1. Add -zc and -zp to your boot args. It's not documented but it greatly helps catch zone allocation problems (OSMalloc/OSFree/buffer overrun) issues.

2. Use snapshots instead of rebooting. It's much faster to revert back to a snapshot than to reboot your virtual machine instance after a crash.

3. If you use a shared directory between your instance and host machine for moving your KEXTs/other code, make sure to MD5 the files before loading. It's very common for the instance to be using stale cached blocks.

4. While not necessary, I create a separate network interface that's host only to debug on. I give it a static ip and add an entry to my host's host file. It makes debugging instances easier to do since I can connect by name, i.e. kdb-remote vm0.