Other Parts Discussed in Thread: HALCOGEN
Tool/software: Code Composer Studio
Hello,
with a new HALCoGen-generated project for the RM46L852, the behaviour of the CCS debugger became "unreasonable": It fails to stop at breakpoints in sys_startup.c or at the begin of main. The only active buttons related to the processor state are 'Terminate' and 'Reset'.
I found several forum threads about these symptoms 'solved' by hinting at watchdog reset and time for initialization (therefore the breakpoints in sys_startup). I learned that, with the Hercules MCUs, the internal WD is inactive by default and the TPS65381 would silently reach and keep in the SAFE STATE without resetting the RM46, isn't it?
I found a thread 'solved' by stating that the 'Debug' Tab must be visible. In my case it is and shows the project name in a line holding the focus (marked blue) and a 2nd line with the debug probe ("TI XDS100v2 ... (Running)"). Only when I accidentally moved the focus to the 2nd line, the 'Pause' Button became active. Pausing reveals that the programm was running in the infinite loop at the end of main() (my USER CODE in sys_main.c). So it does not relate to a watchdog preventing main to be reached.
I also found 2181126 about a HW failure (my board is designed in-house and still under test), but that is also unrelated to my case since I have an older, heavily edited HALCoGen project where the debugger behaves reasonably.
Then I switched to Google for searching the forum and found 1306092, where the customer found out that the problem vanishes by deactivating the CPU Self test (on HALCoGen's 'SAFETY INIT' Tab). That worked for me.
Now, I want to know how the CPU Self test causes this behaviour (may be safety relevant) and whether TI plans to improve on that. (If it is not regarded as a bug, a warning in HALCoGen would prevent wasting time.)
Best regards
Rainald
