STM32L4: clear error programming flags before erase & program operations

[juhoeskeli]

Description

On STM32L486 when using dual bank operation where the program code was in bank 0 and BlockDevice in bank1 the erase and program operations would fail with Flash_error: 32.

From the reference manual: Bit 7 PGSERR: Programming sequence error Set by hardware when a write access to the Flash memory is performed by the code while PG or FSTPG have not been set previously. Set also by hardware when PROGERR, SIZERR, PGAERR, MISSERR or FASTERR is set due to a previous programming error. Cleared by writing 1.

The suggested write sequence can be found from reference: en.DM00083560.pdf 3.3.7 Flash main memory programming sequences - 2. Check and clear all error programming flags due to a previous programming.

The code change performs just that, clears the flags as per instructions. With the change it is possible to use bank 0 for running the program code while writing data to bank 1. I have tested this change on STM32L486 where OTA FW image was loaded to bank 1 and successfully updated from therein to bank 0.

Further reference: MBEDOSTEST-206

Pull request type

[X] Fix
[ ] Refactor
[ ] Target update
[ ] Functionality change
[ ] Breaking change

[jeromecoutant]

Flash test results with this patch for STM32L4 NUCLEO boards:

target	platform_name	test suite	result	elapsed_time (sec)	copy_method
NUCLEO_L432KC-ARM	NUCLEO_L432KC	tests-mbed_drivers-flashiap	OK	21.36	default
NUCLEO_L432KC-ARM	NUCLEO_L432KC	tests-mbed_hal-flash	OK	18.25	default
NUCLEO_L433RC_P-ARM	NUCLEO_L433RC_P	tests-mbed_drivers-flashiap	OK	18.13	default
NUCLEO_L433RC_P-ARM	NUCLEO_L433RC_P	tests-mbed_hal-flash	OK	17.99	default
NUCLEO_L476RG-ARM	NUCLEO_L476RG	tests-mbed_drivers-flashiap	OK	18.09	default
NUCLEO_L476RG-ARM	NUCLEO_L476RG	tests-mbed_hal-flash	OK	17.96	default
NUCLEO_L486RG-ARM	NUCLEO_L486RG	tests-mbed_drivers-flashiap	OK	18.37	default
NUCLEO_L486RG-ARM	NUCLEO_L486RG	tests-mbed_hal-flash	OK	18.1	default
NUCLEO_L496ZG-ARM	NUCLEO_L496ZG	tests-mbed_drivers-flashiap	OK	18.08	default
NUCLEO_L496ZG-ARM	NUCLEO_L496ZG	tests-mbed_hal-flash	OK	18.11	default

[jeromecoutant]

If you clear all flash error, maybe you can remove next __HAL_FLASH_CLEAR_FLAG ?

[juhoeskeli]

Should be OK now. FLASH_FLAG_ALL_ERRORS includes also FLASH_FLAG_OPTVERR

[jeromecoutant]

Need to check other STM32 families

[0xc0170]

Need to check other STM32 families

Any update?

[jeromecoutant]

Any update?

Sorry, this is "only" a reminder for my todo list
You can start CI for this PR

[jeromecoutant]

Need to check other STM32 families

For information, it seems that L4 FLASH driver is different from other STM32 families.
So this fix has not to be applied everywhere.
Regards,

[juhoeskeli]

To next patch release, please.

[NirSonnenschein]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 3293
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/8263/

Triggering tests

/morph test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : FAILURE

Build number : 2917
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/8263/

[NirSonnenschein]

exporter / mbed2 build seems to have failed because of IAR license shortage - re-=running:
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2919
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/8263/

[mbed-ci]

Test : SUCCESS

Build number : 3099
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/8263/3099