This commit enables some significant optimisations for esp32:
- move the VM to iRAM
- move hot parts of the runtime to iRAM (map lookup, load global/name,
mp_obj_get_type)
- enable MICROPY_OPT_LOAD_ATTR_FAST_PATH
- enable MICROPY_OPT_MAP_LOOKUP_CACHE
- disable assertions
- change from -Os to -O2 for compilation
It's hard to measure performance on esp32 due to external flash and
hardware caching. But this set of changes improves performance compared to
master by (on a TinyPICO with the GENERIC build, using IDF 4.2.2, running
at 160MHz):
diff of scores (higher is better)
N=100 M=100 esp32-master -> esp32-perf diff diff% (error%)
bm_chaos.py 71.28 -> 268.08 : +196.80 = +276.094% (+/-0.04%)
bm_fannkuch.py 44.10 -> 69.31 : +25.21 = +57.166% (+/-0.01%)
bm_fft.py 1385.27 -> 2538.23 : +1152.96 = +83.230% (+/-0.01%)
bm_float.py 1060.94 -> 3900.62 : +2839.68 = +267.657% (+/-0.03%)
bm_hexiom.py 10.90 -> 32.79 : +21.89 = +200.826% (+/-0.02%)
bm_nqueens.py 1000.83 -> 2372.87 : +1372.04 = +137.090% (+/-0.01%)
bm_pidigits.py 288.13 -> 664.40 : +376.27 = +130.590% (+/-0.46%)
misc_aes.py 102.45 -> 345.69 : +243.24 = +237.423% (+/-0.01%)
misc_mandel.py 1016.58 -> 2121.92 : +1105.34 = +108.731% (+/-0.01%)
misc_pystone.py 632.91 -> 1801.87 : +1168.96 = +184.696% (+/-0.08%)
misc_raytrace.py 76.66 -> 281.78 : +205.12 = +267.571% (+/-0.05%)
viper_call0.py 210.63 -> 273.17 : +62.54 = +29.692% (+/-0.01%)
viper_call1a.py 208.45 -> 269.51 : +61.06 = +29.292% (+/-0.00%)
viper_call1b.py 185.44 -> 228.25 : +42.81 = +23.086% (+/-0.01%)
viper_call1c.py 185.86 -> 228.90 : +43.04 = +23.157% (+/-0.01%)
viper_call2a.py 207.10 -> 267.25 : +60.15 = +29.044% (+/-0.00%)
viper_call2b.py 173.76 -> 209.42 : +35.66 = +20.523% (+/-0.00%)
Five tests have more than 3x speed up (200%+).
The performance of the tests bm_fft, bm_pidigits and misc_aes now scale
with CPU frequency (eg changing frequency to 240MHz boosts the performance
of these by 50%), which means they are no longer influenced by timing of
external flash access. (The viper_call* tests did previously scale with
CPU frequency, and they still do.)
Turning off assertions reduces code size by about 80k, and going from -Os
to -O2 costs about 100k, so the net change in code size (for the GENERIC
board) is about +20k.
If a board wants to enable assertions, or use -Os instead of -O2, that's
still possible by overriding the sdkconfig parameters.
Signed-off-by: Damien George <damien@micropython.org>
To match network_lan.c and network_ppp.c, and make it clear what code is
specifically for WLAN support.
Also provide a configuration option MICROPY_PY_NETWORK_WLAN which can be
used to fully disable network.WLAN (it's enabled by default).
Signed-off-by: Damien George <damien@micropython.org>
To do this the board must define MICROPY_BOARD_STARTUP, set
MICROPY_SOURCE_BOARD then define the new start-up code.
For example, in mpconfigboard.h:
#define MICROPY_BOARD_STARTUP board_startup
void board_startup(void);
in mpconfigboard.cmake:
set(MICROPY_SOURCE_BOARD
${MICROPY_BOARD_DIR}/board.c
)
and in a new board.c file in the board directory:
#include "py/mpconfig.h"
void board_startup(void) {
boardctrl_startup();
// extra custom startup
}
This follows stm32's boardctrl facilities.
Signed-off-by: Damien George <damien@micropython.org>
Because vPortCleanUpTCB is called by the FreeRTOS idle task, and it checks
thread, but didn't check the thread_mutex.
And if thread is not NULL, but thread_mutex not ready then it will crash
with an error when calling mp_thread_mutex_lock(&thread_mutex, 1).
As suggested by @dpgeorge, move the thread = &thread_entry0 line to the end
of mp_thread_init().
Signed-off-by: leo chung <gewalalb@gmail.com>
This callback allows detecting if there is a USB host connected to the CDC
or not, in which case the stdout_tx should skip CDC TX writing and
flushing or the system will block.
Fixes issue #7820.
This commit allows using all the available PWM timers (up to 8) and
channels (up to 16), without affecting the PWM API.
If a new frequency is set, first it checks if another timer is using the
same frequency. If yes, then it uses this timer, otherwise, it creates a
new one. If all timers are used, the user should set an already used
frequency, or de-init a channel.
This work is based on #6276 and #3608.
This commit refactors machine.PWM and creates extmod/machine_pwm.c. The
esp8266, esp32 and rp2 ports all use this and provide implementations of
the required PWM functionality. This helps to reduce code duplication and
keep the same Python API across ports.
This commit does not make any functional changes.
Signed-off-by: Damien George <damien@micropython.org>
The zephyr port doesn't support SoftI2C so it's not enabled, and the legacy
I2C constructor check can be removed.
Signed-off-by: Damien George <damien@micropython.org>
To keep things neat and tidy, we ensure that each file has 1 and only 1
newline at the end of each file.
Signed-off-by: David Lechner <david@pybricks.com>
Add a new board type for ESP32-C3 revision 3 and up that implement the USB
serial/JTAG port on pin 18 and 19. This variant uses the USB serial for
programming and console, leaving the UART free.
- Pins 18 and 19 are correctly reserved for this variant. Also pins 14-17
are reserved for flash for any ESP32-C3 so they can't be reconfigured
anymore to crash the system.
- Added usb_serial_jtag.c and .h to implement this interface.
- Interface was tested to work correctly together with webrepl.
- Interface was tested to work correctly when sending and receiving
large files with ampy.
- Disconnecting terminal or USB will not hang the system when it's
trying to print.
Reverse operations are supported on stm32 and rp2, and esp32 has enough
space to also enable inplace operations, to make it complete.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This enables optional support for the hardware UART to use the RTS and/or
CTS pins for flow control.
The new "flow" constructor keyword specifies a bitmask of RTS and/or CTS.
This matches the interface used by machine.UART on stm32 and rp2.
Previously on ESP32 it was possible to specify which pins to use for the
RTS and CTS signals, but hardware flow control was never functional: CTS
was not checked before transmitting bytes, and RTS was always driven high
(signalling no buffer space available). With this patch, CTS and RTS both
operate as expected.
This also includes an update to the machine.UART documentation.
Signed-off-by: Will Sowerbutts <will@sowerbutts.com>
This helps the OS switch to and give other threads processing time during
the sleep. It also ensures that pending events are handled, even when
sleeping for 0ms.
Fixes issue #5344.
Signed-off-by: Damien George <damien@micropython.org>
Using a 2-item transaction queue instead of 1 allows long transfers to
be executed with the minimum inter-transaction delay. Limit maximum
transaction length to ensure an integer multiple of the SPI `bits`
setting are transferred. Fixes#7511.
This commit adds I2S protocol support for the esp32 and stm32 ports, via
a new machine.I2S class. It builds on the stm32 work of blmorris, #1361.
Features include:
- a consistent I2S API across the esp32 and stm32 ports
- I2S configurations supported:
- master transmit and master receive
- 16-bit and 32-bit sample sizes
- mono and stereo formats
- sampling frequency
- 3 modes of operation:
- blocking
- non-blocking with callback
- uasyncio
- internal ring buffer size can be tuned
- documentation for Pyboards and esp32-based boards
- tested on the following development boards:
- Pyboard D SF2W
- Pyboard V1.1
- ESP32 with SPIRAM
- ESP32
Signed-off-by: Mike Teachman <mike.teachman@gmail.com>
This change allows specification of the idle level and TX carrier output
level (through changed initialisation API), and more flexible specification
of pulses for write_pulses.
This is a breaking change for the esp32.RMT constructor API. Previous code
of this form:
esp32.RMT(..., carrier_duty_percent=D, carrier_freq=F)
will now raise an exception and should be changed to:
esp32.RMT(..., tx_carrier=(F, D, 1))
When looping, now disable the TX interrupt after calling rmt_write_items()
function to handle change in IDF behaviour (since v4.1). Also check length
of pulses to ensure it fits hardware limit.
Fixes issue #7403.
Dynamically generate/loaded native code (eg from @micropython.native or
native .mpy files) needs to be able allocate from IRAM, and the memory
protection feature must be disabled for that to work. Disabling it is
needed to get native code working on ESP32-S2 and -C3.
Signed-off-by: Damien George <damien@micropython.org>
This adds a wlan.config(reconnects=N) option to set the number of reconnect
attempts that will be made if the WLAN connection goes down. The default
is N=-1 (infinite retries, current behavior). Setting
wlan.config(reconnects=0) will disable the reconnect attempts.
A nice side effect of reconnects=0 is that wlan.status() will report the
disconnect reason now. See related issue #5326.
Ethernet-PHYs from ESP-IDF (LAN8720, IP101, RTL8201, DP83848) are now
supported in IDF v4.1 and above. PHY_KSZ8041 is only for ESP-IDF 4.3 and
above. ESP32S2 is not supported.
Signed-off-by: Tobias Eydam <eydam-prototyping@outlook.com>
Changes introduced are:
- the application offset is now loaded from the partition table instead of
being hard-coded to 0x10000
- maximum size of all sections is computed using the partition table
- an error is generated if any section overflows its allocated space
- remaining bytes are printed for each section
Signed-off-by: Damien George <damien@micropython.org>
Adds support for NeoPixels on GPIO32 and GPIO33 on ESP32. Otherwise,
NeoPixels wired to GPIO32/33 wll silently fail without any hints to the
user.
With thanks to @robert-hh.
Fixes issue #7221.
ATOM is a very small ESP32 development board produced by M5Stack, with a
size of 24mm * 24mm, with peripherals such as WS2812, IR, button, MPU6886
(Only Matrix), and 8 GPIO extensions. It also has a plastic shell.
Improvements made:
- PSRAM support for S2
- partition definition for 16MiB flash
- correct ADC and DAC pins
- correct GPIO and IRQ pins
- S3 components in CMakeLists
Based on original commit made by Seon Rozenblum aka @UnexpectedMaker.
Signed-off-by: Damien George <damien@micropython.org>
So a lock can be acquired on one Python thread and then released on
another. A test for this is added.
Signed-off-by: Damien George <damien@micropython.org>
Because vPortCleanUpTCB runs on the FreeRTOS idle task and cannot execute
any VM or runtime related code like freeing memory.
Signed-off-by: Damien George <damien@micropython.org>
This helper is added to properly set a pending exception, to mirror
mp_sched_schedule(), which schedules a function.
Signed-off-by: Damien George <damien@micropython.org>
This commit re-enables the command-line make option "FROZEN_MANIFEST". The
boards/*/mpconfigboard.cmake will now use the command-line FROZEN_MANIFEST
value if supplied.
Usage: make FROZEN_MANIFEST=~/foo/my-manifest.py
Because "find_package(Python3 ...)" requires at least this version of
CMake. And other features like GREATER_EQUAL and COMMAND_EXPAND_LISTS need
at least CMake 3.7 and 3.8 respectively.
Signed-off-by: Damien George <damien@micropython.org>
Commit 8a917ad252 added the gpio_reset_pin()
call to make sure that pins that were used as ADC inputs could subsequently
be used as digital IO. But calling gpio_reset_pin() will enable the
pull-up on the pin and so pull it high for a brief period. Instead use
rtc_gpio_deinit() which will just reconfigure the pin as a digital IO and
do nothing else.
Fixes issue #7079 (see also #5771).
Signed-off-by: Damien George <damien@micropython.org>
For an unconnected TCP socket, poll should return WR|HUP and read should
raise ENOTCONN. This is implemented by this commit and now the following
tests pass on esp32: extmod/usocket_tcp_basic.py,
net_hosted/connect_poll.py.
Signed-off-by: Damien George <damien@micropython.org>
It was noticed that the esp32 port didn't build ulab correctly. The
problem was a multiple defintion of the 'mp_hal_stdout_tx_str' and
'mp_hal_stdout_tx_strn_cooked' functions.
They were defined in stdout_helpers.c but also in the
ports/esp32/mphalport.c.
Fixed by removing stdout_helpers.c from the build.
Signed-off-by: Michael O'Cleirigh <michael.ocleirigh@rivulet.ca>
Support for User C and C++ modules was lost due to upgrading the esp32 to
the latest CMake based IDF from the GNUMakefile build process.
Restore the support for the esp32 port by integrating with the approach
recently added for the rp2 port.
Signed-off-by: Michael O'Cleirigh <michael.ocleirigh@rivulet.ca>
This commit fixes two issues on the esp32:
- it enables machine.soft_reset() to be called in main.py;
- it enables machine.reset_cause() to correctly identify a soft reset.
The former is useful in that it enables soft resets in applications that
are started at boot time. The support is patterned after the stm32 port.
This commit implements basic NVS support for the esp32. It follows the
pattern of the esp32.Partition class and exposes an NVS object per NVS
namespace. The initial support provided is only for signed 32-bit integers
and binary blobs. It's easy (albeit a bit tedious) to add support for
more types.
See discussions in: #4436, #4707, #6780
This enables -Os for compilation, but still keeps full assertion messages.
With IDF v4.2, -Os changes the GENERIC firmware size from 1512176 down to
1384640, and the GENERIC_SPIRAM firmware is now 1452320 which fits in the
allocated partition.
Signed-off-by: Damien George <damien@micropython.org>
The underlying OS (the ESP-IDF) uses it's own internal errno codes and so
it's simpler and cleaner to use those rather than trying to convert
everything to the values defined in py/mperrno.h.
