Link similar things together. #2163

Passing the number of instructions to be executed seems to be the right approach. However, after inspecting the call stack of wasm_runtime_call_wasm, we need to propagate this "instruction count" down to the following functions:

• For the interpreter:
  1. wasm_call_function
  2. interp_call_wasm (a.k.a. wasm_interp_call_wasm / call_wasm_with_hw_bound_check)
  3. wasm_interp_call_func_bytecode

In this final function, we can use the instruction count to exit the interpreter loop after executing N instructions and return (by propagating an additional variable to the initial call). However, I’m not exactly sure where the loop condition is checked within this goto mess.

• For AOT:
  I'm not entirely sure, but I believe this concerns the compilation process, where a "return" statement must be added every N instructions ? (maybe far less du to branching).