Erasing flash when there is a BLE Connection

RE: Erasing flash when there is a BLE Connection

Gordon — Mon, 18 Apr 2016 11:21:02 GMT

edit: solved.

RE: Erasing flash when there is a BLE Connection

Gordon — Mon, 18 Apr 2016 10:52:52 GMT

Perfect, thanks! That'll almost certainly be the problem then - I was using nrf_nvmc_page_erase

RE: Erasing flash when there is a BLE Connection

Susheel Nuguru — Mon, 18 Apr 2016 10:47:39 GMT

Accessing flash directly while BLE activity is not safe and might lead to flash corruptions and softdevice assertions.

You can use sd_flash_page_erase API that synchronizes with softdevice and makes sure that it is not interrupted by BLE activity.

RE: Erasing flash when there is a BLE Connection

Gordon — Mon, 18 Apr 2016 10:20:11 GMT

Ahh, thanks! I'd got a ble_app_error_handler but not a app_error_fault_handler. Adding that definitely changes the behaviour, but it isn't able to print anything - which makes me think it is being called from within an IRQ (which would be something to do with BLE, since at that point I don't have any active IRQs).

Any idea what could actually be causing this, or how it could be avoided? Unfortunately I don't have debugging set up. Presumably the BLE stack interrupts in the middle of a flash write, and doesn't like it for some reason?

RE: Erasing flash when there is a BLE Connection

Susheel Nuguru — Mon, 18 Apr 2016 08:48:06 GMT

Seems like there is some error returned in one of the functions in your code. You know that APP_ERROR_CHECK default behavior is that it if the condition inside it does not pass, then it will reset you chip. Try to look at this thread. First try to find exact place where the system resets.