Consider this a bug report as I'm going back to version 10 - way too many problems on version 11
Working with an LPC1517, when using the debug launch the resultant binary on the chip has a section of 0xFFFFFFFF in the flash. This results in an immediate hard fault for obvious reasons.
The first flash on an unprogrammed chip is OK, after that this problem occurs. The disassembly of the binary on disk shows that it was compiled correctly. Using the separate GUI flash tool and choosing 'mass erase' results in a proper programmed chip. Using the same setting in the launch profile results in a broken flash
Additionally after the first flash with version 11 I can no longer flash the chip using the serial boot loader.
As stated in the original problem description, using 'mass erase' in the launch profile does not work.
It appears to me that it IS erasing the chip. An erased eprom is all 1's, i.e. 0xFFFFFFFF, since that is what ends up in the on chip binary, it appears that the chip is being erased but certain sectors are just not being re-programmed. This has happened across several chips so it is not related to a specific chip failure.
What is more frustrating (as if this is not bad enough) is that once it happens to a chip, that chip cannot be programmed via serial boot loader any longer. It is NOT the CRP value. I can take the chip back to version 10 and program it via SWD - from that I can see that the CRP value is not locking it down. But even at this point it is no longer accessible via serial.