From: SourceForge.net <no...@so...> - 2005-08-19 09:22:10
|
Patches item #1257328, was opened at 2005-08-12 11:56 Message generated for change (Comment added) made by ssant You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=312694&aid=1257328&group_id=12694 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Pending Resolution: None Priority: 5 Submitted By: Nobody/Anonymous (nobody) Assigned to: Robert Story (rstory) Summary: Patch for counter32 overflow problem Initial Comment: On s390 64 bit system running Net SNMP 5.1.1 i came across this counter32 overflow problem for some of the IBM OSA mibs. On NET-SNMP on 2.6 (intfc 7 is CHPID 2E, 12 is CHPID 1C): # snmpwalk -OS -c public -v 2c localhost ibmOsaExp10GigEthInPackets IBM-OSA-MIB::ibmOsaExp10GigEthInPackets.7 = Counter32: 158745200 IBM-OSA-MIB::ibmOsaExp10GigEthInPackets.12 = Counter32: 18446744072074027879 ^^^^^^^^^^^^^^^^^^^^^^^^^^ The above value is very large because of counter32 value overflow. The problem seems to be with the code in snmplib/mib.c sprint_realloc_counter( ) function. While copying the data to a tmp buffer %lu is used which causes this overflow. Given is the patch to fix this problem. Submitted by sa...@in... ---------------------------------------------------------------------- Comment By: ssant (ssant) Date: 2005-08-19 14:52 Message: Logged In: YES user_id=1328139 Unfortunately i don't have the machine on which i had recreated the problem. I have been trying to get access to the machine. The problem was on a 64 bit s390 system. At this point of time i suggest we close this ticket. I will try to recreate this on a AMD system and try out the patch you have sugested. If the problem can be recreated will update this report. ---------------------------------------------------------------------- Comment By: Robert Story (rstory) Date: 2005-08-18 19:38 Message: Logged In: YES user_id=76148 I finally got a 64 bit test machine, and the patch has been updated to handle many more 64bit overflow cases. Make sure to try the most recent version. http://sourceforge.net/tracker/index.php?func=detail&aid=1259966&group_id=12694&atid=456380 ---------------------------------------------------------------------- Comment By: Robert Story (rstory) Date: 2005-08-18 18:06 Message: Logged In: YES user_id=76148 actually, there are two issues here - the client side and the agent side. Is the agent with the IBM-OSA-MIB a net-snmp agent? ---------------------------------------------------------------------- Comment By: ssant (ssant) Date: 2005-08-16 10:08 Message: Logged In: YES user_id=1328139 I had submitted the patch against 5.1.3.1 release. I will try out the patch in the above mentioned link and post the results. ---------------------------------------------------------------------- Comment By: Robert Story (rstory) Date: 2005-08-15 20:49 Message: Logged In: YES user_id=76148 I believe this has been fixed by this official patch: http://sourceforge.net/tracker/index.php?func=detail&aid=1259966& group_id=12694&atid=456380 Can you see if this helps? (The patch is for more recent versions, hopefully it will still apply cleanly). ---------------------------------------------------------------------- Comment By: Thomas Anders (tanders) Date: 2005-08-15 03:34 Message: Logged In: YES user_id=848638 Can the problem be reproduced with 5.1.3.1? If so, can you please resubmit the patch against that version? There have been *lots* of fixes between 5.1.1 and 5.1.3. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=312694&aid=1257328&group_id=12694 |