Well, I'm doing the same. Just that my simulator is in C, and runs the original firmware (in C). Just on amd64, with all the IO and physical layers simulated (via memcpy).
Mesh networks are extremely tricky to test, as simple changes in the routing (or e.g. physical problems) can cause huge disruptions, exponentially. Unlike everything else. That's why you need a simulator.
But the easy part is, that with mesh testing you don't need a multithreaded simu. With more IO oriented firmware you unfortunately need an IO thread, separated from the original firmware. To simulate it properly, without qemu/renode emulation or even real HW. (Which does not scale).
Mesh networks are extremely tricky to test, as simple changes in the routing (or e.g. physical problems) can cause huge disruptions, exponentially. Unlike everything else. That's why you need a simulator.
But the easy part is, that with mesh testing you don't need a multithreaded simu. With more IO oriented firmware you unfortunately need an IO thread, separated from the original firmware. To simulate it properly, without qemu/renode emulation or even real HW. (Which does not scale).