An attempt to build the coverage module into the nanbox binary failed, but
pointed out that these sites needed explicit conversion from pointer to
object.
Signed-off-by: Jeff Epler <jepler@gmail.com>
Works in the usual USB DFU mode, and can program external SPI flash. It
will enable XSPI memory-mapped mode before jumping to the application
firmware in the external SPI flash.
Signed-off-by: Damien George <damien@micropython.org>
This commit adds preliminary support for ST's new STM32N6xx MCUs.
Supported features of this MCU so far are:
- basic clock tree initialisation, running at 800MHz
- fully working USB
- XSPI in memory-mapped mode
- machine.Pin
- machine.UART
- RTC and deepsleep support
- SD card
- filesystem
- ROMFS
- WiFi and BLE via cyw43-driver (SDIO backend)
Note that the N6 does not have internal flash, and has some tricky boot
sequence, so using a custom bootloader (mboot) is almost a necessity.
Signed-off-by: Damien George <damien@micropython.org>
Attempting to configure SPI3 and SPI4 for the STM32H5 would fail with a
linker error. This patch resolves that, ensuring that appropriate DMA
channels are assigned to those SPI resources.
Signed-off-by: Matt Trentini <matt.trentini@gmail.com>
On STM32H5/STM32H7, SPI flash cannot use as storage device with DMA. SPI
interruption may not be genearated even if DMA transfer has been done.
This is due to lower priority of SPI interruption than DMA.
This commit changes SPI interrupt priority more higher than DMA's priority.
Signed-off-by: Yuuki NAGAO <wf.yn386@gmail.com>
These MCUs only clear the RX idle IRQ if the data register is read, which
won't occur if the only IRQ is the RX idle IRQ (because then reading and
discarding the DR may lead to lost data).
To work around this, explicitly suppress the RX idle IRQ so that it's only
passed through to the Python callback once.
Signed-off-by: Damien George <damien@micropython.org>
It needs a divisor of 100 because the calibration temperatures are 30 and
130 degrees, similar to the H5.
Signed-off-by: Damien George <damien@micropython.org>
Some targets like frdm_k64f don't support GPIO_OUTPUT|GPIO_INPUT, so just
use GPIO_OUTPUT in those cases (it seems they still support reading the
current output state even when configured only as GPIO_OUTPUT, unlike other
targets which require both settings).
Signed-off-by: Damien George <damien@micropython.org>
With a default of -1, for soft timer. This matches other ports, and the
`extmod/machine_timer.c` implementation.
This change allows the `tests/extmod/machine_soft_timer.py` test to pass.
Signed-off-by: Damien George <damien@micropython.org>
Although the rpi_pico can already build and run with the zephyr port, this
configuration improves it in a number of ways:
- Use the USB CDC ACM as the REPL, rather than just a UART.
- Enable I2C and SPI, and add I2C1.
- Enable a filesystem, which matches exactly the rp2 port's RPI_PICO
configuration. So switching between zephyr and rp2 is possible and will
retain the filesystem.
- Make the MicroPython GC heap make the most use of the available RAM.
Signed-off-by: Damien George <damien@micropython.org>
Bluetooth works well now on this board, so enable all supported features.
Also increase the MicroPython GC heap size to make use of the available
RAM.
Unfortunately the filesystem does not match the stm32 port's NUCLEO_WB55
configuration. That's not possible to do because stm32 uses a 512 byte
flash erase size, while zephyr uses 4096 bytes. But at least here in
zephyr there's now a sizable and usable filesystem.
Signed-off-by: Damien George <damien@micropython.org>
Support for importing .mpy files is quite fundamental to MicroPython these
days, eg it allows installing more efficient .mpy code via "mip install"
(and installing `unittest` only works with the .mpy version because the .py
version uses f-strings, which are not enabled on the zephyr port). So
enable it generally for use by all boards.
As part of this, also enable:
- min/max: needed by `micropython/import_mpy_invalid.py`, and widely used
- sys.modules: needed by `run-tests.py` to run .mpy tests with --via-mpy
- io module: needed to run .mpy tests, and useful for `io.IOBase`
- array slice assign: needed to run .mpy tests, and generally useful as a
way to do a memory copy.
Signed-off-by: Damien George <damien@micropython.org>
There are two changes here:
1. Increase the UART input bufffer to 512 bytes. That's necessary to get
basic REPL reliability tests working, and helps improve `mpremote`
usage, eg copying large files.
2. Remove `uart_sem` semaphore. This is no longer needed because
`zephyr_getchar()` should be fully non-blocking and have as low a
latency as possible. `mp_hal_stdin_rx_chr()` (which calls
`zephyr_getchar`) already uses `MICROPY_EVENT_POLL_HOOK` to get
an efficient wait, and doing an extra wait and check for the
semaphore in `zephyr_getchar()` just introduces unnecessary latency and
can lead to slower input, and potentially overflowing the UART input
buffer.
Signed-off-by: Damien George <damien@micropython.org>
This change enables `sys.stdin`, `sys.stdout` and `sys.stderr` objects.
They are useful for general IO, and also help with testing zephyr boards.
Signed-off-by: Damien George <damien@micropython.org>
If there is a filesystem available, this change makes sure there is a "lib"
in `sys.path`, eg so that "mip install" works correctly.
Signed-off-by: Damien George <damien@micropython.org>
Changes here make the zephyr port act the same as other ports for the
start up and shut down sequence:
- `boot.py` is executed if it exists, and can force a soft reset
- `main.py` is only executed if in friendly REPL and if `boot.py` executed
successfully; and it can also force a soft reset
- print "MPY: " before "soft reboot" on soft reset
Signed-off-by: Damien George <damien@micropython.org>
Most boards enable the UART console because it's needed for USB (where USB
CDC creates a virtual UART), and for ctrl-C to work.
The `prj_minimal.conf` settings still use CONSOLE_SUBSYS.
Fixes issue #17608.
Signed-off-by: Damien George <damien@micropython.org>
Zephyr allows setting both GPIO_OUTPUT and GPIO_INPUT on a pin, which means
it's an output pin that can have its current value read.
Fixes issue #17596.
Signed-off-by: Damien George <damien@micropython.org>
Before this commit the UART would only work in very simple use cases.
Receiving large amounts of data would result in lost bytes. Plus the print
function would crash due to `uart_config_get()` returning incorrect values.
Additionally, receiving data with `timeout==0` would fail even if data was
already available in the internal UART Rx FIFO.
This commit fixes those issues. The non-implemented functions have also
been made usable.
Signed-off-by: Daniel Campora <danicampora@gmail.com>
The IDF panic handler resets the watchdog timeout to prevent the printing
of the error message from being cut off by a WDT reset. We use the exact
same function call in our wrapper function for the same purpose.
In IDFv5.4.2 the function used for this was changed from
`esp_panic_handler_reconfigure_wdts` to `esp_panic_handler_feed_wdts`,
specifically in this commit:
cd887ef59a
Signed-off-by: Daniël van de Giessen <daniel@dvdgiessen.nl>
When `MICROPY_PY_SYS_SETTRACE` was enabled, a crash was seen in the
qemu_mips build. It seems likely that this was due to these added fields
not being initialized.
Signed-off-by: Jeff Epler <jepler@gmail.com>
The unix coverage variant should have all features enabled, so they can be
tested for coverage. Therefore, enabled `MICROPY_PY_SYS_SETTRACE`.
Signed-off-by: Jeff Epler <jepler@gmail.com>
JavaScript code uses "Symbol in object" to brand check its own proxies, and
such checks should also work on the Python side.
Signed-off-by: Andrea Giammarchi <andrea.giammarchi@gmail.com>
Poll events during SPI transfer (USB fails during long transfers
otherwise). And add a timeout for the blocking transfer.
Signed-off-by: iabdalkader <i.abdalkader@gmail.com>
