Under some circumstances, deleting an ARP entry can cause VPP to seg fault. This happens if I do either of the following after starting up VPP:

• Leave interfaces down, attempt to delete a non-existent ARP entry using vppctl
• Bring up an interface, immediately attempt to delete a non-existent ARP entry using vppctl

If I bring up an interface and allow it to resolve some ARP entries prior to attempting to delete, this doesn't happen. Ideally users wouldn't try to delete an ARP entry that doesn't exist, but the daemon shouldn't crash if it happens.

Attaching a stack trace.