nRF52840 Cordio stack has some unwanted GPIO control

[marcemmers]

Description

We have created a custom board based on the nRF52840 and are using the Cordio BLE stack as the radio driver. It seems however that the Cordio stack is controlling some of the GPIO which is interfering with our application since we have all GPIO in use.
At least P0.13 and P0.15 are affected but we are experiencing other problems as well when we enable the stack.

Why is the radio driver controlling pins in the first place?

We are using mbed-os release 5.11.5. We have exported to a custom target and are using IAR 8.30.

I've noticed that there is a pull request open for a Cordio stack update: #10272. Are there any changes to the GPIO usage in the update?

Issue request type

[ ] Question
[ ] Enhancement
[x] Bug

[ciarmcom]

Internal Jira reference: https://jira.arm.com/browse/MBOCUSTRIA-1121

[donatieng]

Hi @marcemmers and thanks for raising the issue - we'll make sure this is addressed in the upcoming PR

[paul-szczepanek-arm]

this is fixed in the 19.2 cordio LL, PR pending preceding PR

[yogeshk19]

@donatieng , @paul-szczepanek-arm is this slated to be fixed in the 5.12.2 release?

[shane-taurean]

I'm trying the mbed-os-example-ble-LED and mbed-os-example-ble-BatteryLevel cordio example codes on the nRF52_DK (nRF52832). I get unwanted pulses on the analog pins, around 3 pulses repeated every second during advertising and 1 pulse every 45 ms after connecting via nRFConnect.

I tried mbed-os releases 5.11.5, 5.12.1 and also the master branch.

Is this related to the same issue and any resolution expected soon?

[donatieng]

@shane-taurean that looks like the same issue

Work is being done here to resolve it: #10272