What Exactly Does a Function’s Call Stack Store? A Deep Dive into Stack Frames

Today we take a simple look at what information a function’s call stack saves.

Assume three functions where function 1 calls function 2 and function 2 calls function 3. Each active function receives its own memory block called a stack frame . The three frames together form the stack area .

The boundaries of each stack frame are defined by the CPU registers ESP (stack pointer) and EBP (base pointer). EBP points to the bottom of the frame, while ESP points to the top. Moving ESP down allocates more stack space; moving it up releases space.

The first action of a function call is to push the return address onto the stack, so that after the callee finishes the CPU knows where to continue execution.

When the callee starts, it saves the caller’s EBP (the previous frame’s base pointer) and then copies the current ESP value into EBP, establishing a new frame for the callee.

If necessary, the callee also pushes the values of certain registers onto the stack. This is required because registers are shared among all functions and are limited in number; preserving them prevents the callee from corrupting data the caller still needs.

Local variables (e.g., y and z) can be kept in registers, but when registers run out they are stored on the stack. Likewise, arguments for any further function calls are placed on the stack.

Finally, the return address of the callee is again saved on the stack so the CPU can resume execution after the callee returns.

In summary, a function’s stack frame typically stores four categories of information: the saved registers, local variables, function parameters, and the return address.