At this moment, f9-kernel Kprobe approach was using function address in kernel,
we can saw it at kernel/sampling-kdb.c for the example.
Compare to Linux kernel Kprobe, it was using symbol name for user space program to insert the kprobe into kernel:
struct kprobe kp = {
.symbol_name = "_do_fork",
};
After reviewing, f9-kernel already provide the infrastructure ksym for this approach, we can take time and improve Kprobe for real dynamic tracing in user space.