Has anyone seen this start-up fault in DFU from SDK 8.1?

[Dan Danknick]

I compiled SDK810\examples\dfu\bootloader\pca10028\dual_bank_ble_s110\arm5_no_packs without error and slapped it and the SoftDevice 8.0.0 into my erased target using nRFgo Studio 1.21.0.2. There is NO application loaded so I expect to sit in the bootloader and see "DfuTarg" advertising.

But instead, I'm getting stuck in the HardFault handler. Keil's call tree shows the following:

I'm not quite sure what is happening. If I set a breakpoint at the entry of crc16_compute() all of the parameters look fine but if I run at full speed, and then break in once in the HardFault handler, the length parameter is 1000x too big and surely runs off the end of RAM.

This is my first project using the newfangled DFU that passes bonding info around. I appreciate all of the R&D the Nordic engineers have done and given to us for free, I just wish bring-up was a bit easier.

Thanks in advance, Dan

[Dan Danknick]

The same firmware load worked on the new PCA10031 dongle. The customer's target HW uses the older QFAAG0 chip so I think there is a genuine incompatibility with a modern bootloader plus SD 8. Sorry for the false alarm.