Hi misc@,
        while testing the to be released 4.1 I found a problem with the 
snmpd daemon (package is net-snmp-5.1.3p5).

Trying, from another machine a command like this:

        snmptable -c public -v 1 1.2.3.4 HOST-RESOURCES-MIB::hrSWRunTable

where 1.2.3.4 is the ip address of the OpenBSD server, the snmpd daemon 
hangs eating all the cpu it can find.

I tried running the daemon as:

        snmpd -d -D -f -q -u nobody -g nobody

to see the debug output. The last lines are

snmp_agent: tp->start HOST-RESOURCES-MIB::hrSWRunType, tp->end 
HOST-RESOURCES-MIB::hrSWRunStatus, 
trace: netsnmp_add_varbind_to_cache(): snmp_agent.c, 1806:
snmp_agent: add_vb_to_cache(0x87eab780, 7, 
HOST-RESOURCES-MIB::hrSWRunStatus, 0x872d2180)
trace: snmp_call_callbacks(): callback.c, 176:
callback: START calling callbacks for maj=1 min=12
trace: snmp_call_callbacks(): callback.c, 184:
callback: calling a callback for maj=1 min=12
trace: vacm_in_view(): mibII/vacm_vars.c, 747:
mibII/vacm_vars: vacm_in_view: ver=0, community=public
trace: netsnmp_udp_getSecName(): snmpUDPDomain.c, 744:
netsnmp_udp_getSecName: resolve <"public", 0x2d06bc0a>
trace: netsnmp_udp_getSecName(): snmpUDPDomain.c, 749:
netsnmp_udp_getSecName: compare <"public", 0x4a0110ac/0xffffffff>... nope
trace: netsnmp_udp_getSecName(): snmpUDPDomain.c, 749:
netsnmp_udp_getSecName: compare <"public", 0x2c05bc0a/0xffffffff>... nope
trace: netsnmp_udp_getSecName(): snmpUDPDomain.c, 749:
netsnmp_udp_getSecName: compare <"public", 0x2d06bc0a/0xffffffff>... 
SUCCESS
trace: netsnmp_subtree_find_first(): agent_registry.c, 156:
subtree: looking for subtree for context: ""
trace: netsnmp_subtree_find_first(): agent_registry.c, 160:
subtree: found one for: ""
trace: vacm_in_view(): mibII/vacm_vars.c, 854:
mibII/vacm_vars: vacm_in_view: sn=anonymousSecName002, 
gn=anonymousGroupName002, vn=anonymousView002
trace: vacm_checkSubtree(): vacm.c, 526:
vacm:checkSubtree: , included
trace: snmp_call_callbacks(): callback.c, 196:
callback: END calling callbacks for maj=1 min=12 (1 called)
trace: netsnmp_add_varbind_to_cache(): snmp_agent.c, 1871:
snmp_agent: tp->start HOST-RESOURCES-MIB::hrSWRunStatus, tp->end 
HOST-RESOURCES-MIB::hrSWRunEntry.8, 
trace: netsnmp_call_handlers(): agent_handler.c, 443:
handler:calling: main handler bulk_to_next
trace: netsnmp_call_handler(): agent_handler.c, 381:
handler:calling: calling handler bulk_to_next for mode GETNEXT
trace: netsnmp_call_handler(): agent_handler.c, 381:
handler:calling: calling handler old_api for mode GETNEXT
trace: header_hrswrunEntry(): host/hr_swrun.c, 378:
host/hr_swrun: var_hrswrunEntry: HOST-RESOURCES-MIB::hrSWRunIndex 0
(index 20 (entry #1) ....HOST-RESOURCES-MIB::hrSWRunIndex
 saved

at this time there is no other output and the daemon is running at full 
speed.

The same happens on another 4.1 so I don't think it's hw related.

# dmesg | head -2
OpenBSD 4.1-current (GENERIC) #1466: Fri Apr  6 01:36:13 MDT 2007
    deraadt@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC

# snmpd -v

NET-SNMP version:  5.1.3
Web:               http://www.net-snmp.org/
Email:             net-snmp-coders@lists.sourceforge.net

Any ideas?

D.