18.1.9. The call stack

Application code uses the call stack to pass parameters, store local data and store return addresses. The data each function pushes on the stack is organized into a stack frame. When a debugger stops a core, it might be able to analyze the data on the stack to provide you with a call stack, that is, a list of function calls leading up to the current situation. This can be extremely useful when debugging, as it enables you to determine why the application has reached a particular state.

To reconstruct the call stack, the debugger must be able to determine which entries on the stack contain return address information. This information might be contained in debugger information (DWARF debug tables) if the code was built with these included, or by following a chain of frame pointers pushed on the stack by the application. To do this, the code must be built to use frame pointers. If neither of these types of information are present, the call stack can not be constructed.

In multi-threaded applications, each thread has its own stack. The call stack information therefore only relates to the particular thread being examined.

Copyright © 2015 ARM. All rights reserved.ARM DEN0024A
Non-ConfidentialID050815