I am seeing a deadly embrace occur in a TCP connection immediately after an ARP request.
Attached is a zip file with 5 examples captured via tcpdump.
In all cases, a TCP connection is continuing normally, until an ARP request is made for the default gateway (the firmware behind the CC3000 is talking to a cloud app running on Amazon Web Services.)
Immediately after the ARP request, the CC3000 appears to get into a deadly embrace with it's peer.
This is the first ARP request made by the CC3000 for the default gateway.
This is only broken if/when the firmware behind the CC3000 initiates the transmission of another packet, which seems to break the log-jam.
Setting aucARP to 0 in netapp_timeout_values() does not seem to avoid the arp request being made by the CC3000.
I am waiting for confirmation in another thread (http://e2e.ti.com/support/low_power_rf/f/851/t/315682.aspx) that I can generate non-corrupted debug logs, and will post those here as soon as I know that they are good.
Meanwhile, does this behaviour sound familiar to anyone, is it a known problem etc . We are running 1.24.