Cellular: HSI set to be source clock for WISE_1570

[mirelachirica]

Description

LSE as LPUART source clock is causing WISE_1570 to have stability issues(framing errors) for AT commands communication over LPUART. Changing clock to HSI is fixing the problem.

Internal defect: IOTCELL-988

Pull request type

[x] Fix
[ ] Refactor
[ ] New target
[ ] Feature
[ ] Breaking change

[JanneKiiskila]

To next patch release, please. @adbridge @0xc0170

[cmonr]

@JanneKiiskila @mirelachirica We're having some issues with the testing CI atm. Will start once they're resolved.

[jeromecoutant]

Hi

I would suggest to follow #7469
Then, it should be possible to keep SystemClock with HSE, and configure LPUART clock with HSI

@bcostm

[bcostm]

Can you check PR #7498 ? Thanks.

[cmonr]

/morph build

[cmonr]

@bcostm Who are you referring to when you say "you"?

[cmonr]

@bcostm @jeromecoutant @JanneKiiskila @mirelachirica
Going to try and bring this in w/o the changes in #7498.
Bringing in #7498 before this PR creates a dependency that @JanneKiiskila and @mirelachirica would have to re-verify before moving this forward. The time constraint is that Code Freeze for the patch is about to occur in 24 hours, which would make timing everything incredibly difficult.

@bcostm @jeromecoutant Do y'all have any strong arguments as to why this needs to wait on #7498?

[jeromecoutant]

Hi
It is up to you...
I can see in this PR, that you are downgrading SystemClock from HSE to HSI, in order to get higher LPUART clock.
What @bcostm is proposing is to take #7498, then keep HSE SystemClock, and set LPUART clock to HSI as requested.

[jeromecoutant]

This can be done in 2 steps.
Merge this one, and push another one later with HSE setting

[mbed-ci]

Build : SUCCESS

Build number : 2586
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7489/

Triggering tests

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

[mbed-ci]

Exporter Build : FAILURE

Build number : 2229
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7489/

[cmonr]

Export failure due to multiply defined symbols, but symbol does not appear related tio PR.
Retesting: /morph export-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2232
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7489/

[mbed-ci]

Test : SUCCESS

Build number : 2340
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7489/2340

[juhoeskeli]

Hello, I'm having an issue with this fix. The board MTB_ADV_WISE_1570 fails in SetSysClock_PLL_HSI in function HAL_RCC_OscConfig. More precisely this happens when waiting the PLL to stabilize (stm32l4xx_hal_rcc.c:813). As a workaround increasing PLL_TIMEOUT_VALUE from 2ms to 5ms solves the issue. Then it works 10/10 tries. With 4ms value it works 1/20 times.

If changing the PLL timeout value is a valid fix, maybe the timeout value could be as parameter to HAL_RCC_OscConfig?

[cmonr]

@juhoeskeli Would you mind opening an issue and referencing this PR if you haven't already?

Comments left in PRs after they're merged tend not to get the best attention.

[juhoeskeli]

@cmonr Sounds fair.

[juhoeskeli]

@cmonr I noticed that the issue was due to bootloader using different clock and then switching to HSI in application. By using the same clock in bootloader there is no issue. I consider this working as intended.