Hi,
I am building Net-SNMP-5.4.1 with openssl0.9.7m and ipv6 enabled.

Running SNMPv3 and snmpd.conf has
master agentx
agentXTimeout 10
agentXRetries 6
agentXSocket tcp:127.0.0.1:705

snmpd is run as follows
$MY_SNMP_PATH/snmpd -x tcp:127.0.0.1:705 udp:8500 -c /$MY_SNMP_PATH/
snmpd.conf
ie. master is listening on 8500 and to subagent (agentX) on 705

I'm aware of the harmless duplicate registration traces,
and have confimed that my subagent MIBS register with the master
[ via both REGISTER_MIB() and register_mib() ].
Subagent code does
- netsnmp_ds_set_boolean(NETSNMP_DS_APPLICATION_ID,
NETSNMP_DS_AGENT_ROLE, 1);
- netsnmp_ds_set_string(NETSNMP_DS_APPLICATION_ID,
NETSNMP_DS_AGENT_X_SOCKET,
"tcp:127.0.0.1:705");
prior to init_agent.

QUESTION
--------
An snmpwalk on .1.3.6.1.2.1 snmpd responds no problems.
Snmpwalk for my subagent OID results in packets piling on the 705
connection
and it eventually closes with logs
- Timeout: No Response from localhost:8500

netstat traces
udp 0 0 0.0.0.0:8500 0.0.0.0:*
....
tcp 0 0 127.0.0.1:705 0.0.0.0:*
LISTEN
tcp 0 0 127.0.0.1:705 127.0.0.1:41817
ESTABLISHED
tcp 3168 0 127.0.0.1:41817 127.0.0.1:705
ESTABLISHED
....
tcp 0 0 127.0.0.1:705 0.0.0.0:*
LISTEN
tcp 4513 0 127.0.0.1:41817 127.0.0.1:705
CLOSE_WAIT
tcp 0 0 127.0.0.1:705 127.0.0.1:41817
FIN_WAIT2


This has been working fine with NetSNMP-5.1.4 .
Really appreciate any clues as to what i'm missing, doing wrong or if
i need a patch.