Suricata

On monitoring my bpf map `flow_table_v4`, I realized the timeouts did not follow the yaml configuration I had in the `suricata.yaml` config

```
flow-timeouts:
default:
emergency-established: 100
emergency-new: 10
established: 300
new: 30
bypassed: 5
emergency-bypassed: 5
```

I realised it was because the function used to get the flow timeouts `FlowGetTimeoutPolicy` and `FlowGetFlowTimeoutDirect` were not using the config value but a hardcoded constant (which is FLOW_STATE_CAPTURE_BYPASSED`) in the case of a `FLOW_STATE_CAPTURE_BYPASSED`.

We could apply this diff to use the `flow-timeouts.bypassed` config value:

Not sure if this patch is okey or if we want to add a config field  `flow-timeouts.capture-bypassed` to distinguish timeout between `local_bypass` vs `capture_bypass`.

I can push the above patch if you want.