So, it's great that CrashReporter exists and it is possible to see a stack trace of where the program crashed. However, there is a problem. The usual advice to use addr2line only seems to work well if you don't put any functions into only flash. Usually the address given by the crashdump then works with addr2line and you can figure out where the crash was. However, it seems if you use flash only functions then the way things are arranged gets all messed up and addr2line no longer is a reliable way to find the function. Some programs are much larger than the available RAM1 space and so there comes a need to keep some functions out of RAM.
Am I not doing something right or is this a real problem? Is there a good way to mitigate the issue so that complex programs that are comprised of a mixture between RAM and FLASH resident code can be debugged via the crash handler?
Am I not doing something right or is this a real problem? Is there a good way to mitigate the issue so that complex programs that are comprised of a mixture between RAM and FLASH resident code can be debugged via the crash handler?