In the pwndbg integration code, I ask for function data in all the live stack frames in the current thread, so I can annotate stack addresses from higher function calls properly as well (https://github.com/pwndbg/pwndbg/blob/738dfb763840f1ac0353fb66f5db78a4c46cf9bc/pwndbg/integration/__init__.py#L850-L860). Currently this implies multiple RPC calls. It would be significantly nicer if I could batch all those requests into RPC calls and shave off the performance overhead.
related: pwndbg/pwndbg#3553