March 2006 - Libsmi - lists.ibr.cs.tu-bs.de

libsmi reporting parse error, expecting '$'
by Juergen Schoenwaelder 21 Mar '07

21 Mar '07

Folks, I was getting this error IF-MIB:1: parse error, expecting `$' on several differnt platforms and it finally turned out to be an issue with bison/flex. I do not know the details, but I do know that the combination bison 1.875d and flex 2.5.31 does not work while the combination bison 2.1 and flex 2.5.4 seems to be fine. So if someone runs into similar trouble, make sure you upgrade your build tools... /js -- Juergen Schoenwaelder International University Bremen <http://www.eecs.iu-bremen.de/> P.O. Box 750 561, 28725 Bremen, Germany

2 3

minor memory leak
by Juergen Schoenwaelder 23 Mar '06

23 Mar '06

Hi, below is a very simple program which simply initializes libsmi and afterwards cleans up again. #include <smi.h> int main(int argc, char **argv) { smiInit(NULL); smiExit(); return 0; } Running this program with valgrind on an Intel Linux box reveals that this program leaks some memory. It seems that smiExit() does not cleanup all memory that got allocated. 44 bytes in 1 blocks are definitely lost in loss record 1 of 1 at 0x401E88E: calloc (vg_replace_malloc.c:279) by 0x402D949: smiMalloc (util.c:88) by 0x4028026: addNode (data.c:1029) by 0x402973A: smiInitData (data.c:3363) by 0x402EB5D: smiInit (smi.c:208) by 0x804847B: main (in /tmp/aa) It is not a big deal, still I thought I at least document the issue on this mailing list. ;-) /js -- Juergen Schoenwaelder International University Bremen <http://www.eecs.iu-bremen.de/> P.O. Box 750 561, 28725 Bremen, Germany

1 0

configure error
by Michael Friddell 14 Mar '06

14 Mar '06

I get the following error when I try to run ./configure: configure:2620: gcc -c -g -O2 conftest.c >&5 conftest.c:1: error: syntax error before '__cplusplus' configure:2626: $? = 1 configure: failed program was: | ifndef __cplusplus | choke me | endif The section of the configure script that appears to cause this is: cat >conftest.$ac_ext <<_ACEOF ifndef __cplusplus choke me endif _ACEOF I've tried commenting this section out but then the configure script fails when it tries to use this command. I'm at a loss. ----------------------------------------- Michael Friddell friddell(a)earthlink.net 2721 Airwood Dr, Nashville, TN 37214 cell: 615-330-8677 home: 615-872-7443 work: 615-263-8398 YahooIM/AIM ID: ClearBlueLou

1 0

RE: [libsmi] smiquery not showing correct MAX-ACCESS.
by John Wiebe 06 Mar '06

06 Mar '06

Thanks, that clears it up. I can use the Entry node to verify the read-create access. -----Original Message----- From: Juergen Schoenwaelder [mailto:j.schoenwaelder@iu-bremen.de] Sent: Monday, March 06, 2006 2:14 PM To: John Wiebe Cc: libsmi(a)ibr.cs.tu-bs.de Subject: Re: [libsmi] smiquery not showing correct MAX-ACCESS. On Mon, Mar 06, 2006 at 01:56:18PM -0500, John Wiebe wrote: > I've been using smiquery to get various information from MIB's. > However, I've found that it isn't reporting the correct MAX-ACCESS > if the MAX-ACCESS is read-create, it always returns read-write. Libsmi has internally a different understanding of access levels than the SMIv2. If you have an SMIv2 table which supports row creation, you have to mark all writable columns read-create. Libsmi instead marks them as read-write and sets a creatable flag on the table's entry node. See for example smiquery node SNMP-VIEW-BASED-ACM-MIB::vacmAccessEntry and check the "Creatable:" line. Since smiquery reports the internal representation (very useful for debugging), this is just consistent. /js -- Juergen Schoenwaelder International University Bremen <http://www.eecs.iu-bremen.de/> P.O. Box 750 561, 28725 Bremen, Germany

1 0

smiquery not showing correct MAX-ACCESS.
by John Wiebe 06 Mar '06

06 Mar '06

I've been using smiquery to get various information from MIB's. However, I've found that it isn't reporting the correct MAX-ACCESS if the MAX-ACCESS is read-create, it always returns read-write. I noticed in the changelog that this was fixed once before and I'm courious if I am missing something obvious when I try to get this value. In the example below, I am using the default mibs shipped with libsmi. I am using libsmi-0.4.3 on a linux machine. Any help or work around would be appreciated. Thanks, libsmi-0.4.3 ChangLog: 2000-02-07 Frank Strauss <strauss(a)ibr.cs.tu-bs.de> * tools/dump-mosy.c: fixed determination of types of %tc's and objects; added special type `ObjectID'. Added distinction of read-write/read-create. Checked output against mosy output for some standard modules (IF-MIB, RMON2-MIB, ...). from the IF-MIB MAX-ACCESS ifStackStatus ifStackStatus OBJECT-TYPE SYNTAX RowStatus MAX-ACCESS read-create STATUS current DESCRIPTION ... > smiquery -p IF-MIB node ifStackStatus MibNode: IF-MIB::ifStackStatus OID: 1.3.6.1.2.1.31.1.2.1.3 ParentNode: IF-MIB::ifStackEntry Type: SNMPv2-TC::RowStatus Declaration: OBJECT-TYPE NodeKind: column Access: read-write Status: current Description: The status of the relationship between two sub-layers. Changing the value of this object from 'active' to 'notInService' or 'destroy' will likely have consequences up and down the interface stack. Thus, write access to this object is likely to be inappropriate for some types of interfaces, and many implementations will choose not to support write-access for any type of interface.

2 1

Does smidump support 64 bit addressing.?
by Pamidipati Suresh-G20238 06 Mar '06

06 Mar '06

Hi All, Can smidump be used on 64 bit platforms..? Do we need to specify any option at the time of configuration for this. Please clarify this. Your help will be greatly appreciated. Thanks, Suresh.

3 2

RE: [libsmi] Does smidump support 64 bit addressing.?
by Pamidipati Suresh-G20238 03 Mar '06

03 Mar '06

Thanks for the information. Regards, Suresh. -----Original Message----- From: Frank Strauß [mailto:strauss@ibr.cs.tu-bs.de] Sent: Friday, March 03, 2006 1:52 PM To: Pamidipati Suresh-G20238 Cc: libsmi(a)ibr.cs.tu-bs.de Subject: Re: [libsmi] Does smidump support 64 bit addressing.? Pamidipati Suresh-G20238 wrote: > Hi All, > > Can smidump be used on 64 bit platforms..? Do we need to specify any > option at the time of configuration for this. Please clarify this. > Your help will be greatly appreciated. It compiles and runs on Linux x86_64 without any special intervention and the test suite succeeds (except for currently outdated test cases). However, I don't regularly use it on any 64bit system.

1 0