STM32F4xx 25Mhz Flash emulation

[samitray]

For Stm32F4xx EEPROM_ENABLE 0 in my_machine.h on a custom stm32F446RE board with 25MHz crystal no eeprom chip, (EEPROM arrives after a week) Using Release F446 25MHz configuration, getting build errors because _EEPROM_Emul_Start and _EEPROM_Emul_Sector remain undefined,
So I change to Release F446 12MHz and now build completes ok because defines are found in the linker script, but in this
configuration there is no space for bootloader, since my knowledge of microprocessors is pretty limited and before I nuke the
bootloader on the chip... i presume that, since I use stlink to flash the program, i do not require the bootloader and all will be ok
if I change the HSE_VALUE value to 25M in the Project Settings / Paths and Symbols and flash..... or may be there is more to it....

[terjeio]

i presume that, since I use stlink to flash the program, i do not require the bootloader and all will be ok

if I change the HSE_VALUE value to 25M in the Project Settings / Paths and Symbols and flash.....

Correct.

[samitray]

Ok....I changed configuration to STM32F446RETX_FLASH.ld and the compile threw up a bunch of warnings possibly due to the symbols defined in project settings. So I made a copy of STM32F446RETX_NO_BL_FLASH.ld and replaced the memory map like below
`MEMORY
{
RAM (xrw) : ORIGIN = 0x20000000, LENGTH = 128K
FLASH (rx) : ORIGIN = 0x8000000, LENGTH = 512K - 128K
EEPROM_EMUL(xrw) : ORIGIN = 0x8060000, LENGTH = 128K
}

_FLASH_VectorTable = ORIGIN(FLASH);
_EEPROM_Emul_Start = ORIGIN(EEPROM_EMUL);
_EEPROM_Emul_Sector = 11;`

to
`MEMORY
{
RAM (xrw) : ORIGIN = 0x20000000, LENGTH = 128K
BOOT_FLASH (rx) : ORIGIN = 0x8000000, LENGTH = 16K
EEPROM_EMUL(xrw) : ORIGIN = 0x8004000, LENGTH = 16K
FLASH (rx) : ORIGIN = 0x8008000, LENGTH = 512K - 16K - 16K
}

_EEPROM_Emul_Start = ORIGIN(EEPROM_EMUL);
_EEPROM_Emul_Sector = 1;`

and modified the isr_vector section to BOOT_FLASH
Then flashed the bin file at location 0x8008000 and connected to grblHAL and changed settings worked kind of... earlier it was failing, now it stores the values but prints a bunch of MSG codes

[samitray]

On settings change grblHAL is alway sending 3 MSG codes before ok.... I looked at documentation $EA , $ES, but nothing for MSG codes.... Candle expects ok response, but because of the message codes, Candle thinks that settings change has failed....
$1=26
[MSG:2]
[MSG:1]
[MSG:0]
ok

[terjeio]

... but prints a bunch of MSG codes

These are stray debug messages, removed in recent commits.

[samitray]

Any tips on how to locate the MSG codes... I looked at nvs_buffer.c line 372 where there is a retry loop for flash writing of settings, which prints the failed msg if the loop fails 4 times....so nothing to indicate how the MSG codes are generated....

[samitray]

ok, let me try to refresh the code....

[samitray]

@terjeio
In STM32F446RETX_NO_BL_FLASH.ld related to configuration Release F446 12Mhz , After changing
_EEPROM_Emul_Sector = 11;
to
_EEPROM_Emul_Sector = 7;
as per this table in Ref manual

The settings change fail message goes away

[samitray]

@terjeio Thanks for the commit