Improve network fuzzing mode target health detection

[jzebor]

trafficstars

Using manul for blackbox network fuzzing mode works well for sending fuzzed data, but it does not appear to have any instrumentation to detect and/or recover from a remote target crash.

Provided the remote side blackbox is up and running manul will continue sending the fuzzed data until the remote side crashes. Once a remote side crash happens, the manul threads all die and even if the remote side comes back online manul is still considers the target as dead. It also doesn’t detect remote side crashes in any way I could tell other than executions stop in the UI.

It would be useful to have a "health check" functionality along with some logic on how to handle a remote side crash like saving the last know fuzzed input upon failed health check along with restarting the fuzz run when the remote side is available again.

[mxmssh]

Thanks for opening this issue. I guess this functionality will be added with fork server.