If the _IRQ_L2CAP_RECV handler does the actual consumption of the incoming
data (i.e. via l2cap_recvinto), rather than setting a flag for
non-scheduler-context to handle it later, then two things can happen:
- It can starve the VM (i.e. the scheduled task never terminates). This is
because calling l2cap_recvinto will empty the rx buffer, which will grant
more credits to the channel (an HCI command), meaning more data can
arrive. This means that the loop in hal_uart.c that keeps reading HCI
data from the uart and executing NimBLE events as they are created will
not terminate, preventing other VM code from running.
- There's no flow control (i.e. data will arrive too quickly). The channel
shouldn't be given credits until after we return from scheduler context.
It's preferable that no work is done in scheduler/IRQ context. But to
prevent this being a problem this commit changes l2cap_recvinto so that if
it is called in IRQ context, and the Python handler empties the rx buffer,
then don't grant credits until the Python handler is complete.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Don't clear the IPCC channel flag until we've actually handled the incoming
data, or else the wireless firmware may clobber the IPCC buffer if more
data arrives. This requires masking the IRQ until the data is handled.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This commit adds a new port "rp2" which targets the new Raspberry Pi RP2040
microcontroller.
The build system uses pure cmake (with a small Makefile wrapper for
convenience). The USB driver is TinyUSB, and there is a machine module
with most of the standard classes implemented. Some examples are provided
in the examples/rp2/ directory.
Work done in collaboration with Graham Sanderson.
Signed-off-by: Damien George <damien@micropython.org>
These args are already bounds checked and clipped, and using unsigned ints
can be more efficient. It also eliminates possible issues and compiler
warnings with shifting of signed integers.
Signed-off-by: Damien George <damien@micropython.org>
Adds a new compile-time option MICROPY_EMIT_THUMB_ARMV7M which is enabled
by default (to get existing behaviour) and which should be disabled (set to
0) when building native emitter support (@micropython.native) on ARMv6M
targets.
This returns a reference to the globals dict associated with the function,
ie the global scope that the function was defined in. This attribute is
read-only but the dict itself is modifiable, per CPython behaviour.
Signed-off-by: Damien George <damien@micropython.org>
It's enabled by default to retain the existing behaviour. A board can
disable this option if it manages mounting the filesystem itself, for
example in frozen code.
Signed-off-by: Damien George <damien@micropython.org>
Changes are:
- refactor to use new _create_element function
- support extended version of MOUNT element with block size
- support STATUS element
Signed-off-by: Damien George <damien@micropython.org>
This new element takes the form: (ELEM_TYPE_STATUS, 4, <address>). If this
element is present in the mboot command then mboot will store to the given
address the result of the filesystem firmware update process. The address
can for example be an RTC backup register.
Signed-off-by: Damien George <damien@micropython.org>
Instead it is now passed in as an optional parameter to the ELEM_MOUNT
element, with a compile-time configurable default.
Signed-off-by: Damien George <damien@micropython.org>
The superblock for littlefs is in block 0 and 1, but block 0 may be erased
or partially written, so block 1 must be checked if block 0 does not have a
valid littlefs superblock in it.
Prior to this commit, if block 0 did not contain a valid littlefs
superblock (but block 1 did) then the auto-detection would fail, mounting a
FAT filesystem would also fail, and the system would reformat the flash,
even though it may have contained a valid littlefs filesystem. This is now
fixed.
Signed-off-by: Damien George <damien@micropython.org>
The superblock for littlefs is in block 0 and 1, but block 0 may be erased
or partially written, so block 1 must be checked if block 0 does not have a
valid littlefs superblock in it.
Prior to this commit, the mount of a block device which auto-detected the
filysystem type would fail for littlefs if block 0 did not contain a valid
superblock. That is now fixed.
Signed-off-by: Damien George <damien@micropython.org>
This commit adds many new sections to the existing "Developing and building
MicroPython" chapter to make it all about the internals of MicroPython.
This work was done as part of Google's Season of Docs 2020.
Changes are:
- Use ubuntu-20.04 so that gcc-multilib installs without error.
- Use "fetch-depth: 100" to get history prior to pull request.
Signed-off-by: Damien George <damien@micropython.org>
* Fix a typo in the Makefile that prevented the debug build to be actually
enabled when BTYPE=debug is used.
* Add a missing header in modmachine.c that is used when a debug build is
created.
This commit improves some FTP implementation details for better
compatibility with FTP clients:
* The PWD command now puts quotes around the directory name before
returning it. This fixes BBEdit’s FTP client, which performs a PWD after
each CWD and gets confused if the returned directory path is not
surrounded by quotes.
* The FEAT command is now allowed before logging in. This fixes the lftp
client, which send FEAT first and gets confused (tries to use TLS) if the
server responds with 332.
With MICROPY_FLOAT_IMPL_FLOAT the results of utime.time(), gmtime() and
localtime() change only every 129 seconds. As one consequence
tests/extmod/vfs_lfs_mtime.py will fail on a unix port with LFS support.
With this patch these functions only return floats if
MICROPY_FLOAT_IMPL_DOUBLE is used. Otherwise they return integers.
According to documentation time() has a precision of at least 1 second.
This test runs for 2.5 seconds and calls all utime functions every 100ms.
Then it checks if they returned enough different results. All functions
with sub-second precision will return ~25 results. This test passes with
15 results or more. Functions that do not exist are skipped silently.
This allows sending arbitrary HCI commands and getting the response. The
return value of the function is the status of the command.
This is intended for debugging and not to be a part of the public API, and
must be enabled via mpconfigboard.h. It's currently only implemented for
NimBLE bindings.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
To match the definition of GENERATE_PACK_DFU, so a board can customise the
location/name of this file if needed.
Signed-off-by: Damien George <damien@micropython.org>
To have at least one board configured with MBOOT_ENABLE_PACKING, for CI
testing purposes and demonstration of the feature.
Signed-off-by: Damien George <damien@micropython.org>
This commit adds support to stm32's mboot for signe, encrypted and
compressed DFU updates. It is based on inital work done by Andrew Leech.
The feature is enabled by setting MBOOT_ENABLE_PACKING to 1 in the board's
mpconfigboard.mk file, and by providing a header file in the board folder
(usually called mboot_keys.h) with a set of signing and encryption keys
(which can be generated by mboot_pack_dfu.py). The signing and encryption
is provided by libhydrogen. Compression is provided by uzlib. Enabling
packing costs about 3k of flash.
The included mboot_pack_dfu.py script converts a .dfu file to a .pack.dfu
file which can be subsequently deployed to a board with mboot in packing
mode. This .pack.dfu file is created as follows:
- the firmware from the original .dfu is split into chunks (so the
decryption can fit in RAM)
- each chunk is compressed, encrypted, a header added, then signed
- a special final chunk is added with a signature of the entire firmware
- all chunks are concatenated to make the final .pack.dfu file
The .pack.dfu file can be deployed over USB or from the internal filesystem
on the device (if MBOOT_FSLOAD is enabled).
See #5267 and #5309 for additional discussion.
Signed-off-by: Damien George <damien@micropython.org>
Prior to this fix, the final piece of data in a compressed file may have
been lost when decompressing.
Signed-off-by: Damien George <damien@micropython.org>
This library is a small and easy-to-use cryptographic library which is well
suited to embedded systems.
Signed-off-by: Damien George <damien@micropython.org>
The original logic of reducing a full path to a relative one assumes
"tests/misc" is in the filename which is limited in usage: it never works
for CPython on Windows since that will use a backslash as path separator,
and also won't work when the filename is a path not relative to the tests
directory which happens for example in the common case of running
"./run-tests -d misc".
Fix all cases by printing only the bare filename, which requires them all
to start with sys_settrace_ hence the renaming.
Changes are:
- Remove include of stm32's adc.h because it was recently changed and is
no longer compatible with teensy (and not used anyway).
- Remove define of __disable_irq in mpconfigport.h because it was clashing
with an equivalent definition in core/mk20dx128.h.
- Add -Werror to CFLAGS, and change -std=gnu99 to -std=c99.
Signed-off-by: Damien George <damien@micropython.org>
Mboot builds do not use the external SPI flash in caching mode, and
explicitly disabling it saves RAM and a small bit of flash.
Signed-off-by: Damien George <damien@micropython.org>
This only needs to be enabled if a board uses FAT FS on external SPI flash.
When disabled (and using external SPI flash) 4k of RAM can be saved.
Signed-off-by: Damien George <damien@micropython.org>
When littlefs is enabled extended reading must be supported, and using this
function to read the first block for auto-detection is more efficient (a
smaller read) and does not require a cached SPI-flash read.
Signed-off-by: Damien George <damien@micropython.org>
These functions enable SDRAM data retention in stop mode. Example usage,
in mpconfigboard.h:
#define MICROPY_BOARD_ENTER_STOP sdram_enter_low_power();
#define MICROPY_BOARD_LEAVE_STOP sdram_leave_low_power();
