sda_hold_count to sda_tx_hold_count
* use assert rather than invalid_params_if for internal consistency checks
* i2c: use a more appropriate sda tx hold time at higher baudrates
* i2c: reduce 120/1e9 to the smallest possible integer numerator and denominator
* Don't use implicit 0 for field LSB
Co-authored-by: Luke Wren <wren6991@gmail.com>
* move PLL reset code from clocks driver to pll driver
* Don't clear PLL PWR/FBDIV after reset as unnecessary. Call out in runtime.c why USB/syscfg aren't reset.
Co-authored-by: Peter Lawrence <12226419+majbthrd@users.noreply.github.com>
Co-authored-by: Luke Wren <wren6991@gmail.com>
* build: allow compile time - and hence via board config header - choice of boot stage2 via a default compile_time_choice.S boot stage 2 which conditionally includes others
adafruit_rp2040_feather: fix flash size/clkdiv
* fixup GENERIC_03H #define names
* Add pico_get_unique_board_id_string API
Add a new API to pico_unique which will turn the unique ID into a
canonical text string.
Use this API to update the USB device serial number in stdio_usb.
Supercedes #280
* Clean up -Wconversion=error issues
* Address review comments, fix api typing
Use cleaner binary-to-hex conversion. Update the length parameter to
use uint per the SDK standard .
In order for the returned value to accurately reflect a single moment in time,
ensure the registers are read just once and in the datasheet order.
Before this change, the RTC registers would each be read multiple times,
leading (infrequently) to the returned fields not all reflecting the same
moment in time.
The rp2040 datasheet has what I believe is an incorrect example (embedding
the source of this function); will the datasheet be updated if this function is
fixed?
This problem is only a speculative one; I did not actually observe it in the
wild.
* pico_stdio_usb: be more explicit about includes, fix warning (#257)
* pico_base: NDEBUG backwards for absolute_time_t (#255)
* pico_util: missing extern C in queue.h (#249)
* build: remove -march which was masking -mcpu, now SVC available (#253)
* Add documentation for gpio_irq_callback_t and rtc_callback_t (fixes#175)
* Hook up pico_binary_info and pico_bootsel_via_double_reset to the Doxygen index
* Add link to new Raspberry Pi Pico FAQ
* Consistently add parentheses for \sa or \see links to function-names
* Use consistent capitalisation for 'GitHub'
* Small typos and markup errors
* Board definition header files for the iniital set of SparkFun rp2040 boards
* Add default PICO_DEFAULT_I2C*, allow no PICO_DEFAULT_LED_PIN, no PICO_DEFAULT_UART* (instead of -1)
Fixup SparkFun headers
* Pimoroni board headers
* Add LED related board defines PICO_CONFIGs (to pico_stdlib for now)
* more board config changes
* add Adafruit feather, itsybitsy, qtpy board headers
* add PICO_DEFAULT_WS2812_POWER_PIN define
* MOSI/MISO -> TX/RX, some UART cleanup.. make vgaboard.h defines take preference over pico.h ones
* local change to tinyusb to cope with no default LED or UART
* fix review issues
Co-authored-by: Kirk Benell <github-stuff@accvec.com>
Co-authored-by: ZodiusInfuser <christopher.parrott2@gmail.com>
Co-authored-by: hathach <thach@tinyusb.org>
- setting baud rate to magic value (default=1200) will cause a reset to BOOTSEL mode
- a VENDOR interface along side the CDC interface can be used to reset via refular flash boot, or into
BOOTSEL mode with control for the reset_usb_boot parameters for the latter
either method can be configured/enabled/disabled via #define
* remove broken and fairly pointless deep sleep API, use multicore_reset_core1 instead
* strict-prototype for multicore_fifo_drain() in host header
* Also for multicore_lockout_victim_init()
Co-authored-by: Luke Wren <wren6991@gmail.com>
* Move PICO_FLASH_SIZE_BYTES out of platform_defs.h since it precedes board config; also default to -1 for unknown
* make default for PICO_FLASH_SIZE_BYTES undefined - i.e. no value