Tracking bug for Net-SNMP Counter32 two's complement notation BER bugs
Description
Certain versions of the Net-SNMP agent have trouble following the BER with respect to two's complement notation when marshaling large positive values for Counter32 objects. This problem is suspected to affect versions as old as 5.2, but has been positively identified with version 5.2.1.2 of the Net-SNMP agent when running on the x86_64 architecture. According to the Net-SNMP project's changelogs, the problem was addressed in the very next released version (5.2.2).
Messages like the following one in snmp4j-internal.log are probably indicative of this problem:
org.snmp4j.MessageDispatcherImpl: java.io.IOException: Only 32bit unsigned integers are supported at position 52
The byte sequence referenced in the message above is:
41:05:01:0f:1c:a5:26
See _SNMP, SNMPv2, SNMPv3, and RMON 1 and 2_, 3rd Ed., William Stallings (Addison Wesley, 1999), p. 591 for an explanation of why that sequence is not valid for a positive integer.
Certain versions of the Net-SNMP agent have trouble following the BER with respect to two's complement notation when marshaling large positive values for Counter32 objects. This problem is suspected to affect versions as old as 5.2, but has been positively identified with version 5.2.1.2 of the Net-SNMP agent when running on the x86_64 architecture. According to the Net-SNMP project's changelogs, the problem was addressed in the very next released version (5.2.2).
Messages like the following one in snmp4j-internal.log are probably indicative of this problem:
org.snmp4j.MessageDispatcherImpl: java.io.IOException: Only 32bit unsigned integers are
supported at position 52
The byte sequence referenced in the message above is:
41:05:01:0f:1c:a5:26
See _SNMP, SNMPv2, SNMPv3,
and RMON 1 and 2_, 3rd Ed., William Stallings (Addison Wesley, 1999), p. 591 for an explanation of why that sequence is not valid for a positive integer.