Boosted performance, board.json metadata, more mimxrt, rp2, samd features
This release of MicroPython sees a boost to the overall performance of the
VM and runtime. This is achieved by the addition of an optional cache to
speed up general hash table lookups, as well as a fast path in the VM for
the LOAD_ATTR opcode on instance types. The new configuration options are
MICROPY_OPT_MAP_LOOKUP_CACHE and MICROPY_OPT_LOAD_ATTR_FAST_PATH. As part
of this improvement the MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE option has
been removed, which provided a similar map caching mechanism but with the
cache stored in the bytecode, which made it not useful on bare metal ports.
The new mechanism is measured to be at least as good as the old one,
applies to more map lookups, has a constant RAM overhead, and applies to
native code as well as bytecode.
These performance options are enabled on the esp32, mimxrt, rp2, stm32 and
unix ports. For esp32 and mimxrt some code is also moved to RAM to further
boost performance. On stm32, performance increases by about 20% for
benchmarks that are heavy on name lookups, like misc_pystone.py and
misc_raytrace.py. On esp32 performance can increase by 2-3x, and on mimxrt
it is up to 6x.
All boards in all ports now have a board.json metadata file, which is used
to automatically build firmware and generate a webpage for that board
(among other possibilities). Auto-build scripts have been added for this
purpose and they build all esp32, mimxrt, rp2, samd and stm32 boards. The
generated output is available at https://micropython.org/download.
Support for FROZEN_DIR and FROZEN_MPY_DIR has been deprecated for some time
and was finally removed in this release. Instead of these, FROZEN_MANIFEST
can be used. The io.resource_stream() function is also removed, replaced
by the pure Python version in micropython-lib.
The search order for importing frozen Python modules is now controlled by
the ".frozen" entry in sys.path. This string is added by default in the
second position in sys.path. User code should adjust sys.path depending on
the desired behaviour. Putting ".frozen" first in sys.path will speed up
importing frozen modules.
A bug in multiple precision integers with bitwise of -0 was fixed in commit
2c139bbf4e.
The platform module has been added to allow querying the compiler and
underlying SDK/HAL/libc version. This is enabled on esp32, mimxrt and
stm32 ports.
The mpremote tool now supports seek, flush, mkdir and rmdir on PC-mounted
filesystems. And a help command has been added.
The documentation has seen many additions and improvements thanks (for a
second time) to the Google Season of Docs project. The rp2 documentation
now includes a reference for PIO assembly instructions, a PIO quick
reference and a PIO tutorial. The random and stm modules have been
documented, along with sys.settrace, manifest.py files and mpremote. There
is also now more detail about the differences between MicroPython and
standard Python 3.5 and above.
The esp32 port sees support for ESP32-S3 SoCs, and new boards GENERIC_S3,
ESP32_S2_WROVER, LOLIN_S2_MINI, LOLIN_S2_PICO and UM_FEATHERS2NEO. The PWM
driver has been improved and now supports all PWM timers and channels, and
the duty_u16() and duty_ns() methods, and it keeps the duty constant when
changing frequency. The machine.bitstream() function has been improved to
use RMT, with an option to select the original bit-banging implementation.
The mimxrt port gained new hardware features: SDRAM and SD card support, as
well as network integration with a LAN driver. The machine.WDT class was
added along with the machine.reset_cause(), machine.soft_reset(),
machine.unique_id() add machine.bitstream() functions. DHT sensor support
was added, and f-strings were enabled.
The rp2 port now has support for networking, and bluetooth using NimBLE.
The Nina-W10 WiFi/BT driver is fully integrated and supported by the new
Arduino Nano RP2040 connect board. I2S protocol support is added along
with a machine.bitstream() driver and DHT sensor support. The PWM driver
had a bug fix with the accuracy of setting/getting the frequency, and the
duty value is now retained when changing the frequency.
On the samd port there is now support for the internal flash being a block
device, and for filesystems and the os module. Pin and LED classes have
been implemented. There are more time functions, more Python features
enabled, and the help() function is added. SEEED_WIO_TERMINAL and
SEEED_XIAO board definitions are now available.
The stm32 port now has support for F427, F479 and H7A3(Q)/H7B3(Q) MCUs, and
new board definitions for VCC_GND_H743VI, OLIMEX_H407, MIKROE_QUAIL,
GARATRONIC_PYBSTICK26_F411, STM32H73B3I_DK. A bug was fixed in the SPI
driver where a SPI transfer could fail if the CYW43 WiFi driver was also
active at the same time.
On the windows port the help() function has been enabled, and support for
build variants added, to match the unix port.
The zephyr port upgraded Zephyr to v2.7.0.
The change in code size since the previous release for various ports is
(absolute and percentage change in the text section):
bare-arm: -1520 -2.605%
minimal x86: -2256 -1.531%
unix x64: -457 -0.089%
unix nanbox: -925 -0.204%
stm32: +312 +0.079% PYBV10
cc3200: -176 -0.096%
esp8266: +532 +0.076% GENERIC
esp32: +27096 +1.820% GENERIC
nrf: -212 -0.121% pca10040
rp2: +9904 +2.051% PICO
samd: +35332 +33.969% ADAFRUIT_ITSYBITSY_M4_EXPRESS
The changes that dominate these numbers are:
- bare-arm, minimal: use of new MICROPY_CONFIG_ROM_LEVEL_MINIMUM option and
subsequent disabling of remaining optional features
- unix, cc3200, nrf: general code size reductions of the core
- stm32: performance improvements, addition of platform module
- esp8266: enabling f-strings
- esp32: use of -O2 instead of -Os
- rp2: machine.I2S and other new hardware features
- samd: filesystem support and other new hardware features
Thanks to everyone who contributed to this release: Alan Dragomirecký,
Alexey Shvetsov, Andrew Leech, Andrew Scheller, Antoine Aubert, Boris
Vinogradov, Chris Boudacoff, Chris Fiege, Christian Decker, Damien George,
Daniel Gorny, Dave Hylands, David Michieli, Emilie Feral, Frédéric Pierson,
gibbonsc, Henk Vergonet, iabdalkader, Ihor Nehrutsa, Jan Hrudka, Jan Staal,
jc_.kim, Jim Mussared, Jonathan Hogg, Laurens Valk, leo chung, Lorenzo
Cappelletti, Magnus von Wachenfeldt, Matt Trentini, Matt van de Werken,
Maureen Helm, Michael Bentley, Michael Buesch, Mike Causer, Mike Teachman,
Mike Wadsten, Ned Konz, NitiKaur, oli, patrick, Patrick Van Oosterwijck,
Peter Boin, Peter Hinch, Peter van der Burg, Philipp Ebensberger, Pooya
Moradi, retsyo, robert-hh, roland van straten, Scott Armitage, Sebastian
Wicki, Seon Rozenblum, Sergei Silnov, Simon Baatz, Stewart Bonnick, stijn,
Tobias Thyrrestrup, Tomas Vanek, YoungJoon Chun.
What follows is a detailed list of changes, generated from the git commit
history, and organised into sections.
Main components
===============
all:
- remove MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE
- update Python formatting to latest Black version 21.12b0
- remove support for FROZEN_DIR and FROZEN_MPY_DIR
py core:
- parse: simplify parse nodes representing a list
- emitnative: ensure load_subscr does not clobber existing REG_RET
- mpconfig.h: define initial templates for "feature levels"
- vm: add a fast path for LOAD_ATTR on instance types
- map: add an optional cache of (map+index) to speed up map lookups
- builtinimport: forward all debug printing to MICROPY_DEBUG_PRINTER
- add wrapper macros so hot VM functions can go in fast code location
- runtime: fix crash when exc __new__ doesn't return an exc instance
- mpconfig.h: define the "extra" feature level
- mpconfig.h: revert MICROPY_REPL_INFO to disabled at all levels
- gc: add hook to run code during time consuming GC operations
- showbc: print unary-op string when dumping bytecode
- modsys: replace non-ASCII quote char with ASCII char
- runtime: allow types to use both .attr and .locals_dict
- lexer: support nested [] and {} characters within f-string params
- objfun.h: remove obsolete comments about entries in extra_args
- builtinimport: refactor module importing
- showbc: fix printing of raw bytecode header on nanbox builds
- modio: remove io.resource_stream function
- only search frozen modules when '.frozen' is found in sys.path
- mkrules.cmake: set frozen preprocessor defs early
- runtime: allow initialising sys.path/argv with defaults
- mpstate.h: only include sys.path/argv objects in state when enabled
- mpz: fix bugs with bitwise of -0 by ensuring all 0's are positive
- qstr: reset mpstate.qstr_last_chunk before raising an error
- modbuiltins: add additional macro for extending builtins
- mpconfig.h: define MICROPY_PY_USSL_FINALISER only if not defined
extmod:
- machine_i2c: make SoftI2C configurable via macro option
- machine_spi: make SoftSPI configurable via macro option
- modonewire: make _onewire module configurable via macro option
- machine_pwm: factor out machine.PWM bindings to common code
- move modnetwork and modusocket from stm32 to extmod
- modnetwork: add STA_IF and AP_IF constants
- modnetwork: add extended socket state
- modusocket: add read/write stream methods to socket object
- modnetwork: define network interfaces in port config files
- network_cyw43: make consistent use of STA and AP constants
- modnetwork: remove STM32 references
- modnetwork: remove modnetwork socket u_state member
- mpbthci.h: add mp_bluetooth_hci_uart_any prototype
- nimble: add nimble CMake fragment file
- add platform module
- moduplatform: improve implementation for PC ports
- vfs_posix_file: support MP_STREAM_POLL in vfs_posix_file_ioctl
- modbluetooth: add connection interval to gap_connect
- nimble: update to NimBLE v1.4
- nimble: remove workaround for OS_ENOMEM
- uasyncio: fix gather returning exceptions from a cancelled task
- uplatform: remove unused definitions
- uplatform: use generic custom platform string
- network_ninaw10: fix scan list order to match other NICs
- modbluetooth: support gap_connect(None) to cancel a connection
- modure: redirect regex debug printing to mp_printf
- network_ninaw10: fix config of AP mode
- network_ninaw10: disable active connections before connecting
- network_ninaw10: make NIC state persistent
- network_ninaw10: return -1 on timeout from recv/send
- network_ninaw10: make recv/recvfrom interchangeable
- moduplatform: detect xtensa arch
- modusocket: allow setting timeout on unbound sockets
- modusocket: initialise accepted socket state
- network_ninaw10: use socket timeout preset in modusocket
- modbluetooth: fix conditional compilation of ringbuf_put_uuid
- modbluetooth: put declaration of connect_cancel in correct place
shared:
- libc/string0: don't include string.h, and provide __memcpy_chk
- runtime/pyexec: cleanup EXEC_FLAG flag constants
drivers:
- ninaw10: add ublox Nina-W10 WiFi/BT module driver
- lsm6dsox: add LSM6DSOX driver and examples
- neopixel: avoid heap alloc in fill()
- ninaw10: fix BSSID byte order, and add null byte to ESSID
- ninaw10/nina_wifi_drv: fix DNS resolution
mpy-cross: no changes specific to this component/port
lib:
- mynewt-nimble: switch to the MicroPython fork of NimBLE
- asf4: point submodule to latest commit on circuitpython branch
- update pico-sdk to 1.3.0 and tinyusb to 0.12.0
- stm32lib: update library for L4 v1.17.0, new G4, WL, and MMC fixes
- stm32lib: update library for fix to F7 USB HS
Support components
==================
docs:
- library/os.rst: clarify littlefs requirements for block erase
- library/bluetooth.rst: update incorrect link to gatts_write
- make.bat: change Windows output dir from '_build' to 'build'
- library/machine.I2S.rst: specify that I2S.shift args are kw-only
- esp32: explain ESP32 PWM modes, timers, and channels
- rp2: add reference for PIO assembly instructions, and PIO tutorial
- library/random.rst: document the random module
- reference/mpremote.rst: add docs for mpremote
- reference/manifest.rst: add docs for manifest.py files
- library/stm.rst: document the stm module
- esp32/tutorial: add an example of peripheral control via regs
- rp2/general.rst: fix typo with missing spaces
- library/framebuf.rst: adjust dimensions in example
- library/rp2.rst: update function asm_pio_encode to add sideset_opt
- reference/filesystem.rst: add detail on how to use littlefs fuse
- rp2/quickref.rst: add section on PIO
- library/sys.rst: add docs for sys.settrace
- esp8266/tutorial: fix comments of FrameBuffer examples
- library/uasyncio.rst: detail exception behaviour in cancel/timeout
- library/machine.Timer.rst: document 'id' as positional-only arg
- library/machine.SPI.rst: add example SPI usage
- library/machine.Timer.rst: document `period` and `callback` args
- library/machine.Pin.rst: add Pin.ANALOG mode constant
- remove trailing spaces and convert tabs to spaces
- library/sys.rst: add note about '.frozen' as an entry in sys.path
- differences: document details of new PEPs/features in Python 3.5+
- update copyright year range to include 2022
- esp32: update RMT quickref example to match latest code
examples: no changes specific to this component/port
tests:
- perf_bench: use math.log instead of math.log2
- basics: add tests for type-checking subclassed exc instances
- micropython/const.py: add comment about required config for test
- cpydiff: clarify f-string diffs regarding concatenation
- basics/int_big_cmp.py: add more tests for big-int comparison
- extmod: skip uselect_poll_udp when poll() is not available
tools:
- autobuild: add auto build for GENERIC_C3_USB
- ci.sh: use IDF v4.4 as part of esp32 CI and build GENERIC_S3
- autobuild: add the MIMXRT1010_EVK board to autobuild
- ci.sh: use a specific ESP IDF v4.4 commit
- autobuild: add script to generate website board metadata
- dfu.py: make tool work with python3 when parsing DFU files
- autobuild: automatically build all mimxrt, rp2 and samd boards
- autobuild: automatically build all stm32 boards
- mpremote: implement seek and flush in ioctl method
- autobuild: automatically build all esp32 boards
- upip.py: support == to specify exact package version
- makemanifest.py: make str conversion compatible with Python 2
- makemanifest.py: merge make-frozen.py
- mpremote: add mkdir and rmdir to RemoteFS
- mpremote: add help command
- mpremote: add link to mpremote docs URL in help message
- upip.py: skip '.frozen' entry in sys.path for install path
- autobuild: build esp8266 OTA image with GENERIC_1M board
- ci.sh: upgrade Zephyr docker image to v0.21.0
- ci.sh: build zephyr nucleo_wb55rg to test zephyr bluetooth build
CI:
- workflows: use Python 3.8 for macos workflow
- workflows: add new workflow to build ports download metadata
The ports
=========
all ports:
- add board.json for all boards
- add images, features and urls to board.json
- add '.frozen' as the first entry in sys.path
- move '.frozen' to second entry in sys.path
bare-arm port:
- mpconfigport.h: use MICROPY_CONFIG_ROM_LEVEL_MINIMUM
- mpconfigport.h: disable remaining optional features
cc3200 port: no changes specific to this component/port
esp8266 port:
- boards/GENERIC: enable f-strings
- extract qstr from object when comparing keys in config()
- etshal.h: remove unneeded function declarations
- allow building a board to any dest directory
esp32 port:
- boards: add new FeatherS2-Neo board definition
- machine_timer: use tx_update member for IDF 4.4 and above
- add support for ESP32-S3 SoCs
- boards: add new GENERIC_S3 board definition
- machine_hw_spi: fix hardware SPI DMA channels for S2/S3
- boards: add board definition for ESP32-S2-WROVER module
- boards: add LOLIN_S2_MINI ESP32-S2 board
- machine_pwm: add support for all PWM timers and channels
- README: updated readme with req IDF vers for ESP32-S2, C3 and S3
- usb: add USB host connection detection for CDC serial output
- machine_pin: block out IO16 and IO17 when using SPIRAM on ESP32
- mpthreadport: fix TCB cleanup function so thread_mutex is ready
- main: add option for a board to hook code into startup sequence
- split out WLAN code from modnetwork.c to network_wlan.c
- enable optimisations and move code to iRAM to boost performance
- usb: improve speed of USB CDC output
- add specific deploy_s2.md instructions for esp32-s2
- boards/LOLIN_S2_MINI: add image to board.json
- boards: update board and deploy metadata for UM_xxx boards
- usb: further improve speed of USB CDC output
- boards/LOLIN_S2_PICO: add LOLIN_S2_PICO board definition files
- boards/ESP32_S2_WROVER: link to specific deploy_s2 instructions
- support building with latest IDF v5
- in machine_i2s, send null samples in underflow situations
- in machine_i2s, make object reference arrays root pointers
- add SDCard support for S3, and a GENERIC_S3_SPIRAM board
- boards/GENERIC_S3: enable BLE on ESP32 S3
- machine_pwm: implement duty_u16() and duty_ns() PWM methods
- extract qstr from object when comparing keys in config()
- machine_pin: make GPIO 26 usable for S2,S3 if SPIRAM not config'd
- machine_hw_spi: fix SPI default pins reordering on ESP32-S2/S3
- machine_hw_spi: set proper default SPI(id=1) pins on S2,S3 and C3
- machine_hw_spi: set proper default SPI(id=2) pins on S2 and S3
- boards: remove SPI pin defaults from GENERIC S2/S3 boards
- modnetwork: synchronize WiFi AUTH_xxx constants with IDF values
- machine_pwm: keep duty constant when changing frequency
- machine_bitstream: replace bit-bang code with RMT-based driver
- machine_i2s: add support for ESP-IDF 4.4
- machine_bitstream: fix signal duplication on output pins
- esp32: enable platform module with IDF version
- boards/GENERIC_D2WD: build with -Os optimisation
- esp32_rmt: install RMT driver on core 1
- machine_bitstream: reinstate bitstream bit-bang implementation
javascript port: no changes specific to this component/port
mimxrt port:
- sdcard: implement SDCard driver
- machine_bitstream: add bitstream function to machine module
- rework flash configuration
- sdram: add SDRAM support
- eth: add LAN support and integrate the network module
- modmachine: implement machine.WDT() and machine.reset_cause()
- boards: fix the D14/D15 pin assignment of MIMXRT1050/60/64_EVK
- hal: remove duplicate definitions from flexspi_hyper_flash.h
- dma_channel: fix the DMA channel management
- fix cycle counter for time.ticks_cpu() and machine.bitstream()
- add dht_readinto() to the mimxrt module, and freeze dht.py
- extend the help() message and README.md
- mpconfigport.h: enable f-strings
- modmachine: implement soft_reset() and unique_id() functions
- boards/make-pins.py: allow empty lines and comments in pins.csv
- optimize the runtime speed
- enable the platform module
- boards: add the Seeed ARCH MIX board
- boards: update the board.json files and add deploy_xx.md files
- fix mp_hal_quiet_timing_enter()/exit() so timer still runs
- support PWM using the FLEXPWM and QTMR modules
- define UART 0 on MIMXRT boards
- support selection of PHY type and address
- re-enable eth checksum creation by HW
- fix a tiny unnoticed bug in sdcard.c
- add a driver for the DP83848 PHY device
- refactor the reading of the machine id
- enable ticks_cpu at boot time for NDEBUG builds only
- use -Og instead of -O0 for DEBUG builds
- tidy up the board flash related files
- hal: allow readSampleClkSrc to be configured by a board
- enable MICROPY_PY_USSL_FINALISER
minimal port:
- mpconfigport.h: use MICROPY_CONFIG_ROM_LEVEL_MINIMUM
- Makefile: don't force a 32-bit build
- mpconfigport.h: disable features that are not needed
nrf port:
- Makefile: improve Black Magic Probe commands
- main: use VFS helper function to mount fs and chdir
pic16bit port: no changes specific to this component/port
powerpc port: no changes specific to this component/port
qemu-arm port: no changes specific to this component/port
rp2 port:
- mpconfigport.h: enable heapq module
- add support for bluetooth module using NimBLE
- add framework for networking
- mpconfigport.h: use the "extra" feature level
- enable optimisations (comp goto, map cache, fast attr)
- machine_i2s: add I2S protocol support
- add support for Nina-W10 WiFi/BT module
- boards: add support for Arduino Nano RP2040
- machine_bitstream: implement the machine.bitstream driver
- boards: add neopixel.py to manifest.py
- rp2_pio: support exec with sideset
- boards/PIMORONI_PICOLIPO_16MB: fix 16MB flash size
- boards: add PYBSTICK26 RP2040 board definition
- machine_uart: handle and clear UART RX timeout IRQ
- boards/ARDUINO_NANO_RP2040_CONNECT: set default I2C pins
- machine_pwm: fix PWM frequency setting
- machine_pwm: keep duty value when changing the frequency
- add support for DHT11 and DHT22 sensors
- CMakeLists.txt: allow a board to override PICO_BOARD
- boards/GARATRONIC_PYBSTICK26_RP2040: use correct pico-sdk board cfg
samd port:
- integrate latest asf4, add help, more time funcs and uPy features
- samd_soc: allow a board to configure the low-level MCU config
- add internal flash block device, filesystem and uos support
- add Pin and LED classes, and machine.unique_id
- boards/ADAFRUIT_FEATHER_M0_EXPRESS: update for flash and pins
- boards/ADAFRUIT_ITSYBITSY_M4_EXPRESS: update for flash and pins
- boards/MINISAM_M4: update for flash and pins
- boards/ADAFRUIT_TRINKET_M0: update for flash and pins
- boards/SAMD21_XPLAINED_PRO: update for flash and pins
- boards/SEEED_WIO_TERMINAL: add new board definition
- boards/SEEED_XIAO: add new board definition
- README.md: update README to reflect new features and boards
stm32 port:
- pin: enable GPIO clock of pin if it's constructed without init
- main: don't unconditionally enable GPIO A,B,C,D clocks
- boards/VCC_GND_H743VI: add board definition for VCC_GND_H743VI
- boards/OLIMEX_E407: add Ethernet RMII support
- boards/LEGO_HUB_NO6: remove user paths from cc2564 init file
- boards: remove trailing spaces, and add newline at end of file
- add basic support for STM32H750
- add support for H7A3(Q)/H7B3(Q), and STM32H73B3I_DK board defn
- suggest putting code in main.py not boot.py
- boards/make-pins.py: allow a CPU pin to be hidden
- boards/make-pins.py: allow empty lines and comments in pins.csv
- dma: add functions for external users of DMA to enable clock
- enable LOAD_ATTR fast path, and map lookup caching on >M0
- boards: add OLIMEX H407 board definition
- enable platform module
- extended flash filesystem space to 512K on H743 boards
- boards/NUCLEO_H743ZI: enable VfsLfs2 on NUCLEO_H743ZI(2) boards
- boards: add PF11-BOOT0 to stm32f091_af.csv
- machine_i2c: use hardware I2C for STM32H7
- sdram: enforce gcc opt, and use volatile and DSB in sdram_test
- usbd_cdc_interface: allow a board to hook into USBD CDC RX events
- mpbthciport: allow a board to hook BT HCI poll functions
- pendsv: allow a board to add entries for pendsv_schedule_dispatch
- boards: add images to board.json for Adafruit and VCC_GND boards
- uart: fix race conditions and clearing status in IRQ handler
- mpconfigport.h: use the "extra" feature level
- in machine_i2s, send null samples in underflow situations
- in machine_i2s, make object reference arrays root pointers
- led: support an extra 2 LEDs in board configuration
- boards/MIKROE_CLICKER2_STM32: add more detail to board.json
- boards: add new board MikroElektronika Quail, and F427 support
- main: run optional frozen module at boot
- sdio: don't explicitly disable DMA2 on deinit of SDIO
- dma: make DMA2_Stream3 exclusive to SDIO when CYW43 enabled
- boards: build NUCLEO_WB55 and STM32F769DISC without mboot enabled
- boards: add PYBSTICK26 F411 board definition
- boards/NADHAT_PYBF405: rename board to GARATRONIC_NADHAT_F405
- usb: use a table of allowed values to simplify usb_mode get/set
- boards/NUCLEO_WB55: update rfcore_firmwre for new WS
- flashbdev: support generic flash storage config via link symbols
- boards: convert F413,F439,H743,L4xx,WB55 to new flash FS config
- add support for F479 MCUs
- include HAL MMC code in F4 builds
- boards/make-pins.py: use cpu pins to define static alt-fun macros
- boards/NUCLEO_WB55: fix LED ordering
- boards/LEGO_HUB_NO6: set filesystem label as HUB_NO6
- boards: remove stray '+' characters at start of lines in ld files
- boards: remove unused MICROPY_HW_ENABLE_TIMER config
- boards: enable MICROPY_HW_ENABLE_SERVO on various boards
- update L4 code to build with latest stm32lib and L4 HAL 1.17.0
- main: call sdcard_init when only MICROPY_HW_ENABLE_MMCARD enabled
- sdcard: support 8-bit wide SDIO bus
- sdcard: add config option to force MM card capacity
- factoryreset: init vfs flags before calling pyb_flash_init_vfs
- qspi: fix typo in address comment
- boards/make-pins.py: generate empty ADC table if needed
- boards/OLIMEX_H407: fix typo in OLIMEX H407 board.json
- network_wiznet5k: fix build error with wiznet5k and lwip enabled
- enable MICROPY_PY_USSL_FINALISER
teensy port:
- switch to use manifest.py instead of FROZEN_DIR
unix port:
- enable LOAD_ATTR fast path, and map lookup caching
- modusocket: support MP_STREAM_POLL in unix socket_ioctl
- modos: add support for uos.urandom(n)
- coverage: change remaining printf to mp_printf
- Makefile: use -Og instead of -O0 for debug builds
windows port:
- README: remove unsupported Python instructions for Cygwin
- mpconfigport.h: enable help and help("modules")
- add support for build variants to windows port
- run tests via Makefile
- appveyor: build both standard and dev variants
- appveyor: build mpy-cross only once for mingw-w64
- msvc: run qstr preprocessing phase in parallel
zephyr port:
- mphalport.h: remove unused and unimplemented C-level pin API
- increase minimum CMake version to 3.20.0
- update include path to reboot.h
- get UART console device from devicetree instead of Kconfig
- use CONFIG_USB_DEVICE_STACK for conditional USB device support
- upgrade to Zephyr v2.7.0
- modbluetooth_zephyr: provide dummy connect_cancel function
This allows the compiler to merge strings: e.g. "update",
"difference_update" and "symmetric_difference_update" will all point to the
same memory.
No functional change.
The size reduction depends on the number of qstrs in the build. The change
this commit brings is:
bare-arm: -4 -0.007%
minimal x86: +150 +0.092% [incl +48(data)]
unix x64: -608 -0.118%
unix nanbox: -572 -0.126% [incl +32(data)]
stm32: -1392 -0.352% PYBV10
cc3200: -448 -0.244%
esp8266: -1208 -0.173% GENERIC
esp32: -1028 -0.068% GENERIC[incl -1020(data)]
nrf: -440 -0.252% pca10040
rp2: -1072 -0.217% PICO
samd: -368 -0.264% ADAFRUIT_ITSYBITSY_M4_EXPRESS
Performance is also improved (on bare metal at least) for the
core_import_mpy_multi.py, core_import_mpy_single.py and core_qstr.py
performance benchmarks.
Originally at adafruit#4583
Signed-off-by: Artyom Skrobov <tyomitch@gmail.com>
This changes makemanifest.py & mpy-tool.py to merge string and mpy names
into the same list (now mp_frozen_names).
The various paths for loading a frozen module (mp_find_frozen_module) and
checking existence of a frozen module (mp_frozen_stat) use a common
function that searches this list.
In addition, the frozen lookup will now only take place if the path starts
with ".frozen", which needs to be added to sys.path.
This fixes issues #1804, #2322, #3509, #6419.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This commit removes all parts of code associated with the existing
MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE optimisation option, including the
-mcache-lookup-bc option to mpy-cross.
This feature originally provided a significant performance boost for Unix,
but wasn't able to be enabled for MCU targets (due to frozen bytecode), and
added significant extra complexity to generating and distributing .mpy
files.
The equivalent performance gain is now provided by the combination of
MICROPY_OPT_LOAD_ATTR_FAST_PATH and MICROPY_OPT_MAP_LOOKUP_CACHE (which has
been enabled on the unix port in the previous commit).
It's hard to provide precise performance numbers, but tests have been run
on a wide variety of architectures (x86-64, ARM Cortex, Aarch64, RISC-V,
xtensa) and they all generally agree on the qualitative improvements seen
by the combination of MICROPY_OPT_LOAD_ATTR_FAST_PATH and
MICROPY_OPT_MAP_LOOKUP_CACHE.
For example, on a "quiet" Linux x64 environment (i3-5010U @ 2.10GHz) the
change from CACHE_MAP_LOOKUP_IN_BYTECODE, to LOAD_ATTR_FAST_PATH combined
with MAP_LOOKUP_CACHE is:
diff of scores (higher is better)
N=2000 M=2000 bccache -> attrmapcache diff diff% (error%)
bm_chaos.py 13742.56 -> 13905.67 : +163.11 = +1.187% (+/-3.75%)
bm_fannkuch.py 60.13 -> 61.34 : +1.21 = +2.012% (+/-2.11%)
bm_fft.py 113083.20 -> 114793.68 : +1710.48 = +1.513% (+/-1.57%)
bm_float.py 256552.80 -> 243908.29 : -12644.51 = -4.929% (+/-1.90%)
bm_hexiom.py 521.93 -> 625.41 : +103.48 = +19.826% (+/-0.40%)
bm_nqueens.py 197544.25 -> 217713.12 : +20168.87 = +10.210% (+/-3.01%)
bm_pidigits.py 8072.98 -> 8198.75 : +125.77 = +1.558% (+/-3.22%)
misc_aes.py 17283.45 -> 16480.52 : -802.93 = -4.646% (+/-0.82%)
misc_mandel.py 99083.99 -> 128939.84 : +29855.85 = +30.132% (+/-5.88%)
misc_pystone.py 83860.10 -> 82592.56 : -1267.54 = -1.511% (+/-2.27%)
misc_raytrace.py 21490.40 -> 22227.23 : +736.83 = +3.429% (+/-1.88%)
This shows that the new optimisations are at least as good as the existing
inline-bytecode-caching, and are sometimes much better (because the new
ones apply caching to a wider variety of map lookups).
The new optimisations can also benefit code generated by the native
emitter, because they apply to the runtime rather than the generated code.
The improvement for the native emitter when LOAD_ATTR_FAST_PATH and
MAP_LOOKUP_CACHE are enabled is (same Linux environment as above):
diff of scores (higher is better)
N=2000 M=2000 native -> nat-attrmapcache diff diff% (error%)
bm_chaos.py 14130.62 -> 15464.68 : +1334.06 = +9.441% (+/-7.11%)
bm_fannkuch.py 74.96 -> 76.16 : +1.20 = +1.601% (+/-1.80%)
bm_fft.py 166682.99 -> 168221.86 : +1538.87 = +0.923% (+/-4.20%)
bm_float.py 233415.23 -> 265524.90 : +32109.67 = +13.756% (+/-2.57%)
bm_hexiom.py 628.59 -> 734.17 : +105.58 = +16.796% (+/-1.39%)
bm_nqueens.py 225418.44 -> 232926.45 : +7508.01 = +3.331% (+/-3.10%)
bm_pidigits.py 6322.00 -> 6379.52 : +57.52 = +0.910% (+/-5.62%)
misc_aes.py 20670.10 -> 27223.18 : +6553.08 = +31.703% (+/-1.56%)
misc_mandel.py 138221.11 -> 152014.01 : +13792.90 = +9.979% (+/-2.46%)
misc_pystone.py 85032.14 -> 105681.44 : +20649.30 = +24.284% (+/-2.25%)
misc_raytrace.py 19800.01 -> 23350.73 : +3550.72 = +17.933% (+/-2.79%)
In summary, compared to MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE, the new
MICROPY_OPT_LOAD_ATTR_FAST_PATH and MICROPY_OPT_MAP_LOOKUP_CACHE options:
- are simpler;
- take less code size;
- are faster (generally);
- work with code generated by the native emitter;
- can be used on embedded targets with a small and constant RAM overhead;
- allow the same .mpy bytecode to run on all targets.
See #7680 for further discussion. And see also #7653 for a discussion
about simplifying mpy-cross options.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
* modframebuf: _mp_framebuf_p_t is not "really" a protocol, but the
QSTR assignment caused problems when building as a dynamic module
* modure: str_index_to_ptr is not in the natmod API, disable URE match
spans when dynamic. mp_obj_len() is a bugfix, we should throw here
if the object is not string-like
* moduzlib: Correct paths to uzlib headers & sources. this relative
path (from moduzlib.c to the referenced file) works in all cases,
the other only worked from ports/PORTNAME.
* dynruntime: Handle 2-arg m_malloc, assert_native_inited, add a
micropythonish mp_arg_check_num_mp, fix mp_raise_msg to use dumb
strings, add mp_raise_arg1
* nativeglue: ad assert_native_inited
* translate: MP_ERROR_TEXT evaluates to its argument for DYNRUNTIME
* mpy-tool: A straggling magic number change
* mpy_ld: Have to renumber manually after dynruntime change
* import_mpy_native_gc.py: Update copy of features0 baked into this test
After discussing with danh, I noticed that `a/**/b` would not match `a/b`.
After correcting this and re-running "pre-commit run --all", additional
files were reindented, including the codeformat script itself.
This allows the compiler to merge strings: e.g. "update",
"difference_update" and "symmetric_difference_update"
will all point to the same memory.
Shaves ~1KB off the image size, and potentially allows
bigger savings if qstr attrs are initialized in qstr_init(),
and not stored in the image.
MP_BC_CALL_FUNCTION will leave the result on the Python stack, so that
result must be discarded by MP_BC_POP_TOP.
Signed-off-by: Damien George <damien@micropython.org>
Usage:
mpy-tool.py -o merged.mpy --merge mod1.mpy mod2.mpy
The constituent .mpy files are executed sequentially when the merged file
is imported, and they all use the same global namespace.
Instead of encoding 4 zero bytes as placeholders for the simple_name and
source_file qstrs, and storing the qstrs after the bytecode, store the
qstrs at the location of these 4 bytes. This saves 4 bytes per bytecode
function stored in a .mpy file (for example lcd160cr.mpy drops by 232
bytes, 4x 58 functions). And resulting code size is slightly reduced on
ports that use this feature.
This patch compresses the second part of the bytecode prelude which
contains the source file name, function name, source-line-number mapping
and cell closure information. This part of the prelude now begins with a
single varible length unsigned integer which encodes 2 numbers, being the
byte-size of the following 2 sections in the header: the "source info
section" and the "closure section". After decoding this variable unsigned
integer it's possible to skip over one or both of these sections very
easily.
This scheme saves about 2 bytes for most functions compared to the original
format: one in the case that there are no closure cells, and one because
padding was eliminated.
The start of the bytecode prelude contains 6 numbers telling the amount of
stack needed for the Python values and exceptions, and the signature of the
function. Prior to this patch these numbers were all encoded one after the
other (2x variable unsigned integers, then 4x bytes), but using so many
bytes is unnecessary.
An entropy analysis of around 150,000 bytecode functions from the CPython
standard library showed that the optimal Shannon coding would need about
7.1 bits on average to encode these 6 numbers, compared to the existing 48
bits.
This patch attempts to get close to this optimal value by packing the 6
numbers into a single, varible-length unsigned integer via bit-wise
interleaving. The interleaving scheme is chosen to minimise the average
number of bytes needed, and at the same time keep the scheme simple enough
so it can be implemented without too much overhead in code size or speed.
The scheme requires about 10.5 bits on average to store the 6 numbers.
As a result most functions which originally took 6 bytes to encode these 6
numbers now need only 1 byte (in 80% of cases).
Prior to this patch mp_opcode_format would calculate the incorrect size of
the MP_BC_UNWIND_JUMP opcode, missing the additional byte. But, because
opcodes below 0x10 are unused and treated as bytes in the .mpy load/save
and freezing code, this bug did not show any symptoms, since nested unwind
jumps would rarely (if ever) reach a depth of 16 (so the extra byte of this
opcode would be between 0x01 and 0x0f and be correctly loaded/saved/frozen
simply as an undefined opcode).
This patch fixes this bug by correctly accounting for the additional byte.
.
Previously, when linking qstr objects in native code for ARM Thumb, the
index into the machine code was being incremented by 4, not 8. It should
be 8 to account for the size of the two machine instructions movw and movt.
This patch makes sure the index into the machine code is incremented by the
correct amount for all variations of qstr linking.
See issue #4829.
Fixes errors in the tool when 1) linking qstrs in native ARM-M code; 2)
freezing multiple files some of which use native code and some which don't.
Fixes issue #4829.
The qstr window size is not log-2 encoded, it's just the actual number (but
in mpy-tool.py this didn't lead to an error because the size is just used
to truncate the window so it doesn't grow arbitrarily large in memory).
Addresses issue #4635.
When encoded in the mpy file, if qstr <= QSTR_LAST_STATIC then store two
bytes: 0, static_qstr_id. Otherwise encode the qstr as usual (either with
string data or a reference into the qstr window).
Reduces mpy file size by about 5%.
Instead of emitting two bytes in the bytecode for where the linked qstr
should be written to, it is now replaced by the actual qstr data, or a
reference into the qstr window.
Reduces mpy file size by about 10%.
This is an implementation of a sliding qstr window used to reduce the
number of qstrs stored in a .mpy file. The window size is configured to 32
entries which takes a fixed 64 bytes (16-bits each) on the C stack when
loading/saving a .mpy file. It allows to remember the most recent 32 qstrs
so they don't need to be stored again in the .mpy file. The qstr window
uses a simple least-recently-used mechanism to discard the least recently
used qstr when the window overflows (similar to dictionary compression).
This scheme only needs a single pass to save/load the .mpy file.
Reduces mpy file size by about 25% with a window size of 32.
POP_BLOCK and POP_EXCEPT are now the same, and are always followed by a
JUMP. So this optimisation reduces code size, and RAM usage of bytecode by
two bytes for each try-except handler.
If you happen to only have a really simple frozen file that doesn't contain
any new qstrs then the generated frozen_mpy.c file contains an empty
enumeration which causes a C compile time error.
Following an equivalent fix to py/bc.c. The reason the incorrect values
for the opcode constants were not previously causing a bug is because they
were never being used: these opcodes always have qstr arguments so the part
of the code that was comparing them would never be reached.
Thanks to @malinah for finding the problem and providing the initial patch.
The first dynamic qstr pool is double the size of the 'alloc' field of
the last const qstr pool. The built in const qstr pool
(mp_qstr_const_pool) has a hardcoded alloc size of 10, meaning that the
first dynamic pool is allocated space for 20 entries. The alloc size
must be less than or equal to the actual number of qstrs in the pool
(the 'len' field) to ensure that the first dynamically created qstr
triggers the creation of a new pool.
When modules are frozen a second const pool is created (generally
mp_qstr_frozen_const_pool) and linked to the built in pool. However,
this second const pool had its 'alloc' field set to the number of qstrs
in the pool. When freezing a large quantity of modules this can result
in thousands of qstrs being in the pool. This means that the first
dynamically created qstr results in a massive allocation. This commit
sets the alloc size of the frozen qstr pool to 10 or less (if the number
of qstrs in the pool is less than 10). The result of this is that the
allocation behaviour when a dynamic qstr is created is identical with an
without frozen code.
Note that there is the potential for a slight memory inefficiency if the
frozen modules have less than 10 qstrs, as the first few dynamic
allocations will have quite a large overhead, but the geometric growth
soon deals with this.
This patch allows the following code to run without allocating on the heap:
super().foo(...)
Before this patch such a call would allocate a super object on the heap and
then load the foo method and call it right away. The super object is only
needed to perform the lookup of the method and not needed after that. This
patch makes an optimisation to allocate the super object on the C stack and
discard it right after use.
Changes in code size due to this patch are:
bare-arm: +128
minimal: +232
unix x64: +416
unix nanbox: +364
stmhal: +184
esp8266: +340
cc3200: +128
With caching of map lookups in the bytecode, frozen bytecode can still
work but must be stored in RAM, not ROM. This patch allows mpy-tool.py to
generate code that works with this optimisation, but it's not recommended
to use it on embedded targets (because of lack of RAM).
When an mpy file is frozen it must know the values of certain
configuration variables. This patch provides an explicit check in the
generated C file that the configuration variables are what they are
supposed to be.
