With public declarations moved to extmod/modmachine.h. It's now mandatory
for a port to define MICROPY_PY_MACHINE_PWM_INCLUDEFILE if it enables
MICROPY_PY_MACHINE_PWM. This follows how extmod/machine_wdt.c works.
All ports have been updated to work with this modified scheme.
Signed-off-by: Damien George <damien@micropython.org>
There are currently 7 ports that implement machine.WDT and a lot of code is
duplicated across these implementations. This commit factors the common
parts of all these implementations to a single location in
extmod/machine_wdt.c. This common code provides the top-level Python
bindings (class and method wrappers), and then each port implements the
back end specific to that port.
With this refactor the ports remain functionally the same except for:
- The esp8266 WDT constructor now takes keyword arguments, and accepts the
"timeout" argument but raises an exception if it's not the default value
(this port doesn't support changing the timeout).
- The mimxrt and samd ports now interpret the argument to WDT.timeout_ms()
as signed and if it's negative truncate it to the minimum timeout (rather
than it being unsigned and a negative value truncating to the maximum
timeout).
Signed-off-by: Damien George <damien@micropython.org>
When using malloc and free there were out-of-memory situations depending on
the arm-none-eabi package version. This commit changes malloc/free to use
the MicroPython GC heap instead.
Signed-off-by: robert-hh <robert@hammelrath.com>
Signed-off-by: Damien George <damien@micropython.org>
That can be caused e.g. by an exception. This feature is implemented in
some way already for the stm32, renesas-ra, mimxrt and samd ports. This
commit adds it for the rp2, esp8266, esp32 and nrf ports. No change for
the cc3200 and teensy ports.
Signed-off-by: robert-hh <robert@hammelrath.com>
Currently on rp2 the time.time_ns() function has only seconds resolution.
This commit makes it have microsecond resolution, by using the output of
time_us_64() instead of the RTC.
Tested that it does not drift from the RTC over long periods of time.
Signed-off-by: Damien George <damien.p.george@gmail.com>
Allows using gdb, addr2line, etc. on a "release" ELF file.
No impact to .bin or .uf2 size, only the .elf will get bigger.
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
This removes the duplicate code in cyw43, esp32, esp8266 that implements
the same logic as network.hostname.
Renames the `mod_network_hostname` (where we store the hostname value in
`.data`) to `mod_network_hostname_data` to make way for calling the shared
function `mod_network_hostname`.
And uses memcpy for mod_network_hostname_data, because the length of source
is already known and removes reliance on string data being null-terminated.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This was previously hard-coded to "Micropy" / "Mass Storage" / "1.0".
Now allow it to be overridden by a board.
Also change "Micropy" to "MicroPy" and "1.0" to "1.00" to match stm32.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
While cyw43 is deinitialized, an interrupt occurs. That is handled with
these lines: ports/rp2/mpnetworkport.c#L59-L61 and as pendsv is disabled
while in network code, the poll function then just waits there.
When deinit has finished, the poll func is executed, but skipped:
src/cyw43_ctrl.c#L222-L225 this skips the `CYW43_POST_POLL_HOOK` which
would re-enable interrupts, but also reset `cyw43_has_pending`.
And in that state, the lightsleep code, will skip sleeping as it thinks
there is a network packet pending to be handled.
With this change applied, lightsleep works as expected when the wifi chip
is enabled, and when it's powered off.
Fixes are:
- The baudrate argument is a keyword arg, it was passed before as a
positional arg.
- Use the port and baudrate arguments passed from higher level code instead
of the hard-coded port ID and baudrate, which would allow HCI drivers to
change baudrates.
- Increase UART char timeout and RX buffer size.
Signed-off-by: iabdalkader <i.abdalkader@gmail.com>
In CPython, `_thread.start_new_thread()` returns an ID that is the same ID
that is returned by `_thread.get_ident()`. The current MicroPython
implementation of `_thread.start_new_thread()` always returns `None`.
This modifies the required functions to return a value. The native thread
id is returned since this can be used for interop with other functions, for
example, `pthread_kill()` on *nix. `_thread.get_ident()` is also modified
to return the native thread id so that the values match and avoids the need
for a separate `native_id` attribute.
Fixes issue #12153.
Signed-off-by: David Lechner <david@pybricks.com>
Showing the period alway as microsecond quantities, since tick_hz is
assumed as 1_000_000 if the period is given by freq=xxx. If the period is
larger than 0xffffffff, the value is divided by 1000 and "000" is appended
in the display. That works for periods up to about 50 days.
Signed-off-by: robert-hh <robert@hammelrath.com>
PICO might not always be a unique name across all ports, and the
convention generally for other boards is to do VENDOR_BOARD.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This allows switching between variants without clobbering the build
output.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This was used to override the firmware filename generated by the build
server (to match the historical name before board definitions existed).
Now we're making everything use the board definition name (i.e. the
directory name).
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This is to support a future change to add the variant name to the build
directory and therefore should be the same style as the board name.
This only affects the WEACTSTUDIO board. Also standardises on a convention
for naming flash-size variants. Normally we would write e.g. 2MiB, but in
uppercase, it's awkward to write 2MIB, so instead use 2M, 512K, etc for
variant names, but use 2MiB when not constrained by case (e.g. a regular
filename).
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This is difficult to implement on cmake-based ports, and having the list
of variants in mpconfigboard.{cmake,mk} duplicates information that's
already in board.json.
This removes the existing query-variants make target from stm32 & rp2
and the definition of BOARD_VARIANTS from the various board files.
Also renames the cmake variable to MICROPY_BOARD_VARIANT to match other
variables such as MICROPY_BOARD. The make variable stays as
BOARD_VARIANT.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This commit:
- Finds a common set of board feature tags and maps existing features to
that reduced set.
- Removes some less-useful board feature tags.
- Ensures all MCUs are specified correctly.
- Ensures all boards have a vendor (and fixes some vendor names).
This is to make the downloads page show a less intimidating set of filters.
Work done in conjunction with Matt Trentini <matt.trentini@gmail.com>.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
And use it in mp_hal_get_pin_obj() and machine_pin_make_new(). That way,
mp_hal_get_pin_obj() accepts both int and str objects as argument, allowing
use of a pin specifier instead of a pin object in the constructor of
devices which need a pin as parameter.
E.g. instead of
uart = UART(0, tx=Pin(0), rx=Pin(1))
one can write:
uart = UART(0, tx=0, rx=1)
Signed-off-by: robert-hh <robert@hammelrath.com>
Mostly updates comments, but also renames the UASYNCIO enum value to
ASYNCIO.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
The asyncio module now has much better CPython compatibility and
deserves to be just called "asyncio".
This will avoid people having to write `from uasyncio import asyncio`.
Renames all files, and updates port manifests to use the new path. Also
renames the built-in _uasyncio to _asyncio.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
In 5fe2a3f1 the ESP32 port underwent a change to how `MICROPY_PORT_DIR`
is defined. This commit normalizes the `rp2` port to use the same
underlying variable mechanism (`CMAKE_CURRENT_LIST_DIR`).
Signed-off-by: Brian 'redbeard' Harrington <redbeard@dead-city.org>
Applies to drivers/examples/extmod/port-modules/tools.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Updates any includes, and references from Makefiles/CMake.
This essentially reverts what was done long ago in commit
136b5cbd76
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This renames the builtin-modules, such that help('modules') and printing
the module object will show "module" rather than "umodule".
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Updates all `help()` output to use the phrase:
`For online docs please visit http://docs.micropython.org/`
Some ports previously used different wording, some pointed to the wrong
link. Also make all ports use `help.c` for consistency.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
These have by default FAT support. The SAMD21 build does not support FAT.
The nrf port also implements os.sync(), but has it's own copy of moduos.c.
Code size increases seen: 40 to 56 bytes.
Signed-off-by: robert-hh <robert@hammelrath.com>
Building the Pico-W needs the MICROPY_PY_NETWORK_CYW43 flag to be set in
order to include building the CYW43 Wifi driver. But then mp_hal_get_mac()
handles the MAC assignment for all nics the "CYW43 way", copying the real
MAC provided by the WiFi hardware. This will fail for all other NIC types,
resulting in an invalid MAC address.
The solution in this commit is to add a check for the NIC type parameter
idx and handle the MAC address respectively.
Convert to an absolute path to always reliably locate manifest.py. This is
already done in Makefile, but is also needed in CMakeLists.txt if cmake is
invoked directly.
Signed-off-by: Phil Howard <phil@pimoroni.com>
Currently rp2.StateMachine.exec(instr_in) requires that the instr_in
parameter be a string representing the PIO assembly language instruction
to be encoded by rp2.asm_pio_encode(). This commit allows the parameter
to also be of integral type. This is useful if the exec() method is
being called often where the use of pre-encoded machine code is
desireable.
This commit still supports calls like:
sm.exec("set(0, 1)")
It also now supports calls like:
# Performed once earlier, maybe in __init__()
assembled_instr = rp2.asm_pio_encode("out(y, 8)", 0)
# Performed multiple times later as the PIO state machine is
# configured for its next run.
sm.exec(assembled_instr)
The existing examples/rp2/pio_exec.py and examples/rp2/pio_pwm.py that
exercise the rp2.StateMachine.exec() method still work with this change.
Signed-off-by: Adam Green <adamgrym@yahoo.com>
All ports that enable MICROPY_PY_MACHINE_PWM now enable these two
sub-options, so remove these sub-options altogether to force consistency in
new ports that implement machine.PWM.
Signed-off-by: Damien George <damien@micropython.org>
Changes in this commit:
- Limit duty_u16() to 65535 and duty_ns() to the period duration.
- Return 0 for pwm.freq() if the frequency has not been set yet.
- Return 0 for pwm.duty_us16() and duty_ns() unless both frequency and
duty cycle have been set.
- Initialize the pin to PWM at the very end of the constructor, to avoid
possible glitches on the pin when setting up the PWM.
This adds support for freq/duty_u16/duty_ns keyword arguments in the PWM
constructor, and adds the PWM.init() method. Using init() without
arguments enables a previously deinit-ed PWM again.
Further changes in this commit:
- Do not start PWM output if only duty was set.
- Stop all PWM slices on soft-reset.
- Fix a bug when changing the freq on a channel pair with duty_ns set.
Based on extmod/utime_mphal.c, with:
- a globals dict added
- time.localtime wrapper added
- time.time wrapper added
- time.time_ns function added
New configuration options are added for this module:
- MICROPY_PY_UTIME (enabled at basic features level)
- MICROPY_PY_UTIME_GMTIME_LOCALTIME_MKTIME
- MICROPY_PY_UTIME_TIME_TIME_NS
Signed-off-by: Damien George <damien@micropython.org>
Since c80e7c14e6 changed the GC heap to use
all unused RAM, there is no longer any RAM available for the traditional C
heap (which is not used by default in MicroPython but may be used by C
extensions). This commit adds a provision for a board to reserve RAM for
the C heap, by defining MICROPY_C_HEAP_SIZE.
Signed-off-by: Damien George <damien@micropython.org>
This adds a mechanism to track a pending notify/indicate operation that
is deferred due to the send buffer being full. This uses a tracked alloc
that is passed as the content arg to the callback.
This replaces the previous mechanism that did this via the global pending
op queue, shared with client read/write ops.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This commit adds support for the `timeout` keyword argument to machine.I2C
on the rp2 port, following how it's done on other ports.
The main motivation here is avoid the interpreter crashing due to infinite
loops when SDA is stuck low, which is quite common if the board gets reset
while reading from an I2C device.
A default timeout of 50ms is chosen because it's consistent with:
- Commit a707fe50b0 which used a timeout of
50,000us for zero-length writes on the rp2 port.
- The machine.SoftI2C class which uses 50,000us as the default timeout.
- The stm32 port's hardware I2C, which uses 50,000us for
I2C_POLL_DEFAULT_TIMEOUT_US.
This commit also fixes the default timeout on the esp32 port to be
consistent with the above, and updates the documentation for machine.I2C to
document this keyword argument.
Helps prevent the filesystem from getting formatted by mistake, among other
things. For example, on a Pico board, entering Ctrl+D and Ctrl+C fast many
times will eventually wipe the filesystem (without warning or notice).
Further rationale: Ctrl+C is used a lot by automation scripts (eg mpremote)
and UI's (eg Mu, Thonny) to get the board into a known state. If the board
is not responding for a short time then it's not possible to know if it's
just a slow start up (eg in _boot.py), or an infinite loop in the main
application. The former should not be interrupted, but the latter should.
The only way to distinguish these two cases would be to wait "long enough",
and if there's nothing on the serial after "long enough" then assume it's
running the application and Ctrl+C should break out of it. But defining
"long enough" is impossible for all the different boards and their possible
behaviour. The solution in this commit is to make it so that frozen
start-up code cannot be interrupted by Ctrl+C. That code then effectively
acts like normal C start-up code, which also cannot be interrupted.
Note: on the stm32 port this was never seen as an issue because all
start-up code is in C. But now other ports start to put more things in
_boot.py and so this problem crops up.
Signed-off-by: David Grayson <davidegrayson@gmail.com>
This is a best-effort implementation of write polling. It's difficult to
do correctly because if there are multiple output streams (eg UART and USB
CDC) then some may not be writeable while others are. A full solution
should also have a return value from mp_hal_stdout_tx_strn(), returning the
number of bytes written to the stream(s). That's also hard to define.
The renesas-ra and stm32 ports already implement a similar best-effort
mechanism for write polling.
Fixes issue #11026.
Signed-off-by: Damien George <damien@micropython.org>
Prior to this change, setting of UART parameters like parity, stop bits or
data bits did not work correctly. As suggested by @iabdalkader, adding
__DSB() fixes the problem, making sure that changes to the UART LCR_H
register are seen by the peripheral.
Note: the FIFO is already enabled in the call to uart_init(), so the call
to uart_set_fifo_enabled() is not required, but kept for visibility.
Fixes issue #10976.
For builds with DEBUG=1 and MICROPY_HW_ENABLE_UART_REPL=1, calling
stdio_init_all() in main() detaches the UART input from REPL. This change
suppresses calling stdio_init_all() then.
Previously, setting MICROPY_HW_ENABLE_USBDEV to 0 caused build errors. The
change affects the nrf and samd ports as well, so MICROPY_HW_ENABLE_USBDEV
had to be explicitly enabled there.
The configuration options MICROPY_HW_ENABLE_USBDEV and
MICROPY_HW_ENABLE_UART_REPL are independent, and can be enabled or disabled
by a board.
Signed-off-by: Damien George <damien@micropython.org>
Borrowing an idea from the mimxrt port (also stm32 port): in the loader
input file memmap_mp.ld calculate __GcHeapStart and __GcHeapEnd as the
unused RAM. Then in main.c use these addresses as arguments to gc_init().
The benefits of this change are:
1) When libraries are added or removed in the future changing BSS usage,
main.c's sizing of the GC heap does not need to be changed.
2) Currently these changes make the GC area about 30 KBytes larger, eg on
PICO_W the GC heap increases from 166016 to 192448 bytes. Without that
change this RAM would never get used.
3) If someone wants to disable one or more SRAM blocks on the RP2040 to
reduce power consumption it will be easy: just change the MEMORY section
in memmap_mp.ld. For instance to not use SRAM2 and SRAM3 change it to:
MEMORY
{
FLASH(rx) : ORIGIN = 0x10000000, LENGTH = 2048k
RAM(rwx) : ORIGIN = 0x21000000, LENGTH = 128k
SCRATCH_X(rwx) : ORIGIN = 0x20040000, LENGTH = 4k
SCRATCH_Y(rwx) : ORIGIN = 0x20041000, LENGTH = 4k
}
Then to turn off clocks for SRAM2 and SRAM3 from MicroPython, set the
appropriate bits in WAKE_EN0 and SLEEP_EN0.
Tested by running the firmware.uf2 file on PICO_W and displaying
micropython.mem_info(). Confirmed GC total size approximately matched the
size calculated by the loader.
Signed-off-by: cpottle9 <cpottle9@outlook.com>
This function seems to work fine in multi-core applications now.
The delay is now in units of microseconds instead of depending on the clock
speed, and is adjustable by board configuration headers.
Also added documentation.
This removes the previous WiFi driver from drivers/cyw43 (but leaves behind
the BT driver), and makes the stm32 port (i.e. PYBD and Portenta) use the
new "lib/cyw43-driver" open-source driver already in use by the rp2 port.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This provides a standard interface to setting the global networking config
for all interfaces and interface types.
For ports that already use either a static hostname (mimxrt, rp2) they will
now use the configured value. The default is configured by the port
(or optionally the board).
For interfaces that previously supported .config(hostname), this is still
supported but now implemented using the global network.hostname.
Similarly, pyb.country and rp2.country are now deprecated, but the methods
still exist (and forward to network.hostname).
Because ESP32/ESP8266 do not use extmod/modnetwork.c they are not affected
by this commit.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This matches the behavior of the makefile ports but implemented for CMake,
making it easy to specify custom board definitions.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This further aligns the features available on Pico and Pico W boards.
os.dupterm is generally useful, but can still be disabled by a board if
needed. hashlib.sha1 requires mbedtls for the implementation, but that's
always available (due to ucryptolib's requirements). The entire hashlib
module can still be disabled by an individual board if needed.
Fixes issue #7881.
Signed-off-by: Damien George <damien@micropython.org>
The default now includes all sub-components (security, l2cap, etc)
and using the kwarg options is no longer supported.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Prior to this commit, on Pico W (where the CYW43 driver is enabled) the PIO
instruction memory was not released on soft reset, so using PIO after a
soft reset would eventually (after a few soft resets) lead to ENOMEM when
allocating a PIO program.
This commit fixes that by tracking the use of PIO memory by this module and
freeing it on soft reset.
Similarly, use of the state machines themselves are tracked and released on
soft reset.
Fixes issue #9003.
Signed-off-by: Damien George <damien@micropython.org>
This will ensure that any board with networking support gets:
- webrepl
- mip
- urequests
- ntptime
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This drops the `.cpu` directive from the ARM gchelper_*.s files. Having
this directive breaks the linker when targeting older CPUs (e.g. `-mthumb
-mthumb-interwork` for `-mcpu=arm7tdmi`). The actual target CPU should be
determined by the compiler options.
The exact CPU doesn't actually matter, but rather the supported assembly
instruction set. So the files are renamed to *_thumb1.s and *thumb2.s to
indicate the instruction set support instead of the CPU support.
Signed-off-by: David Lechner <david@pybricks.com>
Prior to this commit, Pin(Pin.OPEN_DRAIN, value=0) would not set the
initial value of the open-drain pin to low, instead it would be high.
Signed-off-by: Damien George <damien@micropython.org>
The mp_plat_print output is already being used by the subsequent call to
mp_obj_print_exception(). And this eliminates all references to printf for
this port (at least in non-debug builds).
Signed-off-by: Damien George <damien@micropython.org>
Pin defines are:
- For Pico define board pins and the default LED pin (WL_GPIO25).
- For Pico-W define board pins, external pins and the default
LED pin (WL_GPIO0).
- For the Nano-RP2040, define board pins, external pins and
the default LED pin (GPIO25)
- For all other boards, the pins.csv defines the LED pin (if any)
for backwards compatibility with code that assumes there's always
an LED pin.
This commit adds support for generating named pin mappings for all pins
including CPU, board-defined, LED and externally controlled pins. CPU pins
are mapped to `pin_GPIO<n>`, externally-controlled pins are mapped to
`pin_EXT_GPIO<n>`, and defined conditionally (up to 10 pins, and can be
expanded in the future), and they are non-const to allow `machine-pin.c` to
write the pin object fields. Both CPU and externally controlled pins are
generated even if there's no board CSV file; if one exists it will just be
added to board pins.
Handle externally controlled GPIO pins more generically, by removing all
CYW43-specific code from `machine_pin.c`, and adding hooks to initialise,
configure, read and write external pins. This allows any driver for an
on-board module which controls GPIO pins (such as CYW43 or NINA), to
provide its own implementation of those hooks and work seamlessly with
`machine_pin.c`.
To allow the USB to work in cases where there is a lot of filesystem
access, in particular on boot.
For example, registering of the USB CDC interface may fail if:
- the board file system is lfs2 (default), and
- sys.path contains entries for the local file system (default), and
- files are imported by boot.py or main.py from frozen bytecode of the file
system (common) and the file system contains many files, like 100.
In that case the board is very busy with scanning LFS, and registering the
USB interface seems to time out. This commit fixes this by allowing the
USB to make progress during filesystem reads.
Also switch existing MICROPY_EVENT_POLL_HOOK uses in this file to
MICROPY_EVENT_POLL_HOOK_FAST now that the latter macro exists.
When switching from a special function like SPI to an input or output,
there was a brief period after the function was disabled but before the
pin's I/O state was configured, in which the state would be poorly defined.
This fixes the problem by switching off the special function after fully
configuring the I/O state.
Fixes#10226.
Signed-off-by: Paul Grayson <pdg@alum.mit.edu>
There were several places where 32-bit integer could overflow with
frequencies of 2^28 Hz or above (~268 MHz). This fixes those overflows and
also introduces rounding for more accurate duty_ns computations.
Signed-off-by: Paul Grayson <pdg@alum.mit.edu>
This changes the freq() and duty_u16() functions to use more simpler, more
accurate formulas, in particular increasing the frequency accuracy from a
few percent to a fraction of a percent in many cases.
Signed-off-by: Paul Grayson <pdg@alum.mit.edu>
This commit prevents the device from "hanging" when using lightsleep while
the WiFi chip is active.
Whenever the WiFi chip wants to interrupt the microcontroller to notify it
for a new package, it sets the CYW43_PIN_WL_HOST_WAKE pin to high,
triggering an IRQ. However, as polling the chip cannot happen in an
interrupt handler, it subsequently notifies the pendsv-service to do a poll
as soon as the interrupt handler ended. In order to prevent a new
interrupt from happening immediately afterwards, even before the poll has
run, the IRQ handler disables interrupts from the pin.
The first problem occurs, when a WiFi package arrives while the main loop
is in cyw43-code. In order to prevent concurrent access of the hardware,
the network code blocks pendsv from running again while entering lwIP code.
The same holds for direct cyw43 code (like changing the cyw43-gpios, i.e.
the LED on the Pico W). While the pendsv is disabled, interrupts can still
occur to schedule a poll (and disable further interrupts), but it will not
run. This can happen while the microcontroller is anywhere in rp2040 code.
In order to preserve power while waiting for cyw43 responses,
cyw43_configport.h defines CYW43_DO_IOCTL_WAIT and
CYW43_SDPCM_SEND_COMMON_WAIT to __WFI(). While this might work in most
cases, there are 2 edge cases where it fails:
- When an interrupt has already been received by the cyw43 stack, for
example due to an incoming ethernet packet.
- When the interrupt from the cyw43 response comes before the
microcontroller entered the __WFI() instruction.
When that happens, wfi will just block forever as no further interrupts are
received. The only way to safely use wfi to wake up from an interrupt is
inside a critical section, as this delays interrupts until the wfi is
entered, possibly resuming immediately until interrupts are reenabled and
the interrupt handler is run. Additionally this critical section needs to
check whether the interrupt has already been disabled and pendsv was
triggered, as in such a case, wfi can never be woken up, and needs to be
skipped, because there is already a package from the network chip waiting.
Note that this turns cyw43_yield into a nop (and thereby the cyw43-loops
into busy waits) from the second time onwards, as after the first call, a
pendsv request will definitely be pending. More logic could be added, to
explicitly enable the interrupt in this case.
Regarding lightsleep, this code has a similar problem. When an interrupt
occurs during lightsleep, the IRQ and pendsv handler and thereby poll are
run immediately, with the clocks still disabled, causing the SPI transfers
to fail. If we don't want to add complex logic inside the IRQ handler we
need to protect the whole lightsleep procedure form interrupts with a
critical section, exiting out early if an interrupt is pending for whatever
reason. Only then we can start to shut down clocks and only enable
interrupts when the system is ready again. Other interrupt handlers might
also be happy, that they are only run when the system is fully operational.
Tested on a Pico W, calling machine.lightsleep() within an endless loop and
pinging from the outside.
This was previously implemented by adding additional members to the
mp_obj_type_t defined for each NIC, which is difficult to do cleanly with
the new object type slots mechanism. The way this works is also not
supported on GCC 8.x and below.
Instead replace it with the type protocol, which is a much simpler way of
achieving the same thing.
This affects the WizNet (in non-LWIP mode) and Nina NIC drivers.
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
App the mp_ prefix to usbd_ symbols and files which are defined here and
not in TinyUSB.
rp2 only for now. This includes some groundwork for dynamic USB devices
(defined in Python).
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
This commit executes __WFI() on core 0 only to avoid core1 locking up since
it doesn't enable any interrupts by default (except for `SIO_IRQ_PROC1`).
This fixes a lockup when calling `cyw43_do_ioctl` from core1.
Fixes issue #9597.
Damien George
iabdalkader
robert-hh
Angus Gratton
Jim Mussared
Wanlin Wang
Felix Dörre
David Lechner
Peter Harper
Brian 'redbeard' Harrington
Ondrej Wisniewski
Phil Howard
Adam Green
Christian Clauss
David (Pololu)
David Grayson
cpottle9
Tomofumi Inoue
Jay Greco
Paul Grayson
Blake Felt