Trace-based Debugging (CTS)

Standard Trace-based Debugging requires:

• Continuous instruction flow trace
• Continuous data flow trace (at least read accesses)

Any record in the trace can be selected as starting point for Trace-based Debugging.





As soon as the starting point is selected TRACE32 reconstructs in its GUI the context of the target system as it was when the instruction at the
starting point was executed.





When the reconstruction is finished the program steps previously recorded in real-time in the trace memory can be debugged again in TRACE32 PowerView.




The following debugging commands are available for forward debugging:

• Single step
• Step over call
• Go to the next instruction
• Go to the end of the current function
• Return to the caller function

The following debugging commands are available for backward debugging:

• Single step back
• Step back over call
• Go to the entry of the current function

Trace display including all local variable



The fact that TRACE32 can reconstruct the register and memory contents for any trace sampling points allows the user to display a full high-level language trace including all register and stack variables.



Function nesting display





Function run-time timing display

If more trace information is generated than the trace interface can output, trace information can be lost. The result are trace gaps
(TARGET FIFO OVERFLOWs).



The Trace-based Debugging technology enables TRACE32 to fill in most of these gaps. These records are marked as "s"ynthetically reconstructed (exes instead of exec in the example above).