Why is a HardFault exception occuring in sd_ble_gatts_hvx (SD 132 V5.1)

[Kyle Krueger]

After already having sent a few notifications with a custom BLE characteristic, I am getting a HardFault in the softdevice.

Initially I had added a log call to print the result of the command with NRF_LOG_FLUSH() called immediately afterwards. However, this log message was never printed.

So I booted up Ozone, and found in the Call Stack that the device is getting choked up on HardFault Exception within the call to sd_ble_gatts_hvx.

Anyone else having crashes like this within the softdevice on the new S132 V5.1 release?

[Kyle Krueger]

I had the issue again with S132 V5.0, but noticed a cause on our side. A buffer overflow in a part of our firmware had overwritten some memory that was important to the sd_ble_gatts_hvx function.

This had not happened with SDK 12.0 and SD 132 V3 due to sheer luck. The memory that was being overwritten had before been non-critical.

Thanks for your your help, happy to say there is no error in the SDK this time!

[Kyle Krueger]

Note: we still ended up having some issues later that were causing a hardfault in the same place. It had to do with priority of SPI being higher than BLE events. It would be great if warnings could be given in these cases.