Add workaround in IPv6.pm to deal with possibly incorrect IPV6-MIB implementations

This commit is contained in:
Jeroen van Ingen
2016-03-01 08:37:26 +01:00
parent e9b4eaaeb9
commit add86bc566

View File

@@ -75,7 +75,7 @@ $VERSION = '3.31';
'ip_addr6_pfx' => 'ipAddressPrefix', # IP-MIB
'c_addr6_pfx' => 'cIpAddressPrefix', # CISCO-IETF-IP-MIB
'ip_addr6_index' => 'ipAddressIfIndex', # IP-MIBw
'ip_addr6_index' => 'ipAddressIfIndex', # IP-MIB
'c_addr6_index' => 'cIpAddressIfIndex', # CISCO-IETF-IP-MIB
'ip_addr6_type' => 'ipAddressType', # IP-MIB
@@ -142,6 +142,12 @@ sub ipv6_n2p_addr {
# so what we've collected in that variable is actually the first byte of the address.
$v6_packed = pack('C', $addrsize) . $v6_packed;
}
if (length($v6_packed) == 17) {
# Workaround for IPV6-MIB on Windows 2012: if the address is one byte too long, the SNMP agent probably has an incorrect
# implementation where a length field precedes the actual IPv6 address.
# In that case, the first character should be chr(16), ie 0x10; strip it if that's the case.
$v6_packed =~ s/^\x10//;
}
if (length($v6_packed) == 16) {
$v6addr = join(':', map { sprintf("%04x", $_) } unpack("n*", $v6_packed) );
$return->{$row} = $v6addr;