Polling mode will cause failures with the mass-erase command due to USB
timeouts, because the USB IRQs are not being serviced. Swiching from
polling to IRQ mode fixes this because the USB IRQs can be serviced between
page erases.
Note that when the flash is being programmed or erased the MCU is halted
and cannot respond to USB IRQs, because mboot runs from flash, as opposed
to the built-in bootloader which is in system ROM. But the maximum delay
in responding to an IRQ is the time taken to erase a single page, about
100ms for large pages, and that is short enough that the USB does not
timeout on the host side.
Recent tests have shown that in the current mboot code IRQ mode is pretty
much the same speed as polling mode (within timing error), code size is
slightly reduced in IRQ mode, and IRQ mode idles at about half of the power
consumption as polling mode.
This is treated more like a "delay before continuing" in the spec and
official tools and does not appear to be really needed. In particular,
downloading firmware is much slower with non-zero timeouts because the host
must pause by the timeout between sending each DFU_GETSTATUS to poll for
download/erase complete.
The implementation internally uses sector erase to wipe everything except
the sector(s) that mboot lives in (by erasing starting from
APPLICATION_ADDR).
The erase command can take some time (eg an STM32F765 with 2MB of flash
takes 8 to 10 seconds). This time is normally enough to make pydfu.py fail
with a timeout. The DFU standard includes a mechanism for the DFU device
to request a longer timeout as part of the get-status response just before
starting an operation. This timeout functionality has been implemented
here.
Before this commit the USB VCP TX ring-buffer used the basic implementation
where it can only be filled to a maximum of buffer size-1. For a 1024 size
buffer this means the largest packet that can be sent is 1023. Once a
packet of this size is sent the next byte copied in goes to the final byte
in the buffer, so must be sent as a 1 byte packet before the read pointer
can be wrapped around to the beginning. So in large streaming transfers,
watching the USB sniffer you basically get alternating 1023 byte packets
then 1 byte packets.
This commit changes the ring-buffer implementation to a scheme that doesn't
have the full-size limitation, and the USB VCP driver can now achieve a
constant stream of full-sized packets. This scheme introduces a
restriction on the size of the buffer: it must be a power of 2, and the
maximum size is half of the size of the index (in this case the index is
16-bit, so the maximum size would be 32767 bytes rounded to 16384 for a
power-of-2). But this is not a big limitation because the size of the
ring-buffer prior to this commit was restricted to powers of 2 because it
was using a mask-based method to wrap the indices.
For an explanation of the new scheme see
https://www.snellman.net/blog/archive/2016-12-13-ring-buffers/
The RX buffer could likely do with a similar change, though as it's not
read from in chunks like the TX buffer it doesn't present the same issue,
all that's lost is one byte capacity of the buffer.
USB VCP TX throughput is improved by this change, potentially doubling the
speed in certain cases.
By passing through the I2C instance to the application callbacks, the
application can implement multiple I2C slave devices on different
peripherals (eg I2C1 and I2C2).
This commit also adds a proper rw argument to i2c_slave_process_addr_match
for F7/H7/WB MCUs, and enables the i2c_slave_process_tx_end callback.
Mboot is also updated for these changes.
Signed-off-by: Damien George <damien@micropython.org>
Mboot now supports FAT, LFS1 and LFS2 filesystems, to load firmware from.
The filesystem needed by the board must be explicitly enabled by the
configuration variables MBOOT_VFS_FAT, MBOOT_VFS_LFS1 and MBOOT_VFS_LFS2.
Boards that previously used FAT implicitly (with MBOOT_FSLOAD enabled) must
now add the following config to mpconfigboard.h:
#define MBOOT_VFS_FAT (1)
Signed-off-by: Damien George <damien@micropython.org>
This commit factors the code for files and streaming to separate source
files (vfs_fat.c and gzstream.c respectively) and introduces an abstract
gzstream interface to make it easier to plug in different filesystems.
Signed-off-by: Damien George <damien@micropython.org>
There's no need to do a directory listing to search for the given firmware
filename, it just takes extra time and code size. Instead this commit
changes it so that the requested firmware file is opened immediately and
will abort if the file couldn't be opened. This also allows to specify
files in a directory.
Signed-off-by: Damien George <damien@micropython.org>
Previously, if FAT was not enabled but LFS1/2 was then MICROPY_PY_IO_FILEIO
would be disabled and file binary-mode was not supported.
Signed-off-by: Damien George <damien@micropython.org>
Commit 8675858465 switched to using the CMSIS
provided SystemInit function which sets VTOR to 0x00000000 (previously it
was 0x08000000). A VTOR of 0x00000000 will be correct on some MCUs but not
on others where the built-in bootloader is remapped to this address, via
__HAL_SYSCFG_REMAPMEMORY_SYSTEMFLASH().
To make sure mboot has the correct vector table, this commit explicitly
sets VTOR to the correct value of 0x08000000.
Signed-off-by: Damien George <damien@micropython.org>
There's no need to duplicate this functionality in mboot, the code provided
in stm32lib/CMSIS does the same thing and makes it easier to support other
MCU series.
Signed-off-by: Damien George <damien@micropython.org>
The flash functions in ports/stm32/flash.c are almost identical to those in
ports/stm32/mboot/main.c, so remove the duplicated code in mboot and use
instead the main stm32 code. This also allows supporting other MCU series.
Signed-off-by: Damien George <damien@micropython.org>
This commit makes the low-level flash C functions usable by code other than
flashbdev.c (eg by mboot). Changes in this commit are:
- flash_erase() and flash_write() now return an errno error code, a
negative value on error.
- flash_erase() now automatically locks the flash, as well as unlocking it.
- flash_write() now automatically unlocks the flash, as well as locking it.
- flashbdev.c is modified for the above changes.
Signed-off-by: Damien George <damien@micropython.org>
irq.h is included by py/mphal.h but it's better to be explicit, eg if mboot
uses powerctrlboot.c.
Signed-off-by: Damien George <damien@micropython.org>
The irq.h file now just provides low-level IRQ definitions and priorities.
All Python binding definitions are moved to modmachine.h, with some
renaming of pyb -> machine, and also the machine_idle definition (was
pyb_wfi) is moved to modmachine.c.
The cc3200 and teensy ports are updated to build with these changes.
Signed-off-by: Damien George <damien@micropython.org>
There are a maximum of 8 USB endpoints and each has 2 buffer slots
(in/out). This commit add support for up to 8 endpoints and adds FIFO
configuration for USB profiles with 2xVCP on MCUs that have device-only USB
peripherals.
Tested on NUCLEO_WB55 in 2xVCP, 2xVCP+MSC and 2xVCP+MSC+HID mode.
Signed-off-by: Damien George <damien@micropython.org>
No functionality change is intended with this commit, it just consolidates
the separate implementations of GC helper code to the lib/utils/ directory
as a general set of helper functions useful for any port. This reduces
duplication of code, and makes it easier for future ports or embedders to
get the GC implementation correct.
Ports should now link against gchelper_native.c and either gchelper_m0.s or
gchelper_m3.s (currently only Cortex-M is supported but other architectures
can follow), or use the fallback gchelper_generic.c which will work on
x86/x64/ARM.
The gc_helper_get_sp function from gchelper_m3.s is not really GC related
and was only used by cc3200, so it has been moved to that port and renamed
to cortex_m3_get_sp.
Note: the uncrustify configuration is explicitly set to 'add' instead of
'force' in order not to alter the comments which use extra spaces after //
as a means of indenting text for clarity.
In mboot, the ability to override the USB vendor/product id's was added
back in 5688c9ba09. However, when the main
firmware is turned into a DFU file the default VID/PID are used there.
pydfu.py doesn't care about this but dfu-util does and prevents its use
when the VID/PID don't match.
This commit exposes BOOTLOADER_DFU_USB_VID/PID as make variables, for use
on either command line or mpconfigboard.mk, to set VID/PID in both mboot
and DFU files.
Add -Wdouble-promotion and -Wfloat-conversion for most ports to ban out
implicit floating point conversions, and add extra Travis builds using
MICROPY_FLOAT_IMPL_FLOAT to uncover warnings which weren't found
previously. For the unix port -Wsign-comparison is added as well but only
there since only clang supports this but gcc doesn't.
Now that error string compression is supported it's more important to have
consistent error string formatting (eg all lowercase English words,
consistent contractions). This commit cleans up some of the strings to
make them more consistent.
These are mainly used by the previous version of uasyncio which is now
replaced by a newer version, with built-in C module _uasyncio. Saves about
1300 bytes of flash.
This function is not used by the core but having it as part of the build
allows it to be used by user C modules, or board extensions. The linker
won't include it in the final firmware if it remains unused.
The "led" argument is always a pointer to the GPIO port, or'd with the pin
that the LED is on, so testing that it is "1" is unnecessary. The type of
"led" is also changed to uint32_t so it can properly hold a 32-bit pointer.
Updating the LED0 state from systick handler ensures LED0 is always
consistent with its flash rate regardless of other processing going on in
either interrupts or main. This improves the visible stability of the
bootloader, rather than LED0 flashing somewhat randomly at times.
This commit also changes the LED0 flash rate depending on the current state
of DFU, giving slightly more visual feedback on what the device is doing.
This makes a cleaner separation between the: driver, HCI UART and BT stack.
Also updated the naming to be more consistent (mp_bluetooth_hci_*).
Work done in collaboration with Jim Mussared aka @jimmo.
Move extmod/modbluetooth_nimble.* to extmod/nimble. And move common
Makefile lines to extmod/nimble/nimble.mk (which was previously only used
by stm32). This allows (upcoming) btstack to follow a similar structure.
Work done in collaboration with Jim Mussared aka @jimmo.
This string is recognised by uncrustify, to disable formatting in the
region marked by these comments. This is necessary in the qstrdef*.h files
to prevent modification of the strings within the Q(...). In other places
it is used to prevent excessive reformatting that would make the code less
readable.
This provides a more consistent C-level API to raise exceptions, ie moving
away from nlr_raise towards mp_raise_XXX. It also reduces code size by a
small amount on some ports.
It's not needed. The C integer implicit promotion rules mean that the
uint8_t of the incoming character is promoted to a (signed) int, matching
the type of interrupt_char. Thus the uint8_t incoming character can never
be equal to -1 (the value of interrupt_char that indicate that interruption
is disabled).
Previous behaviour is when this argument is set to "true", in which case
the function will raise any pending exception. Setting it to "false" will
cancel any pending exception.
This commit improves pllvalues.py to generate PLL values for H7 MCUs that
are valid (VCO in and out are in range) and extend for the entire range of
SYSCLK values up to 400MHz (up to 480MHz is currently unsupported).
This board now has the following 3 build configurations:
- mboot + external QSPI in XIP mode + internal filesystem
- mboot + external QSPI with filesystem (the default)
- no mboot + external QSPI with filesystem
With a SPI flash that has more than 16MB, 32-bit addressing is required
rather than the standard 24-bit. This commit adds support for 32-bit
addressing so that the SPI flash commands (read/write/erase) are selected
automatically depending on the size of the address being used at each
operation.
This modifies the signature of mp_thread_set_state() to use
mp_state_thread_t* instead of void*. This matches the return type of
mp_thread_get_state(), which returns the same value.
`struct _mp_state_thread_t;` had to be moved before
`#include <mpthreadport.h>` since the stm32 port uses it in its
mpthreadport.h file.
PLLM is shared among all PLL blocks on F7 MCUs, and this calculation to
configure PLLSAI to have 48MHz on the P output previously assumed that PLLM
is equal to HSE (eg PLLM=25 for HSE=25MHz). This commit relaxes this
assumption to allow other values of PLLM.
Most types are in rodata/ROM, and mp_obj_base_t.type is a constant pointer,
so enforce this const-ness throughout the code base. If a type ever needs
to be modified (eg a user type) then a simple cast can be used.
This change has the following effects:
- Reduces the resolution of the RTC sub-second counter from 30.52us to
122.07us.
- Allows RTC.calibration() to now support positive values (as well as
negative values).
- Reduces VBAT current consumption in standby mode by a small amount.
For general purpose use 122us resolution of the sub-second counter is
good enough, and the benefits of full range calibration and minor reduction
in VBAT consumption are worth the change.
Instances of the slice class are passed to __getitem__() on objects when
the user indexes them with a slice. In practice the majority of the time
(other than passing it on untouched) is to work out what the slice means in
the context of an array dimension of a particular length. Since Python 2.3
there has been a method on the slice class, indices(), that takes a
dimension length and returns the real start, stop and step, accounting for
missing or negative values in the slice spec. This commit implements such
a indices() method on the slice class.
It is configurable at compile-time via MICROPY_PY_BUILTINS_SLICE_INDICES,
disabled by default, enabled on unix, stm32 and esp32 ports.
This commit also adds new tests for slice indices and for slicing unicode
strings.
For the 3 ports that already make use of this feature (stm32, nrf and
teensy) this doesn't make any difference, it just allows to disable it from
now on.
For other ports that use pyexec, this decreases code size because the debug
printing code is dead (it can't be enabled) but the compiler can't deduce
that, so code is still emitted.
Most stm32 boards can now be built in nan-boxing mode via:
$ make NANBOX=1
Note that if float is enabled then it will be forced to double-precision.
Also, native emitters will be disabled.
- Corrected pin assignments and checked with CubeMX.
- Added additional I2C and UARTs.
- Added Ethernet interface definitions with lwIP and SSL support (but
Ethernet is currently unsupported on H7 MCUs so not fully enabled).
- Removed remarks on DFU/OCD in mpconfigboard.h because deploy-stlink works
fine too.
- Added more UARTs, I2C, corrected SPI, CAN, etc; verified against CubeMX.
- Adapted pins.csv to remove errors, add omissions, etc. according to
NUCLEO-144 User Manual.
- Changed linker file stm32f767.ld to reflect correct size of the Flash.
- Tested with LAN and SD card.
The Nucleo board does not have an SD card slot but does have the requisite
pins next to each other and labelled, so provide the configuration for
convenience.
The default protection for the BLE ringbuf is to use
MICROPY_BEGIN_ATOMIC_SECTION, which disables all interrupts. On stm32 it
only needs to disable the lowest priority IRQ, pendsv, because that's the
IRQ level at which the BLE stack is driven.
qstrs in this file are always included in all builds, even if not used
anywhere. So remove those that are never needed, and make USB names
conditional on having USB enabled.
And return -MP_EIO if calling storage_read_block/storage_write_block fails.
This lines up with the return type and value (negative for error) of the
calls to MICROPY_HW_BDEV_READBLOCKS (and WRITEBLOCKS, and BDEV2 versions).
The pyb.Flash() class can now be used to construct objects which reference
sections of the flash storage, starting at a certain offset and going for a
certain length. Such objects also support the extended block protocol.
The signature for the constructor is: pyb.Flash(start=-1, len=-1).
This commit refactors and generalises the boot-mount routine on stm32 so
that it can mount filesystems of arbitrary type. That is, it no longer
assumes that the filesystem is FAT. It does this by using mp_vfs_mount()
which does auto-detection of the filesystem type.
Using mp_hal_delay_ms allows the scheduler to run, which might result in
another transmit operation happening, which would bypass the sleep (and
fail). Use mp_hal_delay_us instead.
This commit removes the Makefile-level MICROPY_FATFS config and moves the
MICROPY_VFS_FAT config to the Makefile level to replace it. It also moves
the include of the oofatfs source files in the build from each port to a
central place in extmod/extmod.mk.
For a port to enabled VFS FAT support it should now set MICROPY_VFS_FAT=1
at the level of the Makefile. This will include the relevant oofatfs files
in the build and set MICROPY_VFS_FAT=1 at the C (preprocessor) level.
This commit adds an implementation of machine.Timer backed by the soft
timer mechanism. It allows an arbitrary number of timers with 1ms
resolution, with an associated Python callback. The Python-level API
matches existing ports that have a soft timer, and is used as:
from machine import Timer
t = Timer(freq=10, callback=lambda t:print(t))
...
t = Timer(mode=Timer.ONE_SHOT, period=2000, callback=lambda t:print(t))
...
t.deinit()
This commit adds an implementation of a "software timer" with a 1ms
resolution, using SysTick. It allows unlimited number of concurrent
timers (limited only by memory needed for each timer entry). They can be
one-shot or periodic, and associated with a Python callback.
There is a very small overhead added to the SysTick IRQ, which could be
further optimised in the future, eg by patching SysTick_Handler code
dynamically.
For consistency with "umachine". Now that weak links are enabled
by default for built-in modules, this should be a no-op, but allows
extension of the bluetooth module by user code.
Also move registration of ubluetooth to objmodule rather than
port-specific.
This commit implements automatic module weak links for all built-in
modules, by searching for "ufoo" in the built-in module list if "foo"
cannot be found. This means that all modules named "ufoo" are always
available as "foo". Also, a port can no longer add any other weak links,
which makes strict the definition of a weak link.
It saves some code size (about 100-200 bytes) on ports that previously had
lots of weak links.
Some changes from the previous behaviour:
- It doesn't intern the non-u module names (eg "foo" is not interned),
which saves code size, but will mean that "import foo" creates a new qstr
(namely "foo") in RAM (unless the importing module is frozen).
- help('modules') no longer lists non-u module names, only the u-variants;
this reduces duplication in the help listing.
Weak links are effectively the same as having a set of symbolic links on
the filesystem that is searched last. So an "import foo" will search
built-in modules first, then all paths in sys.path, then weak links last,
importing "ufoo" if it exists. Thus a file called "foo.py" somewhere in
sys.path will still have precedence over the weak link of "foo" to "ufoo".
See issues: #1740, #4449, #5229, #5241.
When loading a manifest file, e.g. by include(), it will chdir first to the
directory of that manifest. This means that all file operations within a
manifest are relative to that manifest's location.
As a consequence of this, additional environment variables are needed to
find absolute paths, so the following are added: $(MPY_LIB_DIR),
$(PORT_DIR), $(BOARD_DIR). And rename $(MPY) to $(MPY_DIR) to be
consistent.
Existing manifests are updated to match.
Prior to this commit the systick IRQ priority was set at lowest priority on
F0/L0/WB MCUs, because it was left at the default and never configured.
This commit ensures the priority is configured and sets it to the highest
priority.
On other ports (e.g. ESP32) they provide a complete Nimble implementation
(i.e. we don't need to use the code in extmod/nimble). This change
extracts out the bits that we don't need to use in other ports:
- malloc/free/realloc for Nimble memory.
- pendsv poll handler
- depowering the cywbt
Also cleans up the root pointer management.
STM32F0 has PCLK=48MHz and maximum ADC clock is 14MHz so use PCLK/4=12MHz
to stay within spec of the ADC peripheral. In pyb.ADC set common sampling
time to approx 4uS for internal and external sources. In machine.ADC
reduce sample time to approx 1uS for external source, leave internal at
maximum sampling time.
This commit adds the option to use HSE or MSI system clock, and LSE or LSI
RTC clock, on L4 MCUs.
Note that prior to this commit the default clocks on an L4 part were MSI
and LSE. The defaults are now MSI and LSI.
In mpconfigboard.h select the clock source via:
#define MICROPY_HW_RTC_USE_LSE (0) or (1)
#define MICROPY_HW_CLK_USE_HSE (0) or (1)
and the PLLSAI1 N,P,Q,R settings:
#define MICROPY_HW_CLK_PLLSAIN (12)
#define MICROPY_HW_CLK_PLLSAIP (RCC_PLLP_DIV7)
#define MICROPY_HW_CLK_PLLSAIQ (RCC_PLLQ_DIV2)
#define MICROPY_HW_CLK_PLLSAIR (RCC_PLLR_DIV2)
For use with F0 MCUs that don't have HSI48. Select the clock source
explicitly in mpconfigboard.h.
On the NUCLEO_F091RC board use HSE bypass when HSE is chosen because the
NUCLEO clock source is STLINK not a crystal.
Before this patch the UART baudrate on F0 MCUs was wrong because the
stm32lib SystemCoreClockUpdate sets SystemCoreClock to 8MHz instead of
48MHz if HSI48 is routed directly to SYSCLK.
The workaround is to use HSI48 -> PREDIV (/2) -> PLL (*2) -> SYSCLK.
Fixes issue #5049.
Enabled by default, but disabled when REPL is connected to the VCP (this is
the existing behaviour). Can be configured at run-time with, eg:
pyb.USB_VCP().init(flow=pyb.USB_VCP.RTS | pyb.USB_VCP.CTS)
The new fdcan.c file provides the low-level C interface to the FDCAN
peripheral, and pyb_can.c is updated to support both traditional CAN and
FDCAN, depending on the MCU being compiled for.
According to the schematic, the SDRAM part on this board is a
MT48LC4M32B2B5-6A, with "Row addressing 4K A[11:0]" (per datasheet). This
commit updates mpconfigboard.h from 13 to 12 to match.
- STM32F407VGT6 (1MB of Flash, 192+4 Kbytes of SRAM)
- 5V (via USB) or Li-Polymer Battery (3.7V) power input
- 2 x LEDs
- 2 x user switches
- 2 x mikroBUS sockets
- 2 x 1x26 mikromedia-compatible headers (52 pins)
https://www.mikroe.com/clicker-2-stm32f4
Mboot currently requires at least three LEDs to display each of the four
states. However, since there are only four possible states, the states can
be displayed via binary counting on only 2 LEDs (if only 2 are available).
The existing patterns are still used for 3 or 4 LEDs.
As per PEP 485, this function appeared in for Python 3.5. Configured via
MICROPY_PY_MATH_ISCLOSE which is disabled by default, but enabled for the
ports which already have MICROPY_PY_MATH_SPECIAL_FUNCTIONS enabled.
Before this patch I2C transactions using a hardware I2C peripheral on F0/F7
MCUs would not correctly generate the I2C restart condition, and instead
would generate a stop followed by a start. This is because the CR2 AUTOEND
bit was being set before CR2 START when the peripheral already had the I2C
bus from a previous transaction that did not generate a stop.
As a consequence all combined transactions, eg read-then-write for an I2C
memory transfer, generated a stop condition after the first transaction and
didn't generate a stop at the very end (but still released the bus). Some
I2C devices require a repeated start to function correctly.
This patch fixes this by making sure the CR2 AUTOEND bit is set after the
start condition and slave address have been fully transferred out.
Some SD/MMC breakout boards don't support 4-bit bus mode. This adds a new
macro MICROPY_HW_SDMMC_BUS_WIDTH that allows each board to define the width
of the SD/MMC bus interface used on that board, defaulting to 4 bits.
The previous version did not work on MCUs that only had USB device mode
(compared to OTG) because of the handling of NAK. And this previous
handling of NAK had a race condition where a new packet could come in
before USBD_HID_SetNAK was called (since USBD_HID_ReceivePacket clears NAK
as part of its operation). Furthermore, the double buffering of incoming
reports was not working, only one buffer could be used at a time.
This commit rewrites the HID interface code to have a single incoming
buffer, and only calls USBD_HID_ReceivePacket after the user has read the
incoming report (similar to how the VCP does its flow control). As such,
USBD_HID_SetNAK and USBD_HID_ClearNAK are no longer needed.
API functionality from the user's point of view should be unchanged with
this commit.
This new series of MCUs is similar to the L4 series with an additional
Cortex-M0 coprocessor. The firmware for the wireless stack must be managed
separately and MicroPython does not currently interface to it. Supported
features so far include: RTC, UART, USB, internal flash filesystem.
The new configurations MICROPY_HW_USB_MSC and MICROPY_HW_USB_HID can be
used by a board to enabled or disable MSC and/or HID. They are both
enabled by default.
In a non-thread build, using &_ram_end as the top-of-stack is no longer
correct because the stack is not always at the very top end of RAM. See
eg 04c7cdb668 and
3786592097. The correct value to use is
&_estack, which is the value stored in MP_STATE_THREAD(stack_top), and
using the same code for both thread and non-thread builds makes the code
cleaner.
