Code size report:

   bare-arm:    +0 +0.000% 
minimal x86:    +0 +0.000% 
   unix x64:    +0 +0.000% standard
      stm32:    +0 +0.000% PYBV10
     mimxrt:    +0 +0.000% TEENSY40
        rp2:    +0 +0.000% RPI_PICO_W
       samd:    +0 +0.000% ADAFRUIT_ITSYBITSY_M4_EXPRESS
  qemu rv32:    +0 +0.000% VIRT_RV32

Looks good @andrewleech !

Hey @mcauser, you've worked with these boards a bit haven't you? Would you be able to review?

I have a F411 board here, sold as Weact Studio F411 Blackpill V3.1. There is one difference in the wiring: for my V3.1 board SPI_MISO is at PA6, not PB4 (and the flash size is 8MB). Otherwise the PR works.

Thanks @robert-hh that sounds like something that would work as another variant!

That requires a little bit more awkward approach, since the changes have start in pins.csv and reflected then in a variant-specific mpconfigboard-VARIANT.mk. Mayb by defining two variants of the MISO pin in pins,csv and selecting the right one later.

Thanks. These new definitions seem to work, with two minor hiccups.

• In line 2 of mpconfigvariant_V3_FLASH_8MB.mk the trailing \ is missing
• the board name WEACTV20_F411 indicates, it's for V2 only. Maybe better just use WEACT_F411 or WEACT_F411_BLACKPILL.

Thanks @robert-hh good catch and suggestions, I've updated as such.

@andrewleech I built the firmware using your mpconfigboard.h from:
686a415
and it seems to work. I didn't change any of the other files in the WEACTF411CE folder that were supplied by:
https://github.com/nspsck/STM32F411CEU6_BlackPill_Micropython.

micropython.mem_info()
stack: 476 out of 15360
GC: total: 86976, used: 18304, free: 68672
No. of 1-blocks: 214, 2-blocks: 66, max blk sz: 72, max free sz: 3377

Is there something else I need to do to get the 8MB? I'd be keen to build for the 8MB flash to see if it solves a comms issue for me.

I've restructured the board and variant configuration a bit in the latest push.

v3.1 is now the default as it's what people will get if they buy them new currently, no flash configured.
There are then two extra v3.1 variants for 8MB flash and 8 MHz xtal (no flash) matching the configurations available for sale.

After that there's a variant for v2.0 with 4MB flash - the only difference between v2.0 and v3.1 boards is the pinout of the spiflash I believe so users with v2.0 (no flash) will also be able to use the default build.

I then found the boards I own are actually v1.3 which were the original version released a few years ago. On these the spiflash pinout matches v3.1 but does not have a user switch

I noticed that on v1.3 and v3.1 boards the spiflash pinout aligns with the spi hardware peripheral so that's now being used. I don't know why they changed it in v2.0, regardless it's configured with softspi.

@davefes I'm not quite following if you're trying to copy some of this PR into a different micropython source? This whole PR is the board folder which uses the variant build option to select the alternate configurations. The readme in the stm port folder should give guidance on how this works iirc.

anl@STEP: ~/micropython $ mpremote a1
Connected to MicroPython at /dev/ttyACM1
Use Ctrl-] or Ctrl-x to exit this shell
MicroPython v1.24.0-preview.186.g1548929264 on 2024-08-23; WEACT_F411_BLACKPILL with STM32F411CE
Type "help()" for more information.
>>> import os
>>> s = os.statvfs(".")
>>> s[0] * s[1] / 1024
4096.0
>>>

If anyone can I'd appreciate a test on v2.0 or v3.1 boards!

If anyone can I'd appreciate a test on v2.0 or v3.1 boards!

I tested the V3.1, 25MHz variant with and without flash. It builds and runs. Next week I will get a 8MHz V3.1 board which I then will test. Looking at the config files I would have expected a value of 8 for the 8MHz version at
#define MICROPY_HW_CLK_PLLM (8). I will see.

After the following change:

In file included from ./mpconfigport.h:31,
                 from ../../py/mpconfig.h:91,
                 from ../../py/mpstate.h:31,
                 from ../../py/runtime.h:29,
                 from usrsw.c:29:
usrsw.c: In function 'switch_init0':
boards/WEACTF411CE/mpconfigboard.h:66:38: error: 'pyb_pin_SW' undeclared (first use in this function); did you mean 'pyb_pin_B1'?
   66 | #define MICROPY_HW_USRSW_PIN        (pyb_pin_SW)

it built and runs the base-version.

I don't see a readme in the /stm32 folder. How do I change:
make -j BOARD=WEACT411CE
to build for the 8MB version?

On the 2nd attempt I had to return to the original:
#define MICROPY_HW_USRSW_PIN (pyb_pin_SW)
??

Should I be able to make a LFS2 filesystem?

import make_LFS2
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "make_LFS2.py", line 4, in <module>
AttributeError: 'module' object has no attribute 'VfsLfs2'

The base version works on rsusb but not using mpremote. Probably something flaky at my end.

Still can't work-out how to build for 8MB.

Next week I will get a 8MHz V3.1 board which I then will test.

Looking at the order confirmations, I did not order a 8MHz version of the board, like I thought I did. Shhhh.... But looking into the F411 data sheet, it is obvious that for the 8MHz version the value of MICROPY_HW_CLK_PLLM has to be 8 to get 1 MHz at the input of the PLL.

Note in the latest push I've changed the way the make args are handled to simplify end user building for customised boards, but it's not quite right yet. Don't bother reviewing too much, I'll let everyone here know when it's working.

Built for 8MB flash, made a LFS2 filesystem and the application runs fine. Still, when running:

import micropython
micropython.mem_info()
stack: 468 out of 15360
GC: total: 83072, used: 17120, free: 65952
 No. of 1-blocks: 173, 2-blocks: 60, max blk sz: 72, max free sz: 3183

I have had difficulties getting to see the 8MB flash before so will try my new boards next week.

@davefes micropython.mem_info() shows free ram, not flash. The way to check flash is

>>> os.statvfs('/flash')
(4096, 4096, 1024, 1022, 1022, 0, 0, 0, 0, 255)

The output is somewhat confusing https://docs.python.org/3/library/os.html#os.statvfs
basically the first element is block size (in bytes), the third element is total number of blocks. So for me it's 4096 * 1024 bytes = 4MB

O dear, thanks for the correction.

os.statvfs('/flash')
(4096, 4096, 2048, 2034, 2034, 0, 0, 0, 0, 255)

8MB Great!

I only have 2 boards, at the moment. One board without flash and using the base-version and VFAT works, when I change to LFS2 I get the dreaded timeout error on file transfer. The other board (8MB flash) works properly. Suspect a faulty board.

@andrewleech The 8MHz V3.1 board arrived. The firmware builds and works. USB registers, sleep functions show a reasonable behavior. So I conclude that the clocks are set up properly.

5 out of 5 new 8MB 25MHz boards work. LFS2 as well. Great work.

Thanks for all the effort you put into this.

@dpgeorge I've finished simplifying the pll generation script args and addressed the other comments to date thanks.

I have verified that with this PR:

• PYBV10 firmware is unchanged
• PYBD_SF6 firmware is mostly unchanged (it's the same size, and only around 20 lines of an xxd dump differ), and at least the generated pllfreqtable.h and plli2stable.h files are unchanged

Will this distro end up on the Micropython Downloads page?

I need to fix the previous few code comments to get it into a mergeable state.

@dpgeorge I've updated the scripts, thanks for the suggestions!

Thanks for updating. Looks good now.

@andrewleech unfortunately this board does not build:

ValueError: file:build-WEACT_F411_BLACKPILL/genhdr/qstr.i.last does not contain a definition of micropy_hw_hse_value

It looks like the regex needs to be updated to allow parenthesis within the expression for the value. Eg it can look like this:

static uint32_t __attribute__((unused)) micropy_hw_hse_value = ((25) * 1000000);

A regex like this is simple and seems to work (the uint32_t bit isn't needed):

r"static.* +(micropy_hw_hs[ei]_value) = +([0-9 +-/\*()]+);"

I posted a fix for the above in #16444.

Oh sorry and thanks, my final cleanup of the board configs was more hassle than I expected!